CN120009131A

CN120009131A - A particle detection method, device and system

Info

Publication number: CN120009131A
Application number: CN202311525628.9A
Authority: CN
Inventors: 张瑀健; 张凤波; 徐拓; 王帆; 赵洪志; 付义; 荔栓红; 白羽; 高克京
Original assignee: Beijing Zhongchuang Keyi Technology Co ltd; Petrochina Co Ltd
Current assignee: Beijing Zhongchuang Keyi Technology Co ltd; Petrochina Co Ltd
Priority date: 2023-11-15
Filing date: 2023-11-15
Publication date: 2025-05-16
Anticipated expiration: 2043-11-15
Also published as: CN120009131B

Abstract

The present specification provides a particle detection method, device and system. The method uses two cameras arranged relatively to capture images of freely falling particles, and performs heterochromatic detection based on the images captured by the two cameras. In the process of heterochromatic detection, it is found that some impurities penetrate the particles, which easily leads to repeated counting of the impurity particles. Therefore, it is proposed to repeatedly detect the heterochromatic detection results of the images captured by the two cameras. In the process of repeated detection, the images captured by the camera arranged below in the current cycle are compared with the images of the camera arranged above in the current cycle and the previous cycle, respectively, taking into account the situation that the same particle may appear in the image captured by the lower camera in the current cycle and the image captured by the upper camera in the previous cycle, so that the calculation result is more accurate. The images captured by the two cameras are respectively subjected to heterochromatic recognition, which does not interfere with each other, and can be quickly recognized through two threads to improve the detection efficiency.

Description

Particle detection method, device and system

Technical Field

The present application relates to the field of detection technologies, and in particular, to a method, an apparatus, and a system for detecting particulate matters.

Background

Thermoplastic particulate resins (e.g., polyethylene (PE), polypropylene (PP), polystyrene (PS), impact polystyrene (PS-I), acrylonitrile-butadiene-styrene (ABS), styrene-acrylonitrile (SAN), ethylene-vinyl acetate (EVAC), etc., or modified particulate plastics thereof) have the characteristics of better mechanical and chemical corrosion resistance, higher use temperature, high strength and hardness, mechanical properties (excellent fracture toughness and damage tolerance), excellent fatigue resistance, ability to mold complex geometries and structures, adjustable thermal conductivity, recyclability, good stability in severe environments, repeated molding, weldability, repair, etc. Therefore, the thermoplastic particulate resin is widely used in aspects of human life and work.

The petrochemical industry has strict standard requirements on the appearance quality of plastic particles. For example, in the petrochemical industry standard SH/T1541.1-2019 plastic particle appearance test method (hereinafter referred to as 1541 standard), the defective appearance is defined by classification:

1. Black particles, wherein the visible part of the whole particles is black or dark brown particles;

2. Black speckles-except for black speckles, particles of black or dark brown speckles are visually visible;

3. Color particles, namely particles with other colors besides the black particles, the black specks and the due colors of the resin;

4. Large particles, namely, screen residue particles with the size of more than 5mm, and various continuous particles and elongated particles with two or more particles adhered;

5. small particles, namely, screen residue particles with the size smaller than 2mm, including scraps and crushed particles;

6. snake skin particles are strip resin similar to snake skin;

7. tailing particles, particles with cone angles or burrs, which are generated by poor dicing;

8. Floc, cotton-like, fiber-like or ribbon-like resin with a certain width.

Further, the above plastic particle appearance defects can be classified into three main categories:

1. different colors, namely black particles, black speckles and colored particles, namely particles with other colors besides the due color of the resin;

2. the size is large granules and small granules, namely plastic granules with sieving size of more than 5mm and less than 2 mm;

3. The shape is snake skin particles, tailing particles and floccules, namely the abnormal shape of the plastic particles outside the original shape.

In addition to the classification definition, the 1541 standard gives a visual inspection step of "1000 g of resin pellets were screened out of the large and small granules specified in the definition by means of a pilot screen. Thereafter, at most, 1000g of other types of particles in the pellets were sorted out with tweezers in a period of 10 minutes, and the sorting was counted (the weight of 1 plastic particle was about 0.025g, and the number of 1000g plastic particles was about 4 ten thousand) ".

In practical applications, thermoplastic granular resins are often displayed in front of users in direct natural color, and detection and screening of the resin granules is a necessary option in order to reduce the content of impurities in the granules and improve the yield of the granules. The petrochemical industry has strict standard requirements on the appearance quality of plastic particles. Currently, the main method of verifying the appearance properties of products still relies on manual visual inspection. Relevant regulations are made on the appearance experimental method of the natural-color thermoplastic granular resin, and relevant terms and defined standards are formulated. This standard indicates that besides normal colour, shape standard natural colour particles, other abnormal colour shaped particles are present, such as black particles, black specks, coloured particles, large particles, small particles, snake skin particles and tailing particles, floc and other abnormal particles or impurities.

The visual inspection according to the above standard was performed by manually screening each type of particles with tweezers under good illumination conditions for 10 minutes using a human eye. Along with the increasing requirements of downstream manufacturers on raw material quality, such as the increasing requirements of chemical industry on impurity (heterochromatic special-shaped) and pollutant proportion in raw materials, the manual inspection and counting has the problems of long time consumption, large deviation of different people or the same person on heterochromatic and special-shaped counting at different times, and extremely easy omission of abnormal particles with heterochromatic and abnormal colors below 200 microns.

In the prior art, there are also technical schemes for automatically detecting particulate matters by photographing with a camera, including schemes for detecting different colors according to images photographed by one camera and schemes for detecting different colors according to images photographed by two cameras. However, there is an error between these off-color detection results and the true results. In particular, the error between the result of the different color obtained from the images taken by the two cameras and the actual result is unstable and difficult to eliminate in various ways.

Disclosure of Invention

The specification provides a particle detection method, device and system, which are used for solving the problem that detection errors of the existing particle detection method are difficult to eliminate.

In order to solve the technical problems, a first aspect of the present disclosure provides a method for detecting particles, including acquiring a first color image of particles captured by a first line-up camera and a second color image of particles captured by a second line-up camera during a free falling process of the particles; the first linear array camera and the second linear array camera shoot the particles from two sides of a free falling path of the particles respectively, and the first linear array camera and the second linear array camera are arranged up and down in the direction of the free falling path of the particles; the first linear array camera and the second linear array camera are arranged oppositely on two sides of a free falling path of particles, a period from the moment when a first image is shot to the moment when a second image is shot is taken by the linear array camera is one period, different color detection is carried out on first color images shot in each period of the first linear array camera and different color detection is carried out on second color images shot in each period of the second linear array camera respectively, first operation and/or second operation are carried out on each period of the first linear array camera and the second linear array camera, the first operation comprises judging whether different color detection results of the second color images shot in the current period of the second linear array camera and different color detection results of the first color images shot in the previous period of the first linear array camera all have different color particles, under the condition that different color particles are all available, the different color particles identified in the second color images shot in the current period of the second linear array camera and the first color images shot in the current period of the first linear array camera are compared with the different color particles identified in the corresponding condition, whether the different color particles are correspondingly characterized in the corresponding condition, the second operation comprises judging whether the different color detection result of the second color image shot by the current period of the second linear array camera and the different color detection result of the first color image shot by the current period of the first linear array camera all have different color particles, comparing the different color particles identified in the second color image shot by the current period of the second linear array camera with the different color particles identified in the first color image shot by the current period of the first linear array camera under the condition that the different color particles are all present, determining whether the characteristics of the different color particles are corresponding or not, and discarding the different color particle identification result of one of the corresponding different color particles under the condition that the different color particles corresponding to the characteristics are present.

In some embodiments, determining whether the characteristics of the heterochromatic particles correspond includes determining whether the location of the spot in the heterochromatic particles is a mirror image correspondence.

In some embodiments, determining whether the characteristics of the heterochromatic particles correspond includes mirroring the outer contour shape of the heterochromatic particles.

In some embodiments, the second operation is not performed if the first operation determines that there are different color particles corresponding to the feature, and the first operation is not performed if the second operation determines that there are different color particles corresponding to the feature.

In some embodiments, discarding the identification of one of the corresponding heterochromatic particles includes discarding the identification of the heterochromatic particle having the smaller spot area of the corresponding heterochromatic particle.

In some embodiments, the first color image and the second color image are captured while the diffuse reflective light source irradiates free-falling particulate matter.

In some embodiments, the target image is identified as being heterochromatic when detected; the heterochromatic recognition process includes converting a target image to an HLS color space; acquiring a chromaticity value interval, a brightness value interval and a saturation value interval of N color types of impurities in the particulate matters; the N color types are ordered according to priority, the color type identifiers of the ith priority are set as 255-i, i and N are natural numbers, a first color type identifier is set for each pixel value in a chrominance channel sub-image of a target image according to a chrominance value interval of each color type, a second color type identifier is set for each pixel in a saturation channel sub-image of the target image according to a luminance value interval of each color type, a third color type identifier is set for each pixel in the saturation channel sub-image of the target image according to a saturation value interval of each color type, the first color type identifier, the second color type identifier and the third color type identifier are set to be in a value range of 0, [255-N,255-i ],0 represents the identifier of the color type of the ith priority outside the value interval of each color type, 255-i represents the identifier of the color type of the ith priority, the first color type identifier, the second color type identifier and the third color type identifier are respectively set for each pixel in the saturation channel sub-image of the target image according to a saturation value interval of each color type, whether the first color type identifier, the second color type identifier and the third color type identifier are used as the maximum value of the current color type and the maximum value of the pixel in the current color type are used as the final color type identifiers, the final color type identifier is equal to or the area of the final color type is equal to or greater than a predetermined threshold value, or equal to the area is formed in a predetermined area, and taking the final color type mark of each pixel in the area or the difference value between 255 and the final color type mark as the color mark of the spot area, wherein the target image is the first color image or the second color image.

In some embodiments, the different color identification is performed on the target image during different color detection, wherein the different color identification process further comprises the steps of clipping the target image according to the positions of the spot areas to obtain sub-images, wherein each sub-image corresponds to one spot in the particulate matter, and each sub-image comprises a complete spot image and position information of the spot in the particulate matter.

In some embodiments, the method further comprises performing special-shaped detection by acquiring an image shot by a third linear camera in the process of freely falling the particles, identifying the outline of the particles from the image shot by the third linear camera, calculating the shape characteristic parameters of the particles according to the outline of the particles, and determining whether the particles are special-shaped particles according to the shape characteristic parameters of the particles.

In some embodiments, the shape characteristic parameter comprises at least one of particle area, particle circumference, maximum inscribed circle, minimum circumscribed rectangle, convex hull area, particle size, sieving diameter, surface roughness, roundness, elongation, shape factor, convexity.

In some embodiments, the method further comprises the steps of obtaining a predetermined particle classification model, inputting shape characteristic parameters of the particles into the particle classification model to obtain a particle classification result output by the particle classification model, and determining that the particles are special-shaped particles when the output result of the particle classification model belongs to a predetermined particle type.

In some embodiments, the method further comprises clipping sub-images from the image captured by the third linear camera according to the outline of the identified particulate matter, each sub-image comprising a complete image of the particulate matter, and storing the sub-images in association with a marker indicating whether the particulate matter in the sub-images is a shaped particulate matter.

The second aspect of the specification provides a particulate matter detection system, which comprises an image acquisition unit, a first detection unit and a second detection unit, wherein the image acquisition unit is used for acquiring a first color image of particulate matters shot by a first linear array camera and a second color image of the particulate matters shot by a second linear array camera in the process of freely falling the particulate matters; the first linear array camera and the second linear array camera are used for shooting the particles from two sides of a free falling path of the particles respectively, the first linear array camera and the second linear array camera are arranged up and down in the direction of the free falling path of the particles, the first linear array camera and the second linear array camera are arranged opposite to each other on two sides of the free falling path of the particles, a period from the shooting of a first image to the shooting of a second image of the linear array camera is one period, a different color recognition unit is used for respectively carrying out different color detection on first color images shot in each period of the first linear array camera and respectively carrying out different color detection on second color images shot in each period of the second linear array camera, a summarizing unit is used for executing a first operation and/or a second operation on each period of images of the first linear array camera and the second linear array camera, the first operation comprises judging whether different color detection results of second color images shot by the current period of the second linear array camera and the first color images shot by the first linear array camera and different color detection results of the second color cameras from the first linear array camera to the second linear array camera, and the different color detection results of the first linear array camera different color images shot by the first linear array camera different color images from the first linear array camera different color cameras different color image from the first line camera, the method comprises the steps of determining whether different-color particles exist in different-color detection results of a second color image shot in a current period of a second line-scan camera and different-color detection results of a first color image shot in the current period of the first line-scan camera, comparing the different-color particles identified in the second color image shot in the current period of the second line-scan camera with the different-color particles identified in the first color image shot in the current period of the first line-scan camera under the condition that the different-color particles exist in the different-color detection results of the first color image shot in the current period of the first line-scan camera, determining whether the characteristics of the different-color particles correspond or not, and discarding the different-color particles identified in the corresponding different-color particles under the condition that the different-color particles corresponding to the characteristics exist.

In some embodiments, the image acquisition module further comprises an image of the particulate matter shot by the third linear camera in the process of freely falling the particulate matter, and correspondingly, the system further comprises a special-shaped identification unit, wherein the special-shaped identification unit is used for carrying out special-shaped detection through the following method that the image shot by the third linear camera in the process of freely falling the particulate matter is acquired, the outline of the particulate matter is identified from the image shot by the third linear camera, the shape characteristic parameter of the particulate matter is calculated according to the outline of the particulate matter, and whether the particulate matter is special-shaped or not is determined according to the shape characteristic parameter of the particulate matter.

In some embodiments, the summarizing unit further includes counting the recognition results according to the heterochromatic recognition unit and the heterochromatic recognition unit to obtain a heterochromatic detection result of the particulate matter.

A third aspect of the present specification provides a particulate matter detection device comprising a first hopper for holding particulate matter to be detected; the first conveying mechanism is positioned below the first hopper and is used for conveying the particles falling in the first hopper to the upper part of the second hopper and enabling the particles to freely fall into the second hopper above the second hopper; the second hopper is arranged below the first conveying mechanism and is used for containing particles falling from the first conveying mechanism; the particle detection device comprises a first hopper, a second hopper, a first linear camera, a second linear camera, a third linear camera, an electronic device, a fourth linear camera, a fifth linear camera, a sixth linear camera, a seventh linear camera, a eighth linear camera, a seventh linear camera, a sixth linear camera, a seventh linear camera, a sixth linear camera, a seventh linear camera, a third linear camera, a fourth and a fifth linear camera, and the fourth linear camera, and the first and the third linear camera is arranged in the first and the first is the is and the is and is and.

In some embodiments, the first and second line cameras are each configured with a diffuse light source to capture particulate matter.

In some embodiments, the system further comprises a third linear camera for shooting images of free-falling particles, wherein the images shot by the third linear camera are used for carrying out special-shaped detection, the third linear camera is arranged between the second conveying mechanism and the material collector when the first linear camera and the second linear camera are arranged between the first conveying mechanism and the second hopper, and the third linear camera is arranged between the first conveying mechanism and the second hopper when the first linear camera and the second linear camera are arranged between the second conveying mechanism and the material collector.

A fourth aspect of the present specification provides an electronic device comprising a memory and a processor communicatively coupled to each other, the memory having stored therein computer instructions, the processor implementing the steps of the method of any of the first aspects by executing the computer instructions.

According to the particle detection method, device and system, images of freely falling particles are shot through two cameras which are arranged oppositely, different-color detection is carried out based on the images shot by the two cameras, when some impurities are found to penetrate through the particles in the different-color detection process, the fact that the impurity particles are repeatedly counted is easily caused, therefore, repeated detection is carried out on different-color detection results of the images shot by the two cameras, in the repeated detection process, the images shot by the lower camera in the current period are respectively compared with the images shot by the upper camera in the current period and the images in the previous period, the fact that the same particle possibly appears in the images shot by the lower camera in the current period and the images shot by the upper camera in the previous period is considered, the calculation result is more accurate, different-color identification is carried out on the images shot by the two cameras respectively, the two cameras are not interfered with each other, the detection efficiency is improved through two threads, when the different-color detection results of the images shot by the two cameras are summarized, the same different-color particles are determined according to the characteristics of the different-color particles, the different-color particles are not located between the two images, and the different-color detection results are not accurate, and the different-color particles are not different in the position information are not different.

Drawings

In order to more clearly illustrate the embodiments of the application or the technical solutions in the prior art, the drawings that are necessary for the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the following description are only some of the embodiments described in the application, and that other drawings can be obtained from these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic diagram of a particulate matter detection method provided herein;

FIG. 2 is a schematic diagram of a particulate matter detection device provided herein;

FIG. 3 is a schematic diagram of one embodiment of a detection of a different color;

FIG. 4 is a schematic diagram of a method of heterochromatic identification;

FIG. 5 is a schematic diagram of a method of profile detection;

FIG. 6 is a schematic diagram of a minimum circumscribed circle and a maximum inscribed circle;

FIG. 7 is a schematic diagram of a minimum bounding rectangle;

FIG. 8 is a schematic diagram of convexity calculation basis;

FIG. 9 is a schematic diagram of roughness calculation basis;

FIG. 10 is a schematic diagram of one embodiment of profile detection;

Fig. 11 is a schematic structural diagram of an electronic device provided in the present specification.

Detailed Description

In order to make the technical solution of the present application better understood by those skilled in the art, the technical solution of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application, and it is apparent that the described embodiments are only some embodiments of the present application, but not all embodiments of the present application. All other embodiments, based on the embodiments of the application, which would be apparent to one of ordinary skill in the art without undue burden are intended to be within the scope of the application.

The process of detecting the abnormal color is usually to detect the particulate matters in the image shot by each camera, and if spots (i.e. impurities) in the particulate matters are detected, the number of the abnormal color particles is increased by 1. The inventors found out in the course of studying a technical scheme of detecting a different color from images photographed by two cameras that impurities in some of the particles to be detected penetrate through both sides of the particles, and that such particles are easily repeatedly detected.

For example, the impurities in the target particles penetrate through two sides of the particles, and after the target particles falling freely are photographed by the first linear array camera and the second linear array camera which are arranged oppositely, the target particles photographed by the first linear array camera may contain impurities, and the target particles photographed by the second linear array camera also contain impurities, so that the number of the abnormal-color particles is 2 under the condition that the abnormal-color recognition is carried out on the 'target particles' photographed by the first linear array camera and the second linear array camera only. That is, repeated detection of the target particulate matter is performed.

The conventional target object color detection method aims at an object with uniform surface color such as a ball, and in this case, it is generally not necessary to perform different color detection according to images captured by two cameras which are disposed "opposite". In the case of particles such as thermoplastic granular resins, the proportion of impurities penetrating through both sides of the particles is not high, and the influence of such particles having structural characteristics on the detection result of the abnormal color is easily ignored.

Based on the above findings, the present specification provides a particulate matter detection method. As shown in fig. 1, the method comprises the steps of:

S10, acquiring a first color image of the particulate matters shot by the first linear array camera and a second color image of the particulate matters shot by the second linear array camera in the process of freely falling the particulate matters.

The period of time from when the first image is captured to when the second image is captured is one cycle.

The first linear array camera and the second linear array camera are oppositely arranged at two sides of the free falling path of the particulate matters.

The first linear array camera and the second linear array camera shoot the particles from two sides of a free falling path of the particles respectively, and the first linear array camera and the second linear array camera are arranged up and down in the direction of the free falling path of the particles. As shown in fig. 2, a first line camera a and a second line camera B are provided at both sides of the free falling path of the particulate matter for photographing the particulate matter from both sides of the free falling path of the particulate matter, respectively. In the direction of the free falling path of the particulate matter, the first linear camera A is positioned above the second linear camera B.

In some embodiments, the first and second line cameras are each configured with a diffuse light source. The first linear array camera and the second linear array camera shoot particles by diffuse reflection light sources.

The diffuse reflection light source is a light source that emits light based on the diffuse reflection principle. The diffuse reflection light source has the characteristic of uniform brightness. As shown in fig. 2, 6 and 7 are schematic diagrams of a first linear array camera and a second linear array camera, respectively, and 11 and 12 are schematic diagrams of two diffuse reflection light sources, each of which corresponds to one camera, respectively. The diffuse reflection light source and the camera are arranged at two sides of the falling path of the particulate matters. The diffuse reflection light source includes a light source and a diffuse reflection plate. The light source is positioned on one side of the diffuse reflection plate facing the particulate matters and irradiates towards the diffuse reflection plate. The diffuse reflection plate is a plate with uneven surface, light emitted by the light source is diffusely reflected on the surface of the diffuse reflection plate, and reflected light irradiates the particulate matters.

The images shot by the first linear array camera 6 and the second linear array camera 7 are used for detecting the heterochromatic particles, the colors of the particles are more concerned, and the cambered surfaces on the surfaces of the particles generally form specular reflection under the irradiation of parallel light, so that the heterochromatic detection result is disturbed. This scheme can be so through setting up diffuse reflection light source, can improve the inside light transmissivity of particulate matter on the one hand, is convenient for detect out the inside impurity colour of particulate matter, and on the other hand can make the particulate matter by even light irradiation, prevents that specular reflection from appearing, improves the accuracy that the heterochrosis detected.

S20, respectively carrying out different color detection on the first color images shot in each period of the first linear array camera, and respectively carrying out different color detection on the second color images shot in each period of the second linear array camera.

In the actual execution process, one thread may be used for detecting the different colors of the first color image shot in each period of the first line-scan camera, and another thread may be used for detecting the different colors of the second color image shot in each period of the second line-scan camera.

S30, performing a first operation and/or a second operation on each periodic image of the first line camera and the second line camera:

The first operation comprises the steps of judging whether different-color detection results of a second color image shot in the current period of the second linear array camera and different-color detection results of a first color image shot in the previous period of the first linear array camera all have different-color particles, comparing the different-color particles identified in the second color image shot in the current period of the second linear array camera with the different-color particles identified in the first color image shot in the previous period of the first linear array camera under the condition that the different-color particles are all the different-color particles, determining whether the characteristics of the different-color particles correspond or not, and discarding one of the different-color particle identification results in the corresponding different-color particles under the condition that the different-color particles corresponding to the characteristics are provided.

The second operation comprises the steps of judging whether different-color detection results of a second color image shot in the current period of the second linear array camera and different-color detection results of a first color image shot in the current period of the first linear array camera all have different-color particles, comparing the different-color particles identified in the second color image shot in the current period of the second linear array camera with the different-color particles identified in the first color image shot in the current period of the first linear array camera under the condition that the different-color particles are all the different-color particles, determining whether the characteristics of the different-color particles correspond or not, and discarding one of the different-color particle identification results of the corresponding different-color particles under the condition that the different-color particles corresponding to the characteristics are provided.

The first linear array camera and the second linear array camera are both linear array cameras, the shooting view field of the linear array cameras is in a slender strip shape, each shot frame of image can comprise a plurality of images of particles, and the linear array cameras can shoot the images of each particle and can not repeatedly shoot the images of the particles by adjusting the shooting period of the linear array cameras to be matched with the falling speed of the particles. That is, any particulate matter only appears in one of the images captured by the first line camera, but not in two or more of the images captured by the first line camera, and similarly, any particulate matter only appears in one of the images captured by the second line camera, but not in two or more of the images captured by the second line camera.

The lower the falling height of the particles from the same height is, the faster the speed is, and the greater the probability that the position relationship among the particles is disturbed in the free falling process is, so that the height difference between the first linear array camera and the second linear array camera on the free falling path of the particles is not suitable to be set larger.

However, since the first line camera and the second line camera have a height difference on the path along which the particles freely fall, a target particle in a frame of image captured by the second line camera in the current period may be in an image captured by the first line camera in the current period or may be in an image captured by the first line camera in the current period.

Therefore, after the different color detection is performed on the images shot by the first line camera and the second line camera in step S20, when the different color detection results corresponding to the two cameras are summarized in step S30, the detection result of the image shot by the second line camera in the current period and the detection result of the image shot by the first line camera in the current period are repeatedly detected, and the detection result of the image shot by the second line camera in the current period and the detection result of the image shot by the first line camera in the previous period are repeatedly detected.

According to the method, the first color image shot by the first linear array camera and the second color image shot by the second linear array camera are respectively detected in different colors, and the detection method can realize the rapid detection of different-color particles through two threads and improves the detection efficiency.

In this method of detecting the abnormal color, the correspondence between the particles in the first color image and the particles in the second color image is not recognized, that is, it is not determined which particle in the first color image corresponds to one particle in the second color image. Then, when the detection results of the different colors of the images photographed by the two cameras are repeatedly detected, it is impossible to determine whether or not the detection is repeated according to the correspondence relationship between the particulate matters.

In this way, the method determines whether the first color image and the second color image both detect the different color particles, and if one of the images does not detect the different color particles, the detection results of the two images are not repeated. In the event that both different colored particles are detected, it is determined whether the same different colored particles are present in both images.

When determining whether the first color image and the second color image have the same different-color particles, each different-color particle in the first color image and each different-color particle in the first color image can be compared, and whether the characteristics of the different-color particles correspond or not is determined in the comparison process. In the case where the characteristics of the different-color particles correspond, the different-color particles a in the first color image correspond to the characteristics of the different-color particles b in the second color image, which means that the different-color particles are repeatedly detected, and one different-color detection result may be discarded.

In some embodiments, only the number of different color particles is concerned, and then when one different color detection result is discarded, any one different color detection result may be discarded, for example, the detection result of the different color particle a in the first color image may be discarded, and the detection result of the different color particle b in the second color image may be discarded.

In some embodiments, the size of the impurity in the heterochromatic particles may be of interest in addition to the number of heterochromatic particles, e.g., the impurities are classified according to their size and the number of impurity particles of each class is counted. In this case, the recognition result of the different-color particles having a smaller spot area among the corresponding different-color particles can be discarded.

Since the first line camera and the second line camera are disposed opposite to each other on both sides of the falling path of the particulate matter and capture images of the particulate matter, when determining whether the characteristics of the heterochromatic particles correspond, it may be determined whether the positions of the spots in the heterochromatic particles are mirror-image correspondence. In addition, it is also possible to determine whether the contour shape of the heterochromatic particles is mirror-image-corresponding.

Since one particulate matter only appears in one frame of image captured by the first line camera, there is no need to perform the second operation in the case where the first operation has determined that there is a different color particle corresponding to the feature, and accordingly, there is no need to perform the first operation in the case where the second operation has determined that there is a different color particle corresponding to the feature. Thus, the steps required to be executed by repeated detection can be reduced, and the efficiency of repeated detection, namely the efficiency of heterochromatic detection, can be improved.

One specific embodiment of duplicate detection may be that, as shown in fig. 3, in the summary thread, data is periodically extracted from the recognition result cache obtained from the recognition thread, and the period time T is adjustable. The storage contents are respectively a batch_A_List and a batch_B_List, and the storage contents are identification image cutting images and corresponding speckle parameters.

Batch of the last cycle A List and batch_B_List is marked as Batch u A_List_Old A_List _old. Assuming that camera a is on top and camera B is on bottom as the first line camera, then theoretically the same particle would appear in batch_a_list_old and batch_b_list simultaneously.

The different color information in the Batch_B_List and the different color information in the Batch_A_List and the Batch_A_List_old are cross-compared, when the different color information with the same central position of the different color information is found, the sizes of the different color information and the different color information are compared, the information to be recorded is large in area, the other different color information is recorded as repeated information, and statistics are not recorded in the follow-up.

Note that the T value is generally set as the time difference between images taken by A, B cameras. The larger T is, the worse the effect of real-time display is, but the higher the success rate of repeatability detection is, and the smaller T is, the better the effect of real-time display is, but the lower the success rate of repeatability detection is.

In some embodiments, the target image is identified as being heterochromatic when detected. As shown in fig. 4, the heterochromatic recognition process includes:

and S21, converting the target image into an HLS color space.

In the process of identifying different colors, three channels of RGB (Red, green, blue) color space are all affected by brightness, and in the process of detecting different colors, the brightness is a great influence factor. The accuracy of the heterochromatic identification can be improved by converting the image into the HLS color space, so that the detection result of the particulate matter detection is improved.

S22, acquiring a chromaticity value interval, a brightness value interval and a saturation value interval of N color types of impurities in the particulate matters, wherein the N color types are ordered according to priorities, and the color type mark of the ith priority is set to 255-i, wherein i is a natural number.

S23, setting a first color type identifier for each pixel value in a chrominance channel sub-image of a target image according to a chrominance value interval of each color type, setting a second color type identifier for each pixel in a saturation channel sub-image of the target image according to a luminance value interval of each color type, setting a third color type identifier for each pixel in a saturation channel sub-image of the target image according to a saturation value interval of each color type, wherein the value ranges of the first color type identifier, the second color type identifier and the third color type identifier are integers in 0 and 255-N,255-i, and 0 represents the identifier of the color type of the ith priority outside the value interval of each color type and 255-i represents the identifier of the color type of the ith priority.

When the color type of the impurity is 1, the pixel in the value range of the parameter corresponding to the type is marked as 254, and the pixel outside the value range is marked as 0.

When the color type N >1 of the impurity, there are a plurality of pictures after the process of S23, the pixel marks within the effective area are from 255-1 to 255-N.

For example, the color type of the impurity is 3, if the value of the pixel is in the color type value range of the second priority, the pixel is marked 253, and if the value of the pixel is out of the value range, the pixel is marked 252, and if the value of the pixel is in the color type value range of the third priority, the pixel is marked 0.

One pixel may be provided with a plurality of color type identifications, which are finally fused through S24.

And S24, respectively executing the following operation for each pixel, wherein the maximum value of the first color type identifier, the second color type identifier and the third color type identifier corresponding to the current pixel is used as the final color type identifier of the current pixel.

For example, when the color type of a pixel is identified as 254, 0, 252, then 254 is identified as the final color type for that pixel, i.e., the pixel is considered to be of the 1 st priority color type.

After step S24, a classification map including classification information and position information of the impurity is obtained.

When the classification identifies the classification of the impurity in the graph, 255 may be directly subtracted from the maximum pixel value on the graph to obtain the classification of the impurity, as described in S26.

S25, judging whether the area of each region formed by the pixels with the same final color type identification is larger than or equal to a preset threshold value.

S26, taking a region with the area being greater than or equal to a preset threshold value as a detected spot region, and taking a final color type mark of each pixel in the region or a difference value between 255 and the final color type mark as a color mark of the spot region.

The target image is the first color image or the second color image. The speckle region in the image is the impurity.

For example, in some embodiments, before performing step S23, the validity of each pixel may be verified, e.g., for each of the N color types, a chroma flag may be set for each pixel according to a chroma value interval corresponding to the current color type, a luma flag may be set for each pixel according to a luma value interval corresponding to the current color type, and a saturation flag may be set for each pixel according to a saturation value interval corresponding to the current color type. Wherein the value of the chroma mark, the brightness mark and the saturation mark is 0 or 1, wherein 1 indicates that the corresponding parameters (chroma, brightness and saturation) fall within the value range, and 0 indicates that the parameters fall outside the value range.

Then, the values of three channels (chromaticity, luminance, saturation) of each pixel are and-operated. That is, a pixel is valid when its three HLS channels simultaneously satisfy the range of values of the color type.

In some embodiments, the particulate matter is thermoplastic particulate resin particles and the predetermined particle type (i.e., color type) is at least one of black particles, black specks, colored particles.

FIG. 3 is a schematic diagram of one embodiment of detection of a different color. The term "start scanning" means that the first line camera starts to capture a first color image and the second line camera starts to capture a second color image. The left branch in fig. 3 represents an image acquisition thread. In the image acquisition thread, a first linear camera A periodically acquires images to obtain a first color image and is marked as A to represent the images shot by the first linear camera A, and a second linear camera B periodically acquires images to obtain a second color image and is marked as B to represent the images shot by the second linear camera B. The first color image and the second color image are stored in an image buffer. The middle branch in fig. 3 represents the identified thread. In the identification thread, a first color Image image_A and a second color Image image_B are acquired from an Image buffer, then the acquired images are converted from an RGB color space to an HLS color space to obtain images image_A_HLS and image_B_HLS, pixel values in the images are converted into color type identifications (namely, the step S23) to obtain images image_A_mark and image_B_mark, and binarization operation (namely, the step S24) is further carried out to obtain images image_A_Binry and image_B_Binry, and spot detection is carried out according to the images image_A_Binry and image_B_Binry to obtain spot information Pos_List_A in the second Image and spot information Pos_List_B in the third Image.

In some embodiments, the heterochromatic recognition process further comprises clipping the target image according to the positions of the spot areas to obtain sub-images, wherein each sub-image corresponds to one spot in the particulate matter, and each sub-image comprises a complete spot image and position information of the spot in the particulate matter.

As shown in fig. 3, after determining that a spot exists in the first color Image, a target clipping ratio may be selected from predetermined clipping ratios, and clipping the Image image_a_mark based on pos_list_a to obtain an Image small_image_list_a, and clipping the Image image_a_binary based on pos_list_a to obtain an Image small_image_list_a_mark. Determining particle classification results and spot parameters of particles according to the Image Small_image_List_A and the Image Small_image_List_A_Mark, taking the particle classification results and the spot parameters as labeling information of images, storing the labeling information of the images in a buffer in association with the Image Small_image_List_A and the Image Small_image_List_A_Mark obtained by clipping, and storing the labeling information in a database after the whole detection process is finished.

Similarly, after determining that a blob exists in the second color Image, a target cropping ratio may be selected from the predetermined cropping ratios, and the Image image_b_mark may be cropped based on pos_list_b to obtain an Image small_image_list_b, and the Image image_b_binary may be cropped based on pos_list_b to obtain an Image small_image_list_b_mark. Determining particle classification results and spot parameters of particles according to the Image Small_image_List_B and the Image Small_image_List_B_Mark, taking the particle classification results and the spot parameters as labeling information of images, storing the labeling information of the images in a buffer in association with the cut Image Small_image_List_B and the cut Image Small_image_List_B_Mark, and storing the labeling information and the labeling information in a database after the whole detection process is finished.

When a user checks the detection result, the spot image of the particulate matters and the spot parameters can be seen through the cut image and the label information thereof, and the statistical information of the detection of different colors is checked.

The first color image and the second color image are quite large compared with the particulate matters, and the images of the particulate matters cut out from the first color image can remove useless information, reduce the stored data volume and facilitate quick finding of the images of spots when the images of the particulate matters are inquired.

The speckle parameters may include at least one of area of heterochromatic, heterochromatic equivalent diameter, and the like.

After the profile detection is performed on all the images of the first line camera and the second line camera, a summary thread shown by the right branch in fig. 3 is executed. In the summarizing thread, each image and the label information thereof are read from the buffer, and then whether the second image and the third image are repeated or not is detected, namely whether the spots in the clipping images corresponding to the second image and the third image are spots in the same particulate matter or not is detected. If so, it means that only one piece of heterochromatic particle statistics should be counted, and therefore one of the spot detection results should be discarded, and if not, both spot detection results should be counted, and therefore stored in a database, and the statistics stored in a statistics cache. After the whole detection process is finished, the content stored in the buffer memory in the identification thread and the summarization thread is stored in the database, and all the stored information is related to the identification of one detection test.

The statistical information in fig. 3 may refer to the number of particles belonging to each color type, etc.

According to the particle detection method, images of freely falling particles are shot by two cameras which are arranged oppositely, different-color detection is carried out based on the images shot by the two cameras, when some impurities are found to penetrate through the particles in the different-color detection process, the impurity particles are easily caused to be repeatedly counted, therefore, the repeated detection is carried out on different-color detection results of the images shot by the two cameras, in the repeated detection process, the images shot by the lower camera in the current period are respectively compared with the images shot by the upper camera in the current period and the images in the previous period, the condition that the same particle possibly appears in the images shot by the lower camera in the current period and the images shot by the upper camera in the previous period is considered, the calculation result is more accurate, the images shot by the two cameras are respectively identified in different colors, the two cameras are not interfered with each other, the detection efficiency is improved, when the different-color detection results of the images shot by the two cameras are summarized, the same different-color particles are determined according to the characteristics of the different-color particles, the different-color particles are not the position relationship between the two images is determined, and the different-color particles are not the different-color particles are accurately detected, and the different-color particle detection results are not caused.

In some embodiments, in addition to the off-color detection, as shown in fig. 5, the particulate matter detection method further includes performing the off-shape detection by the following method.

S41, acquiring an image shot by a third linear array camera in the process of freely falling particulate matters.

And S42, recognizing the outline of the particulate matters from the image shot by the third linear array camera.

S43, calculating the shape characteristic parameters of the particulate matters according to the outlines of the particulate matters.

S44, determining whether the particulate matter is a special-shaped particle according to the shape characteristic parameters of the particulate matter.

In some embodiments, the shape characteristic parameter may be obtained by performing a convolution calculation with a predetermined convolution kernel on the first image. The shape characteristic parameters obtained by convolution calculation are generally difficult to correlate with the actual physical meaning of the particulate matter in the first image. Accordingly, the shape characteristic parameters obtained by convolution calculation can be input into a pre-trained artificial intelligent neural network model (such as a fully connected layer), and whether the particulate matter is special-shaped particles is determined through the artificial intelligent neural network model.

In other embodiments, the shape characteristic parameter comprises at least one of particle area, particle circumference, maximum inscribed circle, minimum circumscribed rectangle, convex hull area, particle size, sieving diameter, surface roughness, roundness, elongation, shape factor, convexity. These shape feature parameters have practical physical significance.

In the case of using these shape feature parameters having actual physical significance, the correspondence between these shape feature parameters and the types of particulate matters may be predetermined in advance according to expert experience, and step S43 may determine the types to which the particulate matters belong according to these correspondence. In the case of using these shape feature parameters, an artificial intelligent neural network model may be obtained by training using a deep learning algorithm, and step S43 may input these shape feature parameters into the artificial intelligent neural network model, and determine whether the particulate matter is a special-shaped particulate matter by using the artificial intelligent neural network model.

Fig. 6 is a schematic diagram of a minimum circumscribed circle and a maximum inscribed circle, wherein a circle indicated by a is the minimum circumscribed circle and a circle indicated by B is the maximum inscribed circle.

Fig. 7 is a schematic diagram of a minimum bounding rectangle, wherein the rectangular box indicated by C is the minimum bounding rectangle.

Fig. 8 is a schematic diagram of convexity calculation basis.

FIG. 9 is a schematic diagram of roughness calculation basis.

In the case of adopting the shape characteristic parameter having the meaning of the actual material, in order to calculate the shape characteristic parameter more accurately, step S42 may perform image enhancement processing on the image captured by the third linear camera, then perform binarization processing, and determine the contour of the particle according to the binarization processing result.

FIG. 10 is a schematic diagram of one embodiment of profile detection. Here, "start scanning" means that the third line camera starts shooting an image. The left branch in fig. 10 represents an image acquisition thread. In the Image acquisition thread, the third linear camera periodically acquires images to obtain images image_C shot by the third linear camera, and stores the images into an Image buffer. The middle branch in fig. 10 represents the identified thread. In the identification thread, an Image image_C is acquired from an Image buffer, contrast enhancement is performed first, binarization processing is performed to obtain an Image image_C_Binary, and then the outline of the particle is extracted according to the Image image_C_Binary. And calculating shape characteristic parameters according to the outline of the particles, and determining the category of the particles according to the calculated shape characteristic parameters.

In some embodiments, step S44 may first obtain a predetermined particle classification model, input the shape feature parameter into the particle classification model, obtain a particle classification result output by the particle classification model, and determine that the particle is a special-shaped particle if the output result of the particle classification model belongs to a predetermined particle type.

In the case where the particulate matter is thermoplastic particulate resin particles, the predetermined particle classification type (i.e., shape type) may include at least one of screen shot of less than 2nm in size, screen shot of more than 5nm in size, snake skin shot, tailing shot, flocculent shot.

The particle classification model may be the artificial intelligent neural network model, or may be an expert system based on rules and cases. Expert Systems (ESs) are computer software systems that can solve complex problems like human experts in a particular field. The method can effectively use the experience and expertise accumulated by the expert for many years, and solve the problem that the expert is required to solve by simulating the thinking process of the expert. The expert system needs to store expert knowledge in a knowledge base through a certain knowledge acquisition method, and then works by combining a man-machine interaction interface through an inference engine.

In some embodiments, after the profile detection in step S44, steps S45 and S46 are further included.

And S45, cutting out all sub-images from the image shot by the third linear array camera according to the outline of the identified particulate matter, wherein each sub-image comprises a complete image of the particulate matter.

And S46, storing each sub-graph in association with a mark which indicates whether the particulate matters in the sub-graph are special-shaped particles.

As shown in fig. 10, after the outline of the particulate matter is extracted, a target clipping ratio may be selected from predetermined clipping ratios, and the image of the particulate matter may be clipped from the image as a sub-image. Compared with the particle, the image shot by the third linear array scanning camera is quite large, the image of the particle is cut out from the image shot by the third linear array scanning camera, so that useless information can be removed, the stored data volume is reduced, and the characteristic of the particle can be found out quickly when the image of the particle is inquired.

After the shape characteristic parameters of the particulate matter are calculated according to the outline characteristics of the particulate matter, the shape type of the particulate matter can be determined in addition to determining whether the particulate matter is a special-shaped particle. When the clipping image is stored in the buffer, whether the clipping image is a special-shaped particle, the type to which the particle belongs, and various shape characteristic parameters (such as particle size and the like) of the particle can be stored in the buffer in a correlated manner as the label of the clipping image, and the label is stored in the database after the whole detection process is finished.

After the entire detection process is finished, the summary thread shown in the right branch in fig. 10 is executed. In the summarizing thread, each image and its tag information are read from the buffer, statistical information (for example, the number of particles in each value range of a certain shape characteristic parameter is counted, and the number of particles in each shape class is counted) is calculated, and the statistical information is stored in the buffer. After the whole detection process is finished, the content stored in the buffer memory in the identification thread and the summarization thread is stored in the database, and all the stored information is related to the identification of one detection test.

When a user checks the detection result, the appearance image and each shape characteristic parameter of the particulate matters can be seen through the clipping image and the label information thereof, and each piece of statistical information is checked.

In some embodiments, after the special-shaped detection and the abnormal-color detection process of all the particulate matters are finished, at least one of the following statistical information, namely a special-shaped detection result and an abnormal-color detection result, is obtained, wherein the at least one index of the target number of the particulate matters is counted, namely a abnormal-color thousandth rate, an abnormal-color kilogram rate and an abnormal-color kilogram rate.

As shown in fig. 9, after the abnormal detection and the abnormal detection of all the particles are finished, the current weight (i.e. the total weight of the detected particles) can be obtained by the weight sensor, the data are obtained from the abnormal data cache and the abnormal data cache respectively, and the statistical information is calculated comprehensively according to the data.

The invention also provides a particle detection system which can be used for executing the particle detection method. The system comprises an image acquisition unit, a different-color identification unit and a summarizing unit.

The image acquisition unit is used for acquiring a first color image of particles shot by a first linear array camera and a second color image of the particles shot by a second linear array camera in the free falling process of the particles, the first linear array camera and the second linear array camera respectively shoot the particles from two sides of a free falling path of the particles, the first linear array camera and the second linear array camera are arranged up and down in the direction of the free falling path of the particles, the first linear array camera and the second linear array camera are arranged opposite to each other on two sides of the free falling path of the particles, and a time period from shooting of the first image to shooting of the second image is a period.

The different color identification unit is used for respectively carrying out different color detection on the first color images shot in each period of the first linear array camera and respectively carrying out different color detection on the second color images shot in each period of the second linear array camera.

The summarizing unit comprises a processing unit for performing a first operation and/or a second operation on each periodic image of the first and second line cameras.

In some embodiments, the image acquisition module further comprises a camera for capturing an image of the particulate matter captured by the third linear array camera during free fall of the particulate matter. Correspondingly, the system further comprises a special-shaped identification unit. The special-shaped recognition unit is used for carrying out special-shaped detection by acquiring an image shot by a third linear camera in the free falling process of the particles, recognizing the outline of the particles from the image shot by the third linear camera, calculating the shape characteristic parameters of the particles according to the outline of the particles, and determining whether the particles are special-shaped particles according to the shape characteristic parameters of the particles.

Correspondingly, the summarizing unit further comprises counting according to the identification results of the heterochromatic identification unit and the heterochromatic identification unit to obtain heterochromatic detection results of the particulate matters.

The specification provides a particulate matter detection system, as shown in fig. 2, comprising a first hopper 1, a first conveying mechanism 2, a second hopper 3, a second conveying mechanism 4, a material collector 5, a first line camera 6, a second line camera 7 and an electronic device 8.

The first hopper 1 is used for containing the particulate matter to be detected.

The first conveying mechanism 2 is located below the first hopper 1, and is used for conveying the particles falling in the first hopper 1 to the position above the second hopper 3, and enabling the particles to freely fall into the second hopper 3 above the second hopper 3.

The particulate matter in the present specification may include thermoplastic particulate resins such as Polyethylene (PE), polypropylene (PP), polystyrene (PS), impact polystyrene (PS-I), acrylonitrile-butadiene-styrene (ABS), styrene-acrylonitrile (SAN), ethylene-vinyl acetate (EVAC), etc., or modified particulate plastics thereof.

The first conveyor 2 may be a vibrating table or a motor driven conveyor.

The conveying surface (i.e., the surface on which the particulate matter is placed) in the first conveying mechanism 2 may be a flat surface or an inclined surface. In the case of a planar conveying surface, the planar movement is driven by the motor, so that the particles move from below the first hopper 1 to above the second hopper 3. In the case where the conveying surface is a slope, the particulate matter may be moved above the second hopper 3 under the support and guide of the slope.

"Free fall" in this specification means falling without any object support.

The second hopper 3 is disposed below the first conveying mechanism 2 and is used for containing particles falling from the first conveying mechanism 2. The second hopper 3 and the first hopper 1 are arranged in a staggered manner in the horizontal direction.

The second conveying mechanism 4 is located below the second hopper 3 and is used for conveying the particles falling in the second hopper 3 to the position above the material collector 5 and enabling the particles to freely fall into the material collector 5 above the material collector 5.

The second conveyor 4 may be a vibrating table or a motor driven conveyor.

The conveying surface (i.e., the surface on which the particulate matter is placed) in the second conveying mechanism 4 may be a flat surface or an inclined surface. In the case of a planar conveying surface, the planar movement is driven by the motor so that the particles move from below the second hopper 3 to above the material collector 5. In the case of a ramp, the particles can be moved above the material collector 5 under the support and guidance of the ramp.

"Free fall" in this specification means falling without any object support.

The material collector 5 is arranged below the second conveyor 4 for receiving particulate matter falling from the second conveyor 4. A first line camera 6 is arranged between the first conveyor 2 and the second hopper 3 for taking a first image of the particulate matter falling freely from the first conveyor 2.

A second linear camera 7 is arranged between the second conveyor 4 and the material collector 5 for taking a second image of the particles falling freely from the second conveyor 4.

The electronic device 8 is communicatively connected to the first line camera 6 and the second line camera 7.

In some embodiments, the first and second line cameras 6, 7 are each configured with a diffuse reflective light source. The first linear array camera 6 and the second linear array camera 7 shoot the particles by diffuse reflection light sources.

The diffuse reflection light source is a light source that emits light based on the diffuse reflection principle. The diffuse reflection light source has the characteristic of uniform brightness. As shown in fig. 2, 11 and 12 are schematic diagrams of two diffuse reflection light sources, one for each camera, respectively. The diffuse reflection light source and the camera are arranged at two sides of the falling path of the particulate matters. The diffuse reflection light source includes a light source and a diffuse reflection plate. The light source is positioned on one side of the diffuse reflection plate facing the particulate matters and irradiates towards the diffuse reflection plate. The diffuse reflection plate is a plate with uneven surface, light emitted by the light source is diffusely reflected on the surface of the diffuse reflection plate, and reflected light irradiates the particulate matters.

In some embodiments, the particulate matter detection device further comprises a third linear camera 9. The third line camera 9 is used to take images of the free falling particulate matter. The image taken by the third line camera 9 is used for profile detection.

In the case where the first line camera 6 and the second line camera 7 are both provided between the first conveyor 2 and the second hopper 3, the third line camera 9 is provided between the second conveyor 4 and the material collector 5. In the case where the first line camera 6 and the second line camera 7 are both provided between the second conveying mechanism 4 and the material collector 5, the third line camera 9 is provided between the first conveying mechanism 2 and the second hopper 3.

In some embodiments, as shown in FIG. 2, the particulate matter detection system further includes a backlight source 10. The backlight source 10 and the first camera 6 may be disposed at both sides of the particulate matter falling path such that the parallel light of the backlight source 10 is irradiated toward the first camera 6.

The image shot by the third linear array camera 9 is used for detecting special-shaped particles, the shape of the particles is focused more, and the outline characteristics of the particles can be highlighted by shooting the image particles by adopting the backlight of the third linear array camera 9, so that the accuracy of special-shaped detection is improved.

In some embodiments, as shown in FIG. 2, the particulate matter detection system further includes a load cell 13 disposed below the material collector 5 for detecting the total weight of particulate matter collected in the material collector 5.

The descriptions and functions of the above devices and systems may be understood by referring to the content of the particulate matter detection method, and will not be described in detail.

The present invention also provides an electronic device, as shown in fig. 11, which may include a processor 1101 and a memory 1102, where the processor 1101 and the memory 1102 may be connected by a bus or other means, and in fig. 11, the connection is exemplified by a bus.

The processor 1101 may be a central processing unit (Central Processing Unit, CPU). The Processor 1101 may also be other general purpose processors, digital signal processors (DIGITAL SIGNAL processors, DSPs), application SPECIFIC INTEGRATED Circuits (ASICs), field-Programmable gate arrays (Field-Programmable GATE ARRAY, FPGA) or other Programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or combinations thereof.

The memory 1102 is used as a non-transitory computer readable storage medium for storing non-transitory software programs, non-transitory computer executable programs, and modules, such as program instructions/modules (e.g., an image acquisition unit, a different color identification unit, and a summarization unit) corresponding to the particulate matter detection method in the embodiment of the present invention. The processor 1101 executes various functional applications of the processor and data processing, i.e., implements the particulate matter detection method in the above-described method embodiments, by running non-transitory software programs, instructions, and modules stored in the memory 1102.

The memory 1102 may include a storage program area that may store an operating system, application programs required for at least one function, and a storage data area that may store data created by the processor 1101, etc. In addition, memory 1102 may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, memory 1102 optionally includes memory remotely located relative to processor 1101, which may be connected to processor 1101 via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The one or more modules are stored in the memory 1102, which when executed by the processor 1101, perform the particulate matter detection method described previously.

The specific details of the electronic device may be correspondingly understood by referring to the corresponding related descriptions and effects in the method embodiment, which are not repeated herein.

The present specification also provides a computer storage medium storing computer program instructions that, when executed, implement the steps of the particulate matter detection method described above.

The present specification also provides a computer program product comprising a computer program which, when executed by a processor, performs the steps of the particulate matter detection method described above.

It will be appreciated by those skilled in the art that implementing all or part of the above-described embodiment method may be implemented by a computer program to instruct related hardware, where the program may be stored in a computer readable storage medium, and the program may include the above-described embodiment method when executed. The storage medium may be a magnetic disk, an optical disc, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a Flash Memory (Flash Memory), a hard disk (HARD DISK DRIVE, abbreviated as HDD), a Solid state disk (Solid-state disk STATE DRIVE, SSD), or the like, and may further include a combination of the above types of memories.

In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are referred to each other, and each embodiment is mainly described as different from other embodiments.

The system, apparatus, module or unit set forth in the above embodiments may be implemented in particular by a computer chip or entity, or by a product having a certain function.

For convenience of description, the above devices are described as being functionally divided into various units, respectively. Of course, the functions of each element may be implemented in the same piece or pieces of software and/or hardware when implementing the present application.

From the above description of embodiments, it will be apparent to those skilled in the art that the present application may be implemented in software plus a necessary general hardware platform. Based on such understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a storage medium, such as a ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method of some parts of the various embodiments of the present application.

The application is operational with numerous general purpose or special purpose computer system environments or configurations. Such as a personal computer, a server computer, a hand-held or portable device, a tablet device, a multiprocessor system, a microprocessor-based system, a set top box, a programmable consumer electronics, a network PC, a minicomputer, a mainframe computer, a distributed computing environment that includes any of the above systems or devices, and the like.

The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.

The inventors carried out heterochromatic profile detection on the same first stack of 200g PP particles by visual inspection and the method provided by the invention, respectively. The visual inspection method was carried out for 10 minutes, and the number of each type of particles identified is shown in the following table one.

List one

When the particle detection method provided by the invention is adopted, the detection process is shared for 3 minutes and 45 seconds. The following table II is the comprehensive statistical result of the detection result, the following table III is the special-shaped classification result, the following table IV is the particle size classification result, and the following table V is the heterochromatic classification result.

Watch II

Watch III

Classification category	All of which	Percentage of
			Unknown classification	0.000	0.000
Large particles	20.000	0.256
			Small particles	730.000	9.346
Normal state	6951.000	88.990
			Trailing tail	43.000	0.551
Floc of floc	67.000	0.858
			Data summarization	7811.000	100.00

Table four

Size [ mu m ]	All of which	Percentage of
			[500,1000)	7.000	0.090
[1000,2000)	138.000	1.767
			[2000,2500)	330.000	4.225
[2500,3000)	326.000	4.174
			[3000,3500)	576.000	7.374
[3500,4000)	3329.000	42.619
			[4000,4500)	2490.000	31.878
[4500,5000)	592.000	7.579
			[5000,5500)	22.000	0.282
[5500,+∞)	1.000	0.013
			∑	7811.000	100.00

TABLE five

Size [ mu m ]	All [ number ]	Color particles [ number ]	Black spot particle [ number ]
				[0,50)	192(82%)	0	192
[50,100)	24(10%)	1	23
				[100,150)	9(3%)	1	8
[150,200)	0(0%)	0	0
				[200,250)	3(0%)	0	3
[250,300)	1(0%)	0	1
				[300,350)	1(0%)	0	1
[350,400)	1(0%)	0	1

The inventors carried out heterochromatic profile detection on the same second stack of 200g PP particles by visual inspection and the method provided by the present invention, respectively. The visual inspection method was carried out for 10 minutes, and the number of each type of particles identified is shown in the following table six.

TABLE six

When the particle detection method provided by the invention is adopted, the detection process is shared for 3 minutes and 58 seconds. The following table seven is the comprehensive statistical result of the detection result, the following table eight is the special-shaped classification result, the following table nine is the particle size classification result, and the following table ten is the heterochromatic classification result.

Watch seven

Table eight

Classification category	All of which	Percentage of
			Unknown classification	0.000	0.000
Large particles	22.000	0.283
			Small particles	722.000	9.275
Normal state	6930.000	89.029
			Trailing tail	40.000	0.514
Floc of floc	70.000	0.899
			Data summarization	7784.000	100.00

Table nine

Size [ mu m ]	All of which	Percentage of
			[500,1000)	4.000	0.051
[1000,2000)	134.000	1.721
			[2000,2500)	322.000	4.137
[2500,3000)	330.000	4.239
			[3000,3500)	541.000	6.950
[3500,4000)	3336.000	42.857
			[4000,4500)	2519.000	32.361
[4500,5000)	571.000	7.336
			[5000,5500)	25.000	0.321
[5500,+∞)	2.000	0.026
			∑	7784.000	100.00

Ten meters

Size [ mu m ]	All [ number ]	Color particles [ number ]	Black spot particle [ number ]
				[0,50)	225(82%)	5	220
[50,100)	28(10%)	0	28
				[100,150)	12(4%)	2	10
[150,200)	3(1%)	0	3
				[200,250)	1(0%)	1	0
[250,300)	1(0%)	0	1
				[300,350)	1(0%)	0	1
[350,400)	0(0%)	0	0

The inventors performed heterochromatic profile detection on the same third stack of 200g PE particles using a visual inspection method and the method provided by the present invention, respectively. The number of each type of particle identified is shown in table eleven below, 10 minutes for the visual inspection.

Table eleven

When the particle detection method provided by the invention is adopted, the detection process is shared for 6 minutes and 57 seconds. The following table twelve is the comprehensive statistical result of the detection result, the following table thirteen is the special-shaped classification result, the following table fourteen is the particle size classification result, and the following table fifteen is the heterochromatic classification result.

Twelve watches

Watch thirteen

Classification category	All of which	Percentage of
			Unknown classification	0.000	0.000
Large particles	1.000	0.014
			Small particles	54.000	0.760
Normal state	7004.000	98.620
			Trailing tail	32.000	0.451
Floc of floc	1.000	0.155
			Data summarization	7102.000	100.00

Fourteen watch

Size [ mu m ]	All of which	Percentage of
			[500,1000)	1.000	0.014
[1000,2000)	15.000	0.211
			[2000,2500)	24.000	0.388
[2500,3000)	29.000	0.408
			[3000,3500)	158.000	2.225
[3500,4000)	2189.000	30.822
			[4000,4500)	3891.000	54.787
[4500,5000)	789.000	11.110
			[5000,5500)	3.000	0.042
[5500,+∞)	3.000	0.042
			∑	7102.000	100.00

Table fifteen

Size [ mu m ]	All [ number ]	Color particles [ number ]	Black spot particle [ number ]
				[0,50)	312(81%)	3	309
[50,100)	51(13%)	3	48
				[100,150)	14(3%)	0	14
[150,200)	4(1%)	2	2
				[200,250)	3(0%)	1	2
[250,300)	0(0%)	0	0
				[300,350)	0(0%)	0	0
[350,400)	0(0%)	0	0

The inventors performed heterochromatic profile detection on the same fourth stack of 200g PE particles using a visual inspection method and the method provided by the present invention, respectively. The visual inspection method was carried out for 10 minutes, and the number of each type of particles identified is shown in sixteen tables.

Sixteen watch

When the particle detection method provided by the invention is adopted (the linear array camera light source for detecting the heterochromatic light adopts the diffuse reflection light source), the detection process is shared for 3 minutes and 12 seconds. Seventeen tables below are comprehensive statistical results of the detection results, eighteen tables below are special-shaped classification results, nineteen tables below are particle size classification results, and twenty tables below are heterochromatic classification results.

Seventeen of the table

Watch eighteen

Classification category	All of which	Percentage of
			Unknown classification	0.000	0.000
Large particles	6.000	0.085
			Small particles	48.000	0.677
Normal state	6989.000	98.534
			Trailing tail	24.000	0.338
Floc of floc	26.000	0.367
			Data summarization	7093.000	100.00

Nineteen table

Size [ mu m ]	All of which	Percentage of
			[500,1000)	0.000	0.000
[1000,2000)	25.000	0.352
			[2000,2500)	17.000	0.240
[2500,3000)	25.000	0.352
			[3000,3500)	134.000	1.899
[3500,4000)	2199.000	31.002
			[4000,4500)	3870.000	54.561
[4500,5000)	810.000	11.420
			[5000,5500)	5.000	0.070
[5500,+∞)	8.000	0.113
			∑	7093.000	100.00

Watch twenty

Size [ mu m ]	All [ number ]	Color particles [ number ]	Black spot particle [ number ]
				[0,50)	281(84%)	3	278
[50,100)	34(10%)	2	32
				[100,150)	10(3%)	1	9
[150,200)	3(0%)	0	3
				[200,250)	1(0%)	0	1
[250,300)	1(0%)	1	0
				[300,350)	1(0%)	0	1
[350,400)	1(0%)	1	0

According to the four comparison examples, 1, the particle detection method provided by the invention is more accurate in detection results than the detection results of a visual inspection method for the detected number of various types of particles, can avoid numerical deviation caused by human factors, is high in repeatability, 2, the particle detection method provided by the invention is higher in detection efficiency by one to two times than the visual inspection method, can more rapidly detect and analyze results, and 3, the particle detection method provided by the invention can be used for counting the number of various types of particles, counting the size distribution of special-shaped particles, the average weight of particles, the average diameter of particles, the data such as thousandth special-shaped different rates and kilogram different rates, is more comprehensive in data analysis, can completely embody the shape and quality of products, can guide production operation, and can timely find problems to regulate and control.

While the present application has been described by way of embodiments, those of ordinary skill in the art will recognize that there are many variations and modifications of the present application without departing from the spirit of the application, and it is intended that the appended claims encompass such variations and modifications as do not depart from the spirit of the application.

Claims

1. A particle detection method, comprising:

Acquire a first color image of the particle taken by the first line array camera and a second color image of the particle taken by the second line array camera during the free fall of the particle; the first line array camera and the second line array camera respectively photograph the particle from both sides of the free fall path of the particle, and the first line array camera and the second line array camera are arranged up and down in the direction of the free fall path of the particle; the first line array camera and the second line array camera are arranged opposite to each other on both sides of the free fall path of the particle; the time period from when the line array camera takes the first image to before taking the second image is one cycle;

Performing heterochromatic detection on the first color images taken by the first line array camera in each cycle, and performing heterochromatic detection on the second color images taken by the second line array camera in each cycle;

Performing a first operation and/or a second operation on each periodic image of the first line scan camera and the second line scan camera:

The first operation includes: determining whether the heterochromatic detection result of the second color image captured by the second line array camera in the current cycle and the heterochromatic detection result of the first color image captured by the first line array camera in the previous cycle both have heterochromatic particles; if both have heterochromatic particles, comparing the heterochromatic particles identified in the second color image captured by the second line array camera in the current cycle with the heterochromatic particles identified in the first color image captured by the first line array camera in the previous cycle to determine whether the features of the heterochromatic particles correspond; if there are heterochromatic particles with corresponding features, discarding the recognition result of one heterochromatic particle among the corresponding heterochromatic particles;

The second operation includes: determining whether both the heterochromatic detection result of the second color image shot in the current cycle by the second line array camera and the heterochromatic detection result of the first color image shot in the current cycle by the first line array camera have heterochromatic particles; if both have heterochromatic particles, comparing the heterochromatic particles identified in the second color image shot in the current cycle by the second line array camera with the heterochromatic particles identified in the first color image shot in the current cycle by the first line array camera to determine whether the features of the heterochromatic particles correspond; if there are heterochromatic particles with corresponding features, discarding the recognition result of one of the corresponding heterochromatic particles.

2. The method according to claim 1 is characterized in that determining whether the characteristics of the heterochromatic particles correspond includes: whether the positions of the spots in the heterochromatic particles are mirror-corresponding.

3. The method according to claim 1 is characterized in that determining whether the characteristics of the different-colored particles correspond includes: the outer contour shapes of the different-colored particles correspond in a mirror image.

4. The method according to claim 1, characterized in that, when the first operation is performed first, if the first operation determines that there are heterochromatic particles corresponding to the characteristics, the second operation is no longer performed;

In the case where the second operation is performed first, if the second operation determines that there are heterochromatic particles corresponding to the features, the first operation is no longer performed.

5. The method according to claim 1 is characterized in that discarding the recognition result of one of the corresponding heterochromatic particles comprises: discarding the recognition result of the heterochromatic particle with a smaller spot area among the corresponding heterochromatic particles.

6. The method according to claim 1, characterized in that the first color image and the second color image are taken when a diffuse reflection light source illuminates the freely falling particles.

7. The method according to claim 1 is characterized in that, when detecting heterochromaticity, heterochromaticity recognition is performed on the target image; the heterochromaticity recognition process comprises:

Convert the target image to HLS color space;

Obtaining chromaticity value intervals, brightness value intervals, and saturation value intervals of N color types of impurities in particulate matter; the N color types are sorted according to priority, and the color type identifier of the i-th priority is set to 255-i, where i and N are natural numbers;

According to the chromaticity value interval of each color type, a first color type identifier is set for each pixel value in the chromaticity channel sub-image of the target image; according to the brightness value interval of each color type, a second color type identifier is set for each pixel in the saturation channel sub-image of the target image; according to the saturation value interval of each color type, a third color type identifier is set for each pixel in the saturation channel sub-image of the target image; the value ranges of the first color type identifier, the second color type identifier, and the third color type identifier are: 0, integers in [255-N, 255-i]; 0 indicates that it is outside the value interval of each color type, and 255-i indicates the identifier of the color type of the i-th priority;

For each pixel, the following operations are performed respectively: the maximum value among the first color type identifier, the second color type identifier, and the third color type identifier corresponding to the current pixel is used as the final color type identifier of the current pixel;

Determining whether the areas of the respective regions formed by the pixels with the same final color type identifier are greater than or equal to a predetermined threshold;

A region whose area is greater than or equal to a predetermined threshold is regarded as a detected spot region, and a final color type identifier of each pixel in the region or a difference between 255 and the final color type identifier is regarded as a color mark of the spot region;

The target image is the first color image or the second color image.

8. The method according to claim 1, characterized in that the target image is subjected to heterochromatic recognition during heterochromatic detection: the heterochromatic recognition process further comprises:

The target image is cropped according to the position of each spot area to obtain each sub-image; each sub-image corresponds to a spot in the particle; each sub-image includes a complete spot image and the position information of the spot in the particle.

9. The method according to claim 1, characterized in that, in addition to heterochromatic detection, the method further comprises performing heteromorphic detection by the following method:

Acquire images taken by the third linear array camera during the free fall of particles;

identifying the outline of the particle from the image taken by the third line array camera;

Calculating shape characteristic parameters of the particle according to the contour of the particle;

Whether the particle is a special-shaped particle is determined according to the shape characteristic parameters of the particle.

10. The method according to claim 9 is characterized in that the shape characteristic parameters include at least one of the following: particle area, particle circumference, maximum inscribed circle, minimum circumscribed circle, minimum circumscribed rectangle, convex hull area, particle size, sieve diameter, surface roughness, roundness, elongation, shape factor, convexity.

11. The method according to claim 9, characterized in that the method further comprises:

obtaining a predetermined particle classification model;

Inputting the shape characteristic parameters of the particle into the particle classification model to obtain the particle classification result output by the particle classification model;

In the case that the output result of the particle classification model belongs to a predetermined particle type, the particle is determined to be a special-shaped particle.

12. The method according to claim 9, characterized in that the method further comprises:

According to the outline of the identified particle, cutting out sub-images from the image captured by the third line array camera, each sub-image including a complete image of a particle;

Each sub-graphic is stored in association with a flag indicating whether the particle in the sub-graphic is a special-shaped particle.

13. A particle detection system, comprising:

An image acquisition unit, comprising a first color image of particles photographed by a first line array camera and a second color image of particles photographed by a second line array camera during the free fall of particles; the first line array camera and the second line array camera respectively photograph particles from both sides of the free fall path of particles, and the first line array camera and the second line array camera are arranged up and down in the direction of the free fall path of particles; the first line array camera and the second line array camera are arranged opposite to each other on both sides of the free fall path of particles; the time period from when the line array camera photographs the first image to before photographing the second image is one cycle;

a heterochromatic color recognition unit, configured to perform heterochromatic color detection on the first color image taken by the first line array camera in each cycle, and to perform heterochromatic color detection on the second color image taken by the second line array camera in each cycle;

A summarizing unit, comprising a unit for performing a first operation and/or a second operation on each periodic image of the first line scan camera and the second line scan camera:

14. The system according to claim 13, characterized in that the image acquisition module also includes a device for acquiring an image of the particle taken by the third linear array camera during the free fall of the particle; correspondingly, the system also includes a special-shaped recognition unit;

The shape recognition unit is used to detect the shape by the following method:

15. The system according to claim 14, characterized in that the summary unit also includes obtaining the detection results of the particle's heterogeneous color and shape by statistically analyzing the recognition results of the heterogeneous color recognition unit and the heterogeneous shape recognition unit.

16. A particle detection device, comprising:

A first hopper, used for holding particles to be tested;

A first conveying mechanism, located below the first hopper, is used to convey the particles falling from the first hopper to above the second hopper, and to allow the particles to freely fall from above the second hopper into the second hopper;

The second hopper is arranged below the first conveying mechanism and is used to receive the particles falling from the first conveying mechanism;

A second conveying mechanism, located below the second hopper, is used to convey the particles falling from the second hopper to above the material collector, and to make the particles fall freely from above the material collector into the material collector;

The material collector is arranged below the second conveying mechanism and is used to receive the particles falling from the second conveying mechanism;

The first line array camera and the second line array camera are used to capture color images of the freely falling particles; the color images are used for heterochromatic detection; the first line array camera and the second line array camera are relatively arranged on both sides of the free-falling path of the particles;

An electronic device connected to the first line array camera and the second line array camera; the electronic device is used to perform the particle detection method according to any one of claims 1 to 12;

Wherein, the first line array camera and the second line array camera are both arranged between the first conveying mechanism and the second hopper, or the first line array camera and the second line array camera are both arranged between the second conveying mechanism and the material collector.

17 . The device according to claim 16 , wherein the first line array camera and the second line array camera are respectively configured with diffuse reflection light sources to photograph particulate matter with the diffuse reflection light sources.

18. The device according to claim 16, further comprising: a third line array camera, used to capture images of freely falling particles; the images captured by the third line array camera are used for special-shaped detection;

In the case where the first line array camera and the second line array camera are both arranged between the first conveying mechanism and the second hopper, the third line array camera is arranged between the second conveying mechanism and the material collector;

In the case where the first line array camera and the second line array camera are both arranged between the second conveying mechanism and the material collector, the third line array camera is arranged between the first conveying mechanism and the second hopper.

19. An electronic device, comprising:

A memory and a processor, wherein the processor and the memory are communicatively connected to each other, the memory stores computer instructions, and the processor implements the steps of the method according to any one of claims 1 to 12 by executing the computer instructions.