PCB on-line real-time light source intelligent switching device and method based on feature mapping

Abstract

The invention belongs to the technical field of machine vision detection, and particularly relates to a feature mapping PCB board online real-time light source intelligent switching device and method.The feature mapping PCB board online real-time light source intelligent switching device comprises a rack, a conveying belt, an optical acquisition module, an industrial personal computer and an optical detection module; the optical acquisition module and the optical detection module are connected through the industrial personal computer; the optical acquisition module and the optical detection module are arranged at the front end and the rear end of the conveying belt respectively and are connected through the industrial personal computer, the layout conforms to the sequence of acquisition and detection, and automatic detection of the closed-loop structure of the PCB is achieved conveniently; by constructing a dual-system architecture of the optical acquisition module and the optical detection module, clear and accurate PCB images under an area light source can be obtained by using the optical acquisition module, subsequent image processing and feature extraction are facilitated, data support is provided for automatic switching of light source parameters of the subsequent optical detection module, and then manual switching is replaced.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of machine vision detection, and particularly relates to a feature mapping PCB board online real-time light source intelligent switching device and method. BACKGROUND

[0002] In the field of machine vision detection, the coordination of the light source system and the camera is a key element to guarantee the imaging quality, and is a core link for an automatic optical inspection (AOI) system. The AOI system is widely used in the electronic manufacturing industry and is used for defect detection of printed circuit boards (PCB). In actual production, different PCB boards have diversified material surface characteristics, such as yellow material boards, white material boards, reverse copper foil boards, thick copper boards, blind hole boards, etc. These boards have significant differences in optical reflection, absorption, scattering and other characteristics.

[0003] Currently, the optical adaptation to different PCB board material surface characteristics mainly adopts the manual preset light source parameter mode. Specifically, based on the existing automatic optical inspection equipment, a series of light source parameters are preset to construct a fixed template library. The operator manually switches the corresponding light source parameters from the template library according to the board type to realize the imaging requirements of different boards. However, due to the diversification of PCB boards, there are limitations. Specifically, first, the operator needs to frequently stop the operation during the process of manually switching the preset light source parameters according to experience, which seriously drags down the production line efficiency, resulting in a significant decrease in the comprehensive utilization rate of the equipment. In order to ensure that the operator accurately matches the complex board type, it is necessary to deeply understand the correlation rules of the light source parameters and the board characteristics, and then a high training cost is needed to maintain the professional level of the personnel, increasing the enterprise labor cost burden. In addition, during the manual switching of the light source, it is easy to cause misoperation and lead to poor imaging quality, which in turn increases the defect mis-detection rate, causes the imaging quality to deteriorate and the subsequent detection accuracy to collapse, and seriously affects the product quality and production efficiency.

[0004] Meanwhile, the current industry introduces automatic solutions such as optical flow feature threshold judgment method, image analysis system, etc., which have technical defects in the actual production line. Specifically, first, the image recognition algorithm has poor robustness, large recognition accuracy fluctuation and frequent missed detection when the texture is complex, the light characteristic is reflective and the environment is disturbed, and it cannot stably and accurately realize the board material quality identification; second, the real-time performance and stability are difficult to guarantee, the delay of manual switching leads to unstable imaging quality, and it is difficult to complete the PCB board material quality judgment and light source switching in a millisecond window, causing the first or multiple PCB boards in continuous production to be in a non-optimized imaging state, affecting the imaging stability; finally, the existing technical framework lacks deep modeling of the mapping relationship between "image features-light source parameters", and the light source selection strategy only stays at the level of simple rule matching, which cannot adapt to the nonlinear optical response of new or mixed material PCB boards, resulting in limited system adaptability. This semi-automatic dilemma seriously restricts the adaptability of the system, which cannot meet the trend of board material diversification, and the unstable imaging quality damages the detection reliability, ultimately limiting the universal application of AOI technology in high-precision and high-throughput scenarios, and seriously restricting the development and upgrading of industrial detection towards automation and high precision. SUMMARY

[0005] The application provides a feature-mapped PCB board online real-time light source intelligent switching device and method, which comprises an optical acquisition module and an optical detection module arranged at the front and rear ends of a conveying belt respectively and connected through an industrial computer. The layout conforms to the sequence of acquisition and detection, facilitating the automatic detection of the closed-loop structure of the PCB board. The double-system architecture of the optical acquisition module and the optical detection module can obtain a clear and accurate PCB board image under the area light source, facilitating subsequent image processing and feature extraction and providing data support for the automatic switching of the light source parameters of the subsequent optical detection module, thereby replacing manual switching.

[0006] A feature-mapped PCB board online real-time light source intelligent switching device, comprising a rack, a conveying belt, an optical acquisition module, an industrial computer and an optical detection module, wherein the conveying belt is installed on the rack, and the optical acquisition module and the optical detection module are connected through the industrial computer. The optical acquisition module is installed on the rack through an acquisition module support frame and located at the front end of the conveying belt. The optical acquisition module comprises an area camera and an area light source, and the area camera and the area light source are connected through the acquisition module support frame. The area light source is arranged between the area camera and the conveying belt and parallel to the area camera and the conveying belt. Holes are formed in the area light source, and the holes are axially aligned with the area camera for PCB board image acquisition under the area light source. The industrial computer comprises an image processing module, an identification decision module and a light source switching module; the image processing module is connected to the optical acquisition module; the identification decision module is connected to the image processing module; and the light source switching module is connected to the image processing module and the optical detection module. The optical detection module is installed on the rack through a detection module mounting frame and is located at the rear end of the conveying belt; the optical detection module comprises an upper linear array camera, a lower linear array camera and a linear light source, and the upper linear array camera, the linear light source and the lower linear array camera are connected through the detection module mounting frame; the upper linear array camera, the linear light source, the conveying belt and the lower linear array camera are sequentially arranged vertically downward, and are used for optical detection of the PCB under the switched linear light source.

[0007] By arranging the optical acquisition module and the optical detection module at the front and rear ends of the conveying belt respectively and connecting them through the industrial computer, the layout conforms to the sequence of acquisition and detection, and the automatic detection of the closed-loop structure of the PCB is facilitated; by constructing the double-system architecture of the optical acquisition module and the optical detection module, the optical acquisition module can be used to obtain a clear and accurate image of the PCB under the area light source, which facilitates subsequent image processing and feature extraction and provides data support for automatic switching of the light source parameters of the subsequent optical detection module, thereby replacing manual switching.

[0008] Further, the acquisition module support frame comprises: A fixed support frame is fixedly installed vertically on the rack; The support column comprises two; the support column is fixedly installed on the fixed support frame and is connected perpendicularly to the fixed support frame; The adjustable support frame comprises a connecting plate and an adjusting plate, and the connecting plate is connected perpendicularly to the adjusting plate; the connecting plate is sleeved with the support column; the adjusting plate is in the shape of an inverted T, the top of the adjusting plate is connected to the side wall of the connecting plate, the connecting position of the adjusting plate and the connecting plate is adjusted in the vertical direction through a bolt, the bottom of the adjusting plate is connected to the area light source, and the area light source is parallel to the plane of the support rod.

[0009] By arranging the acquisition module support frame, the face array camera can be used to collect the image of the PCB, and by adjusting the relative positions of the adjusting plate and the connecting plate in the vertical direction, the distance between the face array camera and the area light source can be adjusted and controlled, so that the face array camera can collect a high-precision image of the PCB and ensure the feature integrity of the complex PCB, thereby solving the feature extraction failure problem caused by imaging blur in the traditional scheme.

[0010] Further, the linear light source comprises: The upper line light source comprises an upper first side light source, an upper main line light source and an upper second side light source arranged in a triangular manner, and the upper main line light source is arranged away from the conveying belt; The lower line light source comprises a lower first side light source, a lower main line light source and a lower second side light source arranged in an inverted triangular manner, and the lower main line light source is arranged away from the conveying belt; The upper line light source and the lower line light source are arranged oppositely and on the upper and lower sides of the conveying belt respectively.

[0011] By arranging the upper line light source and the lower line light source and arranging them on the upper and lower sides of the conveying belt, different angles and intensities of light can be provided according to the characteristics of the PCB and the detection requirements, thereby highlighting different characteristics of the PCB and improving the detection efficiency.

[0012] Further, the detection module mounting frame comprises: A fixed mounting frame is vertically upwardly mounted on the rack; A stand is vertically downwardly mounted on the rack and coaxially arranged with the fixed mounting frame; A camera mounting bottom plate comprises an upper camera mounting bottom plate and a lower camera mounting bottom plate 2132; the upper camera mounting bottom plate is mounted on the upper part of the fixed mounting frame and used for connecting the upper line array camera; the lower camera mounting bottom plate 2132 is mounted on the stand and used for connecting the lower line array camera; A light source mounting bottom plate is mounted on the fixed mounting frame and located between the upper camera mounting bottom plate and the rack; the light source mounting bottom plate connects the upper line light source through a triangular support frame; A light source mounting base is mounted on the rack and located at the bottom end of the conveying belt and used for connecting the lower line light source.

[0013] By mounting the detection module mounting frame, the installation of the line light source and the line array camera is facilitated, and the stability and detection accuracy of the optical detection module are ensured.

[0014] Further, lenses are respectively mounted on the area array camera, the upper line array camera and the lower line array camera.

[0015] Further, the recognition and decision module comprises: An AI decision module is used for identifying the mapping relationship between multiple types of PCBs and corresponding light source parameters and constructing a mapping model; An online recognition and decision module maps the image features of the PCB based on the mapping model constructed by the AI decision module, determines the light source parameters corresponding to the current PCB, and sends a light source switching instruction to the optical detection module.

[0016] A feature mapping PCB board online real-time light source intelligent switching device method, comprising: Collecting multiple types of PCB board images by using an optical acquisition module, extracting the image features of each type of PCB board, and manually labeling the corresponding light source parameters of each type of PCB board to construct a labeled data set; Based on the labeled data set, the mapping relationship between the image features of each type of PCB board and the corresponding light source parameters is identified, and a mapping model is generated; Collecting the image of the PCB board to be detected by using the optical acquisition module, extracting the image features of the PCB board to be detected, and determining the light source parameters of the PCB board to be detected according to the mapping model; Based on the determined light source parameters of the PCB board to be detected, the light source of the light source detection module is switched, and optical detection is performed.

[0017] By collecting and pre-identifying the image features of each type of PCB board, a mapping model is constructed which maps the light source parameters, and then the image features of the PCB board to be detected are identified, and the corresponding light source parameters are matched according to the mapping model, the light source of the optical detection module is switched, and optical detection is performed, which can cope with multiple types of PCB boards, thereby significantly reducing the imaging degradation, detection blind area and risk of false and missed judgment caused by light source mismatch.

[0018] Further, the collecting the image of the PCB board to be detected by using the optical acquisition module, extracting the image features of the PCB board to be detected, and determining the light source parameters of the PCB board to be detected according to the mapping model, comprises: Collecting the image of the PCB board to be detected by using the optical acquisition module, and pre-processing by using the image processing module; Based on the pre-processed image of the PCB board to be detected, the color distribution features and the multi-dimensional features of the surface microtexture are extracted, the multi-dimensional PCB board image features are fused to form a region with important features, and the region with important features is generated; Based on the region with important features, combining the mapping model, using a neural network algorithm to learn the mapping relationship between the image features of the PCB board to be detected and the light source parameters, and selecting the light source parameter with the highest probability as the light source parameter of the current PCB board.

[0019] Further, when the light source of the light source detection module is switched, the working time sequence of the line light source and the line array camera is controlled by using the light source synchronous controller to ensure the cooperative work of the light source switching and the line array camera.

[0020] The beneficial effects of the present application are: The application is arranged according to the sequence of collection and detection, and is convenient for realizing the automatic detection of the closed loop structure of the PCB board material; by constructing the double system architecture of the optical collection module and the optical detection module, the optical collection module can be used to obtain the clear and accurate PCB board material image under the surface light source, which is convenient for subsequent image processing and feature extraction, and provides data support for the automatic switching of the light source parameters of the subsequent optical detection module, and then replaces the manual switching; by collecting and pre-recognizing the image features of various types of PCB board materials, a mapping model mapped with the light source parameters is constructed, then the image features of the PCB board to be detected are recognized, and the corresponding light source parameters are matched according to the mapping model, the light source of the optical detection module is switched, and the optical detection is performed, which can cope with various types of PCB board materials, and then significantly reduces the imaging degradation, detection blind area and risk of misjudgment caused by light source mismatch. BRIEF DESCRIPTION OF DRAWINGS

[0021] Figure 1 It is a device structure schematic diagram of the application; Figure 2 It is a device structure schematic diagram of the optical collection module; Figure 3 It is a schematic diagram of the connection relationship between the modules in the application; Figure 4 It is a device side view structure schematic diagram of the optical detection module; Figure 5 It is a flow chart of the application.

[0022] Reference signs: 1, optical collection module; 111, fixed support frame; 112, support column; 113, connecting plate; 114, adjusting plate; 12, area array camera; 13, surface light source; 131, hole; 2, optical detection module; 211, fixed mounting frame; 212, stand column; 2131, upper camera mounting bottom plate; 2132, lower camera mounting bottom plate 2132; 214, light source mounting bottom plate; 215, light source mounting base; 221, upper linear array camera; 222, lower linear array camera; 2311, upper main line light source; 2312, upper first side light source; 2313, upper second side light source; 2321, lower main line light source; 2322, lower first side light source; 2323, lower second side light source; 3, lens; 4, rack 5, conveying belt. DETAILED DESCRIPTION

[0023] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all the other embodiments obtained by a person of ordinary skill in the art without creative effort belong to the scope of protection of the present application.

[0024] It is to be understood that the following description is illustrative only of the aspects of the embodiments described herein within the scope of the appended claims. Numerous other aspects are described herein, and equivalents thereof, and all of which are within the scope of this description. Based on the description herein, one skilled in the art will appreciate that an aspect described herein can be implemented independently of any other aspects and that two or more of these aspects can be combined in any suitable manner. For example, an apparatus can be implemented or a method can be practiced using any number of the aspects described herein. In addition, such an apparatus can be implemented or such a method can be practiced using other structure and / or functionality in addition to or other than one or more of the aspects described herein.

[0025] In addition, in the following description, specific details are provided to thoroughly understand the examples, and the specific meanings of the above-mentioned terms in the present application can be understood by the person of ordinary skill in the art according to the specific circumstances.

[0026] Embodiment 1 Figure 1 As shown, it is a kind of feature mapping PCB board online real-time light source intelligent switching device, including rack 4, conveying belt 5, optical acquisition module 1, industrial computer, optical detection module 2, conveying belt 5 is installed on rack 4, optical acquisition module 1 and optical detection module 2 are connected by industrial computer;By setting optical acquisition module 1 and optical detection module 2 at the two ends before and after conveying belt 5 respectively, and by industrial computer connection, its layout meets the sequential nature of acquisition and detection, to facilitate the automatic detection of PCB board closed loop structure;By constructing the double-system architecture of optical acquisition module 1 and optical detection module 2, clear and accurate PCB board image under area light source 13 can be obtained by optical acquisition module 1, which is convenient for subsequent image processing and feature extraction, and provides data support for subsequent optical detection module 2 light source parameter automatic switching, and then replaces manual switching.

[0027] Specifically, as shown in the drawings, Figure 2 Optical acquisition module 1 is installed on rack 4 by acquisition module support frame and located at the front end of conveying belt 5;Optical acquisition module 1 includes: The collecting module support frame comprises a fixed support frame 111, a support column 112 and an adjustable support frame. The fixed support frame 111 is vertically fixedly installed on the rack 4. The support column 112 comprises two support columns which are fixedly installed on the fixed support frame 111 and are vertically connected with the fixed support frame 111. The adjustable support frame comprises a connecting plate 113 and an adjusting plate 114. The connecting plate 113 is vertically connected with the adjusting plate 114, and the connecting plate 113 is sleeved with the support column 112. The adjusting plate 114 is in an inverted T shape. The top of the adjusting plate 114 is connected with the side wall of the connecting plate 113 and adjusts the connecting position of the adjusting plate 114 and the connecting plate 113 in the vertical direction through a bolt. The bottom of the adjusting plate 114 is connected with the area light source 13, and the area light source 13 is parallel to the plane of the support rod.

[0028] The area array camera 12 is installed on the connecting plate 113 and is arranged towards the conveying belt 5. The area light source 13 is arranged between the area array camera 12 and the conveying belt 5 and is parallel to the area array camera 12 and the conveying belt 5, and is used for collecting the PCB board image under the area light source 13. The area light source 13 is connected with the bottom of the adjusting plate 114 and is parallel to the plane of the support rod. The hole 131 is formed in the area light source 13 and is axially aligned with the area array camera 12, so that the area light source 13 does not block the camera line of sight, and the area array camera 12 can completely and clearly collect the PCB board image.

[0029] Specifically, the industrial computer comprises: The image processing module is connected with the optical collecting module 1 and is used for image preprocessing and image feature extraction of the collected PCB board. The recognition decision module is connected with the image processing module. The recognition decision module comprises an AI decision module and an online recognition decision module. The AI decision module is used for identifying the mapping relationship between the multiple types of PCB boards and the corresponding light source parameters and constructing a mapping model. The online recognition decision module maps the PCB board image features based on the mapping model constructed by the AI decision module, determines the light source parameters corresponding to the current PCB board, and issues a light source switching instruction. The light source switching module is connected with the recognition decision module and the optical detection module 2 and is used for receiving and executing the light source switching instruction of the recognition decision module and switching the light source of the optical detection module 2.

[0030] Figure 3 The connection relationship between the modules is shown in the schematic diagram.

[0031] Specifically, as Figure 4As shown, the optical detection module 2 is installed on the rack 4 through the detection module mounting frame and is located at the rear end of the conveying belt 5; the optical detection module 2 includes an upper linear array camera 221, a lower linear array camera 222 and a linear light source, the upper linear array camera 221, the linear light source and the lower linear array camera 222 are connected through the detection module mounting frame; the upper linear array camera 221, the linear light source, the conveying belt 5 and the lower linear array camera 222 are sequentially arranged vertically downward, and are used for optical detection of the PCB under the switched linear light source.

[0032] The linear light source includes an upper linear light source and a lower linear light source, the upper linear light source and the lower linear light source are arranged oppositely and are respectively located at the upper and lower sides of the conveying belt 5. The upper linear light source includes an upper first side light source 2312, an upper main linear light source 2311 and an upper second side light source 2313 arranged in a triangular manner, and the upper main linear light source 2311 is arranged away from the conveying belt 5; the lower linear light source includes a lower first side light source 2322, a lower main linear light source 2321 and a lower second side light source 2323 arranged in an inverted triangular manner, and the lower main linear light source 2321 is arranged away from the conveying belt 5. The detection module mounting frame includes a fixed mounting frame 211, a column 212, a camera mounting bottom plate, a light source mounting bottom plate 214 and a light source mounting base 215. The fixed mounting frame 211 is installed vertically upward on the rack 4; the column 212 is installed vertically downward on the rack 4 and is coaxially arranged with the fixed mounting frame 211; the camera mounting bottom plate includes an upper camera mounting bottom plate 2131 and a lower camera mounting bottom plate 2132, the upper camera mounting bottom plate 2131 is installed on the upper part of the fixed mounting frame 211 and is used for connecting the upper linear array camera 221, the lower camera mounting bottom plate 2132 is installed on the column 212 and is used for connecting the lower linear array camera 222; the light source mounting bottom plate 214 is installed on the fixed mounting frame 211 and is located between the upper camera mounting bottom plate 2131 and the rack 4, and the light source mounting bottom plate 214 connects the upper linear light source through a triangular support frame; the light source mounting base 215 is installed on the rack 4 and is located at the bottom end of the conveying belt 5, and is used for connecting the lower linear light source.

[0033] In this embodiment, the area array camera 12, the upper linear array camera 221 and the lower linear array camera 222 are respectively provided with a lens 3.

[0034] Embodiment 2 Based on the same design concept, as Figure 5As shown, in the embodiment, an online real-time light source intelligent switching method for feature mapping of PCB board materials is provided. By collecting and pre-identifying image features of various types of PCB board materials, a mapping model mapping light source parameters is constructed, and then image features of the PCB board materials to be detected are identified, and corresponding light source parameters are matched according to the mapping model, the light source of the optical detection module 2 is switched, and optical detection is performed. It can cope with multiple types of PCB board materials, and thus significantly reduce the imaging degradation, detection blind area and risk of false or missed judgment caused by light source mismatch. Specifically, the following steps are included: S1: Collecting multiple types of PCB board material images by using the optical acquisition module 1, extracting image features of various types of PCB board materials, and manually labeling the corresponding light source parameters of each type of PCB board material to construct a labeled data set; S11: Continuously collecting multiple types of PCB board material images under normal production conditions by using the optical acquisition module 1, and performing image preprocessing and image feature extraction by using the image processing module to obtain image features of various types of PCB board materials; In the embodiment, image preprocessing includes image normalization processing, denoising processing, histogram equalization processing, clustering processing, etc.; image feature extraction includes edge feature extraction, histogram statistics, clustering, etc. to extract image features such as image color distribution features, surface microtexture, and solder shape.

[0035] S12: Manually confirming the light source parameters used by various types of PCB board materials; S13: Constructing a labeled data set in combination with image features of various types of PCB board materials and corresponding light source parameters; S2: Based on the labeled data set, identify the mapping relationship between the image features of various types of PCB board materials and the corresponding light source parameters, and generate a mapping model; In the embodiment, based on the labeled data set, based on the mapping relationship between the image features of the PCB board materials and the light source parameters, the mapping model is trained and generated.

[0036] In actual application process, the mapping model can be updated and optimized through the increase of data set, and then the nonlinear optical response of the PCB board materials is realized by self-learning, and then the diversification of the board materials is adapted.

[0037] S3: Collecting images of PCB board materials to be detected by using the optical acquisition module 1, extracting image features of the PCB board materials to be detected, and determining light source parameters of the PCB board materials to be detected according to the mapping model; S31: Collecting images of PCB board materials to be detected by using the optical acquisition module 1, and preprocessing by using the image processing module; S32: Based on the pre-processed PCB material image to be detected, color distribution features, multi-dimensional features of surface micro-texture are extracted, multi-dimensional PCB material image features are fused to form, and regions with important features are generated; S33: Based on the region with important features, combined with the mapping model, the neural network algorithm is used to learn the mapping relationship between the PCB material image features to be detected and the light source parameters, and the light source parameter with the maximum probability is selected as the light source parameter of the current PCB material. In this embodiment, the probability distribution of the PCB material image features to be detected and the light source parameters is identified and calculated, and the light source parameter with the maximum probability is selected as the light source parameter to be switched of the current PCB material to be detected.

[0038] S4: Based on the determined light source parameter of the PCB material to be detected, the light source of the light source detection module is switched, and optical detection is performed.

[0039] In this embodiment, when the light source of the light source detection module is switched, the light source synchronous controller is used to control the working time sequence of the line light source and the line array camera, which is used to ensure the cooperative work of the light source switching and the line array camera.

[0040] In the description of this specification, the description of the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the illustrative description of the above terms does not necessarily mean the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0041] Although the embodiments of the present application have been shown and described above, it should be understood that the above embodiments are exemplary and cannot be understood as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above embodiments within the scope of the present application.

Claims

1. An online real-time light source intelligent switching device for feature mapping PCB board material, characterized in that, The machine frame, the conveying belt, the optical acquisition module, the industrial computer and the optical detection module are connected. The optical acquisition module is installed on the machine frame through the acquisition module support frame and located at the front end of the conveying belt. The optical acquisition module includes a face array camera and a face light source, which are connected through the acquisition module support frame. The face light source is arranged between the face array camera and the conveying belt and parallel to the face array camera and the conveying belt. The face light source is provided with a hole, which is axially aligned with the face array camera, for image acquisition of the PCB under the face light source.

2. The online real-time light source intelligent switching device for feature mapping PCB board material according to claim 1, characterized in that, The industrial computer includes an image processing module, an identification decision module and a light source switching module. The optical detection module is installed on the machine frame through the detection module mounting frame and located at the rear end of the conveying belt. The optical detection module includes an upper linear array camera, a lower linear array camera and a linear light source, which are connected through the detection module mounting frame. The upper linear array camera, the linear light source and the lower linear array camera are vertically arranged in sequence for optical detection of the PCB under the switched linear light source.

3. The device according to claim 1, wherein, The acquisition module support frame includes a fixed support frame, a support column, an adjustable support frame and a face light source. The fixed support frame is vertically fixedly installed on the machine frame. The support column is fixedly installed on the fixed support frame and vertically connected with the fixed support frame. The adjustable support frame includes a connecting plate and an adjusting plate, which are vertically connected.

4. The online real-time light source intelligent switching device for feature mapping PCB board material according to claim 3, characterized in that, The connecting plate is sleeved with the support column. The adjusting plate is in the shape of inverted T, the top of which is connected with the side wall of the connecting plate and adjusts the connection position of the adjusting plate and the connecting plate in the vertical direction through bolts, and the bottom of which is connected with the face light source and parallel to the support rod. The upper linear light source includes an upper first side light source, an upper main linear light source and an upper second side light source arranged in a triangle. The lower linear light source includes a lower first side light source, a lower main linear light source and a lower second side light source arranged in an inverted triangle. The upper linear light source and the lower linear light source are oppositely arranged and respectively located on the upper and lower sides of the conveying belt. The detection module mounting frame includes a fixed mounting frame, a column and a linear light source. The fixed mounting frame is vertically upwardly installed on the machine frame. The column is vertically downwardly installed on the machine frame and coaxially arranged with the fixed mounting frame. A camera mounting base plate, comprising an upper camera mounting base plate and a lower camera mounting base plate 2132; the upper camera mounting base plate is mounted on the upper part of the fixed mounting frame and used for connecting the upper linear array camera; the lower camera mounting base plate 2132 is mounted on the column and used for connecting the lower linear array camera; A light source mounting base plate, which is mounted on the fixed mounting frame and located between the upper camera mounting base plate and the rack; the light source mounting base plate connects the upper linear light source through a triangular support frame; A light source mounting base, which is mounted on the rack and located at the bottom end of the conveying belt and used for connecting the lower linear light source.

5. The device according to claim 1, wherein, The area array camera, the upper linear array camera and the lower linear array camera are respectively provided with a lens.

6. The device according to claim 1, wherein, The recognition decision module comprises: An AI decision module, which is used for identifying the mapping relationship between the multiple types of PCB boards and the corresponding light source parameters and constructing a mapping model; An online recognition decision module, which maps the PCB board image features based on the mapping model constructed by the AI decision module, determines the light source parameters corresponding to the current PCB board, and sends a light source switching instruction to the optical detection module.

7. A method for the feature mapping of the PCB board online real-time light source intelligent switching device of claim 1, characterized in that, It comprises: Collecting multiple types of PCB board images by using an optical acquisition module, extracting image features of each type of PCB board, and manually labeling the light source parameters corresponding to each type of PCB board to construct a labeled data set; Based on the labeled data set, the mapping relationship between the image features of each type of PCB board and the corresponding light source parameters is identified, and a mapping model is generated; Collecting the image of the PCB board to be detected by using the optical acquisition module, extracting the image features of the PCB board to be detected, and determining the light source parameters of the PCB board to be detected according to the mapping model; Based on the determined light source parameters of the PCB board to be detected, the light source of the light source detection module is switched, and optical detection is performed.

8. The method of claim 7, wherein the method further comprises: The use of an optical acquisition module to collect the image of the PCB board to be detected, extract the image features of the PCB board to be detected, and determine the light source parameters of the PCB board to be detected according to the mapping model, comprises: Collecting the image of the PCB board to be detected by using the optical acquisition module, and pre-processing by using the image processing module; Based on the pre-processed PCB board image to be detected, the color distribution features and the multi-dimensional features of the surface micro texture are extracted, fused to form multi-dimensional PCB board image features, and a region with important features is generated; Based on the region with important features, combined with the mapping model, the mapping relationship between the image features of the PCB board to be detected and the light source parameters is learned by using a neural network algorithm, and the light source parameter with the maximum probability is selected as the light source parameter of the current PCB board.

9. The method of claim 7, wherein the method further comprises: When the light source of the light source detection module is switched, the working time sequence of the linear light source and the linear array camera is controlled by using a light source synchronous controller, which is used to ensure the cooperative work of the light source switching and the linear array camera.

Images