An assembled printed circuit board inspection system based on machine vision

Abstract

The present invention discloses an assembled printed circuit board inspection system based on machine vision, which relates to the field of product inspection and comprises: a camera module for collecting printed circuit board surface image data, preprocessing and contour extraction of the printed circuit board surface image data, and outputting a printed circuit board contour image; a derivation module for receiving the printed circuit board contour image, picking up all contours in the printed circuit board contour image, and deriving the printed circuit board contour topology based on the picked contours; the present invention accurately collects the printed circuit board surface image, constructs the contour topology through preprocessing and contour extraction, and then compares it with the standard topology to achieve qualified judgment. At the same time, it can monitor the qualified results in real time, estimate assembly risks, accurately identify defective areas and present them visually. The overall system improves the comparison accuracy through multi-level topology derivation, uses historical data to predict risks, and combines grid segmentation to locate defects, thereby greatly improving the inspection efficiency and accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of product detection, and in particular to an assembled printed circuit board detection system based on machine vision. Background Art

[0002] Fabricated printed circuit boards (PCBs) are rapidly assembled from multiple prefabricated circuit modules using standardized interfaces. They simplify production processes, shorten R&D cycles, facilitate repair and replacement, and accommodate customization and large-scale production needs. They are widely used in electronic equipment manufacturing.

[0003] The invention patent application with application number 202410785382.7 discloses a printed circuit board inspection system, including: an automatic optical circuit measurement device, which obtains circuit size information of a printed circuit board by photographing a circuit image of the printed circuit board; and a circuit characteristic confirmation device, which is connected to the automatic optical circuit measurement device to obtain the circuit size information; wherein, the circuit characteristic confirmation device obtains electrical characteristic information of the printed circuit board based on circuit reference characteristic information of a test sample and the circuit size information of the printed circuit board. This application aims to solve the problem that "when using a contact test method, if the PCB impedance value does not meet the standard, it can only be known that some areas or a certain line have defects, but the actual location of the defect cannot be known to perform repair work. In addition, when using a contact test method, on a printed circuit board or a test sample, each test point needs to be tested, and it takes a lot of time to complete the test of the entire board, resulting in reduced production efficiency."

[0004] However, compared with the appearance inspection of assembled printed circuit boards, the existing technology uses machine vision for inspection, and the inspection and analysis only stays at the circuit board image level. The application of implicit information in the circuit board image is relatively low, resulting in limited room for improvement in the accuracy of existing machine vision inspection technology in such scenarios.

[0005] Therefore, a machine vision-based assembly printed circuit board inspection system is proposed. Summary of the Invention

[0006] In view of the above-mentioned shortcomings of the prior art, the present invention provides an assembled printed circuit board inspection system based on machine vision, which can effectively solve the problems of the prior art.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0008] The present invention discloses a machine vision-based assembly-type printed circuit board detection system, comprising:

[0009] A camera module is used to collect surface image data of a printed circuit board, pre-process the surface image data of the printed circuit board and extract contours to output a contour image of the printed circuit board; a derivation module is used to receive the contour image of the printed circuit board, pick up all contours in the contour image of the printed circuit board, and derive the contour topology of the printed circuit board based on the picked contours; a comparison module is used to receive the contour topology of the printed circuit board output by the derivation module, and determine whether the printed circuit board from which the contour topology of the printed circuit board is derived is qualified based on the comparison of the contour topology of the printed circuit board with the contour topology of the standard printed circuit board; a monitoring module is used to monitor and record the qualified judgment results in the comparison module in real time, and when it is monitored that the printed circuit board is continuously qualified, the recorded historical qualified judgment results are used to estimate the assembly risk of the printed circuit board; an identification module is used to identify defective areas on the printed circuit board; an interaction module is used to receive the identification results of the defective areas on the printed circuit board in the identification module, and to visualize the defective areas on the printed circuit board.

[0010] Furthermore, after the camera module collects the printed circuit board surface image data, the preprocessing of the printed circuit board surface image data includes:

[0011] Denoising printed circuit board surface image data;

[0012] Setting a color segmentation interval, traversing all pixels in the printed circuit board surface image data based on the color segmentation interval, picking pixels that meet the color segmentation interval, using the picked pixels as segmentation targets, and performing a segmentation operation on the printed circuit board surface image data to obtain an image of the printed circuit board surface area;

[0013] During the operation phase of the camera module, the printed circuit board surface area image is used to perform contour extraction to output the printed circuit board contour image;

[0014] When the printed circuit board is manufactured and output by the assembly equipment, it is transmitted on a conveyor belt. The color of the conveyor belt surface is different from the color of the printed circuit board surface. The background of the printed circuit board surface image data is the conveyor belt surface. The segmentation operation performed on the printed circuit board surface image data is a background removal operation on the printed circuit board surface image data.

[0015] Furthermore, all contours in the PCB contour image picked up by the derivation module include: closed figures, contour lines, and contour points. After completing the contour picking operation, the derivation module performs derivation of the PCB contour topology according to the derivation level customized or initially set by the system end user;

[0016] First-level derivation: among all contour points, all combinations of two into a group are used to connect the two contour points contained in each group of contour points; among all contour lines, all combinations of two into a group are used to connect the midpoints of the two contour lines contained in each group of contour lines; among all closed figures, all combinations of two into a group are used to connect the center points of the two closed figures contained in each group of closed figures. The combination of all connecting lines is recorded as the printed circuit board contour topology derivation result;

[0017] Secondary derivation: After completing the first-level derivation, select any two sets from all closed figures, all contour lines, and all contour points, pick up a closed figure, contour line, or contour point in each of the two sets in all possible combinations, and connect the application points again according to the closed figures, contour lines, or contour points in the first-level derivation to complete the newly added connection lines. The combination of all connection lines is recorded as the PCB outline topology derivation result;

[0018] Level 3 derivation: After completing the level 2 derivation, the operation of selecting any two sets from all closed graphics, all contour lines, and all contour points is executed again, except for the set combination selected in the level 2 derivation. Based on the level 2 derivation execution logic, new connection lines are added again, and the combination of all connection lines is recorded as the derivation result of the printed circuit board outline topology.

[0019] Furthermore, the derivative module is connected to an editing module, which is integrated by a two-dimensional drawing application, and the editing module is used to upload printed circuit board components and circuit distribution parameters, and construct a printed circuit board standard outline graphic based on the printed circuit board components and circuit distribution parameters;

[0020] After the standard outline graphic of the printed circuit board is constructed, three standard printed circuit board outline topologies are outputted synchronously based on the three derivative levels;

[0021] The printed circuit board standard outline graphic and the three derived levels corresponding to the standard printed circuit board outline topology are forwarded to the derived module and stored in the derived module.

[0022] Furthermore, during the operation phase of the comparison module, after receiving the printed circuit board outline topology, the derived level of the printed circuit board outline topology source is synchronously identified, and the standard printed circuit board outline topology of the corresponding level is retrieved in the editing module. The offset degree is compared based on the received printed circuit board outline topology and the retrieved standard printed circuit board outline topology, and the comparison result is further compared with a preset qualified judgment threshold. When the comparison result is less than or equal to the qualified judgment threshold, the printed circuit board from which the printed circuit board outline topology originated is judged to be qualified; otherwise, it is judged to be unqualified.

[0023] Furthermore, the judgment logic for monitoring the printed circuit board to be continuously qualified in the monitoring module is: the judgment result output by the comparison module is three times in a row and maintained;

[0024] The printed circuit board assembly risk estimation operation in the monitoring module is:

[0025] When the qualified judgment result in the comparison module of each record is updated, the qualified judgment results in the comparison module of the latest three records are retrieved;

[0026] ;

[0027] Where: The corresponding deviation degree of the qualified judgment results in the comparison module of the latest three records;

[0028] in, Arrange based on record time sequence, The qualified judgment result in the comparison module corresponding to the latest record points to the degree of deviation. When any one or both of the above formulas are true, it means that there is a risk in the printed circuit board assembly process. Otherwise, it means that there is no risk in the printed circuit board assembly process.

[0029] Furthermore, the identification module is controlled and run by a system end user or is triggered to run when the monitoring module monitors that there is a risk in the printed circuit board assembly process.

[0030] Furthermore, the recognition module's logic for identifying defective areas on a printed circuit board is as follows:

[0031] The estimated results are the corresponding printed circuit board outline topology and the standard printed circuit board outline topology of the printed circuit board assembly process that have risks. The printed circuit board outline topology and the standard printed circuit board outline topology are respectively configured with BBoxes of the same specifications so that the printed circuit board outline topology and the standard printed circuit board outline topology are both within the BBox. The BBox is meshed and segmented so that the size of each segmented sub-BBox is equal. Then, the printed circuit board outline topology and the standard printed circuit board outline topology are respectively segmented into several sub-topologies based on the mesh segmentation of the BBox, which are recorded as and ;

[0032] right and Pair and get ,right Perform the offset comparison again for each combination in the , and obtain the combination with the highest offset;

[0033] The sub-topology derived from the printed circuit board outline topology in the combination with the highest degree of deviation points to an area in the printed circuit board outline topology, that is, a defective area on the printed circuit board identified by the operation of the recognition module.

[0034] Furthermore, during the operation stage of the interactive module, the printed circuit board surface image data collected by the camera module, the circuit board outline topology output by the derivative module, and the defective area on the printed circuit board identified by the recognition module are obtained, the circuit board outline topology is overlaid on the corresponding area in the printed circuit board surface image data, and the area corresponding to the defective area on the printed circuit board is rendered within the area covered by the circuit board outline topology in the printed circuit board surface image data.

[0035] Furthermore, the camera module is interactively connected to a derivative module via a wireless network, the derivative module is interactively connected to the editing module and the comparison module via a wireless network, the editing module is interactively connected to the comparison module via a wireless network, the comparison module is interactively connected to the monitoring module via a wireless network, the comparison module and the monitoring module are interactively connected to the identification module via a wireless network, and the identification module is connected to the interaction module via a wireless network.

[0036] Compared with the prior art, the technical solution provided by the present invention has the following beneficial effects:

[0037] The present invention provides an assembled printed circuit board inspection system based on machine vision. During operation, the system accurately captures the surface image of the printed circuit board, constructs the contour topology through preprocessing and contour extraction, and then compares it with the standard topology to achieve qualified judgment. At the same time, it can monitor the qualified results in real time, estimate assembly risks, accurately identify defective areas and visualize them. The system as a whole improves the comparison accuracy through multi-level topology derivation, uses historical data to predict risks, and combines grid segmentation to locate defects, greatly improving the detection efficiency and accuracy, reducing manual intervention, and providing comprehensive and reliable technical support for printed circuit board assembly quality control, effectively reducing the outflow of defective products, and reducing production risks and costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort.

[0039] Figure 1 The figure is a structural diagram of an assembled printed circuit board inspection system based on machine vision. DETAILED DESCRIPTION

[0040] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0041] The present invention will be further described below with reference to the embodiments.

[0042] Example:

[0043] This embodiment is a machine vision-based assembly type printed circuit board detection system, such as Figure 1 As shown, including:

[0044] The camera module is used to collect printed circuit board surface image data, pre-process the printed circuit board surface image data and extract the outline of the printed circuit board to output the printed circuit board outline image;

[0045] After the camera module collects the printed circuit board surface image data, the preprocessing of the printed circuit board surface image data includes:

[0046] Denoising printed circuit board surface image data;

[0047] Setting a color segmentation interval, traversing all pixels in the printed circuit board surface image data based on the color segmentation interval, picking pixels that meet the color segmentation interval, using the picked pixels as segmentation targets, and performing a segmentation operation on the printed circuit board surface image data to obtain an image of the printed circuit board surface area;

[0048] During the camera module operation phase, the printed circuit board surface area image is used to perform contour extraction to output the printed circuit board contour image.

[0049] When the printed circuit board is manufactured and output by the assembly equipment, it is transported on a conveyor belt. The color of the conveyor belt surface is different from the color of the printed circuit board surface. The background of the printed circuit board surface image data is the conveyor belt surface. The segmentation operation performed on the printed circuit board surface image data is a background removal operation on the printed circuit board surface image data.

[0050] A derivation module is used for receiving a printed circuit board outline image, picking up all outlines in the printed circuit board outline image, and deriving a printed circuit board outline topology based on the picked outlines;

[0051] All contours in the PCB outline image picked up by the derivation module include: closed graphics, contour lines, and contour points. After completing the contour picking operation, the derivation module performs derivation of the PCB outline topology according to the system-side user-defined or initially set derivation level;

[0052] First-level derivation: among all contour points, all combinations of two into a group are used to connect the two contour points contained in each group of contour points; among all contour lines, all combinations of two into a group are used to connect the midpoints of the two contour lines contained in each group of contour lines; among all closed figures, all combinations of two into a group are used to connect the center points of the two closed figures contained in each group of closed figures. The combination of all connecting lines is recorded as the printed circuit board contour topology derivation result;

[0053] Secondary derivation: After completing the first-level derivation, select any two sets from all closed figures, all contour lines, and all contour points, pick up a closed figure, contour line, or contour point in each of the two sets in all possible combinations, and connect the application points again according to the closed figures, contour lines, or contour points in the first-level derivation to complete the newly added connection lines. The combination of all connection lines is recorded as the PCB outline topology derivation result;

[0054] Level 3 derivation: After completing the second level derivation, the operation of selecting any two sets from all closed graphics, all contour lines, and all contour points is executed again, excluding the set combination selected in the second level derivation. Based on the second level derivation execution logic, new connection lines are added again, and the combination of all connection lines is recorded as the PCB outline topology derivation result;

[0055] The derivative module is connected to an editing module, which is integrated with a two-dimensional drawing application. The editing module is used to upload printed circuit board components and circuit distribution parameters, and build a standard outline graphic of the printed circuit board based on the printed circuit board components and circuit distribution parameters.

[0056] After the standard outline graphics of the printed circuit board are constructed, three standard printed circuit board outline topologies are outputted synchronously based on three derived levels;

[0057] The printed circuit board standard outline graphic and the three derived levels corresponding to the standard printed circuit board outline topology are forwarded to the derivative module and stored in the derivative module;

[0058] a comparison module, configured to receive the printed circuit board outline topology output by the derivative module, and determine whether the printed circuit board from which the printed circuit board outline topology originates is qualified based on the comparison between the printed circuit board outline topology and the standard printed circuit board outline topology;

[0059] During the operation phase of the comparison module, after receiving the printed circuit board outline topology, the derived level of the printed circuit board outline topology source is synchronously identified, and the standard printed circuit board outline topology of the corresponding level is retrieved in the editing module. A deviation comparison is performed based on the received printed circuit board outline topology and the retrieved standard printed circuit board outline topology. The comparison result is further compared with a preset qualification threshold. When the comparison result is less than or equal to the qualification threshold, the printed circuit board of the source of the printed circuit board outline topology is determined to be qualified; otherwise, it is determined to be unqualified.

[0060] The degree of deviation of the PCB outline topology from the standard PCB outline topology is subject to:

[0061] Taking the standard printed circuit board outline topology as a reference, the printed circuit board outline topology is moved to align with the reference (the alignment condition is the BBox boundary of the topology), that is, the deviation degree of the printed circuit board outline topology A relative to the standard printed circuit board outline topology B is calculated;

[0062] ;

[0063] Where: is the weight coefficient; is the number of nodes in the printed circuit board outline topology A; Outline the nodes in topology A for the printed circuit board Its corresponding node in the standard printed circuit board outline topology B The position offset measure is the Euclidean distance between two points; is the number of edges in the PCB outline topology A; A binary variable used to measure the offset of edge connectivity; is the representation of an edge in the printed circuit board outline topology A, That is, the i-th node in the printed circuit board outline topology A is the basic unit that constitutes the printed circuit board outline topology A. Similarly;

[0064] in, are all positive numbers, and their sum is 1;

[0065] ;

[0066] The set of edges in the standard printed circuit board outline topology B; Represents the standard printed circuit board outline topology B with corresponding edges;

[0067] It should be noted that , which can be understood as a fictitious Completely coincident edges;

[0068] A monitoring module is used to monitor and record the qualified judgment results in the comparison module in real time. When the printed circuit board is monitored to be continuously qualified, the recorded historical qualified judgment results are used to estimate the risk of printed circuit board assembly;

[0069] The judgment logic for the continuous qualification of the printed circuit board monitored in the monitoring module is: the judgment result output by the comparison module is three times in a row and maintained;

[0070] The risk estimation operation of printed circuit board assembly in the monitoring module is:

[0071] When the qualified judgment result in the comparison module of each record is updated, the qualified judgment results in the comparison module of the latest three records are retrieved;

[0072] ;

[0073] Where: The corresponding deviation degree of the qualified judgment results in the comparison module of the latest three records;

[0074] in, Arrange based on record time sequence, The qualified judgment result in the comparison module corresponding to the latest record points to the degree of deviation. When any one or both of the above formulas are true, it means that there is a risk in the printed circuit board assembly process. Otherwise, it means that there is no risk in the printed circuit board assembly process.

[0075] an identification module for identifying defective areas on a printed circuit board;

[0076] The identification module is controlled by the system end user or is triggered to run when the monitoring module detects that there is a risk in the printed circuit board assembly process;

[0077] The recognition module's logic for identifying defective areas on printed circuit boards is as follows:

[0078] The estimated results are the corresponding printed circuit board outline topology and the standard printed circuit board outline topology of the printed circuit board assembly process that have risks. The printed circuit board outline topology and the standard printed circuit board outline topology are respectively configured with BBoxes of the same specifications so that the printed circuit board outline topology and the standard printed circuit board outline topology are both within the BBox. The BBox is meshed and segmented so that the size of each segmented sub-BBox is equal. Then, the printed circuit board outline topology and the standard printed circuit board outline topology are respectively segmented into several sub-topologies based on the mesh segmentation of the BBox, which are recorded as and ;

[0079] right and Pair and get ,right Perform the offset comparison again for each combination in the , and obtain the combination with the highest offset;

[0080] The sub-topology derived from the printed circuit board outline topology in the combination with the highest degree of deviation points to an area in the printed circuit board outline topology, that is, a defective area on the printed circuit board identified by the operation of the recognition module;

[0081] An interaction module is used to receive the defect area recognition result on the printed circuit board from the recognition module and to visualize the defect area on the printed circuit board;

[0082] During the interactive module operation phase, the printed circuit board surface image data collected by the camera module, the circuit board outline topology output by the derivative module, and the defective area on the printed circuit board identified by the recognition module are obtained, the circuit board outline topology is overlaid on the corresponding area in the printed circuit board surface image data, and the area corresponding to the defective area on the printed circuit board is rendered within the area covered by the circuit board outline topology in the printed circuit board surface image data;

[0083] The camera module is interactively connected to the derivative module through a wireless network, the derivative module is interactively connected to the editing module and the comparison module through a wireless network, the editing module is interactively connected to the comparison module through a wireless network, the comparison module is interactively connected to the monitoring module through a wireless network, the comparison module and the monitoring module are interactively connected to the identification module through a wireless network, and the identification module is connected to the interactive module through a wireless network.

[0084] In this embodiment, a camera module operates to collect surface image data of a printed circuit board (PCB), preprocesses and extracts contours of the PCB surface image data, and outputs a PCB contour image. A derivation module operates post-processing to receive the PCB contour image, extracts all contours in the PCB contour image, and derives a PCB contour topology based on the extracted contours. An editing module further uploads PCB component and circuit distribution parameters, and constructs a standard PCB contour graphic based on the PCB component and circuit distribution parameters. A comparison module then receives the PCB contour topology output by the derivation module, compares the PCB contour topology with the standard PCB contour topology, and determines whether the PCB from which the PCB contour topology originated is qualified. The monitoring module monitors and records the qualified determination results in the comparison module in real time. When the PCB is monitored to be continuously qualified, the recorded historical qualified determination results are used to estimate the PCB assembly risk. The identification module simultaneously identifies defective areas on the PCB. Finally, the interaction module receives the defective area identification results on the PCB from the identification module and visualizes the defective areas on the PCB.

[0085] During operation, the system in the above embodiment can rapidly capture images of the circuit board surface and accurately extract its contours. Comparing these contours with standard contours allows for efficient compliance assessment, reducing the time required for manual inspection. Continuous monitoring of compliance status can predict assembly risks, promptly identify defective areas, and visualize them. This significantly shortens inspection cycles, reduces false positives and missed detections, enables early detection of risks, and improves the quality and efficiency of circuit board assembly.

[0086] It should be noted that:

[0087] For the denoising of the printed circuit board surface image data, any one of the existing technologies may be used to perform the denoising process, so as to achieve the purpose of enhancing the image quality of the printed circuit board surface image data and removing noise;

[0088] The selection of the derivative level in the derivative module depends on the accuracy requirements of the system end user. The higher the accuracy requirement, the higher the derivative level can be selected as the operating logic of the derivative module.

[0089] In summary, during operation, the system in the above embodiment accurately captures the surface image of the printed circuit board, constructs the contour topology through preprocessing and contour extraction, and then compares it with the standard topology to achieve qualified judgment. At the same time, it can monitor the qualified results in real time, estimate assembly risks, accurately identify defective areas and visualize them. The system as a whole improves the comparison accuracy through multi-level topology derivation, uses historical data to predict risks, and combines grid segmentation to locate defects, greatly improving detection efficiency and accuracy, reducing manual intervention, and providing comprehensive and reliable technical support for printed circuit board assembly quality control, effectively reducing the outflow of defective products, and reducing production risks and costs.

[0090] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A machine vision-based assembly printed circuit board inspection system, characterized in that: include: The camera module is used to collect printed circuit board surface image data, pre-process the printed circuit board surface image data and extract the outline of the printed circuit board to output the printed circuit board outline image; A derivation module is used for receiving a printed circuit board outline image, picking up all outlines in the printed circuit board outline image, and deriving a printed circuit board outline topology based on the picked outlines; All contours in the PCB contour image picked up by the derivation module include: closed figures, contour lines, and contour points. After completing the contour picking operation, the derivation module performs derivation of the PCB contour topology according to the derivation level customized or initially set by the system end user; First-level derivation: among all contour points, all combinations of two into a group are used to connect the two contour points contained in each group of contour points; among all contour lines, all combinations of two into a group are used to connect the midpoints of the two contour lines contained in each group of contour lines; among all closed figures, all combinations of two into a group are used to connect the center points of the two closed figures contained in each group of closed figures. The combination of all connecting lines is recorded as the printed circuit board contour topology derivation result; Secondary derivation: After completing the first-level derivation, select any two sets from all closed figures, all contour lines, and all contour points, pick up a closed figure, contour line, or contour point in each of the two sets in all possible combinations, and connect the application points again according to the closed figures, contour lines, or contour points in the first-level derivation to complete the newly added connection lines. The combination of all connection lines is recorded as the PCB outline topology derivation result; Level 3 derivation: After completing the second level derivation, the operation of selecting any two sets from all closed graphics, all contour lines, and all contour points is executed again, excluding the set combination selected in the second level derivation. Based on the second level derivation execution logic, new connection lines are added again, and the combination of all connection lines is recorded as the PCB outline topology derivation result; a comparison module, configured to receive the printed circuit board outline topology output by the derivative module, and determine whether the printed circuit board from which the printed circuit board outline topology originates is qualified based on the comparison between the printed circuit board outline topology and the standard printed circuit board outline topology; A monitoring module is used to monitor and record the qualified judgment results in the comparison module in real time. When the printed circuit board is monitored to be continuously qualified, the recorded historical qualified judgment results are used to estimate the risk of printed circuit board assembly; an identification module for identifying defective areas on a printed circuit board; The interactive module is used to receive the defective area recognition result on the printed circuit board from the recognition module and to visually represent the defective area on the printed circuit board.

2. The machine vision-based assembly printed circuit board detection system according to claim 1, characterized in that: After the camera module collects the printed circuit board surface image data, the preprocessing of the printed circuit board surface image data includes: Denoising printed circuit board surface image data; Setting a color segmentation interval, traversing all pixels in the printed circuit board surface image data based on the color segmentation interval, picking pixels that meet the color segmentation interval, using the picked pixels as segmentation targets, and performing a segmentation operation on the printed circuit board surface image data to obtain an image of the printed circuit board surface area; During the operation phase of the camera module, the printed circuit board surface area image is used to perform contour extraction to output the printed circuit board contour image; When the printed circuit board is manufactured and output by the assembly equipment, it is transmitted on a conveyor belt. The color of the conveyor belt surface is different from the color of the printed circuit board surface. The background of the printed circuit board surface image data is the conveyor belt surface. The segmentation operation performed on the printed circuit board surface image data is a background removal operation on the printed circuit board surface image data.

3. The machine vision-based assembly printed circuit board detection system according to claim 1, characterized in that: The derivative module is connected to an editing module, which is integrated with a two-dimensional drawing application program, and is used to upload printed circuit board components and circuit distribution parameters, and construct a standard outline graphic of the printed circuit board based on the printed circuit board components and circuit distribution parameters; After the standard outline graphic of the printed circuit board is constructed, three standard printed circuit board outline topologies are outputted synchronously based on three derivative levels; The printed circuit board standard outline graphic and the three derived levels corresponding to the standard printed circuit board outline topology are forwarded to the derived module and stored in the derived module.

4. The machine vision-based assembly printed circuit board inspection system according to claim 1, characterized in that: During the operation phase of the comparison module, after receiving the printed circuit board outline topology, the derived level of the printed circuit board outline topology source is synchronously identified, and the standard printed circuit board outline topology of the corresponding level is retrieved in the editing module. The offset degree is compared based on the received printed circuit board outline topology and the retrieved standard printed circuit board outline topology, and the comparison result is further compared with a preset qualified judgment threshold. When the comparison result is less than or equal to the qualified judgment threshold, the printed circuit board from which the printed circuit board outline topology originated is judged to be qualified; otherwise, it is judged to be unqualified.

5. The machine vision-based assembly printed circuit board inspection system according to claim 1, characterized in that: The judgment logic for the monitoring module to monitor the printed circuit board to be continuously qualified is: the judgment result output by the comparison module is three times in a row and maintained; The printed circuit board assembly risk estimation operation in the monitoring module is: When the qualified judgment result in the comparison module of each record is updated, the qualified judgment results in the comparison module of the latest three records are retrieved; ； Where: The corresponding deviation degree of the qualified judgment results in the comparison module of the latest three records; in, Arrange based on record time sequence, The qualified judgment result in the comparison module corresponding to the latest record points to the degree of deviation. When any one or both of the above formulas are true, it means that there is a risk in the printed circuit board assembly process. Otherwise, it means that there is no risk in the printed circuit board assembly process.

6. The machine vision-based assembly printed circuit board inspection system according to claim 1, characterized in that: The identification module is controlled and run by the system end user or is triggered to run when the monitoring module monitors that there is a risk in the printed circuit board assembly process.

7. The machine vision-based assembly printed circuit board inspection system according to claim 1, characterized in that: The recognition logic of the defective area on the printed circuit board of the recognition module is as follows: The estimated results are the corresponding printed circuit board outline topology and the standard printed circuit board outline topology of the printed circuit board assembly process that have risks. The printed circuit board outline topology and the standard printed circuit board outline topology are respectively configured with BBoxes of the same specifications so that the printed circuit board outline topology and the standard printed circuit board outline topology are both within the BBox. The BBox is meshed and segmented so that the size of each segmented sub-BBox is equal. Then, the printed circuit board outline topology and the standard printed circuit board outline topology are respectively segmented into several sub-topologies based on the mesh segmentation of the BBox, which are recorded as and ; right and Pair and get ,right Perform the offset comparison again for each combination in the , and obtain the combination with the highest offset; The sub-topology derived from the printed circuit board outline topology in the combination with the highest degree of deviation points to an area in the printed circuit board outline topology, that is, a defective area on the printed circuit board identified by the operation of the recognition module.

8. The machine vision-based assembly printed circuit board inspection system according to claim 1, characterized in that: During the operation phase of the interactive module, the printed circuit board surface image data collected by the camera module, the circuit board outline topology output by the derivative module, and the defective area on the printed circuit board identified by the recognition module are obtained, the circuit board outline topology is overlaid on the corresponding area in the printed circuit board surface image data, and the area corresponding to the defective area on the printed circuit board is rendered within the area covered by the circuit board outline topology in the printed circuit board surface image data.

9. The machine vision-based assembly printed circuit board inspection system according to claim 1, characterized in that: The camera module is interactively connected to the derivative module via a wireless network, the derivative module is interactively connected to the editing module and the comparison module via a wireless network, the editing module is interactively connected to the comparison module via a wireless network, the comparison module is interactively connected to the monitoring module via a wireless network, the comparison module and the monitoring module are interactively connected to the identification module via a wireless network, and the identification module is connected to the interaction module via a wireless network.

Images