CN115791829A - Tobacco shred quality online detection system - Google Patents

Abstract

The invention discloses an online detection system for tobacco shred quality, which comprises: the device comprises a light source, a camera device, an installation frame, a control device and a conveying belt, wherein the light source is arranged above the conveying belt and used for emitting illumination light to the conveying belt, the camera device is arranged above the conveying belt and used for acquiring images of the conveying belt, the installation frame is used for installing the light source and the camera device, the control device is respectively in communication connection with the light source and the camera device and used for processing the images to identify sundries in cut tobacco when the conveying belt is determined to normally run according to the images, and the conveying belt is used for conveying the cut tobacco. The tobacco shred quality online detection system realizes the detection of sundries in tobacco shreds conveyed on the conveying belt, and compared with the prior art, the tobacco shred quality online detection system has a simple structure.

Description

Tobacco quality online detection system

technical field

The invention relates to the field of machine vision, in particular to an online quality detection system for shredded tobacco.

Background technique

As the requirements for the quality of cigarettes become higher and higher, more and more attention is paid to the impurities in the silk production process. At present, the photoelectric impurity removal equipment in the workshop is operating normally, but due to the large amount of impurities in the incoming raw materials, the removal After shredding, it is found that there are still occasional impurities in the shredded tobacco, which affects the subsequent normal production process and the quality of the cut tobacco. Therefore, it is necessary to formulate a detection plan for this phenomenon, and finally remove the impurities to ensure the quality of the shredded tobacco.

The patent literature with the notification number CN214844880U proposes a shredded tobacco debris video detection system. The system includes an industrial camera, an industrial computer, a PLC controller, and a speed sensor. The speed sensor is used to detect the conveying speed of the shredded tobacco conveyor belt. The transmission speed determines the running state of the shredded tobacco conveyor belt and sends it to the industrial computer. When the shredded tobacco conveyor belt is in operation, the industrial computer controls the industrial camera to take pictures of the shredded tobacco on the shredded tobacco conveyor belt to obtain the shredded tobacco image. There is an image processing algorithm running on the industrial computer to process the shredded tobacco image and identify debris. However, the existing shredded tobacco quality detection system has a relatively complex structure, which requires real-time judgment of the running state of the conveyor belt for detection, and the structure and installation flexibility is poor.

Contents of the invention

The object of the present invention is to provide an online detection system for shredded tobacco quality, which has a simpler structure than the prior art.

To achieve the above object, the present invention provides the following technical solutions:

An online detection system for shredded tobacco quality, comprising:

a light source arranged above the conveyor belt to emit illumination light to the conveyor belt;

an imaging device arranged above the conveyor belt to acquire images of the conveyor belt;

A mounting frame for installing the light source and the camera device;

A control device that is communicatively connected with the light source and the camera device and processes the image to identify impurities in shredded tobacco when determining from the image that the conveyor belt is in normal operation, and the conveyor belt is used to transport shredded tobacco.

Optionally, the installation frame includes a vertical support rod and a horizontal support rod, the horizontal support rod is connected to the vertical support rod, and the light source and the camera device are respectively installed on the horizontal support rod.

Optionally, the light source is bar-shaped, and the light source is arranged along the length direction of the horizontal support rod.

Optionally, the light source is connected to the horizontal support bar through a light source connecting piece, the light source connecting piece includes a first T-shaped clip and a first connecting piece, the first T-shaped clip includes a connecting hole, and the horizontal The support rod passes through the connecting hole;

The first connecting piece is connected with the first T-shaped clip, the first connecting piece is provided with a third screw hole and an arc-shaped notch, and one end of the light source is provided with a parallel first screw hole and a second screw holes, the third screw hole is correspondingly connected with the first screw hole through bolts, and the arc-shaped notch is correspondingly connected with the second screw hole through bolts.

Optionally, the camera device is connected to the horizontal support bar through a camera connector, the camera connector includes a second T-shaped clip and a second connector, the second T-shaped clip includes a connection hole, the The horizontal support rod passes through the connecting hole;

The second connecting piece is connected with the second T-shaped clip, and the second connecting piece includes a slot for installing the camera device.

Optionally, the connection hole of the first T-shaped clip is provided with a fracture, and a screw hole passing from one end of the fracture to the other end is provided in the fracture, and a bolt is inserted into the screw hole to adjust the tightness of the connection hole;

Alternatively, the connecting hole of the second T-shaped clip is provided with a fracture, and a screw hole passing from one end of the fracture to the other end is provided in the fracture, and the tightness of the connecting hole is adjusted by passing a bolt into the screw hole.

Optionally, the processing of the image by the control device to identify impurities in shredded tobacco includes that the control device is used to: transform the image into a preset color space, filter out abnormal pixels in the image, and The abnormal pixel refers to the pixel whose first component value, second component value and third component value exceed the first preset range, if a plurality of the abnormal pixels form a connected domain and the connected domain includes abnormal pixels If the number is greater than the first preset value, it is determined that the connected domain corresponds to debris.

Optionally, the control device is further configured to: acquire a preset color channel image of the image, and filter out pixels whose pixel values are in a second preset range in the preset color channel image, if the pixel value is in the If the ratio of the pixels in the second preset range to the number of pixels in the image is greater than or equal to a second preset value, then it is determined that there is no shredded tobacco on the conveyor belt.

Optionally, the control device determining the normal operation of the conveyor belt according to the image includes the control device being used to: binarize the image and perform morphological open operation processing to obtain a binary image, and calculate the The ratio of the total area of pixels with non-zero pixel values in the binary image to the total area of the binary image, if the ratio is within a third preset range, it is determined that there is shredded tobacco on the conveyor belt and the conveyor belt stop running.

Optionally, the camera device is configured to scan the conveyor belt to obtain a continuous multi-frame row image, and sequentially splicing the row images of a predetermined number of consecutive frames to obtain the image of the conveyor belt.

It can be known from the above technical solution that the cut tobacco quality online detection system provided by the present invention includes: a light source arranged above the conveyor belt to emit illuminating light to the conveyor belt; an imaging device arranged above the conveyor belt to obtain images of the conveyor belt; The installation frame for installing the light source and the camera device is connected with the light source and the camera device respectively, and the control device for processing the image to identify impurities in the shredded tobacco is determined according to the image when the conveyor belt is in normal operation. The conveyor belt is used to transport the shredded tobacco. The shredded tobacco quality on-line detection system of the present invention realizes the detection of impurities in the shredded tobacco conveyed on the conveyor belt, and compared with the prior art, the shredded tobacco quality on-line detection system of the present invention has a simple structure.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention. Those skilled in the art can also obtain other drawings based on these drawings without creative work.

Fig. 1 is the schematic diagram of a kind of cut tobacco quality on-line detection system provided by an embodiment of the present invention;

Fig. 2 is a schematic diagram of the connection between the horizontal support bar shown in Fig. 1 and the light source and camera device;

Fig. 3 is the schematic diagram of the control platform of the cut tobacco quality on-line detection system of an embodiment of the present invention;

Fig. 4 is an image of conveying shredded tobacco and operating normally in one embodiment of the present invention;

Fig. 5-1 is the image when conveying shredded tobacco but stopping running in an embodiment of the present invention;

Figure 5-2 is the middle area intercepted from the image shown in Figure 5-1;

Figure 5-3 is the converted grayscale image of the image shown in Figure 5-2;

Figure 5-4 is the image after binarization of the image shown in Figure 5-3;

Figure 5-5 is the image after performing morphological opening operation on the image shown in Figure 5-4;

Fig. 6 is an operation flow chart of the on-line detection system for shredded tobacco quality according to an embodiment of the present invention.

The reference signs in the accompanying drawings of the specification include:

1-base, 2-vertical support rod, 3-horizontal support rod, 4-conveyor belt, 5-cross connection clip, 6-light source connector, 7-light source, 8-camera connector, 9-camera device, 10- Control platform, 11-first T-clip, 12-first connecting piece, 13-third screw hole, 14-arc notch, 15-second T-clip, 16-second connecting piece, 17-control Device, 18-light source controller, 19-alarm device.

Detailed ways

In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described The embodiments are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present invention.

This embodiment provides an online detection system for shredded tobacco quality, including:

a light source arranged above the conveyor belt to emit illumination light to the conveyor belt;

an imaging device arranged above the conveyor belt to acquire images of the conveyor belt;

A mounting frame for installing the light source and the camera device;

A control device that is respectively connected to the light source and the camera device in communication and processes the image to identify impurities in shredded tobacco when the conveyor belt is in normal operation according to the image, and the conveyor belt is used to transport shredded tobacco.

The light source and the camera device are installed through the mounting frame, the light source emits illumination light to the conveyor belt, and the image of the conveyor belt is obtained by the camera device, and the control device processes the image to identify the dust in the shredded tobacco when it is determined that the conveyor belt is in normal operation according to the acquired image of the conveyor belt. sundries. The cut tobacco quality online detection system of this embodiment realizes the detection of impurities in the cut tobacco conveyed on the conveyor belt. Compared with the prior art, the structure of the cut tobacco quality online detection system is simple.

In this embodiment, the type and structure of the light source are not limited, and the light source may be but not limited to a strip light source. The type and structure of the camera device are not limited, and industrial cameras can be used but not limited to.

In some embodiments, the installation frame includes a vertical support rod and a horizontal support rod, the horizontal support rod is connected to the vertical support rod, and the light source and the camera device are respectively installed on the horizontal support rod, so that The light source and camera are located above the conveyor belt.

Optionally, the light source may be in the shape of a bar, and the light source is arranged along the length direction of the horizontal support bar, and the light source is fixedly connected with the horizontal support bar, so as to install the light source stably.

Optionally, the light source can be connected to the horizontal support bar through a light source connecting piece, the light source connecting piece includes a first T-shaped clip and a first connecting piece, the first T-shaped clip includes a connecting hole, and the horizontal support The rod passes through the connecting hole; the first connecting piece is connected with the first T-shaped clip, the first connecting piece is provided with a third screw hole and an arc-shaped notch, and one end of the light source is provided with a juxtaposed The first screw hole and the second screw hole are correspondingly connected to the third screw hole and the first screw hole by bolts, and the arc-shaped notch is correspondingly connected to the second screw hole by bolts. Refer to Figure 1 and Figure 2 for example, Figure 1 is a schematic diagram of an on-line tobacco quality detection system provided by an embodiment, and Figure 2 is a schematic diagram of the connection of the horizontal support rod shown in Figure 1 with the light source and the camera device. As shown in Figure 2, the light source 7 is connected to the horizontal support bar 3 through the light source connector 6. The light source connector 6 includes a first T-shaped clip 11 and a first connector 12, and the first T-shaped clip 11 includes a connecting hole, so that the horizontal The support rod 3 passes through the connecting hole of the first T-shaped clamp 11 . Preferably, the connection hole of the first T-shaped clamp 11 is provided with a fracture, and a screw hole passing through from one end of the fracture to the other end is provided in the fracture, and a bolt is inserted into the screw hole to adjust the tightness of the connection hole, which is also convenient. Then the first T-shaped clamp 11 is removed. The front and rear angles and left and right positions of the light source 7 can be flexibly adjusted by tightening the first T-shaped clip 11 .

The first connecting piece 12 is provided with a third screw hole 13 and an arc-shaped notch 14, and one end of the light source 7 is provided with a first screw hole and a second screw hole parallel up and down, and the third screw hole 13 and one end of the light source 7 are on the upper side. The first screw holes are correspondingly connected by bolts, and the arc-shaped notch 14 corresponds to the second screw hole with one end of the light source 7 below, and is fixed by bolts. On the one hand, the angle and position of the light source 7 can be adjusted by changing the fixed position of the second screw hole and the arc notch 14 by rotating the light source 7; on the other hand, the first T-shaped clamp 11 can be rotated on the horizontal support rod 3 to Adjust the angle and position of the light source 7. The first connecting member 12 and the first T-shaped clip 11 can be fixedly connected by screws.

If the light source is a strip light source, as shown in FIG. 2 , both ends of the light source 7 are respectively connected to the horizontal support bar 3 , and any end of the light source 7 is connected to the horizontal support bar 3 through the light source connector 6 .

Optionally, the camera device 9 is connected to the horizontal support bar 3 through a camera connector, the camera connector includes a second T-shaped clip and a second connector, the second T-shaped clip includes a connection hole, and the horizontal support A rod passes through the connecting hole; the second connecting piece is connected with the second T-shaped clip, and the second connecting piece includes a slot for installing the camera device. Can refer to Fig. 1 and shown in Fig. 2 by way of example, camera device 9 is connected with horizontal support bar 3 by camera connector 8, and camera connector 8 comprises second T-shaped clamp 15 and second connector 16, and the second T-shaped clamp 15 includes a connection hole through which the horizontal support rod 3 passes. The second connecting piece 16 includes a slot, and the camera device 9 is installed in the slot. The camera device 9 can be fixedly connected to the second connecting member 16 by screws. The second connecting piece 16 and the second T-shaped clip 15 can be fixedly connected by screws.

Preferably, the connecting hole of the second T-shaped clamp 15 can be provided with a fracture, and a screw hole passing from one end of the fracture to the other end is provided at the fracture, and a bolt is inserted into the screw hole to adjust the tightness of the connecting hole. It is convenient to remove the second T-shaped clamp 15 . The front and rear angles and left and right positions of the camera device 9 can be flexibly adjusted by tightening the second T-shaped clip 15 .

Optionally, the horizontal support bar 3 and the vertical support bar 2 can be connected through a cross connection clip 5 . The height of the horizontal support bar 3 can be flexibly adjusted by tightening the bolts of the cross connection clamp 5 to adjust the height of the light source 7 and the camera device 9 . Further referring to FIG. 1 , the installation frame may further include a base 1 on which a vertical support rod 2 is fixed. The base 1 can be a gravity base to keep the mounting frame stable.

Preferably, the light source 7 and the imaging device 9 can be located directly above the conveyor belt 4, and the imaging device 9 can be correspondingly located in the middle of the bar-shaped light source 7, so that the light source 7 emits illumination light to the conveyor belt 4, and the image of the conveyor belt acquired by the imaging device 9 is evenly illuminated , which is beneficial to effectively identify impurities in shredded tobacco.

In this embodiment, the type of the control device is not limited, and the control device may be, but not limited to, an industrial computer. Optionally, the control device can be set on the control platform. In some implementations, refer to FIG. 3 , which is a schematic diagram of a control platform of an online shredded tobacco quality detection system according to an embodiment. The control platform 10 may also include a light source controller 18, which is used to control the light source 7 to turn on. , turn off and control the brightness of the illumination light emitted by the light source 7 . The control platform 10 may also include an alarm device 19 communicatively connected with the control device 17, capable of sending out an alarm signal when the control device 17 detects foreign matter in the shredded tobacco according to the image. The control platform 10 may also include a power supply for supplying power to the light source 7 , the camera device 9 , the control device 17 or the alarm device 19 .

In some embodiments, the camera 9 comprises a line scan camera. The camera device 9 is used for line scanning the conveyor belt 4 to obtain continuous multi-frame line images, and sequentially splicing the line images of a predetermined number of consecutive frames to obtain the image of the conveyor belt 4 . The rows of the conveyor belts 4 correspond to the width direction of the conveyor belts 4 . The camera device 9 scans four lines of the conveyor belt in real time, sequentially obtains each frame and line of images, and stitches the continuous preset number of frames and lines of images each time a predetermined number of consecutive frames of lines of images are obtained. In this embodiment, there is no limitation on the value of the preset number, which can be set according to the detection situation in practical applications. As an example, the resolution of a frame of line images obtained by scanning 4 lines of the conveyor belt is 2048*2, and 512 consecutive frames of line images are spliced to obtain an image with a resolution of 2048*1024. The imaging device 9 may adopt a line scan camera.

In some embodiments, the processing of the image by the control device 17 to identify impurities in shredded tobacco includes that the control device 17 is used to: transform the image into a preset color space, and filter out abnormal pixels in the image , the abnormal pixel refers to the pixel whose first component value, second component value and third component value exceed the first preset range, if a plurality of the abnormal pixels form a connected domain and the connected domain includes abnormal If the number of pixels is greater than the first preset value, it is determined that the connected domain corresponds to debris.

The first component value, the second component value and the third component value of the pixel refer to the first component value, the second component value and the third component value of the pixel in the preset color space.

The first component value, the second component value, and the third component value of the pixel are out of the first preset range, which may be that the first component value of the pixel is out of the first component preset range and the second component value is out of the second component preset range. range and the value of the third component is outside the preset range of the third component. In this embodiment, there are no specific limitations on the preset range of the first component, the preset range of the second component, and the preset range of the third component. In practical applications, it can be determined through experiments or based on the detection experience based on shredded tobacco images. Find the range of each component value of the pixel corresponding to shredded tobacco in the image in the preset color space, and set the preset range of each component to detect abnormal pixels.

Optionally, the preset color space may be an HSV color model, and the color parameter components of the HSV color model include hue (Hue), saturation (Saturation) and brightness (Value). For example, in the test, it was found that for the images collected under fixed camera parameters and light source brightness conditions, the experiment showed that more than 99% of the tobacco leaf colors were within the range of 0≤H≤45, 10≤S≤130, 60≤V≤255, Pixels that are not within this range are filtered out.

In this embodiment, the value of the first preset value is not limited. In practical applications, it can be set in combination with the detection accuracy and detection efficiency of impurities in shredded tobacco. For example, the image is converted to the HSV color model, and the pixels in the image that are not in the range of 0≤H≤45, 10≤S≤130, 60≤V≤255 are abnormal pixels, if the area of the connected domain formed by the abnormal pixels is greater than 200 pixels , and set the region as an exception region. In this embodiment, color screening is performed in the image based on the color of the shredded tobacco to identify impurities in the shredded tobacco, and the amount of calculation in the image processing process is small. In this embodiment, the method that the control device 17 processes the image to identify impurities in the shredded tobacco can be applied to the image collected by the line scan camera device through line scanning, and can also be applied to the camera for one exposure to capture a frame with the desired resolution image.

In some implementations, the control device 17 is further configured to: acquire a preset color channel image of the image, filter out pixels in the preset color channel image whose pixel values are in the second preset range, if the pixel value is in the preset color channel image If the ratio of the number of pixels in the second preset range to the number of pixels in the image is greater than or equal to a second preset value, it is determined that there is no shredded tobacco on the conveyor belt.

The ratio of pixels whose pixel values are in the second preset range to the number of pixels in the image refers to the ratio of the number of pixels whose pixel values are in the second preset range to the total number of pixels in the image. In this embodiment, the preset color channels are not limited, and can be selected according to the color of the belt and the color of shredded tobacco in practical applications. There is no limitation on the second preset range, and no limitation on the value of the second preset value.

Optionally, binarization processing may be performed on the preset color channel image of the image, and the pixels of the binarized preset color channel image are screened from the image. For example, the conveyor belt 4 is a white belt, and the shredded tobacco is taupe, as shown in FIG. 4 . FIG. 4 is an image of conveying with shredded tobacco and running normally in one embodiment. When the conveyor belt 4 has no shredded tobacco, there is only the white color of the belt. Binarize the red channel (R channel) image of the image. In the binarized R channel image, when the number of pixels with a pixel value greater than 230 accounts for more than 50%, it is judged that the current conveyor belt 4 has no shredded tobacco. In this embodiment, when detecting debris in the shredded tobacco based on the image, first judge whether there is shredded tobacco on the conveyor belt 4 according to the image, and stop the detection if there is no shredded tobacco, which helps to reduce the amount of calculation and power consumption of the system. In this embodiment, the method that the control device 17 processes the image to detect whether there is shredded tobacco on the conveyor belt 4 can be applied to the image collected by the line scanning camera device through line scanning, or can be applied to the image collected by the camera for one exposure to capture a frame with all image at the desired resolution.

In some implementations, the control device 17 determining the normal operation of the conveyor belt 4 according to the image includes the control device 17 being used to: binarize the image and perform morphological opening processing to obtain a binary image, calculating the ratio of the total area of pixels with non-zero pixel values in the binary image to the total area of the binary image, and if the ratio is within a third preset range, it is determined that there is shredded tobacco on the conveyor belt and The conveyor belt stops running.

Preferably, binarizing the image and performing morphological opening processing to obtain a binary image includes: intercepting the middle region of the image, binarizing the middle region of the image and performing morphological opening operation processing to obtain the binary image. On the conveyor belt 4, shredded tobacco is mainly distributed in the middle area of the conveyor belt 4. Intercept the middle area of the collected image to detect whether the conveyor belt 4 is running, avoiding the influence of the white belt leaking from the edge of the image, which helps to improve accuracy and detection efficiency. .

Optionally, binarizing the image may include: converting the image into a grayscale image, and binarizing the grayscale image. Performing morphological opening processing on the image may include performing morphological opening processing on the image with a preset column matrix. The operation of morphological opening operation on the image is to judge according to the vertical stripes in the image.

The total area of pixels with non-zero pixel values in the binary image refers to the sum of the areas of all pixels with non-zero pixel values in the binary image. In this embodiment, the third preset range is not limited, and may be set according to detection conditions in practical applications. Specifically, the third preset range may be that the ratio is greater than the lower threshold and smaller than the upper threshold, wherein the upper threshold is greater than the lower threshold, and the lower threshold is greater than zero. When the conveyor belt 4 has shredded tobacco and is running normally, some stripes will appear in the image. In order to avoid misjudgment and avoid mistakenly judging the normal situation as the stop of the conveyor belt 4, the lower threshold is set here to be greater than 0, that is, only when the number of stripes in the image is greater than a certain value , it is decided to stop.

As an example, reference can be made to Fig. 5-1, which is an image of conveying shredded tobacco but stopping operation in an embodiment. Further reference can be made to Figure 5-2, Figure 5-3, Figure 5-4 and Figure 5-5, Figure 5-2 is the middle area of the image shown in Figure 5-1, and Figure 5-3 is the image of Figure 5- The converted grayscale image of the image shown in Figure 2, Figure 5-4 is the image after binarization of the image shown in Figure 5-3, and Figure 5-5 is the morphological open operation of the image shown in Figure 5-4 after the image. The collected image is 2048*1024, first intercept the 1024*1024 area in the middle, convert the intercepted image into a grayscale image, then convert it into a black and white image with a limit of 160-255, and then use a matrix of 1*100 The morphological opening operation obtains the final binary image. When the white area of the final binary image is greater than 0.3% and less than 10%, it is determined that the conveyor belt 4 has shredded tobacco but stops running. If the white area of the final binary image is greater than 10%, it indicates that there are many white areas on the conveyor belt 4, and there may be no shredded tobacco. In this embodiment, when detecting debris in shredded tobacco based on the image, judge whether the conveyor belt 4 is running normally according to the image. If the conveyor belt 4 stops running, the detection can be stopped. If the conveyor belt 4 is running normally, the debris will be detected, which helps to reduce the amount of calculation. Reduce system power consumption.

Preferably, in some implementations, the control device 17 is further configured to: perform bilateral filtering on the image, and perform blur processing on the low-frequency part of the image. The image after bilateral filtering is blurred in the low frequency part to reduce the image noise while retaining the image details in the high frequency part.

The control device 17 is also used for: removing the background of the image, specifically including filtering the background of the left edge area and the right edge area of the image respectively, and then processing the image by morphological closing operation to expand the background edge. For example, since the conveyor belt 4 is a belt, the color is white, and the debris also includes white, so the influence of the white background on the detection should be removed. Since the conveyor belt 4 background only appears on both sides of the image, the left edge area of the image, The right edge area is processed, the white background is filtered, and the binary image is output, and then the background edge is expanded by using the morphological closing operation, and the final output binary image is used as a mask for subsequent detection. For example, filter background processing can be performed on the left 1/4 area and the right 1/4 area of the image.

In some embodiments, the control device 17 in the cut tobacco quality online detection system is also used to: mark the area in the image that determines the corresponding sundries, use the marked image as a training sample, and use the training sample to build The deep learning model is trained to obtain a model for detecting impurities in the shredded tobacco on the conveyor belt 4, and the model is used to process the collected images to identify the impurities in the shredded tobacco.

The cut tobacco quality online inspection system of this embodiment uses machine vision algorithms to detect and record sundries and abnormalities in the early stage of system operation, and accumulates sundry data sets during operation. In the later stage of system operation, the accumulated sundries images are used for image annotation, so as to construct the shredded tobacco sundries data set, build a deep learning model, use the shredded tobacco sundries data set as training samples for training, and finally deploy the trained model to the detection software In order to carry out online intelligent detection of cut tobacco quality. The built deep learning model can be but not limited to the YoloV5 model.

The detection algorithm of this cut tobacco quality online detection system adopts the form of transition between the front and the later stage. The machine vision algorithm is used in the early stage, which is simple to implement and does not require a large number of abnormal data sets. At the same time, the data sets can be accumulated during the system operation, but the accuracy is limited; the deep learning is used in the later stage The model can be quickly learned and deployed by using the data set accumulated in the early stage, and the detection accuracy is high. The fast connection between front and back can effectively improve the speed of system launch and reduce time cost.

Reference can be made to Fig. 6, Fig. 6 is an operation flowchart of an online shredded tobacco quality detection system of an embodiment, including the following processes:

After the S1 system starts, connect the industrial camera and start the light source, and enter S2;

S2 sets the relevant parameters of the industrial camera and the brightness of the light source, and enters S3;

S3 starts to collect images and enters into S4;

S4 stitches the collected image from 2048*2 to 2048*1024 to obtain a frame of image, and enters S5;

S5 analyzes the image to determine whether the conveyor belt is running normally, if yes, enter S6, otherwise enter S4;

S6 executes the shredded tobacco quality detection algorithm, judges whether there are sundries or abnormalities according to the returned result, if yes, enters S7, otherwise enters S4;

The S7 industrial computer outputs an alarm signal to the alarm device, waits for manual processing, enters S8, and continues the next detection cycle to enter S4;

S8 The staff checks and processes the alarm information. If there are indeed sundries, after the abnormality is handled, the abnormality processing confirmation is carried out, and the system stores the abnormal pictures and detection results; Detect pictures, store the detection results, and reset the alarm.

The online cut tobacco quality detection system provided by the present invention has been introduced in detail above. In this paper, specific examples are used to illustrate the principle and implementation of the present invention, and the descriptions of the above embodiments are only used to help understand the method and core idea of the present invention. It should be pointed out that for those skilled in the art, without departing from the principle of the present invention, some improvements and modifications can be made to the present invention, and these improvements and modifications also fall within the protection scope of the claims of the present invention.

Claims (10)

1. A shredded tobacco quality online detection system, characterized in that, comprising: a light source arranged above the conveyor belt to emit illumination light to the conveyor belt; an imaging device arranged above the conveyor belt to acquire images of the conveyor belt; A mounting frame for installing the light source and the camera device; A control device that is communicatively connected with the light source and the camera device and processes the image to identify impurities in shredded tobacco when determining from the image that the conveyor belt is in normal operation, and the conveyor belt is used to transport shredded tobacco. 2. The on-line detection system for shredded tobacco quality according to claim 1, wherein the installation frame includes a vertical support rod and a horizontal support rod, the horizontal support rod is connected with the vertical support rod, and the light source 1. The camera devices are respectively installed on the horizontal support rods. 3. The on-line detection system for shredded tobacco quality according to claim 2, characterized in that, the light source is in a bar shape, and the light source is arranged along the length direction of the horizontal support rod. 4. The on-line detection system for shredded tobacco quality according to claim 2, wherein the light source is connected to the horizontal support bar through a light source connector, and the light source connector includes a first T-shaped clip and a first connector , the first T-shaped clamp includes a connection hole, and the horizontal support rod passes through the connection hole; The first connecting piece is connected with the first T-shaped clip, the first connecting piece is provided with a third screw hole and an arc-shaped notch, and one end of the light source is provided with a parallel first screw hole and a second screw holes, the third screw hole is correspondingly connected with the first screw hole through bolts, and the arc-shaped notch is correspondingly connected with the second screw hole through bolts. 5. The on-line detection system for shredded tobacco quality according to claim 2, wherein the camera device is connected to the horizontal support bar through a camera connector, and the camera connector includes a second T-shaped clip and a second connection part, the second T-shaped clamp includes a connection hole, and the horizontal support rod passes through the connection hole; The second connecting piece is connected with the second T-shaped clip, and the second connecting piece includes a slot for installing the camera device. 6. The on-line detection system for shredded tobacco quality according to claim 4 or 5, characterized in that, the connection hole of the first T-shaped clamp is provided with a fracture, and a screw that passes through from one end of the fracture to the other end is provided on the fracture. Hole, through the bolt through the screw hole to adjust the tightness of the connecting hole; Alternatively, the connecting hole of the second T-shaped clip is provided with a fracture, and a screw hole passing from one end of the fracture to the other end is provided in the fracture, and the tightness of the connecting hole is adjusted by passing a bolt into the screw hole. 7. The on-line detection system for shredded tobacco quality according to claim 1, wherein the processing of the image by the control device to identify impurities in the shredded tobacco includes the control device being used to: convert the image into a predetermined The color space is set, and the abnormal pixels in the image are screened out. The abnormal pixels refer to pixels whose first component value, second component value and third component value exceed the first preset range. If the abnormal pixels constitute a connected domain and the number of abnormal pixels included in the connected domain is greater than a first preset value, then it is determined that the connected domain corresponds to debris. 8. The cut tobacco quality online detection system according to claim 1, characterized in that the control device is also used for: acquiring the preset color channel image of the image, and filtering out the pixel values in the preset color channel image For the pixels in the second preset range, if the ratio of the pixels whose pixel values are in the second preset range to the number of pixels in the image is greater than or equal to a second preset value, it is determined that there is no shredded tobacco on the conveyor belt. 9. The on-line detection system for shredded tobacco quality according to claim 1, wherein the control device determines that the conveyer belt is running normally according to the image comprises the control device being used to: binarize the image and Perform morphological opening operation processing to obtain a binary image, and calculate the ratio of the total area of pixels with non-zero pixel values in the binary image to the total area of the binary image, if the ratio is in the third preset Within the range, it is determined that there is shredded tobacco on the conveyor belt and the conveyor belt stops running. 10. The on-line detection system for shredded tobacco quality according to any one of claims 1-9, characterized in that the camera device is used for line scanning the conveyor belt to obtain continuous multi-frame line images, and the continuous preview Assuming that the number of frames of the row images are sequentially spliced to obtain the image of the conveyor belt.

Images