AOI-based micron-sized defect identification and detection system

Abstract

The invention discloses an AOI-based micron-sized defect identification and detection system, and relates to the technical field of defect identification and detection, and the system comprises the following modules: a board type database building module for building a board type database and incorporating images of each bare PCB and each assembled PCB into the board type database; the image value is obtained through the collected image of the PCB, then the image value and the measured value are combined to obtain the judgment coefficient, and whether the PCB is a bare board or an assembled board is judged by comparing the judgment coefficient with the judgment coefficient preset threshold value, so that the recognition accuracy of the PCB is improved, and the recognition accuracy of the PCB is improved by incorporating the judgment coefficient into the board type database. According to the method and the device, the judgment coefficient is obtained, so that the reference plate type image can be subsequently identified in the plate type database according to the judgment coefficient, the PCB can be conveniently subjected to reference comparison, the defects of the PCB can be conveniently identified, the PCB is identified, the defects of the PCB can be subsequently detected in a targeted manner, and the defect identification efficiency can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of defect recognition and detection, and in particular to a micron-level defect recognition and detection system based on AOI. Background Art

[0002] AOI, or automatic optical inspection, is a device that detects common defects encountered in welding production based on optical principles. AOI is a new type of testing technology that has just emerged, but it has developed rapidly. Many manufacturers have launched AOI inspection equipment. During automatic inspection, the machine automatically scans the PCB board through a camera, collects images, and compares the tested solder joints with the qualified parameters in the database. After image processing, the defects on the PCB board are checked and marked through a display or automatic mark for maintenance personnel to repair.

[0003] In the process of using AOI to detect PCB bare board defects, since most of the bare board defects are via position deviation, via size deviation, circuit breakage on the surface of the substrate, etc., the AOI system can give full play to its meticulous defect detection characteristics. However, most of the defects of the PCB assembly board after electronic components are assembled are electronic components that are misassembled or missing, etc. At this time, the AOI system will carefully proofread the scanned image of the PCB assembly board with the qualified product image in the database, which will not only seriously affect the recognition efficiency of PCB assembly board defects, but also easily cause frequent false alarms of defects. In summary, the AOI system lacks the ability of intelligent judgment and recognition in the process of identifying defects, and it is difficult to automatically adjust the detection standard according to the different missing parts of the identified objects, thereby affecting the defect recognition efficiency, and then affecting the production detection efficiency of PCB boards. For this reason, we propose a micron-level defect recognition and detection system based on AOI. Summary of the invention

[0004] 1. Technical issues to be resolved In view of the deficiencies in the prior art, the present invention provides a micron-level defect recognition and detection system based on AOI to solve the above-mentioned problems in the prior art.

[0005] (II) Technical solution To achieve the above objectives, the present invention is implemented through the following technical solutions: a micron-level defect recognition and detection system based on AOI, comprising the following modules: A board type database building module is used to build a board type database, include the images of each PCB bare board and each PCB assembled board into the board type database, define the inspection area on each PCB bare board according to the via position and line position of the PCB bare board, define the inspection area on each PCB assembled board according to the electronic component position of the PCB assembled board, and include the inspection area into the board type database; A distinguishing and identifying module scans the PCB board to obtain a board image, obtains an image value of the PCB board through the board image, obtains a measurement value of the PCB board, obtains a judgment coefficient of the PCB board through the image value of the PCB board and the measurement value of the PCB board, incorporates the judgment coefficient of the PCB board into a board type database, sets a preset threshold of the judgment coefficient, compares the judgment coefficient of the PCB board with the preset threshold of the judgment coefficient, and obtains a PCB board identification result according to the comparison result; The defect recognition module captures the reference board image in the board type database through the judgment coefficient, obtains the reference board detection area, marks the area to be detected on the board to be detected through the reference board detection area, performs defect recognition by calibrating the reference board detection area with the area to be detected, and marks the defect position according to the defect recognition result; The identification and decision module obtains the defect mark position and the defect mark quantity, obtains the scrap value according to the defect mark position and the defect mark quantity, and determines whether to continue defect identification according to the PCB board identification result and the scrap value.

[0006] Preferably, in the identification module, the image value of the PCB board is obtained through the plate image, specifically: Step 1: Obtain the number of color types of the plate image, grayscale the plate image, and obtain the number of silk screen logos, the number of circuits, and the number of via solder points according to the brightness of the plate grayscale image; Step 2: Obtain image values ​​through the number of color types of the plate image, the number of silk-screen logos of the plate grayscale image, the number of lines of the plate grayscale image, and the number of via solder points of the plate grayscale image; The image value is calculated as follows:

[0007] Among them, Tx represents the image value, Yg represents the number of color types, Sy represents the number of silk screen logos, Xs represents the number of lines, and Gx represents the number of via solder points. , , and are weights, , , , .

[0008] Preferably, the measurement value of the PCB board is obtained in the identification module, specifically: Step 1: Use a laser rangefinder to scan the highest point of the PCB board to get the highest point distance, and use a laser rangefinder to scan the lowest point of the PCB board to get the lowest distance. Subtract the lowest distance from the highest distance to get the thickness difference; Step 2: Obtain the weight of the PCB board, and obtain the measured value of the PCB board through the thickness difference and weight of the PCB board; The measurement value is calculated as follows:

[0009] Among them, Cn represents the measured value, Hd represents the thickness difference, and Zn represents the weight. and are weights, , .

[0010] Preferably, the calculation method of the judgment coefficient in the discrimination and recognition module is specifically as follows:

[0011] Among them, Pd represents the judgment coefficient, Tx represents the image value, and Cn represents the measurement value. and are weights, , .

[0012] Preferably, a preset threshold value of the judgment coefficient is set in the identification module, the judgment coefficient of the PCB board is compared with the preset threshold value of the judgment coefficient, and the PCB board identification result is obtained according to the comparison result, specifically: Obtaining a judgment coefficient, setting a preset threshold value of the judgment coefficient, and comparing the magnitude relationship between the judgment coefficient and the preset threshold value of the judgment coefficient; If the judgment coefficient is greater than the preset judgment coefficient threshold, the PCB board is marked as a PCB assembly board, and the recognition result is sent to the device; If the judgment coefficient is less than the preset threshold of the judgment coefficient, the PCB board is marked as a bare PCB board, and the recognition result is sent to the device.

[0013] Preferably, in the defect recognition module, the reference plate image is captured in the plate type database by judging the coefficient, specifically: Step 1: Obtain the judgment coefficient of the PCB board to be detected, set the judgment coefficient floating preset threshold, obtain the maximum judgment coefficient value by summing the judgment coefficient of the PCB board to be detected and the judgment coefficient floating preset threshold, obtain the minimum judgment coefficient value by subtracting the judgment coefficient of the PCB board to be detected from the judgment coefficient floating preset threshold, and obtain the judgment coefficient range by the minimum judgment coefficient value and the maximum judgment coefficient value; Step 2: Capture the judgment coefficients of all board types in the board type database based on the judgment coefficient range, obtain the board images corresponding to all the captured judgment coefficients in the board type database, mark them as optional images, obtain the image of the board to be tested, and overlap and compare the image of the board to be tested with multiple optional images respectively, and select the optional image with the highest overlap from multiple optional images by comparing the overlap of via positions and line positions and mark it as the reference board image.

[0014] Preferably, in the defect recognition module, the area to be detected is marked on the board to be detected by using the detection area of ​​the reference board, specifically: Step 1: Obtain the reference board detection area in the board type database, overlap the reference board image with the image of the board to be detected, and mark the area to be detected in the image of the board to be detected according to the reference board detection area; Step 2: Determine whether there are circuits, vias, or electronic components outside the area to be detected. If there are circuits, vias, or electronic components outside the area to be detected, feedback is given to the management personnel that there is no corresponding reference board image in the database.

[0015] Preferably, the defect mark position and the defect mark quantity are obtained in the identification and decision module, and the scrap value is obtained according to the defect mark position and the defect mark quantity, specifically: Step 1: Get the defect mark position, determine whether the defect position is via misalignment, via size is too large, circuit is broken, or electronic components are installed incorrectly. If the determination result is yes, mark the defect position as a major defect and record the major defect as 500. Determine whether the defect position is via size is too small, via is missing, or electronic components are missing. If the determination result is yes, mark the defect position as a minor defect and record the minor defect as 5. Step 2: Get the current number of major defect marks, get the current number of minor defect marks, and get the scrap value through the number of major defects and the number of minor defect marks; The calculation method of scrap value is as follows:

[0016] Among them, Bf represents the scrap value, Dq represents the number of major defects, and Xq represents the number of minor defects.

[0017] Preferably, in the identification decision module, it is determined whether to continue defect identification according to the PCB board identification result and the scrap value, specifically: Step 1: Get the scrap value, set the preset threshold of the scrap value, compare the size relationship between the scrap value and the preset threshold of the scrap value, if the scrap value is less than the preset threshold of the scrap value, continue to identify defects, if the scrap value is greater than the preset threshold of the scrap value, stop defect identification, and mark the PCB board as a pending board; Step 2: Obtain the PCB recognition result of the pending board. If the PCB recognition result of the pending board is a PCB assembly board, continue to identify defects. If the PCB recognition result of the pending board is a PCB bare board, terminate the defect recognition detection of the board and mark the board as a scrapped board.

[0018] (III) Beneficial effects The present invention provides a micron-level defect recognition and detection system based on AOI, which has the following beneficial effects: (1) In the discrimination and recognition module, the scheme obtains an image value through the captured image of the PCB board, and then combines the image value with the measurement value to obtain a judgment coefficient. The judgment coefficient is compared with a preset threshold value of the judgment coefficient to judge whether the PCB board is a bare board or an assembled board, which is beneficial to improve the recognition accuracy of the PCB board. By incorporating the judgment coefficient into the board type database, it is convenient to identify the reference board type image in the board type database according to the judgment coefficient, and then it is convenient to compare the PCB board with the reference, and then it is convenient to identify the defects of the PCB board. By identifying the PCB board, it is convenient to perform targeted defect detection on the PCB board in the future, which is beneficial to improve the defect recognition efficiency.

[0019] (2) In the recognition role module, the present solution obtains the scrap value through the position and number of defect marks, and then determines whether to continue defect recognition based on the PCB board recognition result and the scrap value, so as to facilitate timely stopping of defect recognition when irreparable defects such as short circuit or open circuit are identified, and then facilitate defect recognition of other PCB boards, which is beneficial to avoid wasting too much defect recognition time on scrapped boards, and further beneficial to intelligently improve the recognition efficiency of PCB board defects. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a structural schematic diagram of a micron-level defect recognition and detection system based on AOI of the present invention; Figure 2 This is a flowchart of a micron-level defect recognition and detection system based on AOI of the present invention. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions 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 creative work are within the scope of protection of the present invention.

[0022] See also Figure 1-Figure 2 The present invention provides a micron-level defect recognition and detection system based on AOI, comprising the following modules: A board type database building module is used to build a board type database, include the images of each PCB bare board and each PCB assembled board into the board type database, define the inspection area on each PCB bare board according to the via position and line position of the PCB bare board, define the inspection area on each PCB assembled board according to the electronic component position of the PCB assembled board, and include the inspection area into the board type database; A distinguishing and identifying module scans the PCB board to obtain a board image, obtains an image value of the PCB board through the board image, obtains a measurement value of the PCB board, obtains a judgment coefficient of the PCB board through the image value of the PCB board and the measurement value of the PCB board, incorporates the judgment coefficient of the PCB board into a board type database, sets a preset threshold of the judgment coefficient, compares the judgment coefficient of the PCB board with the preset threshold of the judgment coefficient, and obtains a PCB board identification result according to the comparison result; The defect recognition module captures the reference board image in the board type database through the judgment coefficient, obtains the reference board detection area, marks the area to be detected on the board to be detected through the reference board detection area, performs defect recognition by calibrating the reference board detection area with the area to be detected, and marks the defect position according to the defect recognition result; The identification and decision module obtains the defect mark position and the defect mark quantity, obtains the scrap value according to the defect mark position and the defect mark quantity, and determines whether to continue the defect identification according to the PCB board identification result and the scrap value; The calculation method of the judgment coefficient in the discrimination and recognition module is as follows:

[0023] Among them, Pd represents the judgment coefficient, Tx represents the image value, and Cn represents the measurement value. and are weights, , .

[0024] In this embodiment, in the board type database building module, this scheme collects the images of each qualified PCB bare board and each qualified PCB assembled board in the board type database, so as to facilitate reference comparison when detecting PCB boards, and delineates the detection area on each PCB bare board according to the via position and line position of the PCB bare board, so as to delineate the detection area according to the detection standard of the PCB bare board, and delineates the detection area on each PCB assembled board according to the electronic component position of the PCB assembled board, so as to facilitate delineating the detection area according to the detection standard of the PCB assembled board, and then facilitates the targeted detection of the delineated detection area after identifying the type of the PCB board, so as to avoid invalid detection in other areas of the PCB board, so as to improve the defect recognition efficiency, so as to improve the detection efficiency and reduce the situation of false alarm of defects; In the resolution and recognition module, the solution obtains an image value through the collected image of the PCB board, and then combines the image value with the measurement value to obtain a judgment coefficient. The judgment coefficient is compared with a preset threshold value of the judgment coefficient to judge whether the PCB board is a bare board or an assembled board, which is beneficial to improving the recognition accuracy of the PCB board. By incorporating the judgment coefficient into the board type database, it is convenient to subsequently identify the reference board type image in the board type database according to the judgment coefficient, and then it is convenient to make a reference comparison of the PCB board, and then it is convenient to identify the defects of the PCB board. By identifying the PCB board, it is convenient to subsequently perform targeted defect detection on the PCB board, which is beneficial to improving the defect recognition efficiency. In the defect recognition module, this solution captures the reference board image in the board type database through the judgment coefficient, and then maps the detection area of ​​the reference board on the board to be detected, so as to facilitate the subsequent targeted detection of the detection area of ​​the board to be detected, which is beneficial to improve the defect detection efficiency of the PCB board and reduce the false alarm of defects. Defect recognition is performed by calibrating the reference board detection area with the area to be detected, and the defect position is marked according to the defect recognition result; In the recognition role module, this solution obtains the scrap value through the position and number of defect marks, and then determines whether to continue defect recognition based on the PCB board recognition result and the scrap value, so as to stop defect recognition in time when unrepairable defects such as line short circuit or open circuit are identified, and then facilitate defect recognition of other PCB boards, which is beneficial to avoid wasting too much defect recognition time on scrapped boards, and further beneficial to intelligently improve the recognition efficiency of PCB board defects; It is worth mentioning that the weight values ​​in this scheme can be obtained by the hierarchical analysis method, and the values ​​of the preset thresholds in this scheme can be obtained by the weight analysis method, which will not be elaborated here. In the defect recognition module, identifying defects by comparing the image to be detected with the image in the database is a disclosed prior art, which will not be elaborated here. For specific operation steps, please refer to the Chinese patents with application numbers CN202210259462.X and CN202110582294.3.

[0025] In the identification module, the image value of the PCB board is obtained through the plate image, specifically: Step 1: Obtain the number of color types of the plate image, grayscale the plate image, and obtain the number of silk screen logos, the number of circuits, and the number of via solder points according to the brightness of the plate grayscale image; Step 2: Obtain image values ​​through the number of color types of the plate image, the number of silk-screen logos of the plate grayscale image, the number of lines of the plate grayscale image, and the number of via solder points of the plate grayscale image; The image value is calculated as follows:

[0026] Among them, Tx represents the image value, Yg represents the number of color types, Sy represents the number of silk screen logos, Xs represents the number of lines, and Gx represents the number of via solder points. , , and are weights, , , , .

[0027] In this embodiment, the plate image is gray-scaled, and the number of silk-screen marks, the number of circuits, and the number of via solder points in the plate image are identified and recorded based on the different characteristics of the positions of silk-screen marks in the gray-scale image and the different brightness of circuits and via solder points. The plate image is then identified based on the characteristics of the different colors of various electronic components and PCB plates in the plate image before gray-scale processing. The image value finally obtained makes it easier to combine multiple parameters to accurately determine whether the board to be inspected is a bare board or an assembled board.

[0028] The measurement values ​​of the PCB board are obtained in the identification module, specifically: Step 1: Use a laser rangefinder to scan the highest point of the PCB board to get the highest point distance, and use a laser rangefinder to scan the lowest point of the PCB board to get the lowest distance. Subtract the lowest distance from the highest distance to get the thickness difference; Step 2: Obtain the weight of the PCB board, and obtain the measured value of the PCB board through the thickness difference and weight of the PCB board; The measurement value is calculated as follows:

[0029] Among them, Cn represents the measured value, Hd represents the thickness difference, and Zn represents the weight. and are weights, , .

[0030] In this embodiment, the measurement value is obtained by the thickness difference and weight of the PCB board, so that it is convenient to distinguish and identify the bare board and the assembled board according to the characteristics of the assembled board with large thickness and large weight, which is beneficial to improve the recognition accuracy of the PCB bare board and assembled board.

[0031] In the identification module, a preset threshold of the judgment coefficient is set, and the judgment coefficient of the PCB board is compared with the preset threshold of the judgment coefficient. The PCB board identification result is obtained according to the comparison result, which is specifically: Obtaining a judgment coefficient, setting a preset threshold value of the judgment coefficient, and comparing the magnitude relationship between the judgment coefficient and the preset threshold value of the judgment coefficient; If the judgment coefficient is greater than the preset judgment coefficient threshold, the PCB board is marked as a PCB assembly board, and the recognition result is sent to the device; If the judgment coefficient is less than the preset threshold of the judgment coefficient, the PCB board is marked as a bare PCB board, and the recognition result is sent to the device.

[0032] In this embodiment, the board to be inspected is identified and judged by comparing the judgment coefficient of the PCB board with the preset threshold of the judgment coefficient, thereby facilitating determination of whether the board to be inspected is a bare board or an assembled board, thereby facilitating subsequent targeted defect identification and inspection of the board to be inspected.

[0033] In the defect recognition module, the reference board image is captured in the board type database through the judgment coefficient, specifically: Step 1: Obtain the judgment coefficient of the PCB board to be detected, set the judgment coefficient floating preset threshold, obtain the maximum judgment coefficient value by summing the judgment coefficient of the PCB board to be detected and the judgment coefficient floating preset threshold, obtain the minimum judgment coefficient value by subtracting the judgment coefficient of the PCB board to be detected from the judgment coefficient floating preset threshold, and obtain the judgment coefficient range by the minimum judgment coefficient value and the maximum judgment coefficient value; Step 2: Capture the judgment coefficients of all board types in the board type database based on the judgment coefficient range, obtain the board images corresponding to all the captured judgment coefficients in the board type database, mark them as optional images, obtain the image of the board to be tested, and overlap and compare the image of the board to be tested with multiple optional images respectively, and select the optional image with the highest overlap from multiple optional images by comparing the overlap of via positions and line positions and mark it as the reference board image.

[0034] In this embodiment, the judgment coefficient range is obtained by judging the coefficient and the judgment coefficient floating preset threshold, and the corresponding image for selection is found in the board type database according to the judgment coefficient range, and then the overlap comparison is performed based on the image for selection and the image of the board to be detected, and the similarity of the via position size and the line direction is compared to select the reference board image with the highest overlap, so as to facilitate the rapid capture of the corresponding reference board image in the board type database, avoid wasting defect detection time by comparing one by one, and further help improve the efficiency of defect detection; It is worth mentioning that in the process of overlapping and comparing the selected image with the image of the plate to be detected, the cosine similarity model can be selected for comparison and judgment.

[0035] In the defect recognition module, the inspection area of ​​the reference board is used to mark the inspection area on the inspection board, specifically: Step 1: Obtain the reference board detection area in the board type database, overlap the reference board image with the image of the board to be detected, and mark the area to be detected in the image of the board to be detected according to the reference board detection area; Step 2: Determine whether there are circuits, vias, or electronic components outside the area to be detected. If there are circuits, vias, or electronic components outside the area to be detected, feedback is given to the management personnel that there is no corresponding reference board image in the database.

[0036] In this embodiment, by judging whether there are circuits, vias, or electronic components outside the area to be inspected, it is convenient to judge whether the selected reference board image is consistent with the model of the board to be inspected, which is beneficial to avoid false defect detection caused by inconsistency between the reference board image and the model of the board to be inspected, and facilitates secondary confirmation of the reference board image before defect detection, which is further beneficial to improving the accuracy and efficiency of defect detection.

[0037] In the identification and decision module, the defect mark position and the defect mark quantity are obtained, and the scrap value is obtained according to the defect mark position and the defect mark quantity, which is specifically: Step 1: Get the defect mark position, determine whether the defect position is via misalignment, via size is too large, circuit is broken, or electronic components are installed incorrectly. If the determination result is yes, mark the defect position as a major defect and record the major defect as 500. Determine whether the defect position is via size is too small, via is missing, or electronic components are missing. If the determination result is yes, mark the defect position as a minor defect and record the minor defect as 5. Step 2: Get the current number of major defect marks, get the current number of minor defect marks, and get the scrap value through the number of major defects and the number of minor defect marks; The calculation method of scrap value is as follows:

[0038] Among them, Bf represents the scrap value, Dq represents the number of major defects, and Xq represents the number of minor defects.

[0039] In this embodiment, by performing defect judgment on the defect mark position, it is convenient to stop defect identification in time when an irreparable defect such as a short circuit or a break is identified, thereby facilitating defect identification of other PCB boards, thereby helping to avoid wasting too much defect identification time on scrapped boards; It is worth mentioning that when judging the cause of the defect by the defect editing position, the target detection network model can be used to detect and identify the cause of the defect.

[0040] In the identification decision module, it is determined whether to continue defect identification based on the PCB board identification result and the scrap value, specifically: Step 1: Get the scrap value, set the preset threshold of the scrap value, compare the size relationship between the scrap value and the preset threshold of the scrap value, if the scrap value is less than the preset threshold of the scrap value, continue to identify defects, if the scrap value is greater than the preset threshold of the scrap value, stop defect identification, and mark the PCB board as a pending board; Step 2: Obtain the PCB recognition result of the pending board. If the PCB recognition result of the pending board is a PCB assembly board, continue to identify defects. If the PCB recognition result of the pending board is a PCB bare board, terminate the defect recognition detection of the board and mark the board as a scrapped board.

[0041] In this embodiment, by comparing the size relationship between the scrap value and the preset threshold value of the scrap value, it is convenient to confirm whether the PCB board is a pending board, and then identify whether the pending board is an assembled board or a bare board according to the PCB identification result. If the pending board is a bare board, the defect detection gate can be stopped to discard the scrapped board. If the pending board is an assembled board, the defect detection will continue. After all defects are detected, it is determined according to the detection situation whether to repair the defective assembled board or to disassemble and recycle the electronic components and then discard the substrate, thereby improving the detection efficiency of PCB boards while fully recycling defective PCB materials.

[0042] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product. A person of ordinary skill in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution.

[0043] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0044] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A micron-level defect recognition and detection system based on AOI, characterized in that: Includes the following modules: A board type database building module is used to build a board type database, include the images of each PCB bare board and each PCB assembled board into the board type database, define the inspection area on each PCB bare board according to the via position and line position of the PCB bare board, define the inspection area on each PCB assembled board according to the electronic component position of the PCB assembled board, and include the inspection area into the board type database; A distinguishing and identifying module scans the PCB board to obtain a board image, obtains an image value of the PCB board through the board image, obtains a measurement value of the PCB board, obtains a judgment coefficient of the PCB board through the image value of the PCB board and the measurement value of the PCB board, incorporates the judgment coefficient of the PCB board into a board type database, sets a preset threshold of the judgment coefficient, compares the judgment coefficient of the PCB board with the preset threshold of the judgment coefficient, and obtains a PCB board identification result according to the comparison result; The defect recognition module captures the reference board image in the board type database through the judgment coefficient, obtains the reference board detection area, marks the area to be detected on the board to be detected through the reference board detection area, performs defect recognition by calibrating the reference board detection area with the area to be detected, and marks the defect position according to the defect recognition result; The identification and decision module obtains the defect mark position and the defect mark quantity, obtains the scrap value according to the defect mark position and the defect mark quantity, and determines whether to continue defect identification according to the PCB board identification result and the scrap value.

2. The micron-level defect recognition and detection system based on AOI according to claim 1, characterized in that: In the identification module, the image value of the PCB board is obtained through the plate image, specifically: Step 1: Obtain the number of color types of the plate image, grayscale the plate image, and obtain the number of silk screen logos, the number of circuits, and the number of via solder points according to the brightness of the plate grayscale image; Step 2: Obtain image values ​​through the number of color types of the plate image, the number of silk-screen logos of the plate grayscale image, the number of lines of the plate grayscale image, and the number of via solder points of the plate grayscale image; The image value is calculated as follows: Among them, Tx represents the image value, Yg represents the number of color types, Sy represents the number of silk screen logos, Xs represents the number of lines, and Gx represents the number of via solder points. , , and are weights, , , , .

3. The micron-level defect recognition and detection system based on AOI according to claim 1, characterized in that: The measurement values ​​of the PCB board are obtained in the identification module, specifically: Step 1: Use a laser rangefinder to scan the highest point of the PCB board to get the highest point distance, and use a laser rangefinder to scan the lowest point of the PCB board to get the lowest distance. Subtract the lowest distance from the highest distance to get the thickness difference; Step 2: Obtain the weight of the PCB board, and obtain the measured value of the PCB board through the thickness difference and weight of the PCB board; The measurement value is calculated as follows: Among them, Cn represents the measured value, Hd represents the thickness difference, and Zn represents the weight. and are weights, , .

4. The micron-level defect recognition and detection system based on AOI according to claim 1, characterized in that: The calculation method of the judgment coefficient in the discrimination and recognition module is as follows: Among them, Pd represents the judgment coefficient, Tx represents the image value, and Cn represents the measurement value. and are weights, , .

5. The micron-level defect recognition and detection system based on AOI according to claim 1, characterized in that: In the identification module, a preset threshold of the judgment coefficient is set, and the judgment coefficient of the PCB board is compared with the preset threshold of the judgment coefficient. The PCB board identification result is obtained according to the comparison result, which is specifically: Obtaining a judgment coefficient, setting a preset threshold value of the judgment coefficient, and comparing the magnitude relationship between the judgment coefficient and the preset threshold value of the judgment coefficient; If the judgment coefficient is greater than the preset judgment coefficient threshold, the PCB board is marked as a PCB assembly board, and the recognition result is sent to the device; If the judgment coefficient is less than the preset threshold of the judgment coefficient, the PCB board is marked as a bare PCB board, and the recognition result is sent to the device.

6. The micron-level defect recognition and detection system based on AOI according to claim 1, characterized in that: In the defect recognition module, the reference board image is captured in the board type database through the judgment coefficient, specifically: Step 1: Obtain the judgment coefficient of the PCB board to be detected, set the judgment coefficient floating preset threshold, obtain the maximum judgment coefficient value by summing the judgment coefficient of the PCB board to be detected and the judgment coefficient floating preset threshold, obtain the minimum judgment coefficient value by subtracting the judgment coefficient of the PCB board to be detected from the judgment coefficient floating preset threshold, and obtain the judgment coefficient range by the minimum judgment coefficient value and the maximum judgment coefficient value; Step 2: Capture the judgment coefficients of all board types in the board type database based on the judgment coefficient range, obtain the board images corresponding to all the captured judgment coefficients in the board type database, mark them as optional images, obtain the image of the board to be tested, and overlap and compare the image of the board to be tested with multiple optional images respectively, and select the optional image with the highest overlap from multiple optional images by comparing the overlap of via positions and line positions and mark it as the reference board image.

7. The micron-level defect recognition and detection system based on AOI according to claim 1, characterized in that: In the defect recognition module, the inspection area of ​​the reference board is used to mark the inspection area on the inspection board, specifically: Step 1: Obtain the reference board detection area in the board type database, overlap the reference board image with the image of the board to be detected, and mark the area to be detected in the image of the board to be detected according to the reference board detection area; Step 2: Determine whether there are circuits, vias, or electronic components outside the area to be detected. If there are circuits, vias, or electronic components outside the area to be detected, feedback is given to the management personnel that there is no corresponding reference board image in the database.

8. The micron-level defect recognition and detection system based on AOI according to claim 1, characterized in that: In the identification and decision module, the defect mark position and the defect mark quantity are obtained, and the scrap value is obtained according to the defect mark position and the defect mark quantity, which is specifically: Step 1: Get the defect mark position, determine whether the defect position is via misalignment, via size is too large, circuit is broken, or electronic components are installed incorrectly. If the determination result is yes, mark the defect position as a major defect and record the major defect as 500. Determine whether the defect position is via size is too small, via is missing, or electronic components are missing. If the determination result is yes, mark the defect position as a minor defect and record the minor defect as 5. Step 2: Get the current number of major defect marks, get the current number of minor defect marks, and get the scrap value through the number of major defects and the number of minor defect marks; The calculation method of scrap value is as follows: Among them, Bf represents the scrap value, Dq represents the number of major defects, and Xq represents the number of minor defects.

9. The micron-level defect recognition and detection system based on AOI according to claim 1, characterized in that: In the identification decision module, it is determined whether to continue defect identification based on the PCB board identification result and the scrap value, specifically: Step 1: Get the scrap value, set the preset threshold of the scrap value, compare the size relationship between the scrap value and the preset threshold of the scrap value, if the scrap value is less than the preset threshold of the scrap value, continue to identify defects, if the scrap value is greater than the preset threshold of the scrap value, stop defect identification, and mark the PCB board as a pending board; Step 2: Obtain the PCB recognition result of the pending board. If the PCB recognition result of the pending board is a PCB assembly board, continue to identify defects. If the PCB recognition result of the pending board is a PCB bare board, terminate the defect recognition detection of the board and mark the board as a scrapped board.

Images