FPC defect detection method, device and system and storage medium

Abstract

The invention relates to an FPC defect detection method, device and system and a storage medium. The method comprises the following steps: controlling an encoder to synchronously send a high-frequency line trigger signal to an optical camera and a light source controller at the same time, so that the light source controller controls a high-uniformity coaxial light source and a multi-angle high-directivity line light source of a multispectral combined light source to work according to the received high-frequency line trigger signal, enabling the optical camera to collect a time-sharing line scanning image of the to-be-detected FPC; carrying out algorithm correction positioning on the obtained time-sharing line scanning image to obtain a correction scanning image; and fusing a traditional machine vision detection algorithm and an AI deep learning detection algorithm, carrying out comparative analysis on the corrected scanning image and the product Gerber image of the to-be-detected FPC, and obtaining defect information of the to-be-detected FPC from the corrected scanning image. According to the invention, the problems of missing detection and insufficient compatibility caused by limitation of a traditional light source and missing detection or detection failure caused by defects of a traditional algorithm can be solved.

Description

Technical Field

[0001] This invention relates to the field of optical inspection technology, and in particular to a method, apparatus, system and storage medium for detecting defects in FPCs. Background Technology

[0002] In recent years, with the development of electronic technology, FPC (Flexible Printed Circuit) has been playing an increasingly important role. Compared with traditional PCB (Printed Circuit Board), FPC has significant advantages in terms of lightweight and bendability, making it suitable for hinges, mechanical joints, and other areas prone to bending. Therefore, for some high-end customized applications, FPC products offer significant advantages and have high added value and profit margins. Strict quality control is required during the FPC manufacturing process. It is crucial to promptly inspect and control defects during the manufacturing process and before shipment (FQC, Final Quality Control) to prevent defective products from reaching customers and causing greater losses. Therefore, in the FPC manufacturing process, the accuracy and defect interception capabilities of AOI (Automated Optical Inspection) of etched circuits and AVI (Automatic Visual Inspection) of the product's appearance before shipment are particularly important.

[0003] In traditional technologies, AOI and AVI inspection of FPC products typically involves using a line scan camera to capture images of the FPC product and then analyzing those images. However, this traditional approach has several drawbacks: the detection light source is monotonous and has a limited angle, leading to missed detections; the light source wavelength is limited (primarily white light), resulting in insufficient compatibility with products manufactured using special processes; traditional optical solutions cannot meet the dual requirements of both directionality and uniformity in actual inspection; and traditional detection algorithms are also prone to missed detections or failures to detect certain components. Summary of the Invention

[0004] This invention provides an FPC defect detection method, apparatus, system, and storage medium, which can solve the problems of missed detection and insufficient compatibility caused by the limitations of traditional light sources, as well as the problems of missed detection or inability to detect due to the defects of traditional algorithms.

[0005] To address the aforementioned technical problems, this invention provides an FPC defect detection method applied to an optical inspection system. The optical inspection system includes an optical camera, an encoder connected to the optical camera, a light source controller connected to the encoder, and a multispectral combined light source connected to the light source controller, consisting of a highly uniform coaxial light source and a multi-angle highly directional line light source. The optical camera and the multispectral combined light source are correspondingly matched. The method includes: The encoder simultaneously sends high-frequency line trigger signals to the optical camera and the light source controller, so that the light source controller controls the high-uniformity coaxial light source and the multi-angle high-point line light source of the multispectral combined light source to work according to the received high-frequency line trigger signals, so that the optical camera can acquire the time-division line scan image of the FPC to be tested. The obtained time-division line scan image is corrected and located using an algorithm to obtain a corrected scan image; By integrating traditional machine vision detection algorithms and AI deep learning detection algorithms, the corrected scan image is compared and analyzed with the product Gerber image of the FPC to be detected, and the defect information of the FPC to be detected is obtained from the corrected scan image.

[0006] Furthermore, the present invention also proposes an FPC defect detection device for use in an optical inspection system; the optical inspection system includes an optical camera, an encoder connected to the optical camera, a light source controller connected to the encoder, and a multispectral combined light source connected to the light source controller, which consists of a highly uniform coaxial light source and a multi-angle highly directional line light source, and the optical camera and the multispectral combined light source are correspondingly matched. The device includes: The image acquisition module is used to control the encoder to simultaneously send high-frequency line trigger signals to the optical camera and the light source controller, so that the light source controller controls the high uniformity coaxial light source and the multi-angle high pointing line light source of the multispectral combined light source to work according to the received high-frequency line trigger signals, so that the optical camera can acquire the time-division line scan image of the FPC to be tested. The image correction module is used to perform algorithmic correction and positioning on the obtained time-division line scan image to obtain a corrected scan image; The defect acquisition module is used to integrate traditional machine vision detection algorithms and AI deep learning detection algorithms to compare and analyze the corrected scan image with the product Gerber image of the FPC to be inspected, and to obtain the defect information of the FPC to be inspected from the corrected scan image.

[0007] Furthermore, the present invention also proposes an optical detection system, comprising: Optical camera; The encoder is connected to the optical camera; The light source controller is connected to the encoder; The multispectral combined light source includes a highly uniform coaxial light source and a multi-angle highly directional line light source connected to the light source controller, and the optical camera is correspondingly matched with both the highly uniform coaxial light source and the multi-angle highly directional line light source; The control processor is connected to the optical camera, the encoder, the light source controller, and the multispectral combined light source. The control processor is used to implement the FPC defect detection method described above.

[0008] Furthermore, the present invention also proposes a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement all or part of the method steps of the FPC defect detection method as described above.

[0009] The beneficial effects of the technical solution provided by this invention include: When photographing and inspecting FPCs under test using an optical camera and a multispectral combined light source, the multispectral combined light source, composed of a highly uniform coaxial light source and multi-angle highly directional linear light sources, can provide targeted imaging of various defects in the FPC at specific angles. This ensures that defects in the inspection image (time-division linear scan image) are clearly presented, preventing some defects from being missed due to their inability to be clearly presented under only one light source (i.e., a single light source). Moreover, by using time-division linear scanning technology to acquire images of the FPC under test, multiple sets of images with different light source combinations can be obtained in a single scan without increasing the working time, providing multi-dimensional defect imaging and ensuring no defects are missed. This solves the problem of traditional streamlined operations that can only acquire images with one light source combination (insufficient defect imaging coverage) or increase working time due to switching between different light source combinations multiple times. Furthermore, the highly uniform coaxial light of the multispectral combined light source... The source and multi-angle highly directional line light source are designed with a multispectral combination, allowing for flexible selection of different spectral combinations to highlight the features to be detected based on the characteristics of the products being inspected. Moreover, by employing an optical scheme that combines a highly uniform linear scan coaxial light source and a multi-angle highly directional line light source, its high directivity can stably highlight the contrast of the detected image and present defect features. At the same time, its uniformity also compensates for the lack of uniformity in the optical schemes of traditional inspection equipment, which can cause interference from product surface texture on the inspection results and lead to a high false alarm rate. Furthermore, by adopting a fusion scheme of traditional detection algorithms based on Gerber comparison and AI deep learning detection algorithms, it can solve the problem of missed detection of some defects with low contrast due to the limitations of traditional detection algorithms, and also solve the problem of missed detection of functional defects caused by pure AI deep learning detection algorithms without Gerber data comparison. Attached Figure Description

[0010] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a simplified flowchart illustrating the steps of the FPC defect detection method according to an embodiment of the present invention. Figure 2 This is a simplified structural diagram of the optical detection system described in an embodiment of the present invention. Figure 1 ; Figure 3 This is a three-dimensional structural diagram of the highly uniform coaxial light source of the optical detection system described in an embodiment of the present invention. Figure 1 ; Figure 4 This is a three-dimensional structural diagram of the highly uniform coaxial light source of the optical detection system described in an embodiment of the present invention. Figure 2 ; Figure 5 This is a three-dimensional structural diagram of the highly uniform coaxial light source of the optical detection system described in an embodiment of the present invention. Figure 3 ; Figure 6 This is a cross-sectional view of the highly uniform coaxial light source of the optical detection system described in an embodiment of the present invention. Figure 7 This is an exploded view of the highly uniform coaxial light source of the optical detection system described in an embodiment of the present invention. Figure 1 ; Figure 8 This is an exploded view of the highly uniform coaxial light source of the optical detection system described in an embodiment of the present invention. Figure 2 ; Figure 9 This is a schematic diagram of the structure of the first light-emitting plate of the highly uniform coaxial light source according to an embodiment of the present invention; Figure 10 This is a three-dimensional structural diagram of the multi-angle high-pointing line light source of the optical detection system described in this embodiment of the invention. Figure 1 ; Figure 11 This is a three-dimensional structural diagram of the multi-angle high-pointing line light source of the optical detection system described in this embodiment of the invention. Figure 2 ; Figure 12 This is an exploded structural diagram of the multi-angle highly directional line light source of the optical detection system described in an embodiment of the present invention; Figure 13 This is a three-dimensional structural diagram of the high-directivity line light source unit described in an embodiment of the present invention. Figure 1; Figure 14 This is a three-dimensional structural diagram of the high-directivity line light source unit described in an embodiment of the present invention. Figure 2 ; Figure 15 This is an exploded structural diagram of the high-pointing line light source unit described in an embodiment of the present invention; Figure 16 This is a schematic diagram of the structure of the second light-emitting lamp board of the high-point line light source unit according to an embodiment of the present invention; Figure 17 This is a schematic diagram illustrating the working principle of the Fresnel lens in the high-point line light source unit according to an embodiment of the present invention; Figure 18 This is a schematic diagram of the driving signals of the optical detection system described in an embodiment of the present invention; Figure 19 This is a simplified block diagram illustrating the structure of the FPC defect detection device according to an embodiment of the present invention. Figure 20 This is a simplified structural diagram of the optical detection system described in an embodiment of the present invention. Figure 2 .

[0012] In the figure: 10, Optical inspection system; 12, Optical camera; 14, Encoder; 16, Light source controller; 18, Multispectral combined light source; 19, Control processor; 100, High-uniformity coaxial light source; 102, First light outlet; 104, First light return outlet; 106, First light source cavity; 110, First light source housing; 112, First end plate; 1122, Vertical slot; 1124, Angled slot; 114, First connecting side plate; 116. First base plate; 118. First top plate; 1182. Limiting block; 120. First light-emitting board; 122. First light-emitting circuit board; 124. First LED bead; 130. Diffuser plate; 140. Semi-transparent and semi-reflective beam splitter; 150. First water-cooling structure; 152. First heat dissipation groove; 154. Sealed side plate; 156. First pipe joint; 160. Window plate; 162. Window mirror; 200. Multi-angle high-direction Linear light source; 202, Second light outlet; 204, Second light return outlet; 210, Light source mounting bracket; 212, Mounting end plate; 2122, Mounting limiting groove; 214, First mounting side plate; 216, Second mounting side plate; 220, High-directivity linear light source unit; 222, Second light source housing; 2222, Second light source cavity; 2224, Second end plate; 22242, Second limiting groove; 2226, Second light source side plate; 22262, Fixed guide groove; 224, Second light-emitting board; 2242, Second light-emitting circuit board; 2244, Second LED bead; 226, Fresnel lens; 228, Second water-cooling heat dissipation structure; 2282, Second heat dissipation groove; 2284, Second pipe connector; 300, Light source connecting bracket; 1000, FPC defect detection device; 1002, Image acquisition module; 1004, Image correction module; 1006, Defect acquisition module. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0014] In traditional technologies, AOI and AVI inspection of FPC products typically employs a line scan camera to acquire images of the FPC product and then analyzes these images. However, traditional methods suffer from several problems: the detection light source is monotonous and has a single angle, leading to missed detections; the detection light source has a limited wavelength (primarily white light), resulting in insufficient compatibility with products manufactured using special processes; traditional optical solutions cannot meet the dual requirements of directionality and uniformity in actual inspection; and traditional detection algorithms are prone to missed detections or failures to detect defects. Therefore, to address these problems, this invention proposes an FPC defect detection method, apparatus, and system.

[0015] like Figure 1 As shown, this invention provides an FPC defect detection method applied to an optical inspection system 10. Furthermore, as... Figure 2 As shown, the optical detection system 10 may include an optical camera 12, an encoder 14 connected to the optical camera 12, a light source controller 16 connected to the encoder 14, and a multispectral combined light source 18 connected to the light source controller 16, which is composed of a highly uniform coaxial light source 100 and a multi-angle highly directional light source 200. The optical camera 12 and the multispectral combined light source 18 are correspondingly matched.

[0016] The encoder 14 can synchronously send high-frequency line trigger signals to the optical camera 12 and the light source controller 16. When the light source controller 16 receives the high-frequency line trigger signal from the encoder 14, it can send the high-frequency line trigger signal to the high-uniformity coaxial light source 100 and the multi-angle high-point line light source 200 of the combined light source without delay. This allows the high-uniformity coaxial light source 100 and the multi-angle high-point line light source 200 to illuminate the FPC under test. Under the action of the high-frequency line trigger signal, the optical camera 12 will take pictures of the FPC under test to obtain a time-division line scan image.

[0017] Specifically, such as Figure 1 As shown, the FPC defect detection method may include the following steps: S100 and the control encoder 14 simultaneously send high-frequency line trigger signals to the optical camera 12 and the light source controller 16, so that the light source controller 16 controls the high uniformity coaxial light source 100 and the multi-angle high pointing line light source 200 of the multispectral combined light source 18 to work according to the received high-frequency line trigger signals, so that the optical camera 12 can acquire the time-division line scan image of the FPC to be tested. S200. Perform algorithmic correction and positioning on the obtained time-division line scan image to obtain the corrected scan image; The S300 integrates traditional machine vision inspection algorithms and AI deep learning inspection algorithms to compare and analyze the corrected scan image with the product Gerber image of the FPC to be inspected, and obtain the defect information of the FPC to be inspected from the corrected scan image.

[0018] When the FPC to be inspected is photographed and inspected using the optical camera 12 and the multispectral combined light source 18, the multispectral combined light source 18, composed of a highly uniform coaxial light source 100 and a multi-angle highly directional line light source 200, can provide targeted imaging of specific angle light sources for various defects of the FPC. This ensures that the defects in the inspection image (time-division line scan image) are clearly presented, preventing some defects from being missed because they cannot be clearly presented under only one set of light sources (i.e., a single light source). Moreover, by using time-division line scan technology to acquire images of the FPC to be inspected, multiple sets of images with different light source combinations can be acquired in one scan without increasing the working time, providing multi-dimensional defect imaging and ensuring that no defects are missed. This solves the problem that traditional streamlined operations can only acquire images with one set of light source combinations (insufficient defect imaging coverage) or require multiple scans with different light source combinations, which increases the working time.

[0019] Furthermore, the highly uniform coaxial light source 100 and the multi-angle highly directional line light source 200 of the multispectral combined light source 18 are designed with a multispectral combination, which can flexibly select different spectral combinations to highlight the features to be detected according to the characteristics of the products being inspected (such as selecting the blue light band to reduce the interference of the underlying circuit through the substrate in AOI inspection of common FPCs; and selecting the IR band to penetrate and highlight the circuit features under the ink in AVI inspection of FPCs). Moreover, by adopting an optical scheme that combines a highly uniform coaxial light source and a multi-angle highly directional line light source, its high directivity can stably highlight the contrast of the inspection image and present defect features. At the same time, its uniformity also makes up for the problem that the lack of uniformity of the optical scheme of traditional inspection equipment can cause interference with the inspection results due to the surface texture of the product, resulting in a high false alarm rate.

[0020] Moreover, by adopting a fusion scheme of traditional detection algorithms based on Gerber comparison and AI deep learning detection algorithms, we can solve the problem that traditional detection algorithms are prone to missing some defects with low contrast due to their limited capabilities. At the same time, we can also solve the problem that pure AI deep learning detection algorithms will miss functional defects due to the lack of Gerber data for comparison.

[0021] Further, in step S100, the light source controller 16 controls the highly uniform coaxial light source 100 and the multi-angle highly directional line light source 200 of the multispectral combined light source 18 to operate according to the received high-frequency line trigger signal, while the optical camera 12 acquires the time-division line scan image of the FPC to be tested. This can further include the following steps: S110, the light source controller 16 drives the high uniformity coaxial light source 100 and the multi-angle high directional line light source 200 of the multispectral combined light source 18 to work according to the received high-frequency line trigger signal and the preset scanning cycle, so that the multispectral combined light source 18 emits uniform light, or / and low-angle light, or / and medium-angle light, or / and high-angle light to the FPC to be tested.

[0022] The high-frequency line trigger signal serves as the synchronization reference for driving the multispectral combined light source 18 to emit light and for driving the optical camera 12 to acquire images. Driven by the high-frequency line trigger signal, the highly uniform coaxial light source 100 and the multi-angle highly directional line light source 200 of the multispectral combined light source 18 emit light, providing at least one of the following: uniform light, low-angle light, medium-angle light, and high-angle light, to illuminate the FPC to be inspected. In a complete inspection process, the highly uniform coaxial light source 100 and the multi-angle highly directional line light source 200 of the multispectral combined light source 18 are not always simultaneously lit; instead, they selectively activate one or more specific angles of light at different scan cycles (or different line trigger times). The multi-angle highly directional line light source 200 may include low-angle, medium-angle, and high-angle light sources.

[0023] Among them, the highly uniform coaxial light source 100 can be used to detect the surface flatness and highly reflective areas of the FPC (such as copper surface scratches and gold finger oxidation). The uniform light can eliminate the glare caused by the surface reflection of the FPC and highlight surface defects. The low-angle light source of the multi-angle highly directional line light source 200 can emit low-angle light, which can be used to detect the three-dimensional features of the FPC (such as component height, pin warping, and foreign objects). The low-angle light will produce obvious shadows when it hits the protrusions, which can be captured by the camera. In addition, the medium-angle light source can emit medium-angle light, and the high-angle light source can emit high-angle light. The medium-angle light or the high-angle light can be used to detect defects in the color, texture, or matte surface of the FPC.

[0024] S120: Control the optical camera 12 to acquire time-division linear scan images of the FPC under test under the action of uniform light, or / and low-angle light, or / and medium-angle light, or / and high-angle light from the multispectral combined light source 18.

[0025] In this embodiment, the optical camera 12 is a line scan camera, which acquires time-division line scan images of the FPC under test under the action of a high-frequency line trigger signal. Moreover, the line scan camera does not acquire a single image, but a series of line images. Since the multispectral combined light source 18 emits light at different angles at different times, the line images acquired by the optical camera 12 at corresponding times correspond to "coaxial light image (uniform light image)," "low-angle light image," "medium-angle light image," "high-angle light image," etc. After acquiring these line images, they are recombined to finally obtain multiple images of the same area of ​​the FPC under test under different lighting conditions (or the information of different lighting conditions is fused into a single multi-channel image).

[0026] In this embodiment, as Figure 2 As shown, the multispectral combined light source 18 may include a highly uniform coaxial light source 100 and a multi-angle highly directional light source 200, as well as a light source connecting bracket 300 connecting the highly uniform coaxial light source 100 and the multi-angle highly directional light source 200. The highly uniform coaxial light source 100 and the multi-angle highly directional light source 200 are connected together by the light source connecting bracket 300 to form a combined light source that corresponds to and cooperates with the optical camera 12. Furthermore, as... Figures 10 to 13 As shown, the multi-angle high-directivity line light source 200 has a low-angle light source, a medium-angle light source, and a high-angle light source, each of which has a second light-emitting port 202. Moreover, the second light-emitting port 202 of the multi-angle high-directivity line light source 200 is used to correspond to the tilt of the FPC to be tested, and its second light-return port 204 is located at the middle position of the multi-angle high-directivity line light source 200. The low-angle, medium-angle, and high-angle light sources of the multi-angle high-point line light source 200 all illuminate the FPC to be tested with high-point line light through their inclined second light outlet 202. The low-angle, medium-angle, and high-angle light sources share a single return port, namely the second return port 204. The second return port 204 is located in the middle of the multiple light sources of the multi-angle high-point line light source 200, and can be perpendicularly aligned with the FPC to be tested and the lens of the optical camera 12. That is, the high-point line light illuminating the FPC to be tested can be perpendicularly reflected to the second return port 204 and enter the lens of the optical camera 12 perpendicularly, which facilitates image acquisition by the optical camera 12.

[0027] Moreover, such as Figures 2 to 5As shown, the highly uniform coaxial light source 100 has a first light-emitting port 102 and a first light-reflecting port 104, both of which are positioned above the second light-reflecting port 204 of the multi-angle highly directional light source 200. The first light-reflecting port 104 of the highly uniform coaxial light source 100 is located above the first light-emitting port 102 and corresponds to the lens of the optical camera 12. The highly uniform coaxial light source 100 emits coaxial uniform light to the FPC under test through the first light-emitting port 102. After being reflected by the FPC under test, the light enters the lens of the optical camera 12 through the first light-reflecting port 104. Furthermore, the first light-emitting port 102 and the first light-reflecting port 104 are both perpendicularly aligned with the FPC under test, ensuring that the coaxial uniform light is perpendicularly incident on the surface of the FPC under test and perpendicularly reflected into the lens of the optical camera 12. Furthermore, by aligning the first light-emitting port 102 and the first light-reflecting port 104 of the highly uniform coaxial light source 100 directly with the second light-reflecting port 204 of the multi-angle highly directional light source 200, not only can the highly directional light be reflected through the second light-reflecting port 204 and the first light-reflecting port 104 into the lens of the optical camera 12, but also the coaxial uniform light can be perpendicularly irradiated onto the FPC to be tested through the first light-emitting port 102 and the second light-reflecting port 204, and the coaxial uniform light can be reflected through the second light-reflecting port 204 and the first light-reflecting port 104 into the lens of the optical camera 12.

[0028] Furthermore, such as Figures 5 to 8 As shown, the highly uniform coaxial light source 100 may include a first light source housing 110 having a first light source cavity 106, and a first light-emitting plate 120, a diffuser plate 130, and a semi-transparent beam splitter 140 sequentially disposed within the first light source cavity 106. The first light-emitting plate 120 is located on one side of the first light source cavity 106, parallel to and corresponding to the diffuser plate 130, and connected to the light source controller 16. The semi-transparent beam splitter 140 is obliquely disposed on the other side of the first light source cavity 106, obliquely corresponding to the diffuser plate 130. Moreover, the bottom surface of the first light source housing 110 has a first light-emitting port 102, corresponding to the bottom surface of the semi-transparent beam splitter 140; the top surface of the first light source housing 110 has a first light-reflecting port 104, corresponding to the top surface of the semi-transparent beam splitter 140, and corresponding to the lens of the optical camera 12.

[0029] Driven by the light source controller 16, the first light-emitting plate 120 emits light to the diffuser plate. After being homogenized by the diffuser plate, uniform coaxial light (i.e., coaxial uniform light) is obtained. The coaxial uniform light illuminates the semi-transparent and semi-reflective mirror (beam splitter) placed at a 45° angle. Most of the light is reflected and then vertically downward (parallel to the optical axis of the lens of the optical camera 12). It then shines vertically onto the FPC to be tested through the first light outlet 102. The surface of the FPC to be tested generates specular reflection light. The specular reflection light returns along the original path (through the first light outlet 102) and passes through the semi-transparent and semi-reflective mirror, then through the first light return port 104 and enters the lens of the optical camera 12. The optical camera 12 only receives specular reflection light (diffuse reflection light generated by the surface of the tested FPC is scattered in all directions and most of it cannot enter the lens), highlighting the micro-undulations and defects of the object surface and suppressing background reflection interference (effectively shielding the interference of product surface texture on imaging detection).

[0030] Furthermore, such as Figures 7 to 8 As shown, the first light source housing 110 may include two parallel first end plates 112, a first connecting side plate 114 connected to one side of the two first end plates 112, and a first bottom plate 116 and a first top plate 118 respectively covering the bottom and top surfaces of the two first end plates 112. The two first end plates 112, the first connecting side plate 114, the first bottom plate 116, and the first top plate 118 form a first light source cavity 106. A first light-emitting plate 120 is located on the other side of the two first end plates 112, opposite to the first connecting side plate 114. Furthermore, end cap limiting grooves can be provided on both sides of the end of each first end plate 112, allowing the first light-emitting lamp plate 120 to be secured between the two end cap limiting grooves on one side of the two first end plates 112, and the first connecting side plate 114 to be secured between the two end cap limiting grooves on the other side of the two first end plates 112. The first light-emitting lamp plate 120 and the first connecting side plate 114 can be fastened together with connecting screws from the first bottom plate 116, the first top plate 118, and the first end plate 112. In addition, the first bottom plate 116 can also be fixedly connected to the bottom surface of the two first end plates 112 and the bottom surface of the first connecting side plate 114 with connecting screws.

[0031] Furthermore, the two ends of the diffuser plate 130 can be respectively disposed on the two first end plates 112, and its upper and lower sides are respectively disposed on the first bottom plate 116 and the first top plate 118, and are parallel to the first light-emitting plate 120; the two ends of the semi-transparent and semi-reflective beam splitter 140 can be respectively disposed on the two first end plates 112, and at least one of its upper and lower sides is disposed on the first bottom plate 116 and / or the first top plate 118. Further, each first end plate 112 has a vertical slot 1122 extending from the top surface to the bottom surface on its inner side, and the two ends of the diffuser plate 130 can be respectively engaged in the vertical slots 1122 on the inner side of the two first end plates 112. Moreover, each first end plate 112 can have multiple parallel vertical slots 1122 on its inner side, thus allowing the engagement position of the diffuser plate 130 to be changed as needed. In addition, each first end plate 112 has an inclined slot 1124 extending from the bottom to the top side on its inner side, and the two ends of the semi-transparent and semi-reflective beam splitter 140 can be respectively located in the inclined slots 1124 on the inner side of the two first end plates 112.

[0032] Furthermore, in order to facilitate the positioning of the bottom and top of the semi-transparent and semi-reflective beam splitter 140, a limiting block 1182 can be provided on the bottom surface of the first top plate 118 (or the top surface of the first bottom plate 116), so that the top (or bottom) of the semi-transparent and semi-reflective beam splitter 140 is positioned between one side of the limiting block 1182 and the bottom surface of the first top plate 118 (or the top surface of the first bottom plate 116), thus preventing the top (or bottom) of the semi-transparent and semi-reflective beam splitter 140 from moving; and the top (or bottom) of the diffuser plate 130 can be limited to the other side of the limiting block 1182, and its bottom (or top) can be limited to the top surface of the first bottom plate 116 (or the bottom surface of the first top plate 118). Furthermore, a side plate limiting groove can be provided on the inner side of the bottom (or top) of the first connecting side plate 114, so that the bottom (or top) of the semi-transparent and semi-reflective beam splitter 140 extends into the side plate limiting groove, thereby limiting the semi-transparent and semi-reflective beam splitter 140 from the bottom (or top).

[0033] Furthermore, in order to facilitate the perpendicular illumination of the coaxial uniform light reflected by the semi-transparent and semi-reflective beam splitter 140 onto the FPC to be tested, the semi-transparent and semi-reflective beam splitter 140 can be tilted at 45°, and the tilt angle of the oblique slot 1124 is correspondingly set to 45°. Furthermore, in order to ensure that coaxial uniform light is perpendicularly irradiated onto the FPC to be tested, an opening can be provided on one side of the first base plate 116 (or the first base plate 116 can be provided only on the side where the first light-emitting plate 120 and the diffuser plate 130 are provided on the two first end plates 112, leaving the other side where the semi-transparent and semi-reflective beam splitter 140 is provided open) to form a first light-emitting port 102 corresponding to the bottom surface of the semi-transparent and semi-reflective beam splitter 140, and an opening can be provided on one side of the first top plate 118 (similarly, the first top plate 118 can be provided only on the side where the first light-emitting plate 120 and the diffuser plate 130 are provided on the two first end plates 112, leaving the other side where the semi-transparent and semi-reflective beam splitter 140 is provided open) to form a first light-reflecting port 104 corresponding to the top surface of the semi-transparent and semi-reflective beam splitter 140.

[0034] Furthermore, the highly uniform coaxial light source 100 may include a window plate 160 with an exit window disposed at the first light return port 104, and a window mirror 162 disposed at the exit window on the window plate 160. The window mirror 162 corresponds to the top surface of the semi-transparent and semi-reflective beam splitter 140, so that the reflected coaxial uniform light passes through the semi-transparent and semi-reflective beam splitter 140 and enters the lens of the optical camera 12 through the window mirror 162 for imaging by the optical camera 12. Moreover, the window plate 160 may be disposed on the top surface of the two first end plates 112 and may be connected to the first top plate 118. An exit window may be provided on the window plate 160 to form the first light return port 104, and then the window mirror 162 may be provided at the first light return port 104. Alternatively, the first top plate 118 may be used as the window plate 160, and an opening may be provided on the side of the first top plate 118 to form an exit window to form the first light return port 104.

[0035] In addition, such as Figures 5 to 8 As shown, the highly uniform coaxial light source 100 may further include a first water-cooled heat dissipation structure 150 disposed at the side opening of the first light source housing 110, and a first light-emitting plate 120 disposed on the side of the first water-cooled heat dissipation structure 150 and located at the side opening of the first light source housing 110. To improve detection efficiency, the light source power and brightness of the first light-emitting plate 120 need to be increased; therefore, it is necessary to enhance the heat dissipation effect of the first light-emitting plate 120. In this embodiment, by placing the first light-emitting plate 120 on the first water-cooled heat dissipation structure 150 and using water cooling to dissipate heat from the first light-emitting plate 120, the heat dissipation capacity of the first light-emitting plate 120 can be enhanced, enabling it to operate stably.

[0036] Furthermore, the first water-cooled heat dissipation structure 150 may include a first heat dissipation groove 152 with an opening on one side, a sealing side plate 154 sealed at the opening side of the first heat dissipation groove 152, a first pipe joint 156 at each end of the first heat dissipation groove 152, and a first circulating water supply structure connecting the two first pipe joints 156. Moreover, the first light-emitting panel 120 may be disposed on one side (the non-opening side) of the first heat dissipation groove 152, so that the first light-emitting panel 120 transfers heat to the first heat dissipation groove 152 during operation. The sealing side plate 154 and the first heat dissipation groove 152 can be closed to form a cooling chamber. Cooling water can be transported into the cooling chamber through one first pipe joint 156 via the first circulating water supply structure, and cooling water in the cooling chamber can be transported back to the first circulating water supply structure through the other first pipe joint 156. During the circulation of cooling water in the cooling chamber of the first heat dissipation groove 152, the heat transferred from the first light-emitting panel 120 to the first heat dissipation groove 152 can be carried away, thus achieving heat dissipation of the first light-emitting panel 120. Furthermore, to enhance heat dissipation, the side of the first heat dissipation groove 152 that contacts the first light-emitting plate 120 can be made of a metal plate, which can enhance heat transfer and dissipation. Further, metal heat dissipation fins connected to the metal plate can be provided in the cooling cavity to further enhance heat dissipation.

[0037] Furthermore, the first circulating water supply structure may include a cooling water tank, a circulating water pump connected to the cooling water tank, and a water supply valve connected to the circulating water pump. One first pipe joint 156 is connected to the water supply valve, and another first pipe joint 156 is connected to the cooling water tank. The components can be connected to each other via water pipes. The circulating water pump can draw cooling water from the cooling water tank, pass it through the water supply valve and one first pipe joint 156, and deliver it to the first heat dissipation tank 152. Then, it can be delivered back to the cooling water tank through the other first pipe joint 156, thus achieving the circulation of cooling water.

[0038] Moreover, such as Figure 9As shown, the first light-emitting board 120 may include a first light-emitting circuit board 122 connected to the light source controller 16, and a plurality of first LED beads arranged in an array on the first light-emitting circuit board 122. To obtain a highly uniform coaxial illumination effect, the plurality of first LED beads on the first light-emitting circuit board 122 of the first light-emitting board 120 can be arranged in a multi-row, multi-column array. The light emitted by the plurality of first LED beads can be uniformly diffused by the diffuser plate 130 to obtain highly uniform coaxial light (i.e., coaxial uniform light), which can effectively shield the interference of product surface texture on imaging detection. Specifically, in this embodiment, the first LED beads can be LED patch beads, and the plurality of LED patch beads can be evenly spaced and densely arranged on the first light-emitting circuit board 122, with a spacing of 2mm between two adjacent LED patch beads. Highly uniform coaxial light is obtained through the diffuser plate 130. Furthermore, in order to ensure high-brightness light source output, the power supply connector of the first light-emitting board 120 can be an aviation connector, which can connect the first light-emitting board 120 and the power supply through the aviation connector, thus ensuring that the first light-emitting board 120 can achieve high-power stable output.

[0039] In addition, such as Figures 10 to 12 As shown, the multi-angle high-point line light source 200 may include a light source mounting bracket 210 and a high-point line light source group symmetrically arranged on both sides of the light source mounting bracket 210. Each high-point line light source group may include at least three high-point line light source units 220 arranged in an arc on the light source mounting bracket 210. Each high-point line light source unit 220 has a second light-emitting port 202 for tilting corresponding to the FPC to be tested. There is a light source gap between two high-point line light source groups, and the light source gap forms a second light-returning port 204 corresponding to the first light-emitting port 102 of the highly uniform coaxial light source 100. Moreover, the sides of the multiple high-point line light source units 220 in each high-point line light source group abut against each other, so that there is no gap between two adjacent high-point line light source units 220, and no light passes through, so that light can only pass through the second light-returning port 204 between the two high-point line light source groups.

[0040] By symmetrically arranging multiple high-point-line light source units 220 in an arc shape on both sides of the light source mounting bracket 210, and arranging the multiple high-point-line light source units 220 on each side in sequence at low, medium, and high angles, a low-angle light source, a medium-angle light source, and a high-angle light source can be formed on each side of the light source mounting bracket 210, and the low-angle light source, medium-angle light source, and high-angle light source on both sides correspond one-to-one. Under the drive of the light source driver, each high-point-line light source unit 220 (low-angle light source, or medium-angle light source, or high-angle light source) can emit oblique high-point-line light towards the surface of the FPC to be tested through its second light outlet 202. The high-point-line light can pass through the second light return port 204 under the vertical reflection of the FPC to be tested, and enter the lens of the optical camera 12 through the first light outlet 102 of the highly uniform coaxial light source 100 corresponding to the second light return port 204, the semi-reflective beam splitter, and the first light return port 104.

[0041] In this embodiment, each high-directivity light source group may include three high-directivity light source units 220 arranged in an arc on the light source mounting bracket 210. The three high-directivity light source units 220 are arranged sequentially at low, medium, and high angles, forming a low-angle light source, a medium-angle light source, and a high-angle light source, respectively. The low-angle light sources on both sides are symmetrically arranged about the center of the light source mounting bracket 210, the medium-angle light sources on both sides are also symmetrically arranged about the center of the light source mounting bracket 210, and the high-angle light sources on both sides are also symmetrically arranged about the center of the light source mounting bracket 210. Further, the setting angles of the low-angle light source, medium-angle light source, and high-angle light source on each side can be 30±5°, 50±5°, and 70±5°, respectively. These setting angles are also the light source angles of each light source, that is, the angle between the emission direction of the high-directivity light emitted by each light source and the horizontal plane, which can accommodate the requirements of different defect characteristics for the light source angle.

[0042] Moreover, such as Figure 12As shown, the light source mounting bracket 210 may include two parallel and aligned arc-shaped mounting end plates 212, a first mounting side plate 214 connected to one side of the two mounting end plates 212, and a second mounting side plate 216 connected to the other side of the two mounting end plates 212. The two mounting end plates 212 and the two mounting side plates (i.e., the first mounting side plate 214 and the second mounting side plate 216) can form a frame-like light source mounting bracket 210 for mounting and arranging multiple high-point-of-sight light source units 220. Furthermore, the tops of the two mounting end plates 212 are provided with multiple mounting limiting grooves 2122 spaced apart, and both ends of each high-point-of-sight light source unit 220 can be respectively engaged in the mutually spaced mounting limiting grooves 2122 on the two mounting end plates 212. In this embodiment, three mounting limiting grooves 2122 can be provided at intervals on both sides of the top of each mounting end plate 212. The setting angles of the three mounting limiting grooves 2122 can be 30±5°, 50±5°, and 70±5° respectively, so that the high-pointing line light source unit 220 installed in the three mounting limiting grooves respectively forms a low-angle light source, a medium-angle light source, and a high-angle light source.

[0043] Moreover, such as Figures 13 to 15 As shown, each high-directivity line light source unit 220 may include a second light source housing 222 with a second light source cavity 2222 disposed on the light source mounting bracket 210, a second light-emitting plate 224 disposed on the top side of the second light source cavity 2222 and connected to the light source controller 16, and a Fresnel lens 226 disposed at the bottom opening of the second light source housing 222. The second light-emitting plate 224 and the Fresnel lens 226 are disposed opposite to each other. The outer side of the Fresnel lens 226 (referring to the side facing away from the second light source cavity 2222) forms a second light-emitting port 202 for tilting and corresponding to the FPC under test. Under the drive of the light source controller 16, the second light-emitting plate 224 emits ordinary line light that illuminates the Fresnel lens 226. Under the action of the Fresnel lens 226, the ordinary line light is focused to form a high-directivity line light, which is then illuminated onto the FPC under test through the second light-emitting port 202.

[0044] Furthermore, the second light source housing 222 may include two parallel second end plates 2224, with a second light source side plate 2226 respectively covering the front and rear sides of the two second end plates 2224. The two second end plates 2224 and the two second light source side plates 2226 surround to form a second light source housing 222 with a second light source cavity 2222. The second light-emitting plate 224 is disposed at the top opening of the second light source housing 222, and the Fresnel lens 226 is disposed opposite to it at the bottom opening of the second light source housing 222. Moreover, in order to make the second light source side plate 2226 tightly connected to the second end plate 2224, a second limiting groove 22242 may be provided on the inner side of the front and rear sides of the second end plate 2224. The two ends of each second light source side plate 2226 may be respectively engaged in the second limiting groove 22242 on the same side of the two second end plates 2224, and may be fastened with connecting screws.

[0045] Furthermore, each of the inner surfaces at the bottom of the two second light source side plates 2226 is provided with a fixing guide groove 22262, which allows the two sides of the Fresnel lens 226 to be movably engaged in the fixing guide grooves 22262 of the two second light source side plates 2226, thus limiting the movement of the Fresnel lens 226. Further, as... Figure 17 As shown, the Fresnel lens 226 can be made of acrylic material. Its surface microstructure allows it to converge light, focusing the width of the radiated light spot emitted by the second light-emitting plate 224 onto the working surface to within 5mm. This results in strong directional light, clearly revealing minute defects on the FPC to be inspected. Unlike traditional cylindrical condenser rods, the Fresnel lens 226 is smaller and has a smaller focused spot, allowing for miniaturization of the combined light source and easier integration into confined spaces. Furthermore, by using a fixed guide groove 22262 on the inner side of the second light source side plate 2226 to position the Fresnel lens 226, it can be easily replaced with a diffuser, polarizer, or other optical accessories with specific functions as needed.

[0046] In addition, such as Figures 13 to 15 As shown, the high-pointing linear light source unit 220 may include a second water-cooled heat dissipation structure 228 disposed at the top opening of the second light source housing 222. The second light-emitting plate 224 may be disposed on the side of the second water-cooled heat dissipation structure 228 and located within the second light source cavity 2222 of the second light source housing 222, opposite to the Fresnel lens 226. To improve detection efficiency, the light source power and brightness of the high-pointing linear light source unit 220 need to be increased. Therefore, water cooling is used to dissipate heat from the second light-emitting plate 224 to ensure stable operation.

[0047] Furthermore, the specific structure of the second water-cooled heat dissipation structure 228 can be similar to that of the first water-cooled heat dissipation structure 150. In addition, to reduce the volume of the high-directivity light source unit 220, the volume of the second water-cooled heat dissipation structure 228 can also be reduced. In this case, the second water-cooled heat dissipation structure 228 may include an integrated second heat dissipation groove 2282, and the second light-emitting plate 224 may be disposed on the side of the second heat dissipation groove 2282. This second heat dissipation groove 2282 forms an integral heat dissipation groove with a water-cooled heat dissipation cavity by opening a through hole in the middle of a long strip, resulting in a simple structure and small size. Besides dissipating heat from the second light-emitting plate 224 during operation, the second heat dissipation groove 2282 also supports and fixes the second light-emitting plate 224. The second light-emitting plate 224 is fixed to the side of the second heat dissipation groove 2282, and to ensure that the second light-emitting plate 224 outputs high-brightness light, its power supply connector can also be an aviation connector, ensuring stable high-power output.

[0048] Furthermore, the second water-cooling heat dissipation structure 228 may also include a second pipe connector 2284 located at each of the two end openings of the second heat dissipation groove 2282, and a second circulating water supply structure connected between the two second pipe connectors 2284. The specific structure of the second circulating water supply structure is similar to that of the first circulating water supply structure, and will not be described again here. Moreover, each high-pointing linear light source unit 220 can use a separate second circulating water supply structure, or multiple high-pointing linear light source units 220 of the multi-angle high-pointing linear light source 200 can share a single second circulating water supply structure. In addition, to facilitate the connection between the second water-cooling heat dissipation structure 228 and the second light source housing 222, the second light source side plates 2226 on both sides of the second light source housing 222 can protrude beyond the second end plate 2224 and respectively overlap the two sides of the second heat dissipation groove 2282 of the second water-cooling heat dissipation structure 228.

[0049] Moreover, such as Figures 15 to 16 As shown, the second light-emitting board 224 may include a second light-emitting circuit board 2242 disposed on the side of the second heat dissipation groove 2282 of the second water-cooled heat dissipation structure 228 and connected to the light source controller 16, and a plurality of second LED beads with different wavelengths arranged in a row on the second light-emitting circuit board 2242. That is, in this embodiment, the plurality of second LED beads on the second light-emitting circuit board 2242 are arranged in a row, but unlike the traditional LED bead arrangement, the plurality of second LED beads in this embodiment are second LED beads with different emission wavelengths arranged at intervals (the emission band of the beads can be red light, blue light, green light, white light, ultraviolet light, infrared light, etc.). One, two or more emission bands of second LED beads can be selected as needed, and each wavelength of second LED beads can be controlled independently to obtain multispectral images of different bands under a set of light sources for specific detection needs.

[0050] In this embodiment, as Figure 16 As shown, multiple white LED beads and multiple infrared LED beads can be arranged at intervals on the second light-emitting circuit board 2242 of the second light-emitting board 224, and the interval between adjacent white LED beads and infrared LED beads can be 2mm. Furthermore, the arrangement of the second LED beads on the second light-emitting circuit board 2242 of the two pairs of high-pointing line light source units 220 (which can be called medium-angle light sources) and the two pairs of high-pointing line light source units 220 (which can be called high-angle light sources) can be selectively made the same, and the white LED beads of these four light sources are simultaneously lit as a light source combination for one channel image, and the infrared LED beads of these four light sources are simultaneously lit as a light source combination for another channel image. As mentioned above, when the optical detection system 10 is working, it outputs images corresponding to the white light channel and the infrared light channel. The white light channel image cannot show the short circuit under the ink, while the infrared light channel image can clearly show the short circuit under the ink.

[0051] In this embodiment, a three-quarter time-line scan is used as an example. See [link / reference]. Figure 18 The three-dimensional linear scan timing diagram shown has a horizontal axis representing a preset scan period T, which can be further divided into three channel periods: 0~T / 3, T / 3~2T / 3, and 2T / 3~T. Each channel period corresponds to a specific combination of light sources. Specifically, 0~T / 3 corresponds to the combination of all low-angle light sources (i.e., low-angle high-pointing line light source units 220), all medium-angle light sources (i.e., medium-angle high-pointing line light source units 220), and all high-angle light sources (i.e., high-angle high-pointing line light source units 220) in the high-uniformity coaxial light source 100 and multi-angle high-pointing line light source 200, which corresponds to a uniform light combination. T / 3~2T / 3 corresponds to the combination of all low-angle light sources in the multi-angle high-pointing line light source 200, which corresponds to a low-angle light combination. 2T / 3~T corresponds to the combination of all medium-angle light sources and all high-angle light sources in the multi-angle high-pointing line light source 200, which corresponds to a medium-high angle light combination. Refer to the three-dimensional linear scan timing diagram. The vertical axis in the diagram represents the high-frequency line trigger signal (trigger pulse signal). The high-level duration of each square wave pulse signal in the diagram corresponds to the working or turning-on time of the encoder 14, optical camera 12, and each light source in each channel cycle.

[0052] In actual operation, within each preset scan cycle T, encoder 14 simultaneously outputs high-frequency line trigger signals to optical camera 12 and multispectral combined light source 18, triggering the corresponding light source combination in each of the three channel cycles, and acquiring one line of images from each, which are then arranged sequentially, resulting in three lines of images: 1, 2, and 3. This motion triggering process is repeated sequentially within each preset scan cycle T. Figure 4 , 5Stack the images 1, 6; 7, 8, 9; ... until the scanning acquisition of the FPC to be detected is completed. Finally, according to the pattern of image acquisition, extract the images 1, 4, 7...; 2, 5, 8...; 3, 6, 9... respectively. Combine each group in sequence to obtain the image of each light source combination channel, and finally obtain the three-part time-division scanning image (i.e., time-division line scanning image).

[0053] Furthermore, in step S200, the obtained time-division line scan image is corrected and positioned using an algorithm to obtain a corrected scan image, which may specifically include the following steps: S210. Based on the obtained time-division line scan image of the FPC to be detected, locate the reference point of the FPC to be detected.

[0054] A positioning algorithm can be used to detect the circular or square solid reference marks (e.g., marks with high contrast and obvious geometric features distributed at the four corners or edges of the FPC) that were preset during the design on the FPC to be detected in the time-division line scan image, and use the reference marks as reference points.

[0055] For example, a grayscale thresholding algorithm can be used first to filter out candidate regions for potential reference points from the time-division line scan image; then, a normalized cross-correlation (NCC) template matching algorithm can be used in conjunction with the product Gerber image (i.e., the design drawing of the FPC) to obtain the approximate location of the reference point; then, a centroid method or Gaussian fitting algorithm can be used to improve the positioning accuracy of the reference point from the pixel level to the sub-pixel level; then, the actual coordinates of multiple reference points can be obtained from the product Gerber image, and a mapping relationship between the image pixel coordinates of the time-division line scan image and the spatial coordinates of the FPC can be established to realize the positioning of the reference point of the FPC to be inspected.

[0056] S220. Based on the location of the reference point of the FPC to be tested, perform geometric distortion correction on the time-division line scan image.

[0057] Based on the positioning results of the reference point of the FPC to be tested, multi-layer geometric corrections (such as lens distortion correction, motion offset and rotation correction, flexible stretching or wrinkle correction, and multi-source image registration correction) can be performed on the time-division line scan image of the FPC to be tested to solve the problems of lens distortion, motion offset and flexible deformation, and eliminate system and deformation errors.

[0058] S230, Correct the brightness unevenness of the time-division line scan image.

[0059] This technology can normalize grayscale values ​​to address brightness fluctuations and edge light loss in time-division linear scan images, improving image consistency and providing a high-contrast image foundation for subsequent defect detection. For example, it can perform global brightness correction, local brightness compensation, and product Gerber image-assisted brightness calibration.

[0060] S240: Output the corrected scan image and perform coordinate mapping.

[0061] This allows for the integration of all time-division linear scan images (including coaxial light source images and multi-angle line light source images) that have undergone geometric and brightness correction, outputting a set of corrected scan images in a unified coordinate system. It establishes a one-to-one correspondence between each pixel in the corrected scan image and the FPC spatial coordinates and Gerber image design coordinates, providing a precise coordinate reference for subsequent comparative analysis between the corrected scan images and the product's Gerber image.

[0062] The main purpose of performing algorithmic correction and positioning on time-division line scan images is to eliminate image deviations caused by various systematic errors and the characteristics of the FPC itself, establish a precise coordinate mapping relationship between the detected image and the FPC design standard (product Gerber image), and provide a highly consistent image benchmark for subsequent defect comparison and analysis.

[0063] Furthermore, in step S300, traditional machine vision detection algorithms and AI deep learning detection algorithms are integrated to compare and analyze the corrected scan image with the product Gerber image of the FPC to be inspected, and to obtain the defect information of the FPC to be inspected from the corrected scan image. This may specifically include the following steps: S310. Perform preprocessing and image registration on the corrected scan image and the product Gerber image of the FPC to be inspected.

[0064] By preprocessing and registering the rectified scan image and the product Gerber image, image noise interference can be eliminated, and a precise coordinate correspondence between the rectified scan image and the product Gerber image can be established. This provides the image foundation for subsequent comparison and avoids false detections caused by registration deviations.

[0065] S320: Using traditional machine vision detection algorithms, the registered corrected scan image and the product Gerber image are compared to determine the basic defects in the corrected scan image and locate all suspected defect areas that deviate from the design standards.

[0066] Traditional machine vision inspection algorithms can quickly identify obvious defects in the FPC to be inspected (such as positional misalignment defects, missing component defects, and color discrepancies), and delineate the suspected defect areas to be verified, thus narrowing the processing scope of AI deep learning inspection algorithms.

[0067] For example, traditional machine vision detection algorithms such as template matching, threshold segmentation, edge detection, and morphological operations can be used to compare the differences in pixel grayscale, line contour, and geometric dimensions between the registered and corrected scanned image and the product Gerber image, and identify basic defects with regularity and high contrast such as excessive line width, open circuit / short circuit, and missing pads. Areas with abnormal grayscale or slight edge deviation that cannot be clearly determined can be marked as suspected defect areas.

[0068] S330. Use an AI deep learning detection algorithm to detect the suspected defect area and determine the actual defect in the suspected defect area.

[0069] AI deep learning detection algorithms can accurately classify suspected defect areas, distinguishing between real defects and noise or normal process deviations. This solves the problem of traditional algorithms easily missing or falsely detecting complex and minute defects, thus improving detection accuracy. For example, a trained AI model (such as a CNN convolutional neural network model, with a training set containing samples of various minute defects in FPC, allowing the model to learn the texture and morphological features of various minute defects) can be used as input to an image patch of a suspected defect area. The AI ​​model can then determine whether the suspected defect area is a real defect (such as a minor scratch, a hidden open circuit, or coating oxidation) or a false defect (such as dust or image noise).

[0070] S340: Combining the detection results of traditional machine vision detection algorithms and AI deep learning detection algorithms, output the judgment result of the defect information of the FPC to be detected.

[0071] Authentic defects identified by traditional algorithms can be directly incorporated into the detection results. Genuine defects confirmed by AI algorithms can be supplemented into the detection results, while false defects excluded by AI algorithms are removed from the suspected area. The output defect information for the FPC to be inspected can include the defect type, location coordinates, size, and confidence level, and is associated with the design standards of the product's Gerber image to determine whether the defect exceeds the process tolerance.

[0072] In addition, such as Figure 19 As shown, the present invention also proposes an FPC defect detection device 1000, applied to an optical inspection system 10. For example... Figure 2 As shown, the optical detection system 10 may include an optical camera 12, an encoder 14 connected to the optical camera 12, a light source controller 16 connected to the encoder 14, and a multispectral combined light source 18 connected to the light source controller 16, which is composed of a highly uniform coaxial light source 100 and a multi-angle highly directional light source 200. The optical camera 12 and the multispectral combined light source 18 are correspondingly matched.

[0073] The encoder 14 can synchronously send high-frequency line trigger signals to the optical camera 12 and the light source controller 16. When the light source controller 16 receives the high-frequency line trigger signal from the encoder 14, it can send the high-frequency line trigger signal to the high-uniformity coaxial light source 100 and the multi-angle high-point line light source 200 of the combined light source without delay. This allows the high-uniformity coaxial light source 100 and the multi-angle high-point line light source 200 to illuminate the FPC under test. Under the action of the high-frequency line trigger signal, the optical camera 12 will take pictures of the FPC under test to obtain a time-division line scan image.

[0074] Specifically, such as Figure 19 As shown, the FPC defect detection device 1000 may include: The image acquisition module 1002 is used to control the encoder 14 to simultaneously send high-frequency line trigger signals to the optical camera 12 and the light source controller 16, so that the light source controller 16 controls the high uniformity coaxial light source 100 and the multi-angle high pointing line light source 200 of the multispectral combined light source 18 to work according to the received high-frequency line trigger signals, so that the optical camera 12 can acquire the time-division line scan image of the FPC to be tested. Image correction module 1004 is used to perform algorithmic correction and positioning on the obtained time-division line scan image to obtain a corrected scan image; The defect acquisition module 1006 is used to integrate traditional machine vision detection algorithms and AI deep learning detection algorithms to compare and analyze the corrected scan image with the product Gerber image of the FPC to be inspected, and to obtain the defect information of the FPC to be inspected from the corrected scan image.

[0075] The FPC defect detection device 1000 described in this embodiment corresponds to the FPC defect detection method described above. The functions of each module in the FPC defect detection device 1000 in this embodiment are described in detail in the corresponding method embodiments, and will not be repeated here.

[0076] In addition, such as Figure 20 As shown, the present invention also proposes an optical detection system 10, including an optical camera 12, an encoder 14 connected to the optical camera 12, a light source controller 16 connected to the encoder 14, a multispectral combined light source 18 connected to the light source controller 16, and a control processor 19 connected to the optical camera 12, the encoder 14, the light source controller 16, and the multispectral combined light source 18. Furthermore, the multispectral combined light source 18 may include a highly uniform coaxial light source 100 and a multi-angle highly directional light source 200 connected to the light source controller 16, with the optical camera 12 corresponding to and cooperating with both the highly uniform coaxial light source 100 and the multi-angle highly directional light source 200.

[0077] Under the control of the control processor 19, the encoder 14 synchronously sends a high-frequency line trigger signal to the optical camera 12 and the light source controller 16. Upon receiving the high-frequency line trigger signal from the encoder 14, the light source controller 16 sends the signal without delay to the combined high-uniformity coaxial light source 100 and the multi-angle high-pointing line light source 200, illuminating the FPC under inspection. The optical camera 12, under the influence of the high-frequency line trigger signal, takes a picture of the FPC under inspection to obtain a time-division line scan image. The control processor 19 can also analyze and process the acquired time-division line scan image to obtain defect information of the FPC under inspection.

[0078] Specifically, the control processor 19 is used to implement the FPC defect detection method described above. The specific implementation method can be found in the details of the FPC defect detection method described above, and will not be repeated here.

[0079] Furthermore, the present invention also proposes a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement all or part of the method steps of the FPC defect detection method as described above.

[0080] Furthermore, the present invention also proposes a computer program product for running on a computer to implement the FPC defect detection method described above.

[0081] The present invention can implement all or part of the processes in the above methods, or it can be accomplished by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content contained in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0082] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the computer device, connecting all parts of the computer device through various interfaces and lines.

[0083] Memory can be used to store computer programs and / or models. The processor performs various functions of the computer device by running or executing the computer programs and / or models stored in the memory, and by accessing data stored in the memory. Memory can primarily include a program storage area and a data storage area. The program storage area can store the operating system and at least one application program required for a function (e.g., sound playback, image playback, etc.); the data storage area can store data created based on the use of the mobile phone (e.g., audio data, video data, etc.). Furthermore, memory can include high-speed random access memory, and can also include non-volatile memory, such as hard disks, RAM, plug-in hard disks, SmartMedia Cards (SMC), Secure Digital (SD) cards, Flash Cards, at least one disk storage device, flash memory device, or other volatile solid-state storage devices.

[0084] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, servers, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0085] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), servers, and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0086] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

Claims

1. An FPC defect detection method, applied to an optical inspection system; the optical inspection system includes an optical camera, an encoder connected to the optical camera, a light source controller connected to the encoder, and a multispectral combined light source connected to the light source controller, consisting of a highly uniform coaxial light source and a multi-angle highly directional line light source, wherein the optical camera and the multispectral combined light source are correspondingly matched; Its features are, The method includes: The encoder simultaneously sends high-frequency line trigger signals to the optical camera and the light source controller, so that the light source controller controls the high-uniformity coaxial light source and the multi-angle high-point line light source of the multispectral combined light source to work according to the received high-frequency line trigger signals, so that the optical camera can acquire the time-division line scan image of the FPC to be tested. The obtained time-division line scan image is corrected and located using an algorithm to obtain a corrected scan image; By integrating traditional machine vision detection algorithms and AI deep learning detection algorithms, the corrected scan image is compared and analyzed with the product Gerber image of the FPC to be detected, and the defect information of the FPC to be detected is obtained from the corrected scan image.

2. The FPC defect detection method according to claim 1, characterized in that, The process involves the light source controller controlling the highly uniform coaxial light source and the multi-angle highly directional line light source of the multispectral combined light source to operate according to the received high-frequency line trigger signal, while simultaneously enabling the optical camera to acquire time-division line scan images of the FPC under test, including: The light source controller controls the high-uniformity coaxial light source and the multi-angle high-point line light source of the multispectral combined light source to work simultaneously according to the received high-frequency line trigger signal and a preset scanning cycle, so that the multispectral combined light source emits uniform light, or / and low-angle light, or / and medium-angle light, or / and high-angle light to the FPC to be tested respectively. The optical camera is controlled to acquire time-division linear scan images of the FPC under test under the action of uniform light, and / or low-angle light, and / or medium-angle light, and / or high-angle light from the multispectral combined light source.

3. The FPC defect detection method according to claim 1, characterized in that, The multispectral combined light source includes the highly uniform coaxial light source and the multi-angle highly directional line light source, as well as a light source connecting frame that connects the highly uniform coaxial light source and the multi-angle highly directional line light source; The second light-emitting port of the multi-angle high-point line light source is used to correspond to the tilt of the FPC to be detected, and its second light-returning port is located at the middle position of the multi-angle high-point line light source. The first light-emitting port of the highly uniform coaxial light source is positioned above the second light-returning port of the multi-angle highly directional line light source, and the first light-returning port of the highly uniform coaxial light source corresponds to the lens of the optical camera.

4. The FPC defect detection method according to claim 3, characterized in that, The highly uniform coaxial light source includes a first light source housing having a first light source cavity, and a first light-emitting plate, a diffuser plate, and a semi-transparent and semi-reflective beam splitter sequentially disposed in the first light source cavity. The first light-emitting plate is located on one side of the first light source cavity, parallel to and corresponding to the diffuser plate, and connected to the light source controller. The semi-transparent and semi-reflective beam splitter is tilted on the other side of the first light source cavity, tilted to correspond to the diffuser plate. The bottom surface of the first light source housing has a first light outlet, which corresponds to the bottom surface of the semi-transparent and semi-reflective beam splitter; the top surface of the first light source housing has a first light return outlet, which corresponds to the top surface of the semi-transparent and semi-reflective beam splitter and is also corresponding to the lens of the optical camera.

5. The FPC defect detection method according to claim 4, characterized in that, The highly uniform coaxial light source includes a first water-cooled heat dissipation structure disposed at the side opening of the first light source housing, and a first light-emitting plate disposed on the side of the first water-cooled heat dissipation structure.

6. The FPC defect detection method according to claim 4, characterized in that, The highly uniform coaxial light source includes a window plate with an exit window disposed at the first light return port, and a window mirror disposed at the exit window on the window plate, the window mirror corresponding to the top surface of the semi-transparent and semi-reflective beam splitter. The first light-emitting board includes a first light-emitting circuit board connected to the light source controller, and a plurality of first LED beads arranged in an array on the first light-emitting circuit board.

7. The FPC defect detection method according to claim 3, characterized in that, The multi-angle high-point line light source includes a light source mounting frame and a high-point line light source group symmetrically arranged on both sides of the light source mounting frame. Each high-point line light source group includes at least three high-point line light source units arranged in an arc on the light source mounting frame. Each of the high-point-direction light source units has a second light-emitting port for tilting in relation to the FPC to be tested; there is a light source gap between two of the high-point-direction light source groups, the light source gap forming a second light-returning port corresponding to the first light-emitting port of the high-uniformity coaxial light source.

8. The FPC defect detection method according to claim 7, characterized in that, Each of the high-pointing linear light source units includes a second light source housing with a second light source cavity disposed on the light source mounting bracket, a second light-emitting plate disposed on the top side of the second light source cavity and connected to the light source controller, and a Fresnel lens disposed at the bottom opening of the second light source housing. The second light-emitting plate is disposed opposite to the Fresnel lens, and the outer side of the Fresnel lens forms a second light-emitting port for tilting corresponding to the FPC to be tested.

9. The FPC defect detection method according to claim 8, characterized in that, The high-pointing linear light source unit includes a second water-cooling heat dissipation structure disposed at the top side opening of the second light source housing, and the second light-emitting board is disposed on the second water-cooling heat dissipation structure; The second light-emitting board includes a second light-emitting circuit board disposed on the side of the second water-cooled heat dissipation structure and connected to the light source controller, and a plurality of second LED beads with different wavelengths arranged in a row on the second light-emitting circuit board.

10. The FPC defect detection method according to claim 1, characterized in that, The method integrates traditional machine vision detection algorithms and AI deep learning detection algorithms to compare and analyze the corrected scan image with the product Gerber image of the FPC to be inspected, and obtains the defect information of the FPC to be inspected from the corrected scan image, including: The corrected scan image and the Gerber image of the FPC to be inspected are preprocessed and registered. The registered corrected scan image and the product Gerber image are compared using traditional machine vision detection algorithms to determine the basic defects in the corrected scan image and locate all suspected defect areas that deviate from the design standards. AI deep learning detection algorithms are used to detect the suspected defect areas and determine the actual defects in the suspected defect areas; Combining the detection results of the traditional machine vision detection algorithm and the AI ​​deep learning detection algorithm, the determination result of the defect information of the FPC to be detected is output.

11. An FPC defect detection device, applied to an optical inspection system; the optical inspection system includes an optical camera, an encoder connected to the optical camera, a light source controller connected to the encoder, and a multispectral combined light source connected to the light source controller, consisting of a highly uniform coaxial light source and a multi-angle highly directional line light source, wherein the optical camera and the multispectral combined light source are correspondingly matched. Its features are, The device includes: The image acquisition module is used to control the encoder to simultaneously send high-frequency line trigger signals to the optical camera and the light source controller, so that the light source controller controls the high uniformity coaxial light source and the multi-angle high pointing line light source of the multispectral combined light source to work according to the received high-frequency line trigger signals, so that the optical camera can acquire the time-division line scan image of the FPC to be tested. The image correction module is used to perform algorithmic correction and positioning on the obtained time-division line scan image to obtain a corrected scan image; The defect acquisition module is used to integrate traditional machine vision detection algorithms and AI deep learning detection algorithms to compare and analyze the corrected scan image with the product Gerber image of the FPC to be inspected, and to obtain the defect information of the FPC to be inspected from the corrected scan image.

12. An optical detection system, characterized in that, include: Optical camera; The encoder is connected to the optical camera; The light source controller is connected to the encoder; The multispectral combined light source includes a highly uniform coaxial light source and a multi-angle highly directional line light source connected to the light source controller, and the optical camera is correspondingly matched with both the highly uniform coaxial light source and the multi-angle highly directional line light source; The control processor is connected to the optical camera, the encoder, the light source controller, and the multispectral combined light source. The control processor is used to implement the FPC defect detection method as described in any one of claims 110.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement all or part of the method steps of the FPC defect detection method as described in any one of claims 1-10.

Images