Multi-light-source combined imaging and detection method and system for FPC (flexible printed circuit) appearance defects

Abstract

The invention provides a multi-light-source combined imaging and detection method and system for FPC appearance defects. The method comprises the following steps: if a to-be-detected state of an FPC is a qualified state, acquiring an FPC appearance image, analyzing and processing the FPC appearance image to obtain FPC identification data, identifying initial FPC defect data, analyzing and processing the initial FPC defect data to obtain adaptive light source parameters, acquiring a first suspected defect area image, analyzing and processing the first suspected defect area image to obtain a defect fusion image, and then obtaining an identification defect area; correlation analysis is carried out through a preset FPC production process database to obtain process adjustment parameters, and the process adjustment parameters are transmitted to an FPC control end; self-adaptive light source parameters are matched through the base material, the circuit density and the defect type, light reflection is restrained through polarization imaging, the detection spectrum is expanded in combination with the ultraviolet light source, forward feedback from defect detection to technological parameter optimization is achieved, and therefore multi-light-source combined imaging and detection of FPC appearance defects are achieved.

Description

Technical Field

[0001] This application relates to the field of industrial machine vision inspection technology, and more specifically, to a multi-source combined imaging and inspection method and system for FPC appearance defects. Background Technology

[0002] FPC production often uses different materials with complex material properties and processes, easily leading to various appearance defects such as open circuits, short circuits, pinholes, burrs, foreign matter adhesion, exposed copper, ink peeling, and interlayer bubbles. For appearance defects, traditional detection techniques often rely on experience and fixed light source combinations and parameter configurations, which cannot dynamically adapt to different substrates, circuit densities, and defect types. Insufficient reflection suppression during image capture results in poor adaptability. Multiple light sources capturing images at different times or simultaneously can easily lead to information redundancy, missing information, or coexistence, making it difficult to accurately integrate the defect characteristics of different light sources, which can easily lead to missed or false detections. After completing defect identification and classification, traditional techniques cannot match the previous generation, cannot achieve defect cause tracing analysis, and cannot optimize the process. Furthermore, conventional light source imaging cannot accurately detect latent or complex defects such as ink peeling and adhesive residue that are sensitive to specific spectra, resulting in insufficient detection effectiveness.

[0003] Effective technical solutions are urgently needed to address the above problems. Summary of the Invention

[0004] The purpose of this application is to provide a multi-source combined imaging and detection method and system for FPC appearance defects. It can achieve multi-source combined imaging and detection of FPC appearance defects by matching adaptive light source parameters with substrate, line density and defect type, suppressing reflection by polarization imaging, expanding the detection spectrum by combining ultraviolet light source, and realizing positive feedback from defect detection to process parameter optimization.

[0005] In a first aspect, this application provides a multi-source combined imaging and detection method for FPC appearance defects, including the following steps: The inspection status of the FPC is obtained. If it is in a qualified state, the appearance image of the FPC is obtained and analyzed to obtain FPC identification data. Initial FPC defect data is identified based on the FPC identification data, and adaptive light source parameters are obtained by analyzing and processing the initial FPC defect data. Image acquisition is performed based on the adaptive light source parameters to obtain an image of the first suspected defect area; The image of the first suspected defect area is analyzed and processed to obtain a defect fusion image; The defect fusion image is analyzed and processed to identify the defect region. Based on the identified defect areas, correlation analysis is performed using a preset FPC production process database to obtain process adjustment parameters, which are then transmitted to the FPC control terminal.

[0006] Optionally, in the multi-source combined imaging and detection method for FPC appearance defects described in this application, the step of acquiring the FPC's inspection state, and if it is in a qualified state, acquiring the FPC appearance image and performing analysis and processing to obtain FPC identification data includes: Obtain the dust removal completion status, stage fixing status, and pose status of the FPC; If the dust removal completion status is qualified, the platform fixation status is qualified in terms of flatness and pose status, then the FPC's inspection status is determined to be qualified. Conversely, the FPC is determined to be in an unqualified state and the abnormality is handled according to the preset abnormality handling method. If the FPC is in a qualified state, an image of the FPC's appearance will be obtained by acquiring the image through a preset imaging light source combination. Based on the FPC appearance image, the preset substrate type identification algorithm, preset line density identification algorithm, and preset suspected defect area identification algorithm are respectively analyzed and processed to obtain FPC identification data, including substrate type feature data, line density level, and first suspected defect area image.

[0007] Optionally, in the multi-source combined imaging and detection method for FPC appearance defects described in this application, the step of identifying initial FPC defect data based on the FPC identification data and analyzing and processing the initial FPC defect data to obtain adaptive light source parameters includes: The image of the first suspected defect area is subjected to feature extraction to obtain a defect feature vector; Based on the substrate type feature data and the line density level, a corresponding preset defect type identification sub-model is obtained by matching with a preset defect type identification model library. The defect feature vector is input into a preset defect type identification sub-model for analysis and processing to obtain initial FPC defect data, including suspected defect category feature data and suspected defect sub-category feature data. The suspected defect category feature data includes concave-convex morphology defect feature data, latent surface defect feature data, fine line defect feature data, material absorption anomaly defect feature data, or unrecorded defect feature data. If the suspected defect category feature data is concave-convex morphology defect feature data, latent surface defect feature data, fine line defect feature data, or material absorption abnormality defect feature data, then adaptive matching of light source parameters is performed based on the suspected defect category feature data and suspected defect sub-category feature data to obtain adaptive light source parameters. If the suspected defect category feature data is unrecorded defect feature data, then the first suspected defect area image and the corresponding defect feature vector are input into a preset light source parameter prediction model for analysis and processing to obtain adaptive light source parameters.

[0008] Optionally, in the multi-source combined imaging and detection method for FPC appearance defects described in this application, the step of acquiring an image based on the adaptive light source parameters to obtain an image of the second suspected defect area includes: Based on the adaptive light source parameters, images of the suspected defect area are acquired to obtain a second suspected defect area image; The second suspected defect area image includes a polarized coaxial image, an annular oblique projection image, a dot matrix backlight image, and an ultraviolet supplemental light image.

[0009] Optionally, in the multi-source combined imaging and detection method for FPC appearance defects described in this application, the step of analyzing and processing the image of the second suspected defect area to obtain a defect fusion image includes: The polarization coaxial image, ring oblique illumination image, dot matrix backlight image and ultraviolet supplementary illumination image are subjected to dark field correction, background subtraction, noise reduction and contrast enhancement image preprocessing to obtain polarization coaxial optimized image, ring oblique illumination optimized image, dot matrix backlight optimized image and ultraviolet supplementary illumination optimized image. Polarization features are extracted from a preset number of polarization coaxial optimized images to obtain polarization degree feature maps and polarization angle feature maps. The polarization angle feature maps include sine polarization angle feature maps and cosine polarization angle feature maps. The local contrast of the polarization degree feature map, sine polarization angle feature map, cosine polarization angle feature map, polarization coaxial optimized image, annular oblique illumination optimized image, dot matrix backlight optimized image and ultraviolet supplementary light optimized image is obtained and normalized to obtain the corresponding image weight values. The polarization degree feature map, sine polarization angle feature map, cosine polarization angle feature map, polarization coaxial optimized image, annular oblique illumination optimized image, dot matrix backlight optimized image, and ultraviolet supplementary light optimized image are weighted and summed with their corresponding image weight values ​​to obtain a defect fusion image.

[0010] Optionally, in the multi-source combined imaging and detection method for FPC appearance defects described in this application, the step of analyzing and processing the defect fusion image to obtain the identified defect region includes: Based on the defect fusion image, a third suspected defect region image is obtained by analyzing and processing it using a preset morphological operation method and a preset connected component analysis method. Feature extraction is performed on the image of the third suspected defect area to obtain defect feature data, including geometric feature data and texture feature data; Based on the polarization degree feature map and polarization angle feature map, feature extraction is performed to obtain polarization feature data; Based on the geometric feature data, texture feature data, and polarization feature data, a matching analysis is performed using a preset defect feature database to obtain defect feature matching degree parameters. The defect feature matching degree parameter is compared with the preset defect feature recognition threshold. If the defect feature matching degree parameter is greater than or equal to the preset defect feature recognition threshold, then the third suspected defect area image is determined to be the defect area to be identified, and the defect category feature data corresponding to the maximum value of the defect feature matching degree parameter is determined to be the final defect category feature data. If the defect feature matching degree parameter is less than the preset defect feature recognition threshold, then the third suspected defect area image is determined to be a non-identifiable defect area.

[0011] Optionally, in the multi-source combined imaging and detection method for FPC appearance defects described in this application, the step of performing correlation analysis based on the identified defect area through a preset FPC production process database to obtain process adjustment parameters and transmitting them to the FPC control terminal includes: Obtain the defect location data of the identified defect area; Based on the defect location data and the final defect category feature data, correlation analysis is performed through a preset FPC production process database to obtain the process parameters corresponding to the generation process in which the source defect was generated. Based on the defect category feature data and defect subcategory feature data, adaptive adjustments are made using a preset process parameter adjustment method to obtain process adjustment parameters, which are then transmitted to the FPC control terminal.

[0012] Secondly, this application provides a multi-source combined imaging and detection system for FPC appearance defects. The system includes a memory and a processor. The memory includes a program for a multi-source combined imaging and detection method for FPC appearance defects. When the program for the multi-source combined imaging and detection method for FPC appearance defects is executed by the processor, it performs the following steps: The inspection status of the FPC is obtained. If it is in a qualified state, the appearance image of the FPC is obtained and analyzed to obtain FPC identification data. Initial FPC defect data is identified based on the FPC identification data, and adaptive light source parameters are obtained by analyzing and processing the initial FPC defect data. Image acquisition is performed based on the adaptive light source parameters to obtain an image of the first suspected defect area; The image of the first suspected defect area is analyzed and processed to obtain a defect fusion image; The defect fusion image is analyzed and processed to identify the defect region. Based on the identified defect areas, correlation analysis is performed using a preset FPC production process database to obtain process adjustment parameters, which are then transmitted to the FPC control terminal.

[0013] Optionally, in the multi-source combined imaging and detection system for FPC appearance defects described in this application, the step of acquiring the FPC's inspection status, and if it is in a qualified state, acquiring the FPC appearance image and performing analysis and processing to obtain FPC identification data includes: Obtain the dust removal completion status, stage fixing status, and pose status of the FPC; If the dust removal completion status is qualified, the platform fixation status is qualified in terms of flatness and pose status, then the FPC's inspection status is determined to be qualified. Conversely, the FPC is determined to be in an unqualified state and the abnormality is handled according to the preset abnormality handling method. If the FPC is in a qualified state, an image of the FPC's appearance will be obtained by acquiring the image through a preset imaging light source combination. Based on the FPC appearance image, the preset substrate type identification algorithm, preset line density identification algorithm, and preset suspected defect area identification algorithm are respectively analyzed and processed to obtain FPC identification data, including substrate type feature data, line density level, and first suspected defect area image.

[0014] Optionally, in the multi-source combined imaging and detection system for FPC appearance defects described in this application, the step of identifying initial FPC defect data based on the FPC identification data and analyzing and processing the initial FPC defect data to obtain adaptive light source parameters includes: The image of the first suspected defect area is subjected to feature extraction to obtain a defect feature vector; Based on the substrate type feature data and the line density level, a corresponding preset defect type identification sub-model is obtained by matching with a preset defect type identification model library. The defect feature vector is input into a preset defect type identification sub-model for analysis and processing to obtain initial FPC defect data, including suspected defect category feature data and suspected defect sub-category feature data. The suspected defect category feature data includes concave-convex morphology defect feature data, latent surface defect feature data, fine line defect feature data, material absorption anomaly defect feature data, or unrecorded defect feature data. If the suspected defect category feature data is concave-convex morphology defect feature data, latent surface defect feature data, fine line defect feature data, or material absorption abnormality defect feature data, then adaptive matching of light source parameters is performed based on the suspected defect category feature data and suspected defect sub-category feature data to obtain adaptive light source parameters. If the suspected defect category feature data is unrecorded defect feature data, then the first suspected defect area image and the corresponding defect feature vector are input into a preset light source parameter prediction model for analysis and processing to obtain adaptive light source parameters.

[0015] As can be seen from the above, the multi-source combined imaging and detection method and system for FPC appearance defects provided in this application achieves multi-source combined imaging and detection of FPC appearance defects by matching adaptive light source parameters with substrate, line density and defect type, suppressing reflection by polarization imaging, expanding the detection spectrum by combining ultraviolet light source, and realizing positive feedback from defect detection to process parameter optimization.

[0016] Other features and advantages of this application will be set forth in the following description and will be apparent in part from the description or may be learned by practicing embodiments of this application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart of a multi-source combined imaging and detection method for FPC appearance defects provided in an embodiment of this application; Figure 2 A flowchart illustrating the process of obtaining FPC identification data using a multi-source combined imaging and detection method for FPC appearance defects, as provided in an embodiment of this application. Figure 3 A flowchart illustrating the process of obtaining adaptive light source parameters for a multi-light source combined imaging and detection method for FPC appearance defects provided in an embodiment of this application; Figure 4 This is a high-level flowchart of a multi-source combined imaging and detection method for FPC appearance defects provided in an embodiment of this application. Detailed Implementation

[0019] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0020] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0021] Please refer to Figure 1 , Figure 1 This is a flowchart of a multi-source combined imaging and detection method for FPC appearance defects according to some embodiments of this application. This multi-source combined imaging and detection method for FPC appearance defects is used in terminal devices, such as computers and mobile terminals. The multi-source combined imaging and detection method for FPC appearance defects includes the following steps: S101. Obtain the inspection status of the FPC. If it is in a qualified state, obtain the appearance image of the FPC and perform analysis and processing to obtain FPC identification data. S102. Identify initial FPC defect data based on the FPC identification data, analyze and process the initial FPC defect data, and obtain adaptive light source parameters. S103. Image acquisition is performed according to the adaptive light source parameters to obtain an image of the first suspected defect area; S104. Analyze and process the image of the first suspected defect area to obtain a defect fusion image; S105. Analyze and process the defect fusion image to obtain the identified defect region; S106. Based on the identified defect area, perform correlation analysis through the preset FPC production process database to obtain process adjustment parameters, and transmit them to the FPC control terminal.

[0022] Further explanation is needed regarding the adaptive combination of light sources and defect identification for different FPC substrates. First, after ensuring the FPC's inspection status is qualified, a rapid pre-scan of the appearance image is performed to identify the substrate type, circuit density, and initial suspected defects. Then, the corresponding defect identification model is matched according to the substrate type and circuit density to identify the defect type of the initial suspected defects. Next, adaptive light source parameters are determined based on the defect type. Then, a precise image set of the suspected defect area is re-acquired based on the determined adaptive light source parameters. The weight value is determined based on the local contrast of each image, and weighted fusion is performed. Finally, the defect type is accurately identified in the fused image, and correlation analysis and adjustments are made to the preceding production process to achieve positive feedback.

[0023] Please refer to Figure 2 , Figure 2 This is a flowchart illustrating the process of obtaining FPC identification data using a multi-source combined imaging and detection method for FPC appearance defects, as described in some embodiments of this application. According to embodiments of the present invention, the step of obtaining the FPC's inspection status, and if it is in a qualified state, acquiring an FPC appearance image and performing analysis and processing to obtain FPC identification data, includes: S21. Obtain the dust removal completion status, stage fixing status, and pose status of the FPC; S221. If the dust removal completion status is qualified, the platform fixation status is qualified in terms of flatness and the pose status is qualified, then the FPC's inspection status is determined to be qualified. S222. Conversely, the FPC is determined to be in an unqualified state and the abnormality is handled according to the preset abnormality handling method. S23. If the FPC is in a qualified state, then the image is acquired by using a preset imaging light source combination to obtain the appearance image of the FPC. S24. Based on the FPC appearance image, the preset substrate type identification algorithm, preset line density identification algorithm, and preset suspected defect area identification algorithm are used to analyze and process the data to obtain FPC identification data, including substrate type feature data, line density level, and first suspected defect area image.

[0024] It should be further explained that after the dust removal operation, a local image is collected by a macro camera to identify the number of visible particles and the particle diameter. The dust removal completion status is determined to be qualified by comparing with a threshold value. If both the number of particles and the particle diameter are less than the corresponding threshold values, it is determined that the dust removal is qualified. The surface of the FPC is scanned by a laser displacement sensor, and the maximum fluctuation value is compared with the corresponding threshold value. If it is less than the threshold value, it is determined that the flatness of the stage fixing state is qualified. The deviation between the theoretical position and the actual position is determined by establishing a transformation matrix between the FPC coordinate system and the stage coordinate system, and then compared with the threshold value to determine whether the pose state is qualified. If it is less than the threshold value, it is qualified. If any item is unqualified, an early warning is output and adjusted. For example, if the dust removal is unqualified, the vacuum pressure is increased or the processing time is extended. If the flatness is unqualified, the pressure is adjusted or replayed. If the pose is unqualified, the position of the stage and the rotation angle of the platform are adjusted. Only when the dust removal completion status is qualified, the stage fixing state is qualified in terms of flatness, and the pose state is qualified in terms of pose, the pending inspection state of the FPC is determined to be a qualified state. By using a preset fast imaging light source combination for fast scanning, an appearance image of the FPC is obtained. Finally, through a preset substrate type recognition algorithm (collecting images by illuminating with red light (630nm), green light (520nm), and blue light (450nm) respectively in the substrate area, calculating the average gray value of each wavelength, and normalizing to obtain the reflectance curve for determination), a preset line density recognition algorithm (using a coaxial white light image, performing Gaussian filtering to eliminate noise, adaptive threshold segmentation to obtain a binary image, morphological closing operation to fill the line gaps, connected component analysis, removing noise with an area <500 pixels, dividing the image into a 10×10 grid, calculating the proportion of the line area in each grid, and comparing with the threshold value for determination), and a preset suspected defect area recognition algorithm (simultaneously analyzing the coaxial white light image and the annular red light image, calculating the normalized cross-correlation difference map of the two images, performing abnormal area detection, texture abnormality detection, and edge discontinuity detection to calibrate the suspected defect area) for analysis and processing, the substrate type characteristic data, the line density level, and the first suspected defect area image are determined. Among them, the substrate type characteristic data is represented by a unique identifier.

[0025] Please refer to Figure 3 , Figure 3 is a flowchart for obtaining adaptive light source parameters of a multi-light source combination imaging and detection method for FPC appearance defects in some embodiments of the present application. According to an embodiment of the present invention, identifying initial FPC defect data according to the FPC identification data, and performing analysis and processing according to the initial FPC defect data to obtain adaptive light source parameters includes: S311. Extract features from the first suspected defect area image to obtain a defect feature vector; ​ S32. Input the defect feature vector into the preset defect type identification sub-model for analysis and processing to obtain initial FPC defect data, including suspected defect category feature data and suspected defect sub-category feature data. The suspected defect category feature data includes concave-convex morphology defect feature data, latent surface defect feature data, fine line defect feature data, material absorption anomaly defect feature data, or unrecorded defect feature data. S331. If the suspected defect category feature data is concave-convex morphology defect feature data, latent surface defect feature data, fine line defect feature data, or material absorption abnormality defect feature data, then adaptive matching of light source parameters is performed based on the suspected defect category feature data and suspected defect sub-category feature data to obtain adaptive light source parameters. S332. If the suspected defect category feature data is unrecorded defect feature data, then the first suspected defect area image and the corresponding defect feature vector are input into the preset light source parameter prediction model for analysis and processing to obtain adaptive light source parameters.

[0026] Further explanation is needed: to achieve adaptive light source combination matching and parameter setting driven by initial defect identification, a person skilled in the art trains a library of preset defect type identification models based on defect feature vectors corresponding to different substrate type feature data and line density levels, along with the corresponding initial FPC defect data. This library is constructed from multiple preset defect type identification sub-models. Then, the corresponding sub-models are matched based on the real-time identified substrate type feature data and line density levels. The corresponding initial FPC defect data is identified based on the real-time extracted defect feature vectors. The suspected defect sub-category feature data includes line gaps, pinholes, pits, ink peeling, foreign matter adhesion, or interlayer bubbles. Defect category feature data and suspected defect subcategory feature data are represented by unique identifiers, such as DB01 for line gaps. For successfully identified defects, the corresponding adaptive light source parameters are matched by a pre-set optimal light source parameter combination database (pre-built by those skilled in the art and dynamically adjustable). For unrecorded defects (including complex defects and unknown defects), a lightweight CNN prediction model is used for analysis and processing to obtain adaptive light source parameters (decoded into specific parameters based on the output 8-dimensional light source parameter vector). The pre-set light source parameter prediction model is obtained by training a large number of historical samples of first suspected defect area images and defect feature vectors and corresponding adaptive light source parameters.

[0027] According to an embodiment of the present invention, the step of acquiring an image of a second suspected defect region based on the adaptive light source parameters includes: Based on the adaptive light source parameters, images of the suspected defect area are acquired to obtain a second suspected defect area image; The second suspected defect area image includes a polarized coaxial image, an annular oblique projection image, a dot matrix backlight image, and an ultraviolet supplemental light image.

[0028] It needs to be further explained that after determining the light source parameters, the suspected defect areas identified are re-acquired according to a preset time sequence, including three frames (polarization directions of 0°, 45° and 90° respectively) of polarized coaxial images, one frame each of the ring oblique illumination image, the dot matrix backlight image and the ultraviolet supplementary illumination image.

[0029] According to an embodiment of the present invention, the step of analyzing and processing the image of the second suspected defect region to obtain a defect fusion image includes: The polarization coaxial image, ring oblique illumination image, dot matrix backlight image and ultraviolet supplementary illumination image are subjected to dark field correction, background subtraction, noise reduction and contrast enhancement image preprocessing to obtain polarization coaxial optimized image, ring oblique illumination optimized image, dot matrix backlight optimized image and ultraviolet supplementary illumination optimized image. Polarization features are extracted from a preset number of polarization coaxial optimized images to obtain polarization degree feature maps and polarization angle feature maps. The polarization angle feature maps include sine polarization angle feature maps and cosine polarization angle feature maps. The local contrast of the polarization degree feature map, sine polarization angle feature map, cosine polarization angle feature map, polarization coaxial optimized image, annular oblique illumination optimized image, dot matrix backlight optimized image and ultraviolet supplementary light optimized image is obtained and normalized to obtain the corresponding image weight values. The polarization degree feature map, sine polarization angle feature map, cosine polarization angle feature map, polarization coaxial optimized image, annular oblique illumination optimized image, dot matrix backlight optimized image, and ultraviolet supplementary light optimized image are weighted and summed with their corresponding image weight values ​​to obtain a defect fusion image.

[0030] Further explanation is needed. First, the acquired multiple frames of images are preprocessed separately. Then, the polarization features of the polarization-coaxial optimized image are extracted, and its local contrast is calculated. Weights are assigned based on the local contrast of each image. Images with higher contrast have a higher weight at that pixel. The multiple images are then weighted and summed to obtain the defect fusion image.

[0031] According to an embodiment of the present invention, the step of analyzing and processing the defect fusion image to obtain the identified defect region includes: Based on the defect fusion image, a third suspected defect region image is obtained by analyzing and processing it using a preset morphological operation method and a preset connected component analysis method. Feature extraction is performed on the image of the third suspected defect area to obtain defect feature data, including geometric feature data and texture feature data; Based on the polarization degree feature map and polarization angle feature map, feature extraction is performed to obtain polarization feature data; Based on the geometric feature data, texture feature data, and polarization feature data, a matching analysis is performed using a preset defect feature database to obtain defect feature matching degree parameters. The defect feature matching degree parameter is compared with the preset defect feature recognition threshold. If the defect feature matching degree parameter is greater than or equal to the preset defect feature recognition threshold, then the third suspected defect area image is determined to be the defect area to be identified, and the defect category feature data corresponding to the maximum value of the defect feature matching degree parameter is determined to be the final defect category feature data. If the defect feature matching degree parameter is less than the preset defect feature recognition threshold, then the third suspected defect area image is determined to be a non-identifiable defect area.

[0032] Further explanation is needed regarding the process of performing morphological operations (such as opening and closing operations) on the fused image to remove noise and small objects, connecting adjacent regions, marking the third suspected defect region through connected component analysis, and extracting geometric feature data (including area, perimeter, aspect ratio, circularity, and minimum bounding rectangle) and texture feature data (including gray-level co-occurrence matrix features and local binary data). Simultaneously, polarization feature data (including average DoP and average AoP within the region) is extracted based on the polarization degree feature map and polarization angle feature map corresponding to the third suspected defect region. The extracted geometric feature data, texture feature data, and polarization feature data are then matched and analyzed using a distance metric against a pre-set defect feature database to obtain defect feature matching parameters. Finally, a threshold comparison is used to determine whether the third suspected defect region is a true defect. For regions identified as defective, matching is performed based on the extracted features to determine the final defect category. The pre-set defect feature database is pre-constructed by those skilled in the art based on numerous historical cases and can be dynamically adjusted.

[0033] According to an embodiment of the present invention, the step of performing correlation analysis based on the identified defect area through a preset FPC production process database to obtain process adjustment parameters and transmitting them to the FPC control terminal includes: Obtain the defect location data of the identified defect area; Based on the defect location data and the final defect category feature data, correlation analysis is performed through a preset FPC production process database to obtain the process parameters corresponding to the generation process in which the source defect was generated. Based on the defect category feature data and defect subcategory feature data, adaptive adjustments are made using a preset process parameter adjustment method to obtain process adjustment parameters, which are then transmitted to the FPC control terminal.

[0034] Further explanation is needed: by correlating the location and type of defects with a pre-set FPC production process database, the process parameters corresponding to the production process that caused the defects are obtained, and the traceability results are fed back to the front-end production equipment to automatically adjust the corresponding process parameters, thereby achieving positive feedback. For example, if an open circuit defect is determined, by correlating the open circuit location with the pre-set FPC production process database, it is found that the etching time of this batch of FPCs is 8% longer than the standard value, and feedback is given to suggest that the etching time be adjusted to the standard range. The pre-set FPC production process database and the pre-set process parameter adjustment method are pre-built by our technical personnel and can be dynamically adjusted.

[0035] Please refer to Figure 4 , Figure 4 This is a high-level flowchart of a multi-source combined imaging and detection method for FPC appearance defects according to some embodiments of this application.

[0036] This invention also discloses a multi-source combined imaging and detection system for FPC appearance defects, including a memory and a processor. The memory includes a program for a multi-source combined imaging and detection method for FPC appearance defects. When the processor executes the program for the multi-source combined imaging and detection method for FPC appearance defects, it performs the following steps: The inspection status of the FPC is obtained. If it is in a qualified state, the appearance image of the FPC is obtained and analyzed to obtain FPC identification data. Initial FPC defect data is identified based on the FPC identification data, and adaptive light source parameters are obtained by analyzing and processing the initial FPC defect data. Image acquisition is performed based on the adaptive light source parameters to obtain an image of the first suspected defect area; The image of the first suspected defect area is analyzed and processed to obtain a defect fusion image; The defect fusion image is analyzed and processed to identify the defect region. Based on the identified defect areas, correlation analysis is performed using a preset FPC production process database to obtain process adjustment parameters, which are then transmitted to the FPC control terminal.

[0037] Further explanation is needed regarding the adaptive combination of light sources and defect identification for different FPC substrates. First, after ensuring the FPC's inspection status is qualified, a rapid pre-scan of the appearance image is performed to identify the substrate type, circuit density, and initial suspected defects. Then, the corresponding defect identification model is matched according to the substrate type and circuit density to identify the defect type of the initial suspected defects. Next, adaptive light source parameters are determined based on the defect type. Then, a precise image set of the suspected defect area is re-acquired based on the determined adaptive light source parameters. The weight value is determined based on the local contrast of each image, and weighted fusion is performed. Finally, the defect type is accurately identified in the fused image, and correlation analysis and adjustments are made to the preceding production process to achieve positive feedback.

[0038] According to an embodiment of the present invention, the step of obtaining the inspection status of the FPC, if it is in a qualified state, involves obtaining an FPC appearance image and performing analysis and processing to obtain FPC identification data, including: Obtain the dust removal completion status, stage fixing status, and pose status of the FPC; If the dust removal completion status is qualified, the platform fixation status is qualified in terms of flatness and pose status, then the FPC's inspection status is determined to be qualified. Conversely, the FPC is determined to be in an unqualified state and the abnormality is handled according to the preset abnormality handling method. If the FPC is in a qualified state, an image of the FPC's appearance will be obtained by acquiring the image through a preset imaging light source combination. Based on the FPC appearance image, the preset substrate type identification algorithm, preset line density identification algorithm, and preset suspected defect area identification algorithm are respectively analyzed and processed to obtain FPC identification data, including substrate type feature data, line density level, and first suspected defect area image.

[0039] It should be further explained that, after the dust removal operation, local images are collected by a macro camera to identify the number of visible particles and the particle diameter. By comparing with thresholds, it is determined whether the dust removal completion status is qualified for dust removal. If both the number of particles and the particle diameter are less than the corresponding thresholds, it is determined that the dust removal is qualified. The surface of the FPC is scanned by a laser displacement sensor, and the maximum fluctuation value is compared with the corresponding threshold. If it is less than the threshold, it is determined that the flatness of the stage fixing state is qualified. By establishing a transformation matrix between the FPC coordinate system and the stage coordinate system, the deviation between the theoretical position and the actual position is determined, and then compared with the threshold to determine whether the pose state is qualified. If it is less than the threshold, it is qualified. If any item is unqualified, an early warning is output and adjusted. For example, if the dust removal is unqualified, the vacuum pressure is increased or the processing time is extended. If the flatness is unqualified, the pressure is adjusted or replayed. If the pose is unqualified, the stage position and the rotation platform adjustment angle are adjusted. Only when the dust removal completion status is qualified for dust removal, the stage fixing state is qualified for flatness, and the pose state is qualified for pose, it is determined that the inspection status of the FPC is a qualified state. By using a preset fast imaging light source combination for fast scanning, the appearance image of the FPC is obtained. Finally, through a preset substrate type recognition algorithm (collecting images by illuminating with red light (630nm), green light (520nm), and blue light (450nm) in the substrate area respectively, calculating the average gray value of each wavelength, and normalizing to obtain the reflectivity curve for determination), a preset line density recognition algorithm (using a coaxial white light image, performing Gaussian filtering to eliminate noise, adaptive threshold segmentation to obtain a binary image, morphological closing operation to fill the line gaps, connected component analysis, removing noise with an area <500 pixels, dividing the image into a 10×10 grid, calculating the proportion of the line area in each grid, and comparing with the threshold for determination), and a preset suspected defect area recognition algorithm (simultaneously analyzing the coaxial white light image and the annular red light image, calculating the normalized cross-correlation difference map of the two images, performing abnormal area detection, texture abnormality detection, and edge discontinuity detection, thereby calibrating the suspected defect area) for analysis and processing, the substrate type characteristic data, the line density level, and the first suspected defect area image are determined. Among them, the substrate type characteristic data is represented by a unique identifier.

[0040] According to an embodiment of the present invention, identifying the initial FPC defect data according to the FPC identification data, and performing analysis and processing according to the initial FPC defect data to obtain adaptive light source parameters, including: Performing feature extraction on the first suspected defect area image to obtain a defect feature vector; Matching according to the substrate type characteristic data and the line density level through a preset defect type recognition model library to obtain a corresponding preset defect type recognition sub-model; The defect feature vector is input into a preset defect type identification sub-model for analysis and processing to obtain initial FPC defect data, including suspected defect category feature data and suspected defect sub-category feature data. The suspected defect category feature data includes concave-convex morphology defect feature data, latent surface defect feature data, fine line defect feature data, material absorption anomaly defect feature data, or unrecorded defect feature data. If the suspected defect category feature data is concave-convex morphology defect feature data, latent surface defect feature data, fine line defect feature data, or material absorption abnormality defect feature data, then adaptive matching of light source parameters is performed based on the suspected defect category feature data and suspected defect sub-category feature data to obtain adaptive light source parameters. If the suspected defect category feature data is unrecorded defect feature data, then the first suspected defect area image and the corresponding defect feature vector are input into a preset light source parameter prediction model for analysis and processing to obtain adaptive light source parameters.

[0041] Further explanation is needed: to achieve adaptive light source combination matching and parameter setting driven by initial defect identification, a person skilled in the art trains a library of preset defect type identification models based on defect feature vectors corresponding to different substrate type feature data and line density levels, along with the corresponding initial FPC defect data. This library is constructed from multiple preset defect type identification sub-models. Then, the corresponding sub-models are matched based on the real-time identified substrate type feature data and line density levels. The corresponding initial FPC defect data is identified based on the real-time extracted defect feature vectors. The suspected defect sub-category feature data includes line gaps, pinholes, pits, ink peeling, foreign matter adhesion, or interlayer bubbles. Defect category feature data and suspected defect subcategory feature data are represented by unique identifiers, such as DB01 for line gaps. For successfully identified defects, the corresponding adaptive light source parameters are matched by a pre-set optimal light source parameter combination database (pre-built by those skilled in the art and dynamically adjustable). For unrecorded defects (including complex defects and unknown defects), a lightweight CNN prediction model is used for analysis and processing to obtain adaptive light source parameters (decoded into specific parameters based on the output 8-dimensional light source parameter vector). The pre-set light source parameter prediction model is obtained by training a large number of historical samples of first suspected defect area images and defect feature vectors and corresponding adaptive light source parameters.

[0042] According to an embodiment of the present invention, the step of acquiring an image of a second suspected defect region based on the adaptive light source parameters includes: Based on the adaptive light source parameters, images of the suspected defect area are acquired to obtain a second suspected defect area image; The second suspected defect area image includes a polarized coaxial image, an annular oblique projection image, a dot matrix backlight image, and an ultraviolet supplemental light image.

[0043] It needs to be further explained that after determining the light source parameters, the suspected defect areas identified are re-acquired according to a preset time sequence, including three frames (polarization directions of 0°, 45° and 90° respectively) of polarized coaxial images, one frame each of the ring oblique illumination image, the dot matrix backlight image and the ultraviolet supplementary illumination image.

[0044] According to an embodiment of the present invention, the step of analyzing and processing the image of the second suspected defect region to obtain a defect fusion image includes: The polarization coaxial image, ring oblique illumination image, dot matrix backlight image and ultraviolet supplementary illumination image are subjected to dark field correction, background subtraction, noise reduction and contrast enhancement image preprocessing to obtain polarization coaxial optimized image, ring oblique illumination optimized image, dot matrix backlight optimized image and ultraviolet supplementary illumination optimized image. Polarization features are extracted from a preset number of polarization coaxial optimized images to obtain polarization degree feature maps and polarization angle feature maps. The polarization angle feature maps include sine polarization angle feature maps and cosine polarization angle feature maps. The local contrast of the polarization degree feature map, sine polarization angle feature map, cosine polarization angle feature map, polarization coaxial optimized image, annular oblique illumination optimized image, dot matrix backlight optimized image and ultraviolet supplementary light optimized image is obtained and normalized to obtain the corresponding image weight values. The polarization degree feature map, sine polarization angle feature map, cosine polarization angle feature map, polarization coaxial optimized image, annular oblique illumination optimized image, dot matrix backlight optimized image, and ultraviolet supplementary light optimized image are weighted and summed with their corresponding image weight values ​​to obtain a defect fusion image.

[0045] Further explanation is needed. First, the acquired multiple frames of images are preprocessed separately. Then, the polarization features of the polarization-coaxial optimized image are extracted, and its local contrast is calculated. Weights are assigned based on the local contrast of each image. Images with higher contrast have a higher weight at that pixel. The multiple images are then weighted and summed to obtain the defect fusion image.

[0046] According to an embodiment of the present invention, the step of analyzing and processing the defect fusion image to obtain the identified defect region includes: Based on the defect fusion image, a third suspected defect region image is obtained by analyzing and processing it using a preset morphological operation method and a preset connected component analysis method. Feature extraction is performed on the image of the third suspected defect area to obtain defect feature data, including geometric feature data and texture feature data; Based on the polarization degree feature map and polarization angle feature map, feature extraction is performed to obtain polarization feature data; Based on the geometric feature data, texture feature data, and polarization feature data, a matching analysis is performed using a preset defect feature database to obtain defect feature matching degree parameters. The defect feature matching degree parameter is compared with the preset defect feature recognition threshold. If the defect feature matching degree parameter is greater than or equal to the preset defect feature recognition threshold, then the third suspected defect area image is determined to be the defect area to be identified, and the defect category feature data corresponding to the maximum value of the defect feature matching degree parameter is determined to be the final defect category feature data. If the defect feature matching degree parameter is less than the preset defect feature recognition threshold, then the third suspected defect area image is determined to be a non-identifiable defect area.

[0047] Further explanation is needed regarding the process of performing morphological operations (such as opening and closing operations) on the fused image to remove noise and small objects, connecting adjacent regions, marking the third suspected defect region through connected component analysis, and extracting geometric feature data (including area, perimeter, aspect ratio, circularity, and minimum bounding rectangle) and texture feature data (including gray-level co-occurrence matrix features and local binary data). Simultaneously, polarization feature data (including average DoP and average AoP within the region) is extracted based on the polarization degree feature map and polarization angle feature map corresponding to the third suspected defect region. The extracted geometric feature data, texture feature data, and polarization feature data are then matched and analyzed using a distance metric against a pre-set defect feature database to obtain defect feature matching parameters. Finally, a threshold comparison is used to determine whether the third suspected defect region is a true defect. For regions identified as defective, matching is performed based on the extracted features to determine the final defect category. The pre-set defect feature database is pre-constructed by those skilled in the art based on numerous historical cases and can be dynamically adjusted.

[0048] According to an embodiment of the present invention, the step of performing correlation analysis based on the identified defect area through a preset FPC production process database to obtain process adjustment parameters and transmitting them to the FPC control terminal includes: Obtain the defect location data of the identified defect area; Based on the defect location data and the final defect category feature data, correlation analysis is performed through a preset FPC production process database to obtain the process parameters corresponding to the generation process in which the source defect was generated. Based on the defect category feature data and defect subcategory feature data, adaptive adjustments are made using a preset process parameter adjustment method to obtain process adjustment parameters, which are then transmitted to the FPC control terminal.

[0049] Further explanation is needed: by correlating the location and type of defects with a pre-set FPC production process database, the process parameters corresponding to the production process that caused the defects are obtained, and the traceability results are fed back to the front-end production equipment to automatically adjust the corresponding process parameters, thereby achieving positive feedback. For example, if an open circuit defect is determined, by correlating the open circuit location with the pre-set FPC production process database, it is found that the etching time of this batch of FPCs is 8% longer than the standard value, and feedback is given to suggest that the etching time be adjusted to the standard range. The pre-set FPC production process database and the pre-set process parameter adjustment method are pre-built by our technical personnel and can be dynamically adjusted.

[0050] The present invention discloses a multi-source combined imaging and detection method and system for FPC appearance defects. By matching adaptive light source parameters with substrate, circuit density and defect type, polarization imaging suppresses reflection, and ultraviolet light source is combined to expand the detection spectrum. It also realizes positive feedback from defect detection to process parameter optimization, thereby achieving multi-source combined imaging and detection of FPC appearance defects.

[0051] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0052] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0053] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0054] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0055] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

Claims

1. A multi-source combined imaging and detection method for FPC appearance defects, characterized in that, Includes the following steps: The inspection status of the FPC is obtained. If it is in a qualified state, the appearance image of the FPC is obtained and analyzed to obtain FPC identification data. Initial FPC defect data is identified based on the FPC identification data, and adaptive light source parameters are obtained by analyzing and processing the initial FPC defect data. Image acquisition is performed based on the adaptive light source parameters to obtain an image of the first suspected defect area; The image of the first suspected defect area is analyzed and processed to obtain a defect fusion image; The defect fusion image is analyzed and processed to identify the defect region. Based on the identified defect areas, correlation analysis is performed using a preset FPC production process database to obtain process adjustment parameters, which are then transmitted to the FPC control terminal.

2. The multi-source combined imaging and detection method for FPC appearance defects according to claim 1, characterized in that, The process involves obtaining the FPC's inspection status. If it is in a qualified state, an image of the FPC's appearance is acquired and analyzed to obtain FPC identification data, including: Obtain the dust removal completion status, stage fixing status, and pose status of the FPC; If the dust removal completion status is qualified, the platform fixation status is qualified in terms of flatness and pose status, then the FPC's inspection status is determined to be qualified. Conversely, the FPC is determined to be in an unqualified state and the abnormality is handled according to the preset abnormality handling method. If the FPC is in a qualified state, an image of the FPC's appearance will be obtained by acquiring the image through a preset imaging light source combination. Based on the FPC appearance image, the preset substrate type identification algorithm, preset line density identification algorithm, and preset suspected defect area identification algorithm are used to analyze and process the data to obtain FPC identification data, including substrate type feature data, line density level, and first suspected defect area image.

3. The multi-source combined imaging and detection method for FPC appearance defects according to claim 2, characterized in that, The step of identifying initial FPC defect data based on the FPC identification data, and analyzing and processing the initial FPC defect data to obtain adaptive light source parameters includes: The image of the first suspected defect area is subjected to feature extraction to obtain a defect feature vector; Based on the substrate type feature data and the line density level, a corresponding preset defect type identification sub-model is obtained by matching with a preset defect type identification model library. The defect feature vector is input into a preset defect type identification sub-model for analysis and processing to obtain initial FPC defect data, including suspected defect category feature data and suspected defect sub-category feature data. The suspected defect category feature data includes concave-convex morphology defect feature data, latent surface defect feature data, fine line defect feature data, material absorption anomaly defect feature data, or unrecorded defect feature data. If the suspected defect category feature data is concave-convex morphology defect feature data, latent surface defect feature data, fine line defect feature data, or material absorption abnormality defect feature data, then adaptive matching of light source parameters is performed based on the suspected defect category feature data and suspected defect sub-category feature data to obtain adaptive light source parameters. If the suspected defect category feature data is unrecorded defect feature data, then the first suspected defect area image and the corresponding defect feature vector are input into a preset light source parameter prediction model for analysis and processing to obtain adaptive light source parameters.

4. The multi-source combined imaging and detection method for FPC appearance defects according to claim 3, characterized in that, The step of acquiring an image of the second suspected defect region based on the adaptive light source parameters includes: Based on the adaptive light source parameters, images of the suspected defect area are acquired to obtain a second suspected defect area image; The second suspected defect area image includes a polarized coaxial image, an annular oblique projection image, a dot matrix backlight image, and an ultraviolet supplemental light image.

5. The multi-source combined imaging and detection method for FPC appearance defects according to claim 4, characterized in that, The step of analyzing and processing the image of the second suspected defect region to obtain a defect fusion image includes: The polarization coaxial image, ring oblique illumination image, dot matrix backlight image and ultraviolet supplementary illumination image are preprocessed with dark field correction, background subtraction, noise reduction and contrast enhancement to obtain polarization coaxial optimized image, ring oblique illumination optimized image, dot matrix backlight optimized image and ultraviolet supplementary illumination optimized image. Polarization features are extracted from a preset number of polarization coaxial optimized images to obtain polarization degree feature maps and polarization angle feature maps. The polarization angle feature maps include sine polarization angle feature maps and cosine polarization angle feature maps. The local contrast of the polarization degree feature map, sine polarization angle feature map, cosine polarization angle feature map, polarization coaxial optimized image, annular oblique illumination optimized image, dot matrix backlight optimized image and ultraviolet supplement light optimized image is obtained and normalized to obtain the corresponding image weight values. The polarization degree feature map, sine polarization angle feature map, cosine polarization angle feature map, polarization coaxial optimized image, annular oblique illumination optimized image, dot matrix backlight optimized image, and ultraviolet supplementary light optimized image are weighted and summed with their corresponding image weight values ​​to obtain a defect fusion image.

6. The multi-source combined imaging and detection method for FPC appearance defects according to claim 5, characterized in that, The step of analyzing and processing the defect fusion image to obtain the identified defect region includes: Based on the defect fusion image, a third suspected defect region image is obtained by analyzing and processing it using a preset morphological operation method and a preset connected component analysis method. Feature extraction is performed on the image of the third suspected defect area to obtain defect feature data, including geometric feature data and texture feature data; Based on the polarization degree feature map and polarization angle feature map, feature extraction is performed to obtain polarization feature data; Based on the geometric feature data, texture feature data, and polarization feature data, a matching analysis is performed using a preset defect feature database to obtain defect feature matching degree parameters. The defect feature matching degree parameter is compared with the preset defect feature recognition threshold. If the defect feature matching degree parameter is greater than or equal to the preset defect feature recognition threshold, then the third suspected defect area image is determined to be the defect area to be identified, and the defect category feature data corresponding to the maximum value of the defect feature matching degree parameter is determined to be the final defect category feature data. If the defect feature matching degree parameter is less than the preset defect feature recognition threshold, then the third suspected defect area image is determined to be a non-identifiable defect area.

7. The multi-source combined imaging and detection method for FPC appearance defects according to claim 6, characterized in that, The step of performing correlation analysis based on the identified defect areas through a preset FPC production process database to obtain process adjustment parameters and transmitting them to the FPC control terminal includes: Obtain the defect location data of the identified defect area; Based on the defect location data and the final defect category feature data, correlation analysis is performed through a preset FPC production process database to obtain the process parameters corresponding to the generation process in which the source defect was generated. Based on the defect category feature data and defect subcategory feature data, adaptive adjustments are made using a preset process parameter adjustment method to obtain process adjustment parameters, which are then transmitted to the FPC control terminal.

8. A multi-source combined imaging and detection system for FPC appearance defects, characterized in that, The system includes a memory and a processor. The memory contains a program for a multi-source combined imaging and detection method for FPC appearance defects. When the program for the multi-source combined imaging and detection method for FPC appearance defects is executed by the processor, it performs the following steps: The inspection status of the FPC is obtained. If it is in a qualified state, the appearance image of the FPC is obtained and analyzed to obtain FPC identification data. Initial FPC defect data is identified based on the FPC identification data, and adaptive light source parameters are obtained by analyzing and processing the initial FPC defect data. Image acquisition is performed based on the adaptive light source parameters to obtain an image of the first suspected defect area; The image of the first suspected defect area is analyzed and processed to obtain a defect fusion image; The defect fusion image is analyzed and processed to identify the defect region. Based on the identified defect areas, correlation analysis is performed using a preset FPC production process database to obtain process adjustment parameters, which are then transmitted to the FPC control terminal.

9. The multi-source combined imaging and detection system for FPC appearance defects according to claim 8, characterized in that, The process involves obtaining the FPC's inspection status. If it is in a qualified state, an image of the FPC's appearance is acquired and analyzed to obtain FPC identification data, including: Obtain the dust removal completion status, stage fixing status, and pose status of the FPC; If the dust removal completion status is qualified, the platform fixation status is qualified in terms of flatness and pose status, then the FPC's inspection status is determined to be qualified. Conversely, the FPC is determined to be in an unqualified state and the abnormality is handled according to the preset abnormality handling method. If the FPC is in a qualified state, an image of the FPC's appearance will be obtained by acquiring the image through a preset imaging light source combination. Based on the FPC appearance image, the preset substrate type identification algorithm, preset line density identification algorithm, and preset suspected defect area identification algorithm are used to analyze and process the data to obtain FPC identification data, including substrate type feature data, line density level, and first suspected defect area image.

10. The multi-source combined imaging and detection system for FPC appearance defects according to claim 9, characterized in that, The step of identifying initial FPC defect data based on the FPC identification data, and analyzing and processing the initial FPC defect data to obtain adaptive light source parameters includes: The image of the first suspected defect area is subjected to feature extraction to obtain a defect feature vector; Based on the substrate type feature data and the line density level, a corresponding preset defect type identification sub-model is obtained by matching with a preset defect type identification model library. The defect feature vector is input into a preset defect type identification sub-model for analysis and processing to obtain initial FPC defect data, including suspected defect category feature data and suspected defect sub-category feature data. The suspected defect category feature data includes concave-convex morphology defect feature data, latent surface defect feature data, fine line defect feature data, material absorption anomaly defect feature data, or unrecorded defect feature data. If the suspected defect category feature data is concave-convex morphology defect feature data, latent surface defect feature data, fine line defect feature data, or material absorption abnormality defect feature data, then adaptive matching of light source parameters is performed based on the suspected defect category feature data and suspected defect sub-category feature data to obtain adaptive light source parameters. If the suspected defect category feature data is unrecorded defect feature data, then the first suspected defect area image and the corresponding defect feature vector are input into a preset light source parameter prediction model for analysis and processing to obtain adaptive light source parameters.

Images