CN116030008A - Multi-patch component detection system and method - Google Patents

Abstract

The invention provides a system and a method for detecting components of multiple patches, wherein the system comprises: the image acquisition module is used for automatically acquiring the original images of the multi-patch components; the image segmentation module is used for segmenting the original image to obtain a plurality of single-chip images only containing single patches; and the image detection module is used for calculating and obtaining the coordinate information and the rotation angle of the single patch contained in the single-chip image.

Description

A multi-chip component detection system and method

technical field

The invention relates to the technical field of intelligent detection, in particular to a multi-chip component detection system and method.

Background technique

In recent years, with the development of the domestic electronics industry, my country has become the world's largest electronics power. Users have higher and higher requirements for the quality of electronic products, and at the same time domestic labor costs continue to increase, prompting more and more electronic companies to introduce fully automatic surface mount production lines. With the blowout growth of demand, domestic companies and research institutions are working hard to localize surface mount production lines. Since the 1980s, the development of placement machines has begun in China. However, so far, although most surface mount equipment has been localized, domestic placement machines that meet market requirements have not yet been developed. At present, domestic manufacturers still need to import placement machines from companies in Europe, America and Japan, and the price of a single placement machine is as high as several million or even tens of millions of dollars. This not only raises the cost of my country's electronics manufacturing enterprises, limits the development of my country's electronics industry, but also causes a large loss of foreign exchange.

With the development of technology and the demand for automated production lines, the global demand for electronic information products continues to increase, and the export volume of surface mount equipment has soared rapidly. At the same time, the demand for equipment related to China's surface mount technology has also increased, and its supply has been imported in large quantities from overseas. According to export data, the number of SMT automation equipment introduced in 2000 was 8,992 units, and China's SMT inventory was 3 More than 10,000 units, and the number of SMT production lines is about 15,000. my country's demand for SMT has made China the world's largest and most important placement machine market. At present, China's placement machine market lacks independent technology brands, resulting in the supply of domestic placement machines almost relying on overseas imports. In this difficult background, our country should improve the research and development of related technologies of placement equipment. As an important part of the SMT production line, the automatic placement machine is a high-tech automatic machine integrating mechanical technology, laser and microelectronics. Therefore, accelerating the upgrading and technology development of SMT equipment has far-reaching significance for the realization of production line automation.

In the manufacturing process of electronic products in our country, SMT surface mount technology is a key part of production line automation, and the application of this technology can also be widely used in the manufacturing of products in other industries. Because SMT surface mount technology has the advantages of high precision and high speed, after years of research, electronic assembly technology has undergone tremendous changes. Traditional manual assembly has been replaced by automated precision assembly. Low-quality assembly technology has been iteratively optimized and automated. Assembly technology will continue to upgrade in the future.

The placement machine has the advantages of high speed and high precision. It is a high-tech automation equipment integrating mechanical technology, laser and microelectronics. It has the characteristics of high speed and high precision. Therefore, it has strict technical requirements and requires high support from relevant hardware configurations. In addition, the vision system of the placement machine needs related supporting digital image algorithms to solve the corresponding visual inspection tasks, so the research and development of the key technology of the placement machine vision is a very challenging task.

High-speed and high-precision placement machine equipment generally adopts visual inspection technology. The visual inspection technology of the placement machine consists of two parts: the pose estimation of the placement component and the positioning of the PCB reference point. Among them, the pose estimation of the placement component It is to accurately and quickly locate the pose state of the patch component, including the center and rotation angle of the patch. PCB reference point positioning requires precise positioning of the PCB

The reference point center on the substrate, by transmitting the positioning information to the motion control system, compensates for the existing pose error, and completes the deviation correction of the chip component and the correction of the reference point center offset. Its positioning accuracy and speed directly affect the placement. The placement accuracy and placement efficiency of the placement machine, so the position visual detection technology of the placement machine is one of the technological breakthroughs to improve the placement accuracy and placement efficiency of the placement machine. In the actual production process of industrial products, manual assembly and inspection of products is inefficient and error-prone. Due to the introduction of machine vision technology, its precise and fast measurement, positioning, detection and identification advantages make automatic assembly The production line replaces the traditional manual inspection, which greatly improves the production efficiency and product quality. However, the current domestic placement machine does not combine advanced technology of high-precision visual aided positioning, so the placement accuracy and speed of the placement machine are always at a low level. The main performance is that the method used in the process of image processing has certain limitations, which affects the efficiency of the solution, and the positioning accuracy needs to be further improved and perfected. Although after more than 30 years of development, there is still a large gap between the level of China's SMT equipment and the world's advanced level. Therefore, it is an urgent task to study the image processing and position detection technology of the placement machine, as the key technology in the placement machine, and to realize the localization of the placement machine.

Therefore, the multi-chip component detection system and method proposed by the present invention are used to accurately detect the coordinate information and rotation angle of the chip, and have high precision, which can meet the actual working requirements of the chip mounter.

Contents of the invention

Aiming at the deficiencies of the prior art, the invention provides a multi-chip component detection system and method, which are used to accurately detect the coordinate information and rotation angle of the chip, and have high precision, which can meet the actual requirements of the chip mounter. Work requirements.

A detection system for multi-chip components, comprising: an image acquisition module for automatically collecting original images of multi-chip components;

An image segmentation module is used to segment the original image to obtain several single-slice images that only include a single patch;

The image detection module is used to calculate the coordinate information and the rotation angle of the single patch contained in the single image.

As an embodiment of the present invention, the image acquisition module includes:

The placement unit is used to place and fix the multi-chip components and send out the first signal;

The light source setting unit is used to set the corresponding light source according to the category and orientation of the multi-chip components after receiving the first signal, and send out the second signal;

The camera unit is used to collect the original image of the multi-chip components after receiving the second signal.

As an embodiment of the present invention, the light source setting unit performs the following operations:

Identify the category of each patch component in the multi-chip component according to the characteristic information, and obtain multiple pieces of category information;

Obtain the orientation information of each patch component on the placement unit, and obtain multiple pieces of orientation information;

Based on multiple pieces of category information and multiple pieces of orientation information, perform association judgments to obtain the degree of association; wherein, the degree of association indicates the degree to which different types of information are affected by light sources in different orientations;

Based on the degree of correlation, determine the state of light to be filled for each patch component;

Carry out one-to-one correspondence between the category information, orientation information and the state of light to be filled for each patch component, and form a demand command for light fill;

Complete the light source setting according to the fill light demand command.

As an embodiment of the present invention, the light source setting is completed according to the supplementary light demand instruction, including:

According to each supplementary light demand instruction, the expected satisfaction degree of supplementary light for each patch component is generated; under the initial condition, the expected satisfaction degree of supplementary light for each patch component is empty; where, when any patch component’s When the fill light demand command is perfectly executed, the corresponding fill light expectation satisfaction is full;

Adaptively generate an initial light source setting scheme based on the total amount of light source scheduling of the light source setting unit and all supplementary light demand instructions;

Obtain the expected satisfaction degree of supplementary light for each patch component under the initial light source setting scheme;

Based on the preset minimum expected satisfaction threshold of fill light, adjust the initial light source setting scheme until the expected fill light satisfaction of each patch component is not lower than the minimum fill light expected satisfaction threshold, and the adjustment is completed, and the target light source setting scheme is obtained;

Complete the light source setting according to the target light source setting scheme.

As an embodiment of the present invention, a multi-chip component detection system also includes:

If the target light source setting scheme cannot be obtained within the preset time, a warning message will be issued;

At the same time, based on the preset minimum fill light expectation satisfaction threshold and fill light maximum rectangle strategy, adjust the initial light source setting scheme until the fill light expectation satisfaction that can satisfy each patch component is not lower than the minimum fill light expectation satisfaction The maximum rectangle formed by the SMD components with the threshold value ends, and the second light source setting scheme is generated;

Complete the light source setting according to the second light source setting scheme;

Obtain the supplementary light demand command ignored in the second light source setting scheme as the ignore command, and send the ignore command to the image segmentation module.

As an embodiment of the present invention, the image segmentation module includes:

an original image preprocessing unit, configured to perform an image preprocessing operation on the original image to obtain the first image;

The image segmentation unit is used to process the first image based on the improved polygon approximation method, calculate the contour distance from the center of gravity of the processed first image at the same time, and segment the first image based on the distance from the contour to the center of gravity to obtain several A monolithic image of the patch.

As an embodiment of the present invention, the image detection module includes:

A patch angle detection unit, configured to determine the rotation angle of a single patch contained in a single image;

The patch position information detection unit is configured to determine coordinate information of a single patch included in the single-slice image.

As an embodiment of the present invention, the patch angle detection unit includes:

The image processing sub-unit is used to perform grayscale processing, filtering processing, binarization processing, Harris corner detection processing, morphological pin removal processing and contour extraction processing on the single-chip image to obtain the main body contour of the single patch ;

The straight line extraction sub-unit is used to perform progressive straight line Hough transform on the outline of the subject to obtain the rotation angle of the single patch included in the single slice image.

As an embodiment of the present invention, the patch position information detection unit performs the following operations:

Based on the edge tracking algorithm, extract the minimum circumscribed rectangle of a single image;

Locate any corner point as the origin of the coordinate system, and calculate the center coordinates of the smallest circumscribed rectangle;

According to the center coordinates, the rotation angle and the preset information of the patch itself, the coordinate information of the single patch included in the single image is calculated.

A method for detecting multi-chip components, comprising: automatically collecting original images of multi-chip components;

Segment the original image to obtain several single-slice images containing only a single patch;

The coordinate information and the rotation angle of the single patch included in the single image are calculated.

The beneficial effects of the present invention are:

The invention provides a multi-chip component detection system and method, which are used to accurately detect the coordinate information and rotation angle of the chip, and have high precision, which can meet the actual working requirements of the chip mounter.

Additional features and advantages of the invention will be set forth in the description which follows, and in part will be apparent from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention may be realized and attained by the structure particularly pointed out in the written description and appended drawings.

The technical solutions of the present invention will be described in further detail below with reference to the accompanying drawings and embodiments.

Description of drawings

The accompanying drawings are used to provide a further understanding of the present invention, and constitute a part of the description, and are used together with the embodiments of the present invention to explain the present invention, and do not constitute a limitation to the present invention. In the attached picture:

1 is a schematic diagram of a system module of a multi-chip component detection system and method in an embodiment of the present invention;

Fig. 2 is a unit schematic diagram of an image acquisition module in a multi-chip component detection system and method in an embodiment of the present invention;

FIG. 3 is a flow chart of a multi-chip component detection system and method in an embodiment of the present invention.

Detailed ways

The preferred embodiments of the present invention will be described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described here are only used to illustrate and explain the present invention, and are not intended to limit the present invention.

Please refer to Fig. 1, an embodiment of the present invention provides a multi-chip component detection system, including: an image acquisition module 1, which is used to automatically collect the original image of the multi-chip component;

The image segmentation module 2 is used to segment the original image to obtain several single-slice images that only include a single patch;

The image detection module 3 is used to calculate the coordinate information and the rotation angle of the single patch contained in the single image;

The working principle and beneficial effects of the above technical solutions are as follows: the multi-chip component detection system is mainly composed of two parts, the first part is the hardware part, mainly the image acquisition module 1, which is mainly responsible for collecting the semiconductor chips that need to be detected within the field of view Image, the second part is the software part, which mainly includes the image segmentation module 2 and the image detection module 3. The main flow of the system during implementation is: firstly, the image acquisition module 1 is used to acquire the image of the chip patch on the platform PCB, and the establishment Automatic image acquisition process, high-quality acquisition of multi-chip images, image preprocessing of the acquired rectangular component images, removal of pin interference processing to obtain the main body outline, and then the use of progressive straight line Hough transform algorithm for contour line extraction, and finally The angle of the component is obtained through the straight line information; an edge tracking algorithm is used for the above-mentioned processed image to find the starting point of the edge and track the mark, then extract the minimum circumscribed rectangle, and finally calculate the center coordinate of the chip patch by defining the coordinate system. Then calculate the coordinates of the chip edge point according to the component angle;

Among them, the progressive straight line Hough transform algorithm is used to extract the straight line of the patch outline, and the method of obtaining the component angle has the advantages of fast processing speed and high efficiency compared with the direct detection of the Hough transform straight line, which can be applied to other high-precision angle detection methods. occasion;

Build a multi-chip component detection system, use the image preprocessing algorithm to remove the image background, noise and pin interference. The method of obtaining the outline of the main body is compared with the traditional method of detecting the center of mass of the pin. This method uses the pre-image processing method. Detection deviation caused by incomplete symmetry or slight deformation of pins can be excluded;

The edge tracking algorithm can extract the enclosing relationship of the connected regions in the binary image, and obtain the topological structure of the image edge. Compared with the general image processing method, it can eliminate the connected regions or holes in the image according to the geometric or topological characteristics. Reduce background interference and improve the accuracy of patch position detection.

Referring to Fig. 2, in one embodiment, the image acquisition module 1 includes:

a placement unit 11, configured to place and fix multi-chip components, and send a first signal;

The light source setting unit 12 is configured to set a corresponding light source according to the type and orientation of the multi-chip components after receiving the first signal, and send a second signal;

The camera unit 13 is used to collect the original image of the multi-chip component after receiving the second signal;

The working principle and beneficial effects of the above technical solutions are as follows: the image acquisition module is an important part of the detection system, and the quality of the collected patch component images has a direct impact on the subsequent image processing and position detection, so it is necessary to use suitable industrial cameras and Appropriate light sources, traditional manual visual detection of patch angles and patch character areas have relatively large limitations, not only the speed of patch detection is slow and the efficiency is low, but also when the detection time is long, it is prone to fatigue, resulting in inaccurate detection results. Accurate, the image acquisition module of this detection system is used in the traditional manual visual inspection method to perform high-quality and high-efficiency positioning detection work on the patch angle and chip character area information. When used for actual collection, the chip is placed in the corresponding model In the tray, use the placement unit 11 to fix the tray and place it horizontally on the placement unit 11. The image acquisition work is completed by the light source setting unit 12 and the camera unit 13. The camera unit 13 includes but not limited to hardware such as industrial cameras and lenses. The parameters include but are not limited to camera frame rate, camera exposure value, camera trigger mode and other parameters. The light source setting unit 12 is used to be responsible for the light source and the initialization of the light source. After the image is acquired through the camera SDK function, the original image is sent by the first sending unit Import the image segmentation module and establish an efficient image access mechanism to reduce unnecessary memory usage and improve software efficiency.

In one embodiment, the light source setting unit performs the following operations:

Identify the category of each patch component in the multi-chip component according to the characteristic information, and obtain multiple pieces of category information;

Obtain the orientation information of each patch component on the placement unit, and obtain multiple pieces of orientation information;

Based on multiple pieces of category information and multiple pieces of orientation information, perform association judgments to obtain the degree of association; wherein, the degree of association indicates the degree to which different types of information are affected by light sources in different orientations;

Based on the degree of correlation, determine the state of light to be filled for each patch component;

Carry out one-to-one correspondence between the category information, orientation information and the state of light to be filled for each patch component, and form a demand command for light fill;

Complete the light source setting according to the fill light demand command;

The beneficial effect of the above technical solution is: through the above technical solution, it is better to set the light source for each patch component in the multi-chip components, and to ensure that the image acquisition unit can collect the original image of each patch component with high quality. image.

In one embodiment, the light source setting is completed according to the supplementary light requirement instruction, including:

According to each supplementary light demand instruction, the expected satisfaction degree of supplementary light for each patch component is generated; under the initial condition, the expected satisfaction degree of supplementary light for each patch component is empty; where, when any patch component’s When the fill light demand command is perfectly executed, the corresponding fill light expectation satisfaction is full;

Adaptively generate an initial light source setting scheme based on the total amount of light source scheduling of the light source setting unit and all supplementary light demand instructions;

Obtain the expected satisfaction degree of supplementary light for each patch component under the initial light source setting scheme;

Based on the preset minimum expected satisfaction threshold of fill light, adjust the initial light source setting scheme until the expected fill light satisfaction of each patch component is not lower than the minimum fill light expected satisfaction threshold, and the adjustment is completed, and the target light source setting scheme is obtained;

Complete the light source setting according to the target light source setting scheme;

The beneficial effect of the above-mentioned technical solution is: through the above-mentioned technical solution, the configuration of the light source can be reasonably scheduled to prevent the impact of the chip components that first complete the light-filling requirement due to insufficient light source resources from occupying too many resources, resulting in the inability of the subsequent chip components to complete the light-filling requirement High-quality image acquisition.

In one embodiment, a multi-chip component detection system also includes:

If the target light source setting scheme cannot be obtained within the preset time, a warning message will be issued;

At the same time, based on the preset minimum fill light expectation satisfaction threshold and fill light maximum rectangle strategy, adjust the initial light source setting scheme until the fill light expectation satisfaction that can satisfy each patch component is not lower than the minimum fill light expectation satisfaction The maximum rectangle formed by the SMD components with the threshold value ends, and the second light source setting scheme is generated;

Among them, the supplementary light maximum rectangle strategy is to obtain the largest rectangle formed by the original multi-chip components that can satisfy the fill-light expectation satisfaction of each patch component is not lower than the minimum fill-light expectation satisfaction threshold , other SMD components outside the largest rectangle will not be filled with light, and all light source resources will be used to fill light with other SMD components within the largest rectangle;

Complete the light source setting according to the second light source setting scheme;

Obtain the supplementary light demand instruction ignored in the second light source setting scheme as the ignore instruction, and send the ignore instruction to the image segmentation module;

Furthermore, the patch components involved in the instruction are ignored and sent to the workshop to be processed, waiting for subsequent processing;

The beneficial effects of the above technical solution are: through the above technical solution, the continuity of detection is improved, and the fault tolerance rate of the system is enhanced at the same time. When there is a problem with the light source, the collection range is automatically reduced to continue collection, and there is no need to wait for the user to respond to maintenance before continuing to work , and at the same time send an ignore command to the image segmentation module to prevent the subsequent image segmentation module from performing invalid processing on the part of the components and improve the processing efficiency.

In one embodiment, the image segmentation module includes:

an original image preprocessing unit, configured to perform an image preprocessing operation on the original image to obtain the first image;

The image segmentation unit is used to process the first image based on the improved polygon approximation method, calculate the contour distance from the center of gravity of the processed first image at the same time, and segment the first image based on the distance from the contour to the center of gravity to obtain several monolithic image of the patch;

The working principle and beneficial effects of the above technical solution are as follows: the original image preprocessing unit is used to perform image preprocessing operations on the original image, which includes but not limited to operations such as enhancing, smoothing, and denoising the original image to obtain the first image , at the same time, the image segmentation unit is used to process the first image based on the improved polygon approximation method, and at the same time, the contour distance from the center of gravity of the processed first image is calculated, and the first image is segmented based on the distance from the contour to the center of gravity to obtain several The single-slice image of a single patch improves the subsequent accurate detection of each patch.

In one embodiment, the image detection module includes:

A patch angle detection unit, configured to determine the rotation angle of a single patch contained in a single image;

A patch position information detection unit, configured to determine coordinate information of a single patch contained in a single image;

The working principle and beneficial effects of the above technical solution are as follows: through the patch angle detection unit and the patch position information detection unit, the coordinate information and rotation angle of the patch can be accurately detected, so that the system can meet the actual working requirements of the chip mounter.

In one embodiment, the patch angle detection unit includes:

The image processing sub-unit is used to perform grayscale processing, filtering processing, binarization processing, Harris corner detection processing, morphological pin removal processing and contour extraction processing on the single-chip image to obtain the main body contour of the single patch ;

The straight line extraction subunit is used to perform progressive straight line Hough transform on the subject contour to obtain the rotation angle of the single patch contained in the single image;

The working principle and beneficial effects of the above technical solution are as follows: patch angle detection mainly includes image preprocessing and straight line extraction. Image preprocessing includes grayscale, filtering, binarization, morphological pin removal, and Harris corner detection. Processing and contour extraction; straight line extraction performs progressive straight line Hough transformation on the subject contour to obtain the rotation angle value of the component, wherein the image processing subunit performs the following operations:

Grayscale processing, the pictures collected by the visual system are usually color pictures, which cannot reflect the morphological characteristics of the image, so it is necessary to grayscale the image for later processing; in order to improve the clarity of the image, the weighted average method is used to process the components The image is grayscaled, and the color image is weighted and averaged according to the three components of R, G, and B with different weights to obtain the grayscale image of the component;

Filtering processing. Due to the influence of the working environment during the collection process, there may be noise in the image. In order to remove the influence of noise on image processing, filtering processing is required. The Gaussian filtering method is commonly used to process the image. The adaptive Gaussian filtering method can be used according to The local features of the image are filtered with different thresholds for different regions, and the information of the detailed regions can be maintained while removing noise;

Based on Harris corner detection, at present, corner detection methods are mainly divided into two types based on image edge features and image grayscale. The detection performance of corner detection methods based on image edge features is relatively stable, but it is sensitive to local changes in edge contours. An accurate threshold cannot be given to extract corner points; the corner point detection method based on image grayscale is simple to calculate and has low time complexity, but is sensitive to noise and isolated points;

There are 5 common corner types: L-type, T-type, Y-type, X-type, and T-type. In this embodiment, the corner points in the target outline are all L-type corner points, so when pairing chip pins/ In the plastic package image correction process, the L-shaped corners are directly processed, which can greatly reduce the calculation amount of corner detection and image correction;

Compared with other corner detection methods, the Harris corner detection algorithm is simple to calculate and easy to implement, which is beneficial to reduce costs;

There may be some discrete points after image binarization. The present invention uses morphological processing to accurately extract the outline of the main body of the component; uses the opening operation to process the image, and the opening operation is to first corrode and then expand, and has the ability to smooth the target outline and remove isolated points. , The overall position and shape are unchanged; Morphological processing cannot completely remove the pins, so in order to obtain the main body contour of the rectangular component, pin removal and contour extraction are also required;

Pin removal includes: scanning and counting the area of each closed area in the picture, setting the area of the main body area as the parameter S, and removing the closed area with an area smaller than the parameter S, so as to remove the pin to obtain the main body of the component;

Contour extraction includes: using the eight-neighborhood contour tracking method to extract the outline of the main body of the component. The basic idea is that if a point in the image is white and its 8 adjacent points are all white, remove it to obtain the outline of the main body;

The straight line extraction subunit is used to perform progressive straight line Hough transform on the outline of the subject to obtain the rotation angle of a single patch contained in a single image. The traditional Hough transform uses the duality of points and lines to pass a given point in the image space through The form of the curve expression is mapped to a point in the parameter space, and the problem of image line detection is transformed into the problem of finding the peak in the parameter space. In the image space, the gray value of M×N image pixels is set as I( _xi ,y _i ); in the parameter space, H(ρ _q ,θ _k ) is the accumulator unit of the pixel gray value of all points, and the value of ρ The number of samples is q, and a straight line detection with a step distance of Δ is performed on θ in the interval [0, π), where the step distance Δ is inversely proportional to the detection accuracy; in order to solve the problem that the traditional Hough transform algorithm cannot meet the problem of fast processing under high precision , use the progressive method of step distance from large to small to detect the line, and finally obtain the angle of the patch component through the line information;

Wherein, the step of progressive Hough transform is:

(1) Carry out a rough detection of a straight line with a step distance of Δ ₁ in the interval [0, π) for θ;

(2) from the detected straight line, select the optimal straight line, get its angle θ _α with the x-axis;

范围内进行步距为Δ₂的直线精检测，取Δ₁>>Δ₂(相差一到两个数量级)；(3) at θ _α

Within the range, the step distance is Δ ₂ for straight line precision detection, and Δ ₁ >> Δ ₂ (the difference is one to two orders of magnitude);

Among them, the progressive straight line Hough transform algorithm is used to extract the straight line of the patch outline, and the method of obtaining the component angle has the advantages of fast processing speed and high efficiency compared with the direct detection of the Hough transform straight line, which can be applied to other high-precision angle detection methods. occasion;

Build a multi-chip component detection system, use the image preprocessing algorithm to remove the image background, noise and pin interference. The method of obtaining the outline of the main body is compared with the traditional method of detecting the center of mass of the pin. This method uses the pre-image processing method. Detection deviations caused by incomplete symmetry or slight deformation of the pins can be ruled out.

In one embodiment, the patch position information detection unit performs the following operations:

Based on the edge tracking algorithm, extract the minimum circumscribed rectangle of a single image;

Locate any corner point as the origin of the coordinate system, and calculate the center coordinates of the smallest circumscribed rectangle;

According to the center coordinates, the rotation angle and the preset information of the patch itself, the coordinate information of the single patch included in the single image is calculated;

The working principle and beneficial effects of the above technical solution are: based on the edge tracking algorithm, the specific steps of extracting the minimum circumscribed rectangle of the monolithic image are preferably: 1. Find the starting point of the edge, scan the monolithic image, and find the pixel point that meets the starting point of the edge 2. Calculate the parent edge of the newly discovered edge; in the image scanning stage, make the edge serial numbers as close as possible; 3. Start tracking the edge from the starting point; 4, extract the minimum circumscribed rectangle, what the above process obtains is the edge point sequence of the object in the image, and each point in the sequence is arranged clockwise, and extracts the smallest rectangle containing each edge point in the sequence;

Locate any corner point as the origin of the coordinate system, calculate the center coordinates of the smallest circumscribed rectangle, preferably locate the upper left corner of the single image as the origin of the coordinate system, determine the orientation of the object in the image, and calculate its center coordinates through the smallest rectangle where the object is located;

Finally, according to the center coordinates, the rotation angle and the preset information of the patch itself, the coordinate information of the single patch contained in the single image is calculated;

Among them, the edge tracking algorithm can extract the enclosing relationship of the connected regions in the binary image, and obtain the topological structure of the image edge. Compared with the general image processing method, it can eliminate the connected regions or Holes reduce background interference and improve the accuracy of patch position detection.

Please refer to Fig. 3, a method for detecting multi-chip components, including: S101, automatically collecting original images of multi-chip components;

S102. Segment the original image to obtain several single-slice images that only include a single patch;

S103. Calculate and obtain the coordinate information and the rotation angle of the single patch contained in the single image.

The working principle and beneficial effects of this multi-chip component detection method can refer to the corresponding working principle and beneficial effects mentioned in the above-mentioned functional modules of a multi-chip component detection system, and will not be repeated here. Tired.

Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention also intends to include these modifications and variations.

Claims (10)

1. A multi-patch component detection system, comprising:

the image acquisition module is used for automatically acquiring the original images of the multi-patch components;

the image segmentation module is used for segmenting the original image to obtain a plurality of single-chip images only containing single patches;

and the image detection module is used for calculating and obtaining the coordinate information and the rotation angle of the single patch contained in the single-chip image.

2. The multi-chip component detection system of claim 1, wherein the image acquisition module comprises:

the placing unit is used for placing and fixing the multi-patch components and sending out a first signal;

the light source setting unit is used for setting corresponding light sources according to the category and the azimuth of the multi-patch component after receiving the first signal and sending out a second signal;

and the camera shooting unit is used for acquiring an original image of the multi-patch component after receiving the second signal.

3. The multi-chip component inspection system according to claim 2, wherein the light source setting unit performs operations comprising:

carrying out category identification on each patch element in the multi-patch elements according to the characteristic information to obtain a plurality of pieces of category information;

acquiring azimuth information of each patch element on the placement unit to obtain a plurality of pieces of azimuth information;

based on the multiple pieces of category information and the multiple pieces of azimuth information, carrying out association judgment to obtain association degree; the association degree indicates the influence degree of the light source on different directions of different types of information;

determining the light supplementing state of each patch element based on the association degree;

the category information, the azimuth information and the state to be supplemented with light of each patch element are in one-to-one correspondence to form a light supplementing demand instruction;

and finishing the light source setting according to the light supplementing demand instruction.

4. A multi-patch component detection system as claimed in claim 3, wherein the light source setting is performed according to the light supplement demand instruction, comprising:

generating light filling expected satisfaction of each patch component according to each light filling demand instruction; under the initial condition, the light filling expected satisfaction degree of each patch element is null; when the light supplementing demand instruction of any one of the patch components is perfectly executed, the corresponding light supplementing expected satisfaction degree is a full value;

adaptively generating an initial light source setting scheme based on the light source scheduling total and all light supplementing demand instructions of the light source setting unit;

acquiring the light filling expected satisfaction degree of each patch component under the initial light source setting scheme;

based on a preset minimum light filling expected satisfaction threshold, adjusting an initial light source setting scheme until the light filling expected satisfaction of each patch element is not lower than the minimum light filling expected satisfaction threshold, and finishing adjustment to obtain a target light source setting scheme;

and finishing the light source setting according to the target light source setting scheme.

5. The multi-chip component inspection system according to claim 4, further comprising:

if the target light source setting scheme cannot be obtained within the preset time, sending out warning information;

meanwhile, based on a preset minimum light filling expected satisfaction threshold and a light filling maximum rectangle strategy, an initial light source setting scheme is adjusted until the largest rectangle formed by the patch components which can meet that the light filling expected satisfaction of each patch component is not lower than the minimum light filling expected satisfaction threshold is ended, and a second light source setting scheme is generated;

completing the light source setting according to the second light source setting scheme;

and acquiring an neglected light supplementing demand instruction in the second light source setting scheme as an neglected instruction, and sending the neglected instruction to the image segmentation module.

6. The multi-chip component inspection system of claim 1, wherein the image segmentation module comprises:

the original image preprocessing unit is used for performing image preprocessing operation on the original image to obtain a first image;

the image segmentation unit is used for processing the first image based on the improved polygon approximation method, calculating the contour distance center of gravity of the processed first image, and segmenting the first image based on the contour distance center of gravity to obtain a plurality of single-chip images only comprising single patches.

7. The multi-chip component inspection system of claim 1, wherein the image inspection module comprises:

a patch angle detection unit for determining a rotation angle of a single patch included in the single-sheet image;

and the patch position information detection unit is used for determining coordinate information of the single patch contained in the single-chip image.

8. The multi-chip component inspection system according to claim 7, wherein the chip angle inspection unit comprises:

the image processing subunit is used for respectively carrying out gray processing, filtering processing, binarization processing, harris corner detection processing, morphological pin removal processing and contour extraction processing on the single-chip image to obtain a main contour of the single patch;

and the linear extraction subunit is used for carrying out progressive linear Hough transformation on the main body outline to obtain the rotation angle of the single patch contained in the single-chip image.

9. The multi-chip component inspection system according to claim 7, wherein the chip position information inspection unit performs operations comprising:

extracting the minimum circumscribed rectangle of the single-chip image based on an edge tracking algorithm;

positioning any corner point as an origin of a coordinate system, and calculating to obtain a center coordinate of the minimum circumscribed rectangle;

and calculating to obtain the coordinate information of the single patch contained in the single-chip image according to the center coordinate, the rotation angle and the preset patch self information.

10. The method for detecting the multi-patch component is characterized by comprising the following steps of:

automatically collecting an original image of the multi-patch component;

dividing the original image to obtain a plurality of single-chip images only containing single patches;

and calculating to obtain the coordinate information and the rotation angle of the single patch contained in the single-chip image.

Images