AOI detection-based printed circuit board key region positioning method and system

Abstract

The embodiment of the application provides a circuit board key area positioning method and system based on AOI detection, relates to the technical field of electronic manufacturing detection, and first acquires images and generates results based on the association of a circuit board design drawing and detection visual angle information by an AOI detection unit; then generates a feature association set by associating design drawing key area marking information and image area features; further generates positioning rules defining corresponding relations, precision requirements and abnormal processes in combination with the feature association set; the AOI detection unit positions key areas according to the positioning rules and generates results; finally, based on the results, the parameter and the positioning rules are calibrated in reverse linkage, instructions are generated, and an iterative collection mechanism is formed to form a bidirectional driving process. The application improves positioning accuracy and stability.

Description

Technical Field

[0001] This application relates to the field of electronic manufacturing inspection technology, and more specifically, to a method and system for locating key areas of circuit boards based on AOI inspection. Background Technology

[0002] In the electronics manufacturing industry, circuit boards (PCBs) are core components of electronic products, and their quality directly affects the performance and reliability of the entire product. Locating critical areas on PCBs is a crucial step in the PCB inspection process. Traditional methods for locating critical areas mainly rely on manual visual inspection or simple image matching algorithms. Manual visual inspection is not only inefficient but also easily affected by the subjective factors of the inspectors, leading to unstable positioning accuracy and failing to meet the demands of large-scale production. While methods based on simple image matching algorithms improve positioning efficiency to some extent, these algorithms often have high requirements for image quality and shooting angle. In actual production environments, variations in PCB placement and shooting angles can easily lead to image matching failures, thus failing to accurately locate critical areas. Furthermore, most existing methods lack self-calibration and optimization mechanisms. Once positioning deviations occur during inspection, they cannot be adjusted and corrected in a timely manner, further affecting the accuracy and stability of positioning. Summary of the Invention

[0003] In view of this, the purpose of this application is to provide a method and system for locating critical areas of circuit boards based on AOI detection.

[0004] According to a first aspect of this application, a method for locating critical areas of a circuit board based on AOI detection is provided, the method comprising: The AOI detection unit uses the association information between the circuit board design drawings and the detection perspective to collect circuit board images and generate linked acquisition results. The linked acquisition results include images of each area of ​​the circuit board, detection perspective parameters, and the correspondence between the images and the design drawings. Based on the results of the joint acquisition, the key area annotation information in the circuit board design drawings is associated with the area features in the image to generate a set of associated features of the key area of ​​the circuit board. By combining the feature association set of key areas of the circuit board, a key area positioning rule for the circuit board is generated. The key area positioning rule for the circuit board defines the correspondence between image features and key areas, positioning accuracy control requirements, and abnormal positioning processing procedures. The AOI detection unit locates the key areas of the circuit board image in the linkage acquisition results according to the key area positioning rules of the circuit board, and generates the key area positioning results of the circuit board. Based on the location results of the key areas of the circuit board, the linkage acquisition parameters of the AOI detection unit and the location rules of the key areas of the circuit board are calibrated in reverse to generate parameter calibration instructions. The AOI detection unit executes the parameter calibration instructions and iterative linkage acquisition mechanism to form a two-way driven process of positioning and calibration.

[0005] According to a second aspect of this application, a circuit board critical area positioning system based on AOI detection is provided. The circuit board critical area positioning system based on AOI detection includes a processor and a readable storage medium. The readable storage medium stores a program that, when executed by the processor, implements the aforementioned circuit board critical area positioning method based on AOI detection.

[0006] Based on any of the above aspects, the AOI inspection unit acquires circuit board images in conjunction with the circuit board design drawings and inspection perspective association information, generating a linked acquisition result. This linked acquisition result includes images of each region, inspection perspective parameters, and the correspondence between the images and the design drawings. Secondly, based on the linked acquisition result, the key region annotation information in the design drawings is correlated with the regional features in the images to generate a key region feature association set, further refining the feature information of the key regions. Thirdly, the key region feature association set is combined to generate circuit board key region positioning rules, clarifying the correspondence between image features and key regions, positioning accuracy control requirements, and abnormal positioning handling procedures. Then, the AOI inspection unit locates the key regions of the circuit board image according to the positioning rules and generates positioning results. Finally, based on the positioning results, the linked acquisition parameters and positioning rules of the AOI inspection unit are calibrated in reverse, generating parameter calibration instructions and iterating the linked acquisition mechanism. This allows the entire positioning system to continuously optimize and adjust itself according to the actual inspection situation, effectively improving positioning accuracy and stability, and reducing positioning errors caused by environmental changes, equipment errors, and other factors. Attached Figure Description

[0007] Figure 1 A flowchart illustrating the method for locating critical areas of a circuit board based on AOI detection provided in an embodiment of this application is shown. Figure 2 A schematic diagram of the component structure of the AOI-based critical area positioning system for circuit boards provided in an embodiment of this application is shown. Detailed Implementation

[0008] Figure 1 A flowchart illustrating the method for locating critical areas of a circuit board based on AOI detection provided in this application is shown, and the detailed steps are described below.

[0009] Step S110: The AOI detection unit acquires circuit board images and generates linked acquisition results based on the circuit board design drawings and the association information between the detection viewpoints. The linked acquisition results include images of each area of ​​the circuit board, detection viewpoint parameters, and the correspondence between the images and the design drawings.

[0010] The AOI inspection unit retrieves circuit board design drawings stored on a local server via a data interface. These drawings are in CAD format and include the location coordinates of various areas of the circuit board, circuit layout, and annotations of key areas. Simultaneously, the AOI inspection unit acquires inspection viewpoint association information through its own sensors, including deployment location coordinates, inspection angle adjustment range, and image acquisition range under different viewpoints. The AOI inspection unit associates the circuit board design drawings with the inspection viewpoint association information, calculates the coverage and clarity of each key area under different viewpoints, selects suitable inspection viewpoint combinations, adjusts the inspection angle and acquisition distance, and simultaneously initiates multi-view image acquisition operations, acquiring images of various areas of the circuit board and recording the corresponding inspection viewpoint parameters for each acquisition. The acquired images are then matched with the design drawings, and the corresponding circuit board area, inspection viewpoint parameters, and location in the design drawings are labeled for each image. Blurred, occluded, or incomplete areas in the acquired images are identified, and the viewpoint is readjusted to supplement the acquisition. All acquired images, association matching results, and supplementary acquired images are integrated, duplicate images are removed, association information is supplemented, and a linked acquisition result is generated.

[0011] Step S111: Retrieve the circuit board design drawings through the AOI detection unit, extract the key area annotation information, circuit layout and the relationship between the dimensions of each area in the circuit board design drawings, and generate a circuit board design information set.

[0012] The AOI inspection unit retrieves the circuit board design drawings from the server via a data interface. These drawings contain information such as the overall layout of the circuit board, the location of each component, and the connection relationships of the circuits. The AOI inspection unit parses the design drawings, extracting key area annotation information, including the type of key area (such as pads, circuit intersections, chip mounting areas, etc.), location coordinates, and design dimensions; extracting circuit layout information, including the extension direction of the circuits, the connection method between circuits, and the connection relationship between circuits and key areas; and extracting the dimensional relationships between different areas, including the size ratio between key areas and surrounding areas, and the dimensional relationships between sub-areas within key areas. All the extracted information is integrated into a circuit board design information set. Each piece of data in this set contains fields such as key area identifier, type, location coordinates, design dimensions, circuit layout direction, and dimensional relationships.

[0013] Step S112: Obtain detection viewpoint association information, which includes the deployment location of the AOI detection unit, the detection angle adjustment range and the image acquisition range under different viewpoints, and generate a set of detection viewpoint parameters.

[0014] The AOI inspection unit acquires deployment location information, including the X, Y, and Z coordinates of the AOI inspection unit at the inspection station, through its own position sensor; it acquires the inspection angle adjustment range, including the horizontal and vertical angle adjustment ranges, through an angle sensor; and it acquires the image acquisition range from different viewpoints through an image acquisition module, including the X and Y coordinate ranges and imaging resolution of the image region corresponding to each viewpoint. This acquired information is integrated into a detection viewpoint parameter set, where each data point includes fields such as viewpoint identifier, deployment location coordinates, inspection angle adjustment range, image acquisition range, and imaging resolution.

[0015] Step S113: Associate the circuit board design information set with the detection viewpoint parameter set, calculate the image coverage and clarity of each key area of ​​the circuit board under different detection viewpoints, select the detection viewpoint combination that can cover all key areas and meet the preset clarity threshold, and obtain the viewpoint combination determination result.

[0016] The AOI inspection unit associates the circuit board design information set with the inspection viewpoint parameter set. For each key area, it calculates the image coverage of that key area under that viewpoint based on its location coordinates and the image acquisition range of the inspection viewpoint; it also calculates the image sharpness of that key area under that viewpoint based on the imaging resolution of the inspection viewpoint and the design size of the key area. The calculation results are compared with preset coverage and sharpness thresholds to select inspection viewpoint combinations that can cover all key areas and meet the sharpness thresholds. The selected viewpoint combinations are integrated into a viewpoint combination determination result, where each data entry includes fields such as viewpoint combination identifier, included viewpoint identifier, covered key area identifier, and sharpness information.

[0017] Step S1131: Analyze the circuit board design information set, extract the location distribution, size, shape characteristics and their relationship with the surrounding areas of each key area of ​​the circuit board, and obtain the key area collection requirements details.

[0018] The AOI inspection unit parses the circuit board design information set, extracting the location distribution information of key areas, including the X-axis and Y-axis coordinates of the key areas on the circuit board; extracting the size information of the key areas, including the length, width, and height; extracting the shape feature information of the key areas, including the shape type (e.g., circle, rectangle, irregular shape) and shape parameters (e.g., radius of a circle, length and width of a rectangle); and extracting the relationship between the key areas and surrounding areas, including the relative position and size ratio of the key areas to surrounding areas. The extracted information is integrated into a key area acquisition requirement detail. Each data point in this detail includes fields such as key area identifier, location distribution, size, shape features, and relationship with surrounding areas.

[0019] Step S1132: Analyze the set of detection viewpoint parameters, extract the physical position coordinates of the AOI detection unit, the detection viewpoint angle adjustment range, the image acquisition area coordinates and imaging resolution parameters under each viewpoint, and obtain the detailed characteristics of the detection viewpoint.

[0020] The AOI detection unit parses the set of detection viewpoint parameters, extracting the physical location coordinates of the AOI detection unit, including the X-axis, Y-axis, and Z-axis coordinates; extracting the detection viewpoint angle adjustment range, including the minimum and maximum angles in the horizontal and vertical directions; extracting the image acquisition area coordinates under each viewpoint, including the X-axis start coordinate, X-axis end coordinate, Y-axis start coordinate, and Y-axis end coordinate of the image area; and extracting the imaging resolution parameters, including the width and height pixels of the image. The extracted information is integrated into a detection viewpoint characteristic detail, where each data point includes fields such as viewpoint identifier, physical location coordinates, angle adjustment range, image acquisition area coordinates, and imaging resolution parameters.

[0021] Step S1133: Associate the key area acquisition requirements with the detection viewpoint characteristics, and analyze the coverage, imaging clarity and acquisition completeness of the key area by different detection views for each key area of ​​the circuit board to obtain the single-view acquisition evaluation results.

[0022] The AOI detection unit correlates the detailed acquisition requirements of key areas with the detailed characteristics of the detection viewpoint. For each key area and each detection viewpoint, it determines the coverage of the key area by that viewpoint based on the location distribution of the key area and the image acquisition area coordinates of the detection viewpoint; it determines the image clarity of the key area by that viewpoint based on the size of the key area and the imaging resolution parameters of the detection viewpoint; and it determines the acquisition completeness of the key area by that viewpoint based on the shape characteristics of the key area and the angle adjustment range of the detection viewpoint. These judgments are integrated into a single-view acquisition evaluation result. Each data point in this single-view acquisition evaluation result includes fields such as key area identifier, viewpoint identifier, coverage, image clarity, and acquisition completeness.

[0023] Step S1134: For key areas that cannot be fully covered or have blurred images in the single-view acquisition evaluation results, select other detection views that can supplement the coverage of the key areas and improve the imaging quality, label the selected detection views and their supplementary acquisition advantages, and obtain the supplementary view selection results.

[0024] The AOI (Automated Optical Inspection) unit analyzes the single-view acquisition evaluation results to identify key areas that cannot be fully covered or have blurred images. For each of these key areas, other detection views are selected from the detection view characteristic details. These views can cover the uncovered parts of the key area or improve the imaging quality of the key area. The selected detection views are labeled with their supplementary acquisition advantages, such as wider coverage or higher imaging resolution. The above selection results are integrated into a supplementary view selection result. Each data point in this supplementary view selection result includes fields such as key area identifier, description of uncovered or blurred areas, supplementary view identifier, and supplementary acquisition advantages.

[0025] Step S1135: Based on the single-view acquisition evaluation results and the supplementary view selection results, combine the detection views so that each key area has at least one detection view that can completely cover it and produce clear images. Control the overlap of the view coverage range and reduce the repeated acquisition area by adjusting the view angle to obtain the initial view combination.

[0026] The AOI detection unit combines detection perspectives based on single-view acquisition evaluation results and supplementary perspective selection results. First, at least one detection perspective that can completely cover and clearly image each key region is selected. Then, the overlap of coverage between different perspectives is controlled to avoid excessive overlap leading to duplicate acquisition. The area of ​​duplicate acquisition is reduced by adjusting the angle of the perspectives. The combined perspectives are then integrated into an initial perspective combination. Each data entry in this initial perspective combination includes fields such as perspective combination identifier, included perspective identifier, key region identifier, and coverage overlap magnitude.

[0027] Step S1136: Analyze the acquisition sequence and connection logic of each detection view in the initial view combination, adjust the view switching sequence, shorten the view switching time, ensure the continuity of the acquisition process, supplement the acquisition view of the missing key areas, and obtain the view combination optimization result.

[0028] The AOI detection unit analyzes the initial viewpoint combination to determine the acquisition order and connection logic of each detection viewpoint. Based on the time consumption of viewpoint switching and the requirements for the continuity of the acquisition process, the viewpoint switching order is adjusted, prioritizing viewpoints with shorter switching times. The unit checks whether any key areas were missed in the initial viewpoint combination; if so, the corresponding viewpoints are added. The adjusted and supplemented viewpoint combinations are integrated into the viewpoint combination optimization result. Each data point in this result includes fields such as viewpoint combination identifier, included viewpoint identifiers, acquisition order, viewpoint switching time, and supplemented viewpoint identifiers.

[0029] Step S1137: Label the key area coverage, imaging parameter requirements and view switching conditions corresponding to each detection view in the view combination optimization results, determine the acquisition time of each view and the collaborative acquisition logic with other views, and obtain the view combination details.

[0030] The AOI detection unit annotates the optimized viewpoint combination results, specifying the key area coverage, imaging parameter requirements (such as imaging resolution and exposure time), and viewpoint switching conditions (such as switching to the next viewpoint after the current viewpoint acquisition is completed) for each detection viewpoint. The acquisition duration for each viewpoint is determined, calculated based on the size of the key area and the imaging resolution. The collaborative acquisition logic with other viewpoints is determined, such as simultaneously acquiring different areas or acquiring them sequentially. The above annotated and determined information is integrated into a viewpoint combination detail. Each data entry in this detail includes fields such as viewpoint identifier, key area coverage, imaging parameter requirements, viewpoint switching conditions, acquisition duration, and collaborative acquisition logic.

[0031] Step S1138: Associate historical view combination data, extract cases where the view combination was unreasonable, the acquisition time exceeded the preset range, and the imaging could not clearly present the core features of the key area during the historical acquisition process, correct the view selection deviation and unreasonable switching logic in the view combination details, and obtain the view combination correction result.

[0032] The AOI detection unit associates historical viewpoint combination usage data, which includes information such as historical viewpoint combinations, acquisition time, and image quality. Cases of unreasonable viewpoint combinations are extracted from the historical data, such as incorrect viewpoint selection leading to missing key areas; cases where acquisition time exceeds a preset range, such as excessively long viewpoint switching time; and cases where the imaging fails to clearly present the core features of key areas, such as insufficient imaging resolution. Based on these cases, viewpoint selection deviations in the viewpoint combination details are corrected, viewpoint switching logic is adjusted, and acquisition time is optimized. The corrected viewpoint combinations are integrated into a viewpoint combination correction result. Each data point in this result includes fields such as viewpoint combination identifier, corrected viewpoint identifier, corrected switching logic, and optimized acquisition time.

[0033] Step S1139: Supplement the applicable scenario description in the viewpoint combination correction result, covering the adjustment method of viewpoint combination under different types of circuit boards and different testing environments, mark the adjustment basis, and obtain the initial viewpoint combination determination result.

[0034] The AOI inspection unit supplements the viewpoint combination correction results by adding descriptions of applicable scenarios. Applicable scenarios include different types of circuit boards, such as single-layer and multi-layer circuit boards; and different inspection environments, such as environments with varying light intensities and temperatures. For each applicable scenario, the adjustment method for the viewpoint combination is explained, such as adjusting the viewpoint angle and imaging resolution. The basis for adjustment is indicated, such as adjusting the viewpoint angle according to the type of circuit board to adapt to different circuit layouts, and adjusting the exposure time according to the light intensity of the inspection environment to ensure image quality. The supplemented viewpoint combinations are integrated into the initial viewpoint combination determination result, where each data entry includes fields such as viewpoint combination identifier, applicable scenario, adjustment method, and adjustment basis.

[0035] Step S11310: Simulation test on the initial viewpoint combination determination result: Based on the circuit board design information set and the detection viewpoint parameter set, calculate the theoretical image coverage and expected image clarity of each key area under each viewpoint combination; select viewpoint combinations in which the theoretical image coverage of all key areas reaches the set coverage and the expected image clarity exceeds the preset clarity threshold as valid candidate combinations; if there are multiple candidate combinations, select the combination with the fewest viewpoint switching times to generate the final viewpoint combination determination result.

[0036] The AOI detection unit performs simulation tests on the initial viewpoint combination determination results. First, based on the circuit board design information set and the detection viewpoint parameter set, it calculates the theoretical image coverage of each key region under each viewpoint combination, i.e., the proportion of the key region in the image; it also calculates the expected image sharpness, i.e., the detail discernibility of the key regions in the image. Then, viewpoint combinations where the theoretical image coverage of all key regions reaches the set coverage and the expected image sharpness exceeds the preset sharpness threshold are selected as valid candidate combinations. If multiple candidate combinations exist, the number of viewpoint switching times for each candidate combination is compared, and the combination with the fewest switching times is selected. The finally selected viewpoint combinations are integrated into the viewpoint combination determination result. Each data point in this result includes fields such as viewpoint combination identifier, included viewpoint identifier, theoretical image coverage of the key regions, expected image sharpness, and number of viewpoint switching times.

[0037] Step S114: Determine the results by combining the viewing angles using the AOI detection unit, adjust the detection angle and acquisition distance, synchronously start the multi-view image acquisition operation, acquire images of each area of ​​the circuit board, record the detection viewing angle parameters corresponding to each acquisition, and obtain the multi-view image acquisition set.

[0038] The AOI detection unit determines the result based on the viewpoint combination, adjusts the detection angle to match the angles corresponding to each viewpoint in the viewpoint combination, and adjusts the acquisition distance to match the distances corresponding to each viewpoint in the viewpoint combination. Then, it synchronously initiates multi-view image acquisition to acquire images of various areas of the circuit board. During the acquisition process, the detection viewpoint parameters for each acquisition are recorded, including viewpoint identifier, detection angle, acquisition distance, and imaging resolution. The acquired images and recorded parameters are integrated into a multi-view image acquisition set. Each data entry in this multi-view image acquisition set includes fields such as image identifier, circuit board area, detection viewpoint parameters, and acquisition time.

[0039] Step S115: Associate and match the multi-view image acquisition set with the circuit board design information set, label the circuit board area, detection view parameters and corresponding position in the design drawing corresponding to each circuit board image, and obtain the image association matching result.

[0040] The AOI (Automated Optical Inspection) unit correlates and matches the multi-view image acquisition set with the circuit board design information set. First, it extracts features from each image in the multi-view image acquisition set, such as the shape of the circuit and the location of key areas; it also extracts features from each region in the circuit board design information set, such as the shape of the circuit and the location of key areas. Then, it matches the image features with the design information features to determine the corresponding circuit board region for each image. It labels the corresponding circuit board region, detection viewpoint parameters, and the corresponding position in the design drawing for each image. The matching results are then integrated into an image correlation matching result, where each data point includes fields such as image identifier, circuit board region, detection viewpoint parameters, corresponding position in the design drawing, and matching degree.

[0041] Step S116: Identify circuit board images that are blurry, occluded, or incomplete in the multi-view image acquisition set, mark the corresponding areas and detection view parameters, readjust the detection view to supplement the acquisition, generate supplementary acquisition images, and obtain image quality optimization results.

[0042] The AOI (Analog and Optical Inspection) unit performs quality checks on images in the multi-view image acquisition set, identifying images that are blurry, occluded, or incompletely acquired. For blurry images, an image sharpness algorithm is used to detect image sharpness; for occluded images, an image segmentation algorithm is used to detect occluded areas; for incompletely acquired images, an image comparison algorithm is used to compare the image with the design drawings to detect missing areas. The corresponding regions and detection viewpoint parameters of these images are marked. Then, the detection viewpoint is readjusted, and supplementary acquisition images are generated for the marked regions. The supplementary acquisition images are integrated into an image quality optimization result, where each data point includes fields such as supplementary image identifier, corresponding region, detection viewpoint parameters, and acquisition time.

[0043] Step S117: Integrate the multi-view image acquisition set, image association matching results and image quality optimization results, remove duplicate acquired circuit board images, supplement acquired images and corresponding association information, and obtain the initial linkage acquisition results.

[0044] The AOI detection unit integrates multi-view image acquisition sets, image association matching results, and image quality optimization results. First, duplicate images—those with identical content and corresponding regions—are removed. Then, acquired images and their corresponding association information are added, such as circuit board areas, detection viewpoint parameters, and corresponding locations on design drawings. The integrated results are then combined into an initial linked acquisition result. Each data point in this initial linked acquisition result includes fields such as image identifier, circuit board area, detection viewpoint parameters, corresponding location on design drawings, image quality information, and acquisition time.

[0045] Step S118: Associate historical linkage acquisition effect data, extract cases of unreasonable viewpoint combinations and poor image quality during historical acquisition, correct viewpoint parameter labeling errors and image association deviations in the initial linkage acquisition results, and obtain corrected linkage acquisition results.

[0046] The AOI detection unit associates historical linkage acquisition effect data, which includes information such as historical linkage acquisition results and acquisition effect evaluation. Cases of unreasonable viewpoint combinations are extracted from the historical data, such as viewpoint combinations leading to the omission of key areas; cases of poor image quality are also extracted, such as blurry, occluded, or incomplete images. Based on these cases, errors in the viewpoint parameter annotations in the initial linkage acquisition results are corrected, such as incorrect viewpoint identification or detection angle annotation; image association deviations are also corrected, such as incorrect circuit board areas corresponding to images or incorrect corresponding positions on design drawings. The corrected results are integrated into a corrected linkage acquisition result, where each data point includes fields such as image identifier, circuit board area, detection viewpoint parameters, corresponding position on design drawings, image quality information, acquisition time, and correction description.

[0047] Step S119: Supplement the acquisition time and equipment operation status information of the AOI detection unit to the corrected linkage acquisition results, mark the generation basis of each part and the connection relationship with subsequent steps, obtain the linkage acquisition results, and sort the linkage acquisition results according to the circuit board area.

[0048] The AOI detection unit supplements the acquisition time information into the corrected linked acquisition results, including the start and end times of image acquisition; it also supplements the equipment operating status information, including voltage, current, temperature, and other status parameters of the AOI detection unit during the acquisition process. The basis for generating each part is labeled, such as the generation basis of image association matching results being the matching of image features with design information features; the connection relationship with subsequent steps is also labeled, such as the linked acquisition results being used for subsequent key area localization. The supplemented and labeled results are integrated into the linked acquisition results, and the linked acquisition results are categorized and organized according to circuit board area, with images of the same area grouped together.

[0049] Step S120: Based on the linkage acquisition results, associate the key area annotation information in the circuit board design drawings with the area features in the image to generate a set of associated key area features of the circuit board.

[0050] The AOI inspection unit extracts regional features from the linked acquisition results of the circuit board image, including the width of the lines, the shape of the solder joints, the spacing between lines, and the grayscale distribution of the region. It also extracts key area annotation information from the circuit board design drawings, including the type, location coordinates, and design dimensions of the key areas. The regional features are then correlated with the key area annotation information to establish a correspondence between the regional features and the key areas. The correlation results are integrated into a circuit board key area feature association set. Each data entry in this set contains fields such as key area identifier, type, location coordinates, design dimensions, regional features, and correspondence.

[0051] Step S121: Analyze the linkage acquisition results, extract the regional features of each area of ​​the circuit board image, the regional features include line width, solder joint shape, line spacing and regional grayscale distribution features, label the circuit board area and detection viewpoint parameters corresponding to each regional feature, and obtain the image regional feature set.

[0052] The AOI inspection unit analyzes the linked acquisition results and extracts regional features of each area of ​​the circuit board image. For line width, the line width is measured using an image measurement algorithm; for solder joint morphology, the shape and size of the solder joints are identified using an image recognition algorithm; for line spacing, the distance between lines is measured using an image measurement algorithm; and for regional grayscale distribution characteristics, the distribution of grayscale values ​​within the region is analyzed using an image analysis algorithm. The circuit board area and inspection viewing angle parameters corresponding to each regional feature are labeled. The extracted and labeled information is integrated into an image region feature set. Each data point in this image region feature set contains fields such as image identifier, circuit board area, inspection viewing angle parameters, line width, solder joint morphology, line spacing, and regional grayscale distribution characteristics.

[0053] Step S122: Retrieve the circuit board design drawings associated with the linkage acquisition results, and extract the key area annotation information in the circuit board design drawings. The key area annotation information includes the key area type, location coordinates, design dimensions and performance requirements, to obtain the key area information of the design drawings.

[0054] The AOI inspection unit retrieves the associated circuit board design drawings from the linked acquisition results, parses the design drawings, and extracts the annotation information of key areas. Key area types include pads, circuit intersections, chip mounting areas, etc.; location coordinates include the X and Y axis coordinates of the key areas in the design drawings; design dimensions include the length, width, and height of the key areas; performance requirements include the conductivity and heat resistance of the key areas. The extracted information is integrated into the key area information of the design drawings. Each data point in this key area information includes fields such as key area identifier, type, location coordinates, design dimensions, and performance requirements.

[0055] Step S123: Match the image region feature set with the key region information of the design drawing, compare them one by one according to the key region type of the circuit board. When the difference between the image region feature and the key region annotation information is within the preset tolerance range, it is determined that the match is successful, and the successfully matched feature pair and its tolerance are recorded to obtain the feature information matching result.

[0056] The AOI (Analog and Imaging Inspection) unit matches the image region feature set with the key region information in the design drawings. First, it compares the key regions one by one according to their type on the circuit board. For each type of key region, it compares the image region features with the key region annotation information in the design drawings. It calculates the difference between the image region features and the key region annotation information, such as the difference in line width or position coordinates. When the difference is within a preset tolerance range, it is considered a successful match. Successfully matched feature pairs are recorded, such as the correspondence between line width features and design dimension features; tolerances are also recorded, such as the tolerance range for line width. The matching results are integrated into a feature information matching result. Each data point in this result includes fields such as key region identifier, type, image region features, key region information in the design drawings, difference, and tolerance.

[0057] Step S1231: Analyze the image region feature set, extract the line width, solder joint shape, line spacing, regional grayscale distribution features of each circuit board region image and its corresponding circuit board region and detection viewpoint parameters, classify and organize them according to circuit board region to obtain image feature details.

[0058] The AOI detection unit analyzes the image region feature set, extracting the line width, solder joint shape, line spacing, and regional grayscale distribution features of each circuit board region image, along with their corresponding circuit board region and detection viewing angle parameters. This information is categorized and organized according to circuit board region, with information from the same region grouped together. The organized information is then integrated into an image feature detail, where each data point includes fields such as circuit board region, detection viewing angle parameters, line width, solder joint shape, line spacing, and regional grayscale distribution features.

[0059] Step S1232: Analyze the key area information of the design drawings, extract the type, location coordinates, design dimensions, performance requirements and the relationship with the surrounding areas of each circuit board key area, classify and organize them according to the key area type, and obtain the design feature details.

[0060] The AOI inspection unit parses the key area information of the design drawings, extracting the type, location coordinates, design dimensions, performance requirements, and relationships with surrounding areas for each circuit board's key areas. This information is then categorized and organized according to key area type, with information of the same type grouped together. The organized information is integrated into a design feature detail, where each data point contains fields such as key area type, location coordinates, design dimensions, performance requirements, and relationships with surrounding areas.

[0061] Step S1233: Set the priority rules for feature matching. The priority rules stipulate that during the matching process, the size features of the line width and solder joint shape are matched first. When the matching degree of the size features exceeds the first threshold, the regional grayscale distribution features are introduced for auxiliary matching. Among them, the confidence weight of the size feature matching result is higher than the confidence weight of the regional grayscale distribution feature matching result, and the matching priority setting result is obtained.

[0062] The AOI inspection unit sets priority rules for feature matching. The rules stipulate that during the matching process, the dimensional features of the line width and solder joint shape are matched first, and the matching degree of the dimensional features is calculated. When the matching degree exceeds a first threshold, regional grayscale distribution features are then introduced for auxiliary matching, and the matching degree of the regional grayscale distribution features is calculated. Simultaneously, the confidence weight of the dimensional feature matching result is stipulated to be higher than the confidence weight of the regional grayscale distribution feature matching result; that is, the dimensional feature matching result has a greater impact on the final matching result. The set rules are integrated into a matching priority setting result, which includes fields such as matching order, first threshold, and confidence weight.

[0063] Step S1234: Associate the image feature details and design feature details one by one according to the key area type of the circuit board. First, compare the fit of the core design size features, and then further verify with auxiliary features. Mark the feature fit points and differences to obtain the single-type area matching results.

[0064] The AOI inspection unit correlates image feature details with design feature details one by one according to the key area type of the circuit board. First, it compares the fit of core design dimensional features, such as line width and solder joint size, with the design dimensions. Then, it further verifies by combining auxiliary features, such as the fit of regional grayscale distribution features with design requirements. Feature fit points are marked, i.e., the parts where image features match design features; discrepancies are marked, i.e., the parts where image features do not match design features. The comparison and marking results are integrated into a single-type area matching result, where each data point includes fields such as key area type, image features, design features, fit points, and discrepancies.

[0065] Step S1235: For regions with low feature fit in the single-type region matching results, analyze the reasons for the differences, adjust the matching logic, supplement the differential feature descriptions, and re-perform feature matching to obtain the corrected results for the differential region matching.

[0066] The AOI detection unit analyzes the single-type region matching results and identifies regions with low feature fit. It analyzes the reasons for the discrepancies, such as poor image acquisition quality or inaccurate feature descriptions. Based on the reasons for the discrepancies, it adjusts the matching logic, such as adjusting the matching tolerance range and increasing the weight of auxiliary features. It supplements the differential feature descriptions, such as providing detailed descriptions of the features of the discrepancy regions. Feature matching is then performed again, and the adjusted matching degree is calculated. The corrected matching results are integrated into a differential region matching correction result, where each data point includes fields such as key region type, differential region description, reason for the discrepancy, adjusted matching logic, differential feature description, and corrected matching degree.

[0067] Step S1236: Associate the single-type region matching results with the difference region matching correction results, label the image region features, design features and matching basis corresponding to the key regions of each circuit board, determine the range of successfully matched regions and the matching accuracy, and obtain the preliminary feature matching results.

[0068] The AOI detection unit correlates single-type region matching results with difference region matching correction results. It annotates the image region features, design features, and matching criteria corresponding to each key region of the circuit board, such as size feature matching and auxiliary feature matching. It determines the range of successfully matched regions, i.e., the areas in the image that successfully match the design features; and determines the matching accuracy, i.e., how close the matching result is to the design features. The correlated and determined information is integrated into preliminary feature matching results, where each data point includes fields such as key region identifier, type, image region features, design features, matching criteria, range of successfully matched regions, and matching accuracy.

[0069] Step S1237: In the preliminary results of feature matching, there are confusion situations where multiple image regions correspond to the same key region and a single image region corresponds to multiple key regions. By combining the detection viewpoint parameters and the position coordinates of the design drawing, the correspondence is distinguished, the confusion problem is corrected, and the matching confusion correction result is obtained.

[0070] The AOI detection unit analyzes the preliminary feature matching results, identifying confusion situations where multiple image regions correspond to the same key region, and a single image region corresponds to multiple key regions. Combining detection perspective parameters, such as the angle and acquisition distance, it analyzes the correspondence between image regions and key regions; combining the location coordinates on the design drawing, it analyzes the correspondence between the position of the key region on the design drawing and the image region. Based on the analysis results, the correspondence is distinguished, and confusion issues are corrected. The corrected results are integrated into a matching confusion correction result, where each data point includes fields such as key region identifier, confusion description, detection perspective parameters, design drawing location coordinates, and the corrected correspondence.

[0071] Step S1238: Integrate the preliminary feature matching results with the matching confusion correction results, organize them according to the key area type of the circuit board, and mark the image feature correspondence, matching accuracy and abnormal situations in the matching process of each key area to obtain the initial feature information matching results.

[0072] The AOI detection unit integrates the preliminary feature matching results with the matching confusion correction results. The integrated results are then organized according to the type of key regions on the circuit board, grouping key regions of the same type together. The image feature correspondences of each key region are labeled, i.e., the correspondence between image features and design features; the matching accuracy is labeled; and anomalies during the matching process, such as confusion and discrepancies, are labeled. The organized and labeled results are then integrated into the initial feature information matching results, where each data entry includes fields such as key region identifier, type, image feature correspondence, matching accuracy, and anomaly information.

[0073] Step S1239: Supplement the matching condition description in the initial feature information matching result, covering the adjustment method of feature matching under different detection perspectives and the judgment criteria for matching accuracy, to obtain the feature information matching result.

[0074] The AOI detection unit supplements the initial feature information matching results by adding matching condition descriptions. These descriptions cover adjustment methods for feature matching under different detection perspectives, such as adjusting the tolerance range for matching under different perspectives; they also cover the criteria for judging matching accuracy, such as the classification of matching accuracy levels and the corresponding judgment conditions. The supplemented results are then integrated into a feature information matching result, where each data point includes fields such as key region identifier, type, image feature correspondence, matching accuracy, anomalies, and matching condition descriptions.

[0075] Step S12310: Verify the accuracy of the feature information matching results, check the uniqueness of the image region features corresponding to each key region, correct matching errors and feature omissions, and complete the feature information matching process.

[0076] The AOI detection unit verifies the accuracy of the feature information matching results. It checks whether the image region features corresponding to each key region are unique, i.e., whether one key region corresponds to multiple different image region features. If a matching error exists, such as a key region matching an incorrect image region feature, the matching relationship is corrected; if a feature is missing, such as a key region not matching a corresponding image region feature, the corresponding feature is added. After verification and correction, the feature information matching process ends.

[0077] Step S124: Identify the parts in the feature information matching results where the image region features do not match the key region information of the design drawings, analyze the reasons for the differences, distinguish whether it is caused by the acquisition perspective deviation or the image feature extraction deviation, and obtain the feature matching deviation analysis results.

[0078] The AOI detection unit analyzes the feature information matching results to identify areas where the image region features do not match the key area information of the design drawings. It analyzes the reasons for these discrepancies, such as image distortion caused by viewing angle deviation or inaccurate feature extraction due to errors in the image feature extraction algorithm. It distinguishes whether the discrepancy is caused by viewing angle deviation or image feature extraction deviation. The analysis results are then integrated into a feature matching deviation analysis result, where each data point includes fields such as key area identifier, description of the mismatched part, reason for the difference, and deviation type (viewing angle deviation or image feature extraction deviation).

[0079] Step S125: Based on the feature matching deviation analysis results, adjust the matching logic between image region features and key region annotation information, correct the feature association relationship corresponding to the deviation, supplement the differential feature description, and obtain the feature association correction result.

[0080] The AOI detection unit adjusts the matching logic between image region features and key region annotations based on the feature matching deviation analysis results. For differences caused by viewing angle deviations, it adjusts the matching tolerance range or introduces a viewing angle correction algorithm; for differences caused by image feature extraction deviations, it adjusts the parameters of the feature extraction algorithm or increases the dimension of feature extraction. It corrects the feature association relationships corresponding to the deviations, such as re-establishing the correspondence between image region features and key region annotations. It supplements differential feature descriptions, such as providing detailed descriptions of the features in the deviation regions. The adjusted, corrected, and supplemented results are integrated into a feature association correction result, where each data point includes fields such as key region identifier, deviation type, adjusted matching logic, corrected feature association relationship, and differential feature description.

[0081] Step S126: Extract the core features of the key areas of the circuit board, label the core features and feature combinations that can uniquely identify each key area, eliminate interference from non-key features, and obtain the core feature extraction results.

[0082] The AOI inspection unit analyzes the features of key areas on the circuit board, extracting core features—those that uniquely identify key areas. It then labels the core features and feature combinations that uniquely identify each key area, such as identifying key areas through a combination of line width and solder joint shape. Non-critical features are excluded, meaning features with minimal impact on key area identification are ignored. The extracted and labeled information is integrated into the core feature extraction result, where each data point includes fields such as key area identifier, core features, feature combination method, and excluded non-critical features.

[0083] Step S127: The correlation correction results and core feature extraction results are sorted according to the key area type of the circuit board, and the image area features, design annotation information and core feature combination corresponding to each key area are determined to obtain the initial circuit board key area feature correlation set.

[0084] The AOI detection unit correlates the feature association correction results with the core feature extraction results. The correlated results are then organized according to the type of key regions on the circuit board, grouping key regions of the same type together. The image region features, design annotation information, and core feature combinations corresponding to each key region are determined. The organized and determined information is integrated into an initial circuit board key region feature association set, where each data entry includes fields such as key region identifier, type, image region features, design annotation information, and core feature combinations.

[0085] Step S128: Verify the validity of the initial circuit board key area feature association set, check each key area one by one, supplement the missing image feature correspondence and design information association content, correct the problems of unclear association and missing features, and obtain the verified feature association set.

[0086] The AOI inspection unit verifies the validity of the initial feature association set for key areas of the circuit board. Each key area is checked individually to see if there are any missing image feature correspondences or design information associations. If missing information is found, the corresponding content is added; if there are unclear associations or missing features, the associations are corrected or features are added. The verification and correction results are integrated into a verified feature association set, where each data entry includes fields such as key area identifier, type, image area features, design annotation information, core feature combination, and verification description.

[0087] Step S129: Supplement the applicable detection perspective range for feature association of each key region, mark the adjustment method of feature association under different perspectives, and generate the final set of key region feature associations for the circuit board.

[0088] The AOI detection unit supplements the validated feature association set by adding the applicable detection perspective range for each key region feature association, i.e., which detection perspectives the feature association is applicable to. It also annotates the adjustment methods for feature associations under different perspectives, such as adjusting the weights of core feature combinations under different perspectives. The supplemented results are integrated into the final circuit board key region feature association set, where each data entry includes fields such as key region identifier, type, image region features, design annotation information, core feature combination, applicable detection perspective range, and adjustment methods under different perspectives.

[0089] Step S130: Combine the feature association set of the key area of ​​the circuit board to generate the key area positioning rule of the circuit board. The key area positioning rule of the circuit board defines the correspondence between image features and key areas, positioning accuracy control requirements and abnormal positioning processing flow.

[0090] The AOI (Automated Optical Inspection) unit generates key area localization rules for the circuit board based on the feature association set of key areas. These rules define the correspondence between image features and key areas, i.e., specific image features correspond to specific key areas; they also define the localization accuracy control requirements, i.e., the deviation range between the localization result and the design drawings; and they define the abnormal localization handling process, i.e., the steps to take when abnormal localization results occur. The generated rules are integrated into a circuit board key area localization rule, which includes fields such as the correspondence between image features and key areas, localization accuracy control requirements, and abnormal localization handling process.

[0091] Step S131: Analyze the feature association set of key areas of the circuit board, extract the image area features, design annotation information, core feature combination and applicable detection view range corresponding to each key area of ​​the circuit board, and obtain the basic data for rule generation.

[0092] The AOI detection unit parses the feature association set of key areas of the circuit board, extracting image region features, design annotation information, core feature combinations, and applicable detection view range corresponding to each key area of ​​the circuit board. The extracted information is integrated into rule-based data generation. Each piece of data in this rule-based data generation includes fields such as key area identifier, type, image region features, design annotation information, core feature combinations, and applicable detection view range.

[0093] Step S132: Generate basic data based on rules, and set the correspondence rules between image features and key regions. The correspondence rules stipulate that when the extracted image feature value falls within the preset feature value range of any key region, it is determined to be the key region, and the feature correspondence rules are obtained.

[0094] The AOI detection unit generates basic data based on rules, setting the correspondence rules between image features and key regions. The rules stipulate that when the extracted image feature values ​​fall within the preset feature value range of any key region, the image region is determined to correspond to that key region. The preset feature value range is determined based on the core feature combination of the key region. The set rules are integrated into feature correspondence rules, which include fields such as key region identifier, preset feature value range, and correspondence relationship.

[0095] Step S133: Set the positioning accuracy control requirements for key areas of the circuit board, divide the positioning deviation range of key area boundaries, the accuracy of position coordinate marking, and the accuracy difference requirements of different key areas to obtain the accuracy control rules.

[0096] The AOI inspection unit sets the positioning accuracy control requirements for critical areas of the circuit board. This includes defining the critical area boundary positioning deviation range (the deviation range between the critical area boundary in the positioning result and the boundary in the design drawings); defining the position coordinate annotation accuracy (the precision of the position coordinates of the critical area in the positioning result); and defining the accuracy differentiation requirements for different critical areas (different types of critical areas have different accuracy requirements). These requirements are integrated into accuracy control rules, which include fields such as critical area type, boundary positioning deviation range, position coordinate annotation accuracy, and accuracy differentiation requirements.

[0097] Step S134: Construct anomaly localization processing rules. For anomaly types such as feature matching failure, ambiguous localization results, and multi-region feature confusion, divide the processing flow of AOI detection units, image re-acquisition conditions, and localization result correction methods to obtain anomaly processing rules.

[0098] The AOI detection unit constructs anomaly localization processing rules. For anomalies involving feature matching failures, the following processing steps are defined: (e.g., re-extracting image features, adjusting matching logic); (e.g., re-acquiring when the number of feature matching failures exceeds a preset number); (e.g., correcting localization results by combining design drawings). For anomalies involving blurred localization results, the following processing steps are defined: (e.g., enhancing image contrast, adjusting image resolution); (e.g., re-acquiring when the blurriness exceeds a preset threshold); (e.g., correcting blurred areas through image fusion). For anomalies involving feature confusion in multiple regions, the following processing steps are defined: (e.g., increasing feature extraction dimensions, introducing design drawings for differentiation); (e.g., re-acquiring when the confusion exceeds a preset threshold); (e.g., differentiating confused regions through feature combination). These constructed rules are integrated into anomaly processing rules, which include fields for anomaly type, processing steps, image re-acquisition conditions, and localization result correction methods.

[0099] Step S135: Associate feature correspondence rules, accuracy control rules, and anomaly handling rules, classify and organize them according to the key area type of the circuit board, mark the positioning rule details corresponding to different key areas, divide the execution priority of each rule, and obtain the initial key area positioning rules of the circuit board.

[0100] The AOI detection unit associates feature mapping rules, accuracy control rules, and anomaly handling rules. The associated rules are then categorized and organized according to the type of critical area on the circuit board, grouping critical areas of the same type together. Details of the positioning rules for different critical areas are annotated, such as the feature mapping rules, accuracy control requirements, and anomaly handling procedures for specific critical areas. The execution priority of each rule is determined, with feature mapping rules taking precedence over accuracy control rules, and accuracy control rules taking precedence over anomaly handling rules. The organized and annotated rules are then integrated into initial circuit board critical area positioning rules, where each data entry includes fields such as critical area type, feature mapping rule, accuracy control rule, anomaly handling rule, positioning rule details, and execution priority.

[0101] Step S136: Retrieve historical circuit board key area positioning data, analyze cases of unreasonable rules, insufficient positioning accuracy, and untimely anomaly handling in historical positioning, revise the corresponding clauses in the initial circuit board key area positioning rules, optimize rule details, and obtain rule revision results.

[0102] The AOI detection unit retrieves historical key area positioning data for the circuit board, which includes historical positioning results and rule execution information. It analyzes cases of unreasonable rules in historical positioning, such as feature-matching rules leading to positioning errors; cases of insufficient positioning accuracy, such as accuracy control rules failing to meet requirements; and cases of untimely anomaly handling, such as excessively long anomaly handling processes. Based on these cases, it corrects the corresponding clauses in the initial key area positioning rules for the circuit board, such as adjusting the preset feature value range of the feature-matching rules, increasing the requirements of the accuracy control rules, and optimizing the steps of the anomaly handling process. The corrected results are integrated into a rule correction result, where each data entry includes fields such as key area type, corrected clause, reason for correction, and optimized rule details.

[0103] Step S137: Supplement the applicable scenario description of the key area positioning rules of the circuit board, covering the adjustment methods of the rules under different types of circuit boards and different testing environments, mark the basis for adjustment, and obtain the scenario adaptation supplementary results.

[0104] The AOI inspection unit supplements the rule correction results by adding applicable scenario descriptions. These descriptions cover different types of circuit boards, such as single-layer and multi-layer boards, and different inspection environments, such as environments with varying light intensities and temperatures. For each applicable scenario, the method of rule adjustment is explained, including the preset feature value range for the adjusted feature and the requirements for adjusting the accuracy control rules. The basis for adjustment is also indicated, such as adjusting rules based on circuit board type to adapt to different circuit layouts, and adjusting rules based on the inspection environment to adapt to different imaging conditions. The supplemented results are integrated into a scenario adaptation supplementary result, where each data point includes fields such as applicable scenario, rule adjustment method, and adjustment basis.

[0105] Step S138: Integrate the rule correction results and scenario adaptation supplementary results, remove conflicting rule content, supplement rule connection clauses, refine the execution priority ranking of each rule, and obtain the draft rules for locating key areas of the circuit board.

[0106] The AOI detection unit integrates the rule correction results with the scene adaptation supplementary results. It removes conflicting rules that arise during integration, i.e., contradictory parts between different rules. It supplements rule connection clauses, specifying the connection methods between different rules. It refines the execution priority of each rule, such as further clarifying the execution order of feature correspondence rules, accuracy control rules, and anomaly handling rules. The integrated, removed, supplemented, and refined results are combined into a draft of key area positioning rules for the circuit board. This draft includes fields such as key area type, feature correspondence rules, accuracy control rules, anomaly handling rules, applicable scenario descriptions, rule connection clauses, and execution priority order.

[0107] Step S139: Execute the draft rules for locating key areas of circuit boards on the historical dataset, and statistically analyze the success rate and accuracy of the rules on different types of circuit board images; when the success rate is lower than the preset success rate threshold or the accuracy exceeds the allowable deviation range, mark the clauses in the rules that cause failure, and obtain the rule test results.

[0108] The AOI detection unit executes a draft of key area localization rules for circuit boards on a historical dataset. The historical dataset contains images of different types of circuit boards and their corresponding localization results. The system calculates the localization success rate (the ratio of successfully localized images to the total number of images) and the localization accuracy (the degree of deviation between the localization result and the design drawing). When the localization success rate falls below a preset success rate threshold or the localization accuracy exceeds the allowable deviation range, the rules that lead to failure are flagged. The statistical and flagging results are integrated into a rule test result, which includes fields such as circuit board type, localization success rate, localization accuracy, and flagged failure clauses.

[0109] Step S1310: Based on the rule test results, iteratively optimize the marked terms: relax the tolerance threshold for feature matching or add auxiliary judgment features, and then re-execute the test until the positioning success rate and positioning accuracy both meet the preset thresholds, and generate the final circuit board key area positioning rules.

[0110] The AOI detection unit iteratively optimizes the marked clauses based on the rule test results. For clauses leading to low positioning success rates, the tolerance threshold for feature matching is relaxed or auxiliary judgment features are added; for clauses leading to insufficient positioning accuracy, the requirements of the accuracy control rules are increased or the feature extraction algorithm is optimized. After optimization, the rule test is re-executed on the historical dataset. This process is repeated until both the positioning success rate and positioning accuracy meet the preset thresholds. The final optimized rules are integrated into circuit board key area positioning rules, which include fields such as key area type, feature correspondence rules, accuracy control rules, anomaly handling rules, applicable scenario descriptions, rule connection clauses, and execution priority ranking.

[0111] Step S140: The AOI detection unit performs key area localization on the circuit board image in the linkage acquisition result according to the key area localization rules of the circuit board, and generates the key area localization result of the circuit board.

[0112] The AOI detection unit locates key areas of the circuit board image in the linked acquisition results according to the circuit board key area localization rules. First, it extracts the regional features of the image; then, it determines the key areas corresponding to the image areas according to the feature correspondence rules; next, it checks the localization accuracy according to the accuracy control rules; finally, it handles localization anomalies according to the anomaly handling rules. The localization results are integrated into the circuit board key area localization results, where each data entry includes fields such as image identifier, key area identifier, type, location coordinates, localization accuracy, and anomaly handling status.

[0113] Step S141: Retrieve the positioning rules of the key area of ​​the circuit board through the AOI detection unit, extract the feature correspondence rules, accuracy control rules and anomaly handling rules, and combine them with the detection perspective parameters in the linkage acquisition results to generate the basis for the execution of the positioning rules.

[0114] The AOI detection unit retrieves the key area localization rules of the circuit board from the storage device, extracting feature correspondence rules, accuracy control rules, and anomaly handling rules. Combined with the detection viewpoint parameters from the linked acquisition results, such as viewpoint identifiers, detection angles, and acquisition distances, it generates the basis for executing the localization rules. This basis for executing localization rules includes fields such as image identifiers, detection viewpoint parameters, feature correspondence rules, accuracy control rules, and anomaly handling rules.

[0115] Step S142: According to the feature correspondence rule, perform regional feature extraction and matching on the circuit board image in the linkage acquisition result, compare the image region features with the core feature combination of each key region, mark the successfully matched regions and the corresponding key region types, and obtain the feature matching and positioning results.

[0116] The AOI detection unit extracts regional features from the circuit board image in the linked acquisition results according to feature correspondence rules. Extracted features include line width, solder joint shape, line spacing, and regional grayscale distribution. The extracted image region features are compared with the core feature combinations of each key region. When an image region feature matches a core feature combination, the successfully matched region and its corresponding key region type are marked. The matching results are integrated into a feature matching localization result, where each data point includes fields such as image identifier, successfully matched region, key region type, and core feature combination.

[0117] Step S143: Based on the accuracy control rules, verify the accuracy of the boundary and position coordinate annotations of key areas in the feature matching positioning results, adjust the positioning content whose deviation exceeds the allowable range, correct the boundary annotations and coordinate information, and obtain the accuracy verification results.

[0118] The AOI detection unit verifies the accuracy of key area boundaries and location coordinate annotations in the feature matching positioning results according to accuracy control rules. It checks whether the deviation between the key area boundaries and the boundaries in the design drawings is within the allowable range; and whether the accuracy of the location coordinate annotations meets the requirements. For positioning content with deviations exceeding the allowable range, the boundary annotations and coordinate information are adjusted. The verification and adjustment results are integrated into an accuracy verification result, where each data point includes fields such as image identifier, key area identifier, boundary deviation, location coordinate accuracy, and adjusted boundary annotations and coordinate information.

[0119] Step S144: Monitor abnormal situations in the feature matching and localization process, start the corresponding processing flow according to the abnormality handling rules, re-extract image region features and match again when feature matching fails, call the supplementary acquired image from the corresponding detection view to assist in localization when the localization result is blurry, and combine the design drawing annotation information to distinguish when features in multiple regions are confused, and obtain the abnormality handling result.

[0120] The AOI detection unit monitors for anomalies during feature matching and localization. When a feature matching failure occurs, image region features are re-extracted, and the parameters or dimensions of feature extraction are adjusted before re-matching. When a blurry localization result occurs, supplementary images acquired from the corresponding detection viewpoint are retrieved, and features from these supplementary images are extracted to assist in locating the blurry area. When multiple regions exhibit feature confusion, design drawing annotations, such as the location coordinates and design dimensions of key areas, are used to distinguish the confused regions. The processed results are integrated into an anomaly handling result, where each data entry includes fields such as image identifier, anomaly type, processing flow, and the processed localization result.

[0121] Step S1441: During the feature matching and localization process, the success status of image region feature extraction, the success status of feature matching, and the coordinate deviation value of multiple consecutive localization results are detected in real time. When any status is failure or the deviation value exceeds the threshold, the abnormal monitoring result is triggered.

[0122] The AOI detection unit performs real-time detection during feature matching and localization. It detects the success status of image region feature extraction (whether features were successfully extracted), the success status of feature matching (whether features were successfully matched), and the coordinate deviation value of multiple consecutive localization results (the difference in the coordinates of key areas across multiple localization results). When any status is a failure or the deviation value exceeds a threshold, an anomaly monitoring result is triggered. This result includes fields such as image identifier, detection status, coordinate deviation value, and anomaly type.

[0123] Step S1442: For cases where feature matching fails in the anomaly monitoring results, re-extract the image region features of the corresponding area according to the anomaly handling rules, adjust the feature extraction range and extraction method, supplement the missing core features, and match them again with the key region features to obtain the feature re-matching result.

[0124] The AOI detection unit handles feature matching failures in anomaly detection results according to anomaly handling rules. This includes re-extracting image region features from the corresponding area, adjusting the feature extraction range (e.g., expanding or shrinking the extraction area), and adjusting the feature extraction method (e.g., changing the parameters of the feature extraction algorithm). It also supplements the extraction of missed core features, such as key region features not previously extracted. The re-extracted and supplemented features are then combined with the key region features for matching again. The matching results are integrated into a feature re-matching result, which includes fields such as image identifier, re-extracted features, supplemented core features, and the matching result.

[0125] Step S1443: For cases where the positioning result is ambiguous in the anomaly monitoring results, call the supplementary acquisition image of the corresponding detection perspective according to the anomaly handling rules, extract the regional features in the supplementary acquisition image, combine them with the features of the original image to locate the key area, correct the ambiguous positioning boundary and coordinate information, and obtain the ambiguous positioning correction result.

[0126] The AOI inspection unit handles ambiguous localization results in abnormal monitoring results according to anomaly handling rules. It retrieves supplementary images acquired from the corresponding inspection viewpoint. These supplementary images are acquired from the same inspection viewpoint. Regional features, such as line width and solder joint shape, are extracted from the supplementary images. These features are combined with those from the original image to perform key area localization, enhancing the accuracy of the localization using the features of the supplementary image. The ambiguous localization boundaries and coordinate information are corrected to make the localization results clearer and more accurate. The corrected results are then integrated into a fuzzy localization correction result, which includes fields such as image identifier, supplementary image, extracted features, corrected boundary information, and coordinate information.

[0127] Step S1444: In response to the situation of multiple regions having confused features in the anomaly monitoring results, retrieve the key region annotation information in the circuit board design drawings according to the anomaly handling rules, compare the design dimensions, location coordinates and surrounding relationships of each confused region, distinguish different key regions by combining image region features, determine the correspondence of each region, and obtain the confusion region distinction results.

[0128] The AOI detection unit handles cases of feature confusion across multiple regions in the anomaly monitoring results according to anomaly handling rules. It retrieves key region annotation information from the circuit board design drawings, such as design dimensions, location coordinates, and surrounding relationships. It compares the design dimensions of each confused region (length, width, etc.), location coordinates (X-axis, Y-axis coordinates in the design drawings), and surrounding relationships (connections to other key regions). Combining image region features (grayscale distribution, circuit shape, etc.), it distinguishes different key regions. It determines the correspondence between each region, i.e., the correspondence between the image region and the key regions in the design drawings. The distinction results are integrated into a confused region distinction result, which includes fields such as image identifier, confused region, design dimensions, location coordinates, surrounding relationships, image region features, and correspondence.

[0129] Step S1445: For localization anomalies in the anomaly monitoring results that cannot be resolved through feature rematching, fuzzy localization correction, and confusion region differentiation, mark the corresponding anomaly region and anomaly type according to the anomaly handling rules, record the AOI detection unit operating parameters, image acquisition information, and matching process when the anomaly occurs, start the manual auxiliary judgment process, and obtain the manual auxiliary judgment result.

[0130] The AOI detection unit handles localization anomalies that cannot be resolved through the above-mentioned processing in the anomaly monitoring results according to the anomaly handling rules. It marks the corresponding anomaly region and anomaly type, such as feature matching failure, blurred localization results, and feature confusion across multiple regions. It records the AOI detection unit's operating parameters at the time of the anomaly, such as voltage, current, and temperature; it records image acquisition information, such as acquisition time and detection viewpoint parameters; and it records the matching process, such as feature extraction steps and matching steps. A manual assistance judgment process is initiated, sending the anomaly information to a manual judgment terminal for human judgment. The results of the manual judgment are integrated into a manual assistance judgment result, which includes fields such as image identifier, anomaly region, anomaly type, operating parameters, acquisition information, matching process, and manual judgment result.

[0131] Step S1446: Associate the feature rematch results, fuzzy positioning correction results, confused region differentiation results, and manual assistance judgment results, and label the processing method, processing process, and positioning effect after processing for each abnormal situation to obtain the preliminary results of abnormality processing.

[0132] The AOI detection unit correlates feature rematching results, fuzzy localization correction results, confused region differentiation results, and manual-assisted judgment results. It annotates the handling method for each anomaly, such as feature rematching and fuzzy localization correction; the processing steps, such as re-extracting features and calling supplementary images; and the localization effect after processing, such as localization success rate and localization accuracy. The correlated and annotated results are integrated into a preliminary anomaly handling result, which includes fields such as image identifier, anomaly type, handling method, processing steps, and localization effect.

[0133] Step S1447: Analyze the processing efficiency and effect of each abnormal situation in the preliminary results of anomaly handling, adjust the operation steps and timing in the anomaly handling process, optimize the conditions for feature re-extraction and supplementary image retrieval, and obtain the optimized results of anomaly handling.

[0134] The AOI detection unit analyzes the preliminary results of anomaly handling, evaluating the processing efficiency (time taken to handle each anomaly) and the accuracy of the post-processing localization. Based on the analysis, the operation steps and timing in the anomaly handling process are adjusted, such as optimizing the order of feature re-extraction and adjusting the timing of supplementary image retrieval. The conditions for feature re-extraction and supplementary image retrieval are also optimized, such as adjusting the trigger threshold for feature re-extraction and the trigger conditions for supplementary image retrieval. The adjusted and optimized results are integrated into an optimized anomaly handling result, which includes fields such as anomaly type, processing efficiency, processing effect, adjusted operation steps and timing, and optimized conditions.

[0135] Step S1448: Integrate the preliminary results of anomaly handling with the optimization results of anomaly handling, determine the final handling scheme, the positioning accuracy after handling, and the subsequent preventive measures for each anomaly type, and obtain the initial anomaly handling results.

[0136] The AOI detection unit integrates the preliminary and optimized anomaly handling results. It determines the final processing scheme for each anomaly type, such as feature re-matching combined with fuzzy localization correction; it determines the post-processing localization accuracy, i.e., the deviation range between the post-processing localization result and the design drawings; and it determines subsequent preventative measures, such as optimizing image acquisition parameters and adjusting feature extraction algorithms. The integrated and determined information is then combined into an initial anomaly handling result, which includes fields such as anomaly type, final processing scheme, post-processing localization accuracy, and subsequent preventative measures.

[0137] Step S1449: Supplement the anomaly cause analysis in the initial anomaly handling results, mark the root cause of each anomaly, and obtain the anomaly handling results.

[0138] The AOI detection unit supplements the initial anomaly handling results by adding anomaly cause analysis. It analyzes the root cause of each anomaly, such as feature matching failure due to poor image acquisition quality, or blurry localization results due to poor lighting conditions. The supplemented results are then integrated into the anomaly handling results, which include fields such as anomaly type, final processing solution, post-processing localization accuracy, subsequent preventative measures, and root cause.

[0139] Step S14410: Verify the validity of the anomaly handling results, check whether the processed positioning results meet the accuracy control requirements, correct any issues such as incomplete processing or remaining positioning deviations, and complete the anomaly handling process.

[0140] The AOI inspection unit verifies the validity of the anomaly handling results. It checks whether the processed positioning results meet accuracy control requirements, i.e., whether the deviation between the positioning results and the design drawings is within the allowable range. If there are issues with incomplete processing or persistent positioning deviations, the processing plan is revised, such as re-performing feature matching or adjusting positioning parameters. After verification and correction, the anomaly handling process ends.

[0141] Step S145: Associate the feature matching positioning results, accuracy verification results, and anomaly handling results, and annotate the positioning information, accuracy verification status, and anomaly handling details of each key area to obtain the initial circuit board key area positioning results.

[0142] The AOI detection unit correlates feature matching localization results, accuracy verification results, and anomaly handling results. It annotates the localization information of each key area, such as location coordinates and area range; it annotates the accuracy verification status, such as whether the localization accuracy meets requirements; and it annotates anomaly handling details, such as anomaly type, handling method, and handling result. The correlated and annotated results are then integrated into the initial circuit board key area localization results, where each data entry includes fields such as image identifier, key area identifier, type, localization information, accuracy verification status, and anomaly handling details.

[0143] Step S146: Compare the initial key area positioning results of the circuit board with the key area annotation information in the circuit board design drawings, verify the consistency between the positioning position, area range and design information, correct the positioning content corresponding to the deviation, and obtain the positioning result verification and correction result.

[0144] The AOI inspection unit compares the initial location results of critical areas on the circuit board with the critical area annotations in the circuit board design drawings. It verifies the consistency between the located position and the critical area position in the design drawings, and the consistency between the area range and the critical area range in the design drawings. For any deviations in the location, the location position or area range is corrected. The verified and corrected results are integrated into a location result verification and correction result, where each data entry includes fields such as image identifier, critical area identifier, type, location information, design annotation information, deviation details, and corrected location content.

[0145] Step S147: Extract the core information from the key area positioning results of the circuit board. The core information includes the location coordinates, boundary range, type label and positioning accuracy information of each key area. Arrange the information according to the circuit board area order to obtain the core positioning information set.

[0146] The AOI (Automated Optical Inspection) unit extracts core information from the positioning results of key areas on the circuit board, including the location coordinates, boundary range, type label, and positioning accuracy information of each key area. This information is then organized according to the order of the circuit board areas, with information from the same area grouped together. The organized information is integrated into a core positioning information set, where each data entry includes fields such as circuit board area, key area identifier, type, location coordinates, boundary range, and positioning accuracy information.

[0147] Step S148: Supplement the AOI detection unit operating parameters, positioning time and detection perspective information during the positioning process to the core positioning information set, label the image source and rule basis corresponding to each positioning result, and obtain the positioning results of the key area of ​​the circuit board.

[0148] The AOI detection unit supplements the core positioning information set by adding AOI detection unit operating parameters during the positioning process, such as voltage, current, and temperature; positioning time, i.e., the time taken for the positioning process; and detection perspective information, such as perspective markers, detection angles, and acquisition distances. It also labels the image source corresponding to each positioning result, indicating which image the positioning result comes from, and labels the rule basis, i.e., the circuit board key area positioning rule clauses upon which the positioning results are based. The supplemented and labeled results are integrated into the circuit board key area positioning results, where each data entry includes fields such as circuit board area, key area marker, type, location coordinates, boundary range, positioning accuracy information, operating parameters, positioning time, detection perspective information, image source, and rule basis.

[0149] Step S149: Organize the positioning results of the key areas of the circuit board according to the preset format, and output the positioning results of the key areas of the circuit board, and transmit them to the calibration module simultaneously.

[0150] The AOI inspection unit organizes the location results of critical areas of the circuit board according to a preset format, which includes the data arrangement order and field display method. After organization, it outputs the location results of critical areas of the circuit board, which can be displayed on the terminal screen or stored on the server. Simultaneously, the location results are transmitted to the calibration module for subsequent parameter calibration.

[0151] Step S150: Based on the positioning results of the key area of ​​the circuit board, the linkage acquisition parameters of the AOI detection unit and the positioning rules of the key area of ​​the circuit board are reverse-calibrated to generate parameter calibration instructions. The AOI detection unit executes the parameter calibration instructions iterative linkage acquisition mechanism to form a two-way driving process of positioning and calibration.

[0152] The AOI inspection unit reverse-calibrates the linkage acquisition parameters and positioning rules based on the positioning results of key areas on the circuit board. It analyzes deviations in the positioning results to determine the adjustment direction of the linkage acquisition parameters and the execution effect of the positioning rules to determine the optimization direction of the positioning rules. It generates parameter calibration instructions, which include the adjustment values ​​of the linkage acquisition parameters and the optimization content of the positioning rules. The AOI inspection unit executes the parameter calibration instructions, adjusts the linkage acquisition parameters, optimizes the positioning rules, and iterates the linkage acquisition mechanism. This forms a two-way driven process of positioning and calibration: the positioning results are used to calibrate the parameters and rules, and the calibrated parameters and rules are used for more accurate positioning.

[0153] Step S151: Analyze the positioning results of key areas of the circuit board, extract the positioning information, accuracy verification status, anomaly handling details and corresponding AOI detection unit operating parameters of each key area, and obtain the calibration basic data.

[0154] The AOI inspection unit analyzes the positioning results of key areas on the circuit board, extracting positioning information for each key area, such as location coordinates and area range; extracting accuracy verification information, such as whether the positioning accuracy meets requirements; extracting anomaly handling details, such as anomaly type, handling method, and handling result; and extracting corresponding AOI inspection unit operating parameters, such as voltage, current, and temperature. The extracted information is integrated into calibration baseline data, where each data entry includes fields such as key area identifier, positioning information, accuracy verification information, anomaly handling details, and operating parameters.

[0155] Step S152: Associate the calibration base data with the linkage acquisition parameters of the AOI detection unit, calculate the functional relationship between positioning accuracy and detection viewing angle, acquisition distance, and image resolution, identify the linkage acquisition parameter values ​​that cause the positioning accuracy to be lower than the preset threshold or trigger abnormal positioning, and obtain the acquisition parameter deviation analysis results.

[0156] The AOI detection unit correlates the calibration baseline data with the linked acquisition parameters of the AOI detection unit. These linked acquisition parameters include the detection viewing angle, acquisition distance, and image resolution. The system calculates the functional relationship between positioning accuracy and the detection viewing angle (i.e., how positioning accuracy changes with the detection viewing angle); the functional relationship between positioning accuracy and the acquisition distance (i.e., how positioning accuracy changes with the acquisition distance); and the functional relationship between positioning accuracy and image resolution (i.e., how positioning accuracy changes with image resolution). Based on these functional relationships, the system identifies the linked acquisition parameter values ​​that cause positioning accuracy to fall below a preset threshold or trigger abnormal positioning. The analysis results are integrated into an acquisition parameter deviation analysis result, where each data point includes fields such as the linked acquisition parameter type, functional relationship, the parameter value causing the deviation, and the deviation manifestation.

[0157] Step S153: Based on the analysis results of the acquisition parameter deviation, adjust the linkage acquisition parameters of the AOI detection unit, correct the detection viewing angle, acquisition distance and image resolution, set the verification standard after parameter adjustment, and obtain the linkage acquisition parameter calibration results.

[0158] The AOI detection unit adjusts the linked acquisition parameters based on the analysis results of the acquired parameter deviation. For the detection viewing angle, the angle value is adjusted according to a functional relationship to improve positioning accuracy; for the acquisition distance, the distance value is adjusted according to a functional relationship to improve positioning accuracy; for the image resolution, the resolution value is adjusted according to a functional relationship to improve positioning accuracy. Verification standards are set for the adjusted parameters, such as positioning accuracy reaching a preset threshold and the number of abnormal positioning occurrences being lower than a preset number. The results of the adjustments and settings are integrated into the linked acquisition parameter calibration results, where each data entry includes fields such as linked acquisition parameter type, value before adjustment, value after adjustment, and verification standard.

[0159] For example, step S1531: Analyze the analysis results of the acquisition parameter deviation, extract the types of linkage acquisition parameter deviations that lead to insufficient positioning accuracy and abnormal positioning. The deviation types include detection viewing angle deviation, acquisition distance deviation and image resolution deviation. Mark the influence range and manifestation of each deviation to obtain parameter deviation details.

[0160] The AOI detection unit analyzes the results of the acquired parameter deviation analysis, extracting the types of linked acquired parameter deviations that lead to insufficient positioning accuracy and abnormal positioning, including detection viewpoint angle deviation, acquisition distance deviation, and image resolution deviation. The impact range of each deviation is labeled; for example, detection viewpoint angle deviation affects the coverage of key areas, acquisition distance deviation affects image clarity, and image resolution deviation affects the accuracy of feature extraction. The manifestation of each deviation is also labeled; for example, detection viewpoint angle deviation causes key areas to be uncovered, acquisition distance deviation causes image blurring, and image resolution deviation causes feature extraction errors. The extracted and labeled information is integrated into a parameter deviation detail, where each data point includes fields such as deviation type, impact range, and manifestation.

[0161] Step S1532: For the detection viewing angle deviation, based on the position coordinates of the key area of ​​the circuit board and the deployment position of the AOI detection unit, calculate the viewing angle adjustment amount required to make the key area fall into the center of the image through geometric projection relationship, and drive the rotation mechanism to perform the angle adjustment to obtain the viewing angle calibration result.

[0162] The AOI detection unit calibrates to address viewing angle deviations. A geometric projection model is established based on the location coordinates of the critical area on the circuit board and the deployment position of the AOI detection unit. The required viewing angle adjustment to center the critical area in the image is calculated using the geometric projection relationship; this is the difference between the current and target viewing angles. The rotation mechanism is then driven to perform this angle adjustment, bringing the detection viewing angle to the target angle. The adjustment results are integrated into a viewing angle calibration result, which includes fields such as critical area identification, current viewing angle, target viewing angle, adjustment amount, and adjustment result.

[0163] Step S1533: For the acquisition distance deviation, calculate the optimal acquisition distance range that meets the image clarity requirements based on the minimum feature size of the key area and the depth of field range of the lens, and drive the linear module to move the AOI detection unit to a specified distance within the optimal acquisition distance range to obtain the acquisition distance calibration result.

[0164] The AOI detection unit is calibrated to address acquisition distance deviations. A sharpness model is established based on the minimum feature size of the critical area and the lens's depth of field. The optimal acquisition distance range that meets image sharpness requirements is calculated using this model; that is, the range within which image sharpness is highest. The driving linear module moves the AOI detection unit to a specified distance within this optimal acquisition distance range. The adjustment results are integrated into an acquisition distance calibration result, which includes fields such as critical area identification, minimum feature size, depth of field range, optimal acquisition distance range, specified distance, and adjustment result.

[0165] Step S1534: For image resolution deviation, calculate the minimum number of pixels required to meet feature extraction based on the minimum feature size and positioning accuracy requirements of the key area, and set the resolution parameters of the image sensor accordingly. At the same time, limit the image resolution to within the preset maximum resolution threshold to obtain the image resolution calibration result.

[0166] The AOI detection unit is calibrated to address image resolution deviations. A pixel count model is established based on the minimum feature size and positioning accuracy requirements of the key region. This model calculates the minimum number of pixels required for feature extraction; that is, the image sensor resolution needs to reach this number of pixels to clearly extract features from the key region. Based on this, the image sensor resolution parameters are set, while the image resolution is limited to a preset maximum resolution threshold to avoid excessive data volume due to excessively high resolution. The settings are integrated into an image resolution calibration result, which includes fields such as key region identification, minimum feature size, positioning accuracy requirements, minimum pixel count, set resolution parameters, and maximum resolution threshold.

[0167] Step S1535: Correlate the viewing angle calibration results, acquisition distance calibration results, and image resolution calibration results, analyze the synergistic effect of each parameter adjustment, so that the parameter combination can improve the overall positioning accuracy and reduce abnormal positioning situations, and obtain the parameter combination calibration results.

[0168] The AOI detection unit correlates the viewing angle calibration results, acquisition distance calibration results, and image resolution calibration results. It analyzes the synergistic effect of adjusting each parameter, i.e., the impact of adjusting the viewing angle, acquisition distance, and image resolution on the overall positioning accuracy. The parameter combinations are adjusted to improve overall positioning accuracy and reduce abnormal positioning situations. The adjusted parameter combinations are then integrated into a parameter combination calibration result, which includes fields such as viewing angle, acquisition distance, image resolution, synergistic effect analysis, and the adjusted parameter combinations.

[0169] Step S1536: Set the verification criteria after parameter adjustment, covering the requirements for key area coverage integrity, image clarity, feature extraction accuracy and positioning accuracy, determine the verification process and judgment criteria, and obtain the parameter verification criteria.

[0170] The AOI detection unit sets verification standards after parameter adjustments. These standards cover: critical area coverage integrity (whether the critical area is completely covered); image clarity (whether the image is clear and distinguishable); feature extraction accuracy (whether feature extraction is accurate); and positioning accuracy requirements (whether the deviation between the positioning result and the design drawings is within acceptable limits). The verification process is defined, such as first verifying critical area coverage integrity, then image clarity, followed by feature extraction accuracy, and finally positioning accuracy. Judgment criteria are determined, such as critical area coverage integrity reaching a preset proportion, image clarity reaching a preset level, feature extraction accuracy reaching a preset accuracy rate, and positioning accuracy reaching a preset threshold. The set standards and defined processes are integrated into a parameter verification standard, which includes fields such as verification items, verification standards, verification process, and judgment criteria.

[0171] Step S1537: According to the parameter verification standard, simulate the operation effect of the adjusted linkage acquisition parameters, acquire circuit board images and perform key area positioning simulation, record the positioning accuracy and abnormal situation occurrence rate during the simulation process, and obtain the parameter simulation verification results.

[0172] The AOI inspection unit undergoes simulation verification according to parameter validation standards. The simulation examines the operational effect of the adjusted linkage acquisition parameters, specifically the acquisition process after adjustments to the inspection viewing angle, acquisition distance, and image resolution. Circuit board images are acquired, and key area positioning simulations are performed. The positioning accuracy during the simulation is recorded, i.e., the deviation between the positioning result and the design drawings; the anomaly occurrence rate is also recorded, i.e., the ratio of abnormal positioning occurrences to the total number of positioning occurrences. The recorded results are integrated into a parameter simulation verification result, which includes fields such as simulation parameters, positioning accuracy, and anomaly occurrence rate.

[0173] Step S1538: Based on the parameter simulation verification results, fine-tune the linkage acquisition parameters, correct the parameter adjustment range, optimize the parameter combination, so that the positioning effect after parameter adjustment meets the verification standard, and obtain the parameter fine-tuning results.

[0174] The AOI detection unit fine-tunes the linkage acquisition parameters based on the parameter simulation verification results. If the positioning accuracy does not reach the preset threshold, the parameter adjustment range is corrected, such as further adjusting the detection viewing angle, acquisition distance, or image resolution. If the occurrence rate of abnormal situations does not meet the preset requirements, the parameter combination is optimized, such as adjusting the proportional relationship between parameters. The fine-tuned parameters are integrated into a parameter fine-tuning result, which includes fields such as linkage acquisition parameter type, value before fine-tuning, value after fine-tuning, and adjustment basis.

[0175] Step S1539: Integrate the parameter combination calibration results and parameter fine-tuning results, and mark the values ​​before adjustment, after adjustment, adjustment basis and verification effect of the AOI detection unit linkage acquisition parameters to obtain the initial linkage acquisition parameter calibration results.

[0176] The AOI detection unit integrates the parameter combination calibration results with the parameter fine-tuning results. It annotates the pre-adjustment values, post-adjustment values, adjustment basis, and verification results of the AOI detection unit's linked acquisition parameters. For example, the adjustment basis is the parameter simulation verification result, and the verification result is that the positioning accuracy reaches a preset threshold. The integrated and annotated results are combined into the initial linked acquisition parameter calibration results, where each data point includes fields such as linked acquisition parameter type, pre-adjustment value, post-adjustment value, adjustment basis, and verification result.

[0177] Step S15310: Supplement the parameter adjustment sequence in the initial linkage acquisition parameter calibration result, determine the adjustment order and mutual coordination requirements of the viewing angle, acquisition distance, and image resolution, and generate the final linkage acquisition parameter calibration result.

[0178] The AOI detection unit supplements the initial linkage acquisition parameter calibration results by adding parameter adjustment timing. It determines the adjustment order for viewing angle, acquisition distance, and image resolution, such as adjusting the acquisition distance first, then the viewing angle, and finally the image resolution. It also determines the coordination requirements, such as adjusting the viewing angle after adjusting the acquisition distance to ensure coverage of key areas. The supplemented results are then integrated into the final linkage acquisition parameter calibration results, where each data point includes fields such as linkage acquisition parameter type, value before adjustment, value after adjustment, adjustment basis, verification effect, adjustment order, and coordination requirements.

[0179] Step S154: Associate the calibration base data with the positioning rules of the key areas of the circuit board, analyze the adaptability of the feature correspondence logic, accuracy control requirements, anomaly handling process and positioning results in the positioning rules, identify the clauses in the rules that cause positioning deviation and inefficient anomaly handling, and obtain the positioning rule deviation analysis results.

[0180] The AOI inspection unit correlates calibration baseline data with the positioning rules for key areas of the circuit board. It analyzes the compatibility between the feature-correspondence logic in the positioning rules and the positioning results, i.e., whether the feature-correspondence logic can accurately match image features and key areas; it analyzes the compatibility between accuracy control requirements and the positioning results, i.e., whether the accuracy control requirements can effectively guarantee positioning accuracy; and it analyzes the compatibility between the anomaly handling process and the positioning results, i.e., whether the anomaly handling process can effectively handle positioning anomalies. It identifies clauses in the rules that lead to positioning deviations and inefficient anomaly handling. The analysis results are integrated into a positioning rule deviation analysis result, where each data point includes fields such as positioning rule clause, compatibility analysis, cause of deviation, and inefficiency.

[0181] Step S155: Based on the deviation analysis results of the positioning rules, optimize the positioning rules of the key areas of the circuit board, adjust the judgment criteria of the core feature combination in the feature correspondence logic, correct the deviation range in the accuracy control requirements, improve the operation steps in the abnormal handling process, and obtain the optimization results of the positioning rules.

[0182] Based on the deviation analysis results of the positioning rules, the AOI detection unit optimizes the positioning rules for key areas of the circuit board. This includes adjusting the judgment criteria for core feature combinations in the feature correspondence logic, such as increasing or decreasing the number of core features or adjusting their weights; correcting the deviation range in accuracy control requirements, such as expanding or narrowing the allowable deviation range; and improving the operational steps in the anomaly handling process, such as adding anomaly handling steps or optimizing the operation sequence. The optimized rules are then integrated into the positioning rule optimization result, where each data entry includes fields such as positioning rule clauses, content before adjustment, content after adjustment, and adjustment basis.

[0183] Step S156: Integrate the calibration results of the linkage acquisition parameters with the optimization results of the positioning rules, determine the adjustment content of the linkage acquisition parameters of the AOI detection unit, the optimization clauses of the positioning rules, and the collaborative adjustment logic of the two, and obtain the integrated calibration content result.

[0184] The AOI detection unit integrates the calibration results of the linked acquisition parameters with the optimization results of the positioning rules. This involves determining the adjustment content of the linked acquisition parameters, such as the adjustment values ​​for the detection viewing angle, acquisition distance, and image resolution; determining the optimization clauses of the positioning rules, such as the optimization content of feature correspondence logic, accuracy control requirements, and anomaly handling procedures; and determining the collaborative adjustment logic between the two, such as the need to optimize the positioning rules after adjusting the linked acquisition parameters, and vice versa. The integrated and determined information is then combined into a calibration content integration result, which includes fields such as the linked acquisition parameter adjustment content, the positioning rule optimization clauses, and the collaborative adjustment logic.

[0185] Step S157: Generate a parameter calibration instruction framework, annotate the core content of parameter calibration, execution steps and timing requirements, divide the order of linkage acquisition parameter adjustment and positioning rule optimization, and obtain the initial parameter calibration instruction.

[0186] The AOI detection unit generates a parameter calibration instruction framework. It annotates the core content of parameter calibration, such as adjusting linkage acquisition parameters and optimizing positioning rules; it annotates the execution steps, such as adjusting linkage acquisition parameters first, then optimizing positioning rules; it annotates the timing requirements, such as completing linkage acquisition parameter adjustment within a specified time, and completing positioning rule optimization within a specified time after linkage acquisition parameter adjustment; it defines the order of linkage acquisition parameter adjustment and positioning rule optimization, such as linkage acquisition parameter adjustment taking precedence over positioning rule optimization. The annotated and defined information is integrated into an initial parameter calibration instruction, which includes fields such as core content, execution steps, timing requirements, and order of execution.

[0187] Step S158: According to the operating logic of the AOI detection unit, split the calibration content and integrate the results, supplement the specific operation guidelines for adjusting each parameter, the execution details after the optimization of the positioning rules, mark the precautions in the calibration process, and obtain the parameter calibration instructions.

[0188] The AOI detection unit breaks down the integrated calibration results according to its operational logic. It breaks down the linkage acquisition parameter adjustment content, obtaining specific operational guidelines for each parameter adjustment, such as the specific steps for adjusting the detection viewpoint angle, acquisition distance, and image resolution. It also breaks down the positioning rule optimization clauses, obtaining the execution details after optimization, such as the execution steps after optimizing feature correspondence logic, accuracy control requirements, and exception handling procedures. It notes precautions during the calibration process, such as disabling image acquisition when adjusting linkage acquisition parameters and backing up the original rules when optimizing positioning rules. The broken down, supplemented, and labeled information is integrated into a parameter calibration instruction, which includes fields such as parameter adjustment operation guidelines, positioning rule execution details, and precautions.

[0189] Step S159: Verify the feasibility of the parameter calibration command. Based on the hardware capacity of the AOI detection unit and the circuit board detection requirements, correct the command clauses with execution obstacles, adjust the parameter adjustment range and timing requirements, obtain the final parameter calibration command, and output the final parameter calibration command to the AOI detection unit to trigger the linkage acquisition parameter adjustment and positioning rule update, and complete the reverse calibration process.

[0190] The AOI inspection unit verifies the feasibility of the parameter calibration instructions. The instructions are checked against the hardware capabilities of the AOI inspection unit, such as the angle adjustment range of the rotating mechanism, the movement range of the linear module, and the resolution range of the image sensor, to ensure they are within the hardware's limits. The instructions are also checked against the circuit board's inspection requirements, such as inspection speed and accuracy, to ensure they meet the requirements. Instructions with execution obstacles are corrected, such as adjusting parameter adjustments to keep them within the hardware's limits and adjusting timing requirements to meet inspection speed requirements. The verified and corrected instructions are integrated into a final parameter calibration instruction, which is then output and transmitted to the AOI inspection unit. The AOI inspection unit executes the instruction, adjusts the linked acquisition parameters, updates the positioning rules, and completes the reverse calibration process.

[0191] Furthermore, Figure 2 A schematic diagram of the hardware structure of a circuit board critical area positioning system 100 based on AOI detection for implementing the method provided in the embodiments of this application is shown. Figure 2 As shown, the AOI-based circuit board critical area positioning system 100 may include at least one processor 102 (the processor 102 may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.), a memory 104 for storing data, a transmission device 106 for communication functions, and a controller 108. Those skilled in the art will understand that... Figure 2 The structure shown is for illustrative purposes only and does not limit the structure of the AOI-based critical area localization system 100 for circuit boards. For example, the AOI-based critical area localization system 100 for circuit boards may also include components that are more advanced than those shown in the diagram. Figure 2 The more or fewer components shown, or having the same Figure 2 The different configurations shown.

[0192] The memory 104 can be used to store software programs and modules for application software, such as the program instructions corresponding to the method embodiments described above in this application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby realizing the above-described method for locating key areas of a circuit board based on AOI detection. The transmission device 106 is used to acquire or send data via a network.

[0193] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

Claims

1. A method for locating critical areas of a circuit board based on AOI detection, characterized in that, The method includes: The AOI detection unit uses the association information between the circuit board design drawings and the detection perspective to collect circuit board images and generate linked acquisition results. The linked acquisition results include images of each area of ​​the circuit board, detection perspective parameters, and the correspondence between the images and the design drawings. Based on the results of the joint acquisition, the key area annotation information in the circuit board design drawings is associated with the area features in the image to generate a set of associated features of the key area of ​​the circuit board. By combining the feature association set of key areas of the circuit board, a key area positioning rule for the circuit board is generated. The key area positioning rule for the circuit board defines the correspondence between image features and key areas, positioning accuracy control requirements, and abnormal positioning processing procedures. The AOI detection unit locates the key areas of the circuit board image in the linkage acquisition results according to the key area positioning rules of the circuit board, and generates the key area positioning results of the circuit board. Based on the location results of the key areas of the circuit board, the linkage acquisition parameters of the AOI detection unit and the location rules of the key areas of the circuit board are calibrated in reverse to generate parameter calibration instructions. The AOI detection unit executes the parameter calibration instructions iterative linkage acquisition mechanism to form a two-way driving process of positioning and calibration. Based on the results of the linked acquisition, the key area annotation information in the circuit board design drawings is associated with the area features in the image to generate a set of associated key area features of the circuit board, including: The results of the coordinated acquisition are analyzed, and the images of each area of ​​the circuit board, the detection perspective parameters, and the correspondence between the images and the design drawings are extracted. The images are then classified and organized according to the key area types of the circuit board to obtain the image analysis results. The image parsing results are used to extract the regional features of each region of the circuit board image. The regional features include line width, solder joint shape, line spacing and regional grayscale distribution features. The circuit board region and detection viewpoint parameters corresponding to each regional feature are labeled to obtain the image region feature set. Retrieve the associated circuit board design drawings from the linkage acquisition results, extract the key area annotation information from the circuit board design drawings, the key area annotation information includes the key area type, location coordinates, design dimensions and performance requirements, and obtain the key area information of the design drawings; The image region feature set is matched with the key region information of the design drawing. The key regions are compared one by one according to the type of circuit board. When the difference between the image region feature and the key region annotation information is within the preset tolerance range, it is determined that the match is successful. The successfully matched feature pairs and their tolerances are recorded to obtain the feature information matching result. In the feature matching results, identify the parts where the image region features do not match the key region information of the design drawings, analyze the reasons for the differences, and distinguish whether they are caused by the acquisition perspective deviation or the image feature extraction deviation, and obtain the feature matching deviation analysis results. Based on the feature matching deviation analysis results, the matching logic between image region features and key region annotation information is adjusted, the feature association relationship corresponding to the deviation is corrected, and the differential feature description is supplemented to obtain the feature association correction result; Extract the core features of key areas of the circuit board, label the core features and feature combinations that can uniquely identify each key area, eliminate interference from non-key features, and obtain the core feature extraction results. The correlation correction results and core feature extraction results are organized according to the key area type of the circuit board. The image area features, design annotation information and core feature combinations corresponding to each key area are determined to obtain the initial circuit board key area feature correlation set.

2. The method for locating critical areas of a circuit board based on AOI detection according to claim 1, characterized in that, The AOI detection unit, based on the correlation information between the circuit board design drawings and the detection viewpoint, acquires circuit board images in a coordinated manner and generates coordinated acquisition results, including: The AOI detection unit retrieves the circuit board design drawings, extracts the key area annotation information, circuit layout and the relationship between the size of each area in the circuit board design drawings, and generates a set of circuit board design information. Obtain detection viewpoint association information, which includes the deployment location of the AOI detection unit, the detection angle adjustment range, and the image acquisition range under different viewpoints, and generate a set of detection viewpoint parameters; By associating the circuit board design information set with the detection viewpoint parameter set, the image coverage and clarity of each key area of ​​the circuit board under different detection viewpoints are calculated. The detection viewpoint combination that can cover all key areas and meet the preset clarity threshold is selected to obtain the viewpoint combination determination result. The results are determined by the AOI detection unit according to the combination of viewing angles. The detection angle and acquisition distance are adjusted, and the multi-view image acquisition operation is started synchronously to acquire images of each area of ​​the circuit board. The detection viewing angle parameters corresponding to each acquisition are recorded to obtain the multi-view image acquisition set. The multi-view image acquisition set is associated and matched with the circuit board design information set. The circuit board area, detection view parameters and corresponding positions in the design drawings corresponding to each circuit board image are labeled to obtain the image association and matching results. The system identifies blurred, occluded, or incomplete circuit board images in a multi-view image acquisition set, marks the corresponding areas and detection view parameters, readjusts the detection view to supplement the acquisition, generates supplementary acquisition images, and obtains image quality optimization results. By integrating multi-view image acquisition sets, image association matching results, and image quality optimization results, duplicate circuit board images are removed, and acquired images and corresponding association information are added to obtain the initial linkage acquisition results. By associating historical data on the effects of coordinated acquisition, cases of unreasonable viewpoint combinations and poor image quality during historical acquisition are extracted. The viewpoint parameter annotation errors and image association deviations in the initial coordinated acquisition results are corrected to obtain the corrected coordinated acquisition results. Supplement the acquisition time and equipment operation status information of the AOI detection unit to the corrected linkage acquisition results, mark the generation basis of each part and the connection relationship with subsequent steps, obtain the linkage acquisition results, and sort the linkage acquisition results according to the circuit board area.

3. The method for locating critical areas of a circuit board based on AOI detection according to claim 1, characterized in that, The AOI detection unit, according to the circuit board key area localization rules, performs key area localization on the circuit board image in the linked acquisition results, and generates circuit board key area localization results, including: The AOI detection unit retrieves the positioning rules for key areas of the circuit board, extracts feature correspondence rules, accuracy control rules, and anomaly handling rules, and combines them with the detection perspective parameters in the linkage acquisition results to generate the basis for the execution of positioning rules. According to the feature correspondence rule, regional feature extraction and matching are performed on the circuit board image in the linkage acquisition results. The image region features are compared with the core feature combination of each key region. The successfully matched regions and the corresponding key region types are marked to obtain the feature matching and localization results. Based on the accuracy control rules, the accuracy of the boundary and location coordinate annotations of key areas in the feature matching positioning results is verified. Positioning content with deviations exceeding the allowable range is adjusted, and boundary annotations and coordinate information are corrected to obtain the accuracy verification results. Monitor abnormal situations during the feature matching and localization process, start the corresponding processing flow according to the abnormality handling rules, re-extract image region features and match again when feature matching fails, call supplementary acquired images from the corresponding detection viewpoint to assist localization when the localization result is blurry, and combine design drawing annotation information to distinguish when features from multiple regions are confused, and obtain abnormality handling results. By associating feature matching positioning results, accuracy verification results, and anomaly handling results, and annotating the positioning information, accuracy verification status, and anomaly handling details of each key area, the initial positioning results of the key areas of the circuit board are obtained. The initial key area location results of the circuit board are compared with the key area annotation information in the circuit board design drawings. The consistency between the location position, area range and design information is verified, and the location content corresponding to the deviation is corrected to obtain the location result verification and correction result. Extract the core information from the key area positioning results of the circuit board. The core information includes the location coordinates, boundary range, type label and positioning accuracy information of each key area. Arrange the information according to the circuit board area order to obtain the core positioning information set. Supplement the AOI detection unit's operating parameters, positioning time, and detection perspective information during the positioning process into the core positioning information set, and label the image source and rule basis corresponding to each positioning result to obtain the positioning results of the key areas of the circuit board. The positioning results of the key areas of the circuit board are organized according to the preset format, and the positioning results of the key areas of the circuit board are output and synchronously transmitted to the calibration module.

4. The method for locating critical areas of a circuit board based on AOI detection according to claim 1, characterized in that, Based on the location results of key areas of the circuit board, the linkage acquisition parameters and key area location rules of the AOI detection unit are reverse-calibrated to generate parameter calibration instructions, including: Analyze the positioning results of key areas of the circuit board, extract the positioning information, accuracy verification status, anomaly handling details and corresponding AOI detection unit operating parameters of each key area, and obtain the calibration basic data; By linking the basic calibration data with the linkage acquisition parameters of the AOI detection unit, the functional relationship between positioning accuracy and detection angle, acquisition distance, and image resolution is calculated. The linkage acquisition parameter values ​​that cause positioning accuracy to fall below the preset threshold or trigger abnormal positioning are identified, and the acquisition parameter deviation analysis results are obtained. Based on the analysis results of the acquisition parameter deviation, the linkage acquisition parameters of the AOI detection unit are adjusted, the detection viewing angle, acquisition distance and image resolution are corrected, the verification standard after parameter adjustment is set, and the linkage acquisition parameter calibration results are obtained. By correlating calibration baseline data with the positioning rules for key areas of the circuit board, analyzing the compatibility between the feature correspondence logic, accuracy control requirements, anomaly handling procedures and positioning results in the positioning rules, identifying clauses in the rules that lead to positioning deviations and inefficient anomaly handling, and obtaining the positioning rule deviation analysis results; Based on the results of the positioning rule deviation analysis, the positioning rules for key areas of the circuit board are optimized, the judgment criteria for core feature combination in the feature correspondence logic are adjusted, the deviation range in the accuracy control requirements is corrected, and the operation steps in the anomaly handling process are improved to obtain the positioning rule optimization results. By integrating the calibration results of the linkage acquisition parameters and the optimization results of the positioning rules, the adjustment content of the linkage acquisition parameters of the AOI detection unit, the optimization clauses of the positioning rules, and the collaborative adjustment logic of the two are determined, and the integrated calibration content result is obtained. Generate a parameter calibration instruction framework, annotate the core content of parameter calibration, execution steps and timing requirements, divide the order of linkage acquisition parameter adjustment and positioning rule optimization, and obtain the initial parameter calibration instruction; Based on the operating logic of the AOI detection unit, the calibration content is broken down and the results are integrated. Specific operation guidelines for adjusting each parameter, execution details after optimization of the positioning rules are added, and precautions during the calibration process are marked to obtain parameter calibration instructions. To verify the feasibility of the parameter calibration command, the hardware capacity of the AOI detection unit and the circuit board detection requirements were compared. The command clauses with execution obstacles were corrected, the parameter adjustment range and timing requirements were adjusted, and the final parameter calibration command was obtained. The final parameter calibration command was output and transmitted to the AOI detection unit to trigger the linkage acquisition parameter adjustment and positioning rule update, thus completing the reverse calibration process.

5. The method for locating critical areas of a circuit board based on AOI detection according to claim 2, characterized in that, The associated circuit board design information set and detection viewpoint parameter set are used to calculate the image coverage and clarity of each key area of ​​the circuit board under different detection viewpoints. A combination of detection viewpoints that can cover all key areas and meet a preset clarity threshold is selected to obtain the viewpoint combination determination result, including: Analyze the circuit board design information set, extract the location distribution, size, shape characteristics and their relationship with the surrounding areas of each key area of ​​the circuit board, and obtain the key area collection requirements details; The detection viewpoint parameter set is analyzed, and the physical position coordinates of the AOI detection unit, the detection viewpoint angle adjustment range, the image acquisition area coordinates and imaging resolution parameters under each viewpoint are extracted to obtain the detailed characteristics of the detection viewpoint. By associating the key area acquisition requirements with the inspection viewpoint characteristics, the coverage, imaging clarity and acquisition completeness of the key areas of the circuit board by different inspection views are analyzed one by one to obtain the single-view acquisition evaluation results. For key areas that cannot be fully covered or have blurred images in the single-view acquisition evaluation results, other detection views that can supplement the coverage of the key areas and improve the image quality are selected. The selected detection views and their supplementary acquisition advantages are labeled to obtain the supplementary view selection results. Based on the single-view acquisition and evaluation results and the supplementary view selection results, the detection views are combined to ensure that each key area is fully covered and clearly imaged by at least one detection view. The overlap of the view coverage is controlled, and the repeated acquisition areas are reduced by adjusting the view angle to obtain the initial view combination. The acquisition sequence and connection logic of each detection perspective in the initial perspective combination are analyzed, the perspective switching sequence is adjusted, the perspective switching time is shortened, the acquisition process is ensured to be continuous, and the acquisition perspectives of the missing key areas are supplemented to obtain the perspective combination optimization results. The key area coverage, imaging parameter requirements and view switching conditions corresponding to each detection view in the annotation view combination optimization results are determined, the acquisition time of each view and the collaborative acquisition logic with other views are determined, and the view combination details are obtained. By associating historical perspective combination data, we extract cases from the historical acquisition process where the perspective combination was unreasonable, the acquisition time exceeded the preset range, and the imaging could not clearly present the core features of key areas. We then correct the problems of perspective selection deviation and unreasonable switching logic in the perspective combination details to obtain the perspective combination correction results. The supplementary viewpoint combination correction results include descriptions of applicable scenarios, covering adjustment methods for viewpoint combinations under different types of circuit boards and different testing environments, marking the basis for adjustment, and obtaining the initial viewpoint combination determination results; Simulation tests are conducted on the initial viewpoint combination determination results: Based on the circuit board design information set and the detection viewpoint parameter set, the theoretical image coverage and expected image clarity of each key area under each viewpoint combination are calculated; viewpoint combinations in which the theoretical image coverage of all key areas reaches the set coverage and the expected image clarity exceeds the preset clarity threshold are selected as valid candidate combinations; if there are multiple candidate combinations, the combination with the fewest viewpoint switching times is selected to generate the final viewpoint combination determination results.

6. The method for locating critical areas of a circuit board based on AOI detection according to claim 1, characterized in that, The process involves matching the image region feature set with the key region information of the design drawings, comparing them one by one according to the key region type of the circuit board. When the difference between the image region features and the key region annotation information is within a preset tolerance range, a successful match is determined, and the successfully matched feature pairs and their tolerances are recorded to obtain the feature information matching result, including: The image region feature set is analyzed, and the line width, solder joint shape, line spacing, and regional grayscale distribution features of each circuit board region image are extracted, along with their corresponding circuit board regions and detection viewpoint parameters. The features are then classified and organized according to the circuit board regions to obtain detailed image features. Analyze the key area information of the design drawings, extract the type, location coordinates, design dimensions, performance requirements and the relationship with the surrounding areas of each circuit board's key areas, classify and organize them according to the type of key areas, and obtain the design feature details; A priority rule for feature matching is set, which stipulates that: during the matching process, the size features of line width and solder joint shape are matched first. When the matching degree of size features exceeds a first threshold, regional grayscale distribution features are introduced for auxiliary matching. Among them, the confidence weight of the size feature matching result is higher than the confidence weight of the regional grayscale distribution feature matching result, thus obtaining the matching priority setting result. The image feature details and design feature details are associated one by one according to the key area type of the circuit board. First, the fit of the core design size features is compared, and then the auxiliary features are combined for further verification. The feature fit points and differences are marked to obtain the single-type area matching results. For regions with low feature fit in single-type region matching results, analyze the reasons for the differences, adjust the matching logic, supplement the description of the differences, and re-perform feature matching to obtain the corrected results for the difference region matching. By associating the single-type region matching results with the difference region matching correction results, the image region features, design features and matching basis corresponding to the key regions of each circuit board are labeled, the range of successfully matched regions and the matching accuracy are determined, and the preliminary feature matching results are obtained. The initial results of feature matching show confusion between multiple image regions corresponding to the same key region and a single image region corresponding to multiple key regions. By combining the detection viewpoint parameters and the location coordinates of the design drawings, the correspondence is distinguished, the confusion problem is corrected, and the matching confusion correction result is obtained. Integrate the preliminary feature matching results with the matching confusion correction results, organize them according to the key area type of the circuit board, and annotate the image feature correspondence, matching accuracy and abnormal situations in the matching process of each key area to obtain the initial feature information matching results; The matching conditions in the initial feature information matching results are supplemented, covering the adjustment methods for feature matching under different detection perspectives and the criteria for judging matching accuracy, to obtain the feature information matching results; Verify the accuracy of the feature information matching results, check the uniqueness of the image region features corresponding to each key region, correct matching errors and feature omissions, and complete the feature information matching process.

7. The method for locating critical areas of a circuit board based on AOI detection according to claim 3, characterized in that, In the monitoring feature matching and localization process, abnormal situations are handled according to the corresponding processing rules. When feature matching fails, image region features are re-extracted and matched again. When the localization result is blurry, supplementary images acquired from the corresponding detection viewpoint are used to assist in localization. When features from multiple regions are confused, the design drawing annotation information is used to distinguish them, resulting in the following abnormality handling results: During the feature matching and localization process, the success status of image region feature extraction, the success status of feature matching, and the coordinate deviation value of multiple consecutive localization results are detected in real time. When any status is failure or the deviation value exceeds the threshold, the abnormal monitoring result is triggered. In cases where feature matching fails in anomaly monitoring results, the image region features of the corresponding area are re-extracted according to the anomaly handling rules, the feature extraction range and extraction method are adjusted, the missing core features are supplemented, and the features are combined with the key region features again for matching to obtain the feature re-matching result. In response to the fuzzy positioning results in the anomaly monitoring results, supplementary images from the corresponding detection perspective are called according to the anomaly handling rules. Regional features in the supplementary images are extracted and combined with the features of the original images to locate key regions, correct the fuzzy positioning boundaries and coordinate information, and obtain the fuzzy positioning correction results. In response to the situation of multiple regions having confused features in the anomaly monitoring results, the key region annotation information in the circuit board design drawings is retrieved according to the anomaly handling rules. The design dimensions, location coordinates and surrounding relationships of each confused region are compared. The different key regions are distinguished by combining the image region features, the correspondence of each region is determined, and the confused region distinction results are obtained. For localization anomalies in the anomaly monitoring results that cannot be resolved through feature rematching, fuzzy localization correction and confusion region differentiation, the corresponding anomaly region and anomaly type are marked according to the anomaly handling rules. The AOI detection unit operating parameters, image acquisition information and matching process at the time of the anomaly are recorded, and the manual auxiliary judgment process is initiated to obtain the manual auxiliary judgment result. The results of rematching associated features, fuzzy localization correction, confusion region differentiation, and manual assistance judgment are labeled with the processing method, process, and localization effect after processing for each anomaly, thus obtaining preliminary results of anomaly handling. The processing efficiency and effect of each abnormal case in the preliminary results of anomaly handling were analyzed. The operation steps and timing in the anomaly handling process were adjusted, and the conditions for feature re-extraction and supplementary image retrieval were optimized to obtain the optimized results of anomaly handling. By integrating the preliminary and optimized results of anomaly handling, the final handling scheme, post-handling positioning accuracy, and subsequent preventive measures for each anomaly type are determined, and the initial anomaly handling results are obtained. Supplement the initial anomaly handling results with anomaly cause analysis, label the root cause of each anomaly, and obtain the anomaly handling results; Verify the validity of the anomaly handling results, check whether the processed positioning results meet the accuracy control requirements, correct any incomplete processing or positioning deviations, and complete the anomaly handling process.

8. A circuit board critical area positioning system based on AOI detection, characterized in that, The device includes a processor and a readable storage medium storing a program that, when executed by the processor, implements the method for locating critical areas of a circuit board based on AOI detection as described in any one of claims 1-7.

Images