Automobile circuit board image acquisition and reconstruction optimization method with high dynamic range

Abstract

The invention provides an automobile circuit board image acquisition and reconstruction optimization method with a high dynamic range, and the method comprises the steps: obtaining a polarization radiation pattern of an automobile circuit board at different angles through a polarization imaging technology, recording the reflection characteristics of each assembly, calculating the polarization degree and polarization angle of incident light, carrying out the automatic recognition and processing of a residual flare region, and carrying out the reconstruction of the residual flare region. Polarization radiation light intensity suppression of a strong reflection area is achieved by calculating a Stokes vector and a polarization angle of a specific area, exposure parameters are optimized according to reflection characteristics of different materials, an overexposure area is marked through a stripe coding strategy, and a stripe mode is customized by combining the geometrical shape and material characteristics of an automobile circuit board. According to the method, the 2D image, the 3D image and the polarization image are fused through the PCA method, a high-quality fusion image without saturated pixels is generated, the problems of flare, overexposure and the like existing in the automobile circuit board image obtaining process are effectively solved, and the dynamic range and definition of the image are improved.

Description

Technical Field

[0001] The present invention relates to the field of image recognition and processing, and in particular to a method for acquiring and reconstructing an automobile circuit board image with a high dynamic range. Background Art

[0002] In the modern electronics manufacturing industry, quality control of automotive circuit boards is crucial. As the basic component of electronic equipment, the surface quality and correct installation of components of automotive circuit boards are directly related to the performance and reliability of the product. Therefore, the development of efficient automotive circuit board inspection technology is of great significance to improving production efficiency and product quality. Traditional automotive circuit board inspection methods mainly rely on manual visual inspection or simple optical imaging technology. These methods have obvious deficiencies in inspection speed, accuracy and degree of automation. Especially when inspecting complex automotive circuit boards, traditional methods have difficulty in effectively identifying tiny defects and highly reflective areas.

[0003] With the advancement of computer vision and image processing technologies, automotive circuit board inspection technology based on high dynamic range imaging has become a research hotspot. High dynamic range imaging can capture a wider range of brightness, helping to overcome the limitations of traditional imaging techniques in high-contrast scenes. In particular, combining it with polarization imaging can further enhance image quality, reduce glare interference, and improve defect detection accuracy. Polarization imaging, by analyzing the polarization characteristics of light, can effectively distinguish the surface properties of different materials, which is particularly important for the diverse materials found on complex automotive circuit boards. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for acquiring and reconstructing an automobile circuit board image with a high dynamic range.

[0005] The problem to be solved by the present invention is: to solve the problem of image quality degradation caused by different material surface reflective properties during the acquisition of automotive circuit board images, especially the glare phenomenon caused by high reflectivity areas, which affects the contrast and clarity of the image, and thus affects subsequent defect detection and quality assessment. By adopting polarization imaging technology, combined with image processing algorithms and multimodal image fusion technology, glare can be effectively suppressed, image quality can be improved, unsaturated pixels in the image can be ensured, the image's detail expression can be enhanced, and high dynamic range imaging requirements can be met, providing a high-quality image foundation for automated inspection of automotive circuit boards.

[0006] A high dynamic range automotive circuit board image acquisition and reconstruction optimization method adopts the following technical solutions: S1: Initialize the polarization imaging system, calibrate the photodetector, maintain a stable light source, adjust the position of the polarizer, operate under constant lighting conditions, perform beam splitting and attenuation on-site lighting, obtain polarization radiation patterns of glare at 0°, 45°, and 90°, and observe and record the reflectance characteristics of various components on the automotive circuit board. S2: Calculate the Stokes parameters of the incident light based on the recorded light intensity values, and determine the polarization degree P and polarization angle A of the incident light; S3: Select the residual flare region (ROI) in the 90° polarized radiation pattern, calculate the Stokes vector and A parameter of the regional flare, and obtain the corresponding polarized radiation intensity. Use image processing algorithms to automatically identify strong reflection areas. S4: Use the polarization image to divide the automotive circuit board into regions, determine the optimal exposure time based on the reflective properties of different materials, and set exposure parameters based on the functions of different regions; S5: Use fringe coding strategy to mark overexposed areas and adjust the grayscale value of the fringe pattern to set a fringe pattern specifically adapted to the unique geometry and material of automotive circuit boards; S6: 2D, 3D, and polarization images are used as multi-channel inputs to the model for feature extraction and fusion image reconstruction, and the images of different modalities are aligned in the fusion process. S7: Analyze and calculate the covariance matrix and its eigenvalues and eigenvectors, calculate the weighting coefficients, and generate the fused image , the principal component analysis (PCA) method is used to fuse the regional flare suppression polarized radiation map with the original polarized radiation map to obtain a fused map without saturated pixels; S8: Verify the quality of the reconstructed image to ensure that there are no saturated pixels and that the image clarity meets the predetermined standards. Use a combination of objective evaluation indicators such as PSNR and SSIM and subjective evaluation including expert review to verify the image quality.

[0007] Furthermore, the step S2 calculates the Stokes parameters of the incident light according to the recorded light intensity values and determines the polarization degree P and polarization angle A of the incident light, including: S21: Stokes matrix of incident light for , the photodetector receives light intensity for ,in 、 、 、 They are the total light intensity, the difference in polarization components of light along the horizontal and vertical directions, the difference in polarization components of light along the +45° and -45° directions, and the difference in polarization components of light along the left-hand circular polarization and right-hand circular polarization directions. is the angle between the transmission axis of the linear polarizer and the X axis, The phase delay angle introduced by the wave plate; S22: Rotate the polarizer by 0, π / 4, and π / 2, and and Substitute into S21 The calculation formula is , , , ; S23: calculated according to S22 Calculating the degree of polarization and polarization angle .

[0008] Furthermore, the Stokes vector and A parameter of the regional flare are calculated in S3 to obtain the corresponding polarized radiation intensity, and the image processing algorithm is used to automatically identify the strong reflection area, including: S31: For any pixel point, obtain the Stokes vector and A parameter of the pixel point through simultaneous polarization measurement, and obtain the pixel point Polarized radiation intensity corresponding to the direction , ; S32: Select the residual flare region (ROI) in the 90° polarized radiation pattern, calculate the Stokes vector and A parameter of the regional flare, and then obtain the corresponding polarized radiation intensity. For the selected flare region, calculate the average polarization angle of the ROI. , ,in is the polarization angle of the incident light measured at the i-th pixel in the ROI, and m is the number of pixels in the ROI; S33: Preprocess the polarized radiation pattern in the 90° direction, including denoising and normalization. Use the threshold segmentation method to divide the image into background and foreground, preliminarily identify strong reflection areas, and select a threshold to classify the pixels in the image into two categories: strong reflection areas and non-strong reflection areas. Strong reflection areas are areas where the pixel value is greater than the threshold. S34: Use the connected region marking algorithm to mark the detected strong reflection area, give each ROI a unique identifier, and for each marked ROI, repeat the steps in S32 to calculate the Stokes vector and polarization angle A; S35: For each ROI, substitute the calculation formula in S31 The suppressed polarized radiation intensity is calculated, and the calculated suppressed polarized radiation intensity is used to replace the corresponding ROI area in the original image.

[0009] Furthermore, the optimal exposure time is determined according to the reflective characteristics of different materials in S4, and exposure parameters are set according to the functions of different areas, including: S41: For an object with k materials, each of them has k different polarization degrees, which are reflected as k peaks in the histogram of the polarization degree image. The peaks are divided into k regions and then back-mapped back to the polarization degree image, dividing the original image into k regions according to the polarization degree. S42: Calculate the Fresnel reflection coefficients of the object surface in each area for s-light and p-light based on the polarization phase of the projection light source and , , , calculate the incident angle and refraction angle , according to the law of refraction, the refractive index n of the material is obtained and the material type of the object surface is determined; S43: setting k exposures for objects of k types of materials, and synthesizing the optimal exposure time images of k areas into a measurement image.

[0010] Furthermore, in S5, a stripe coding strategy is used to mark the overexposed area, and a stripe pattern is set specifically adapted to the unique geometry and material of the automotive circuit board, including: S51: After the polarization state is obtained, the area with a large vertical polarization component is processed to extract the soft mirror area, that is, the highlight area is segmented and the pixels in the highlight area are marked as 255; S52: Generate a projector version image including all matched saturated pixel clusters, wherein all pixels on and inside the boundary contours of the matched clusters are marked as 255 saturated, and the remaining pixels are marked as 0 unsaturated; S53: Mark the overexposed area by projecting a fringe pattern with a higher maximum input grayscale value, select a brightness value of 120 as the grayscale value of the projected fringe, and use a horizontal and vertical sinusoidal fringe projection method with a maximum input grayscale of 120 to match all projector pixels; S54: After the coefficient of the kth exposure point is determined, all saturated pixels are calculated to obtain the optimal intensity of the image captured by the camera, which is matched with the projection intensity of the projector to obtain the maximum input grayscale value, obtain the final adaptive projection fringe map, obtain the final adaptive projection fringe map, determine the intensity mapping coefficient of each cluster, and optimize the coefficient.

[0011] Furthermore, in S7, the principal component analysis (PCA) method is used to fuse the regional flare suppression polarized radiation image with the original polarized radiation image to obtain a fused image without saturated pixels, including: S71: For the entire flare scene, select different ROI areas to perform flare polarization suppression, and obtain a series of regional flare suppression polarization radiation images. , where j is the ordinal number of ROI and n is the total number of ROIs; S72: Calculate the 2*2 covariance matrix of the flare suppression polarization radiation pattern and the original polarization radiation pattern, and calculate the eigenvector corresponding to the maximum eigenvalue of the eigenvalue diagonal matrix , , and are the eigenvectors corresponding to the largest eigenvalues; S73: According to and Calculate weighting coefficients and , , , calculate the fusion graph , the fused image is normalized and the pixel value range is divided into [0,255].

[0012] The beneficial effects of the present invention are: through polarization imaging technology, the glare phenomenon in the image is effectively reduced, the contrast and clarity of the image are improved, and the subtle reflective characteristics of different materials on the automotive circuit board are accurately captured, making the image more realistic and reliable; By setting different exposure parameters for different materials, it is possible to more accurately present details on automotive circuit boards, especially those subtle structures that are difficult to distinguish using traditional imaging methods, such as solder joints and circuits. The introduction of image processing algorithms, automatic identification of strong reflection areas, and connected area marking algorithms have improved the automation level of image processing, reduced manual intervention, and improved efficiency; The information of three modalities, 2D image, 3D image and polarization image, is integrated to obtain more comprehensive and rich image data, which is helpful for subsequent image analysis and processing. A stripe coding strategy is used to mark overexposed areas, and the exposure parameters are adaptively adjusted to effectively expand the dynamic range of the image, avoid overexposure or underexposure, and ensure the overall quality and consistency of the image. The principal component analysis (PCA) method is used to fuse the images to effectively suppress glare while retaining important image features, ensuring that the fused image has no saturated pixels. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 This is a flow chart of a high dynamic range automotive circuit board image acquisition and reconstruction optimization method. DETAILED DESCRIPTION

[0014] The present invention is further clearly and completely described below, but the protection scope of the present invention is not limited thereto.

[0015] A high dynamic range automotive circuit board image acquisition and reconstruction optimization method adopts the following technical solutions: S1: Initialize the polarization imaging system, calibrate the photodetector, maintain a stable light source, adjust the position of the polarizer, operate under constant lighting conditions, perform beam splitting and attenuation on-site lighting, obtain polarization radiation patterns of glare at 0°, 45°, and 90°, and observe and record the reflectance characteristics of various components on the automotive circuit board. S2: Calculate the Stokes parameters of the incident light based on the recorded light intensity values, and determine the polarization degree P and polarization angle A of the incident light; S3: Select the residual flare region (ROI) in the 90° polarized radiation pattern, calculate the Stokes vector and A parameter of the regional flare, and obtain the corresponding polarized radiation intensity. Use image processing algorithms to automatically identify strong reflection areas. S4: Use the polarization image to divide the automotive circuit board into regions, determine the optimal exposure time based on the reflective properties of different materials, and set exposure parameters based on the functions of different regions; S5: Use fringe coding strategy to mark overexposed areas and adjust the grayscale value of the fringe pattern to set a fringe pattern specifically adapted to the unique geometry and material of automotive circuit boards; S6: 2D, 3D, and polarization images are used as multi-channel inputs to the model for feature extraction and fusion image reconstruction, and the images of different modalities are aligned in the fusion process. S7: Analyze and calculate the covariance matrix and its eigenvalues and eigenvectors, calculate the weighting coefficients, and generate the fused image , the principal component analysis (PCA) method is used to fuse the regional flare suppression polarized radiation map with the original polarized radiation map to obtain a fused map without saturated pixels; S8: Verify the quality of the reconstructed image to ensure that there are no saturated pixels and that the image clarity meets the predetermined standards. Use a combination of objective evaluation indicators such as PSNR and SSIM and subjective evaluation including expert review to verify the image quality.

[0016] Furthermore, the step S2 calculates the Stokes parameters of the incident light according to the recorded light intensity values and determines the polarization degree P and polarization angle A of the incident light, including: S21: Stokes matrix of incident light for , the photodetector receives light intensity for ,in 、 、 、 They are the total light intensity, the difference in polarization components of light along the horizontal and vertical directions, the difference in polarization components of light along the +45° and -45° directions, and the difference in polarization components of light along the left-hand circular polarization and right-hand circular polarization directions. is the angle between the transmission axis of the linear polarizer and the X axis, The phase delay angle introduced by the wave plate; S22: Rotate the polarizer by 0, π / 4, and π / 2, and and Substitute into S21 The calculation formula is , , , ; S23: calculated according to S22 Calculating the degree of polarization and polarization angle .

[0017] Furthermore, the Stokes vector and A parameter of the regional flare are calculated in S3 to obtain the corresponding polarized radiation intensity, and the image processing algorithm is used to automatically identify the strong reflection area, including: S31: For any pixel point, obtain the Stokes vector and A parameter of the pixel point through simultaneous polarization measurement, and obtain the pixel point Polarized radiation intensity corresponding to the direction , ; S32: Select the residual flare region (ROI) in the 90° polarized radiation pattern, calculate the Stokes vector and A parameter of the regional flare, and then obtain the corresponding polarized radiation intensity. For the selected flare region, calculate the average polarization angle of the ROI. , ,in is the polarization angle of the incident light measured at the i-th pixel in the ROI, and m is the number of pixels in the ROI; S33: Preprocess the polarized radiation pattern in the 90° direction, including denoising and normalization. Use the threshold segmentation method to divide the image into background and foreground, preliminarily identify strong reflection areas, and select a threshold to classify the pixels in the image into two categories: strong reflection areas and non-strong reflection areas. Strong reflection areas are areas where the pixel value is greater than the threshold. S34: Use the connected region marking algorithm to mark the detected strong reflection area, give each ROI a unique identifier, and for each marked ROI, repeat the steps in S32 to calculate the Stokes vector and polarization angle A; S35: For each ROI, substitute the calculation formula in S31 The suppressed polarized radiation intensity is calculated, and the calculated suppressed polarized radiation intensity is used to replace the corresponding ROI area in the original image.

[0018] Furthermore, the optimal exposure time is determined according to the reflective characteristics of different materials in S4, and exposure parameters are set according to the functions of different areas, including: S41: For an object with k materials, each of them has k different polarization degrees, which are reflected as k peaks in the histogram of the polarization degree image. The peaks are divided into k regions and then back-mapped back to the polarization degree image, dividing the original image into k regions according to the polarization degree. S42: Calculate the Fresnel reflection coefficients of the object surface in each area for s-light and p-light based on the polarization phase of the projection light source and , , , calculate the incident angle and refraction angle , according to the law of refraction, the refractive index n of the material is obtained and the material type of the object surface is determined; S43: setting k exposures for objects of k types of materials, and synthesizing the optimal exposure time images of k areas into a measurement image.

[0019] Furthermore, in S5, a stripe coding strategy is used to mark the overexposed area, and a stripe pattern is set specifically adapted to the unique geometry and material of the automotive circuit board, including: S51: After the polarization state is obtained, the area with a large vertical polarization component is processed to extract the soft mirror area, that is, the highlight area is segmented and the pixels in the highlight area are marked as 255; S52: Generate a projector version image including all matched saturated pixel clusters, wherein all pixels on and inside the boundary contours of the matched clusters are marked as 255 saturated, and the remaining pixels are marked as 0 unsaturated; S53: Mark the overexposed area by projecting a fringe pattern with a higher maximum input grayscale value, select a brightness value of 120 as the grayscale value of the projected fringe, and use a horizontal and vertical sinusoidal fringe projection method with a maximum input grayscale of 120 to match all projector pixels; S54: After the coefficient of the kth exposure point is determined, all saturated pixels are calculated to obtain the optimal intensity of the image captured by the camera, which is matched with the projection intensity of the projector to obtain the maximum input grayscale value, obtain the final adaptive projection fringe map, obtain the final adaptive projection fringe map, determine the intensity mapping coefficient of each cluster, and optimize the coefficient.

[0020] Furthermore, in S7, the principal component analysis (PCA) method is used to fuse the regional flare suppression polarized radiation image with the original polarized radiation image to obtain a fused image without saturated pixels, including: S71: For the entire flare scene, select different ROI areas to perform flare polarization suppression, and obtain a series of regional flare suppression polarization radiation images. , where j is the ordinal number of ROI and n is the total number of ROIs; S72: Calculate the 2*2 covariance matrix of the flare suppression polarization radiation pattern and the original polarization radiation pattern, and calculate the eigenvector corresponding to the maximum eigenvalue of the eigenvalue diagonal matrix , , and are the eigenvectors corresponding to the largest eigenvalues; S73: According to and Calculate weighting coefficients and , , , calculate the fusion graph , the fused image is normalized and the pixel value range is divided into [0,255].

[0021] The present invention provides a method for acquiring and reconstructing automotive circuit board images with a high dynamic range. The method uses polarization imaging technology to acquire polarized radiation patterns of the automotive circuit board at different angles, record the reflective properties of each component, calculate the degree of polarization and polarization angle of the incident light, automatically identify and process residual glare areas, and suppress the intensity of polarized radiation in strongly reflective areas by calculating the Stokes vector and polarization angle of specific areas. Exposure parameters are optimized based on the reflective properties of different materials, overexposed areas are marked using a fringe coding strategy, and a customized fringe pattern is created based on the geometric shape and material characteristics of the automotive circuit board. Principal component analysis (PCA) is then used to fuse 2D, 3D, and polarized images to generate a high-quality fused image without saturated pixels. This method effectively addresses the problems of glare and overexposure that exist in the automotive circuit board image acquisition process, while improving the dynamic range and clarity of the image.

Claims

1. A high dynamic range automotive circuit board image acquisition and reconstruction optimization method, characterized in that: include: S1: Initialize the polarization imaging system, calibrate the photodetector, maintain a stable light source, adjust the position of the polarizer, operate under constant lighting conditions, perform beam splitting and attenuation on-site lighting, obtain polarization radiation patterns of glare at 0°, 45°, and 90°, and observe and record the reflectance characteristics of various components on the automotive circuit board. S2: Calculate the Stokes parameters of the incident light based on the recorded light intensity values, and determine the polarization degree P and polarization angle A of the incident light; S3: Select the residual flare region (ROI) in the 90° polarized radiation pattern, calculate the Stokes vector and A parameter of the regional flare, and obtain the corresponding polarized radiation intensity. Use image processing algorithms to automatically identify strong reflection areas. S4: Use the polarization image to divide the automotive circuit board into regions, determine the optimal exposure time based on the reflective properties of different materials, and set exposure parameters based on the functions of different regions; S5: Use fringe coding strategy to mark overexposed areas and adjust the grayscale value of the fringe pattern to set a fringe pattern specifically adapted to the unique geometry and material of automotive circuit boards; S6: 2D, 3D, and polarization images are used as multi-channel inputs to the model for feature extraction and fusion image reconstruction, and the images of different modalities are aligned in the fusion process. S7: Analyze and calculate the covariance matrix and its eigenvalues and eigenvectors, calculate the weighting coefficients, and generate the fused image , the principal component analysis (PCA) method is used to fuse the regional flare suppression polarized radiation map with the original polarized radiation map to obtain a fused map without saturated pixels; S8: Verify the quality of the reconstructed image to ensure that there are no saturated pixels and that the image clarity meets the predetermined standards. Use a combination of objective evaluation indicators such as PSNR and SSIM and subjective evaluation including expert review to verify the image quality.

2. The method for acquiring and reconstructing a high dynamic range automotive circuit board image according to claim 1, wherein: The step S2 calculates the Stokes parameters of the incident light according to the recorded light intensity values, and determines the polarization degree P and polarization angle A of the incident light, including: S21: Stokes matrix of incident light for , the photodetector receives light intensity for ,in 、 、 、 They are the total light intensity, the difference in polarization components of light along the horizontal and vertical directions, the difference in polarization components of light along the +45° and -45° directions, and the difference in polarization components of light along the left-hand circular polarization and right-hand circular polarization directions. is the angle between the transmission axis of the linear polarizer and the X axis, The phase delay angle introduced by the wave plate; S22: Rotate the polarizer by 0, π / 4, and π / 2, and and Substitute into S21 The calculation formula is , , , ; S23: calculated according to S22 Calculating the degree of polarization and polarization angle .

3. The method for acquiring and reconstructing a high dynamic range automotive circuit board image according to claim 1, wherein: The S3 calculates the Stokes vector and A parameter of the regional flare and obtains the corresponding polarized radiation intensity, and automatically identifies the strong reflection area using an image processing algorithm, including: S31: For any pixel point, obtain the Stokes vector and A parameter of the pixel point through simultaneous polarization measurement, and obtain the pixel point Polarized radiation intensity corresponding to the direction , ; S32: Select the residual flare region (ROI) in the 90° polarized radiation pattern, calculate the Stokes vector and A parameter of the regional flare, and then obtain the corresponding polarized radiation intensity. For the selected flare region, calculate the average polarization angle of the ROI. , ,in is the polarization angle of the incident light measured at the i-th pixel in the ROI, and m is the number of pixels in the ROI; S33: Preprocess the polarized radiation pattern in the 90° direction, including denoising and normalization. Use the threshold segmentation method to divide the image into background and foreground, preliminarily identify strong reflection areas, and select a threshold to classify the pixels in the image into two categories: strong reflection areas and non-strong reflection areas. Strong reflection areas are areas where the pixel value is greater than the threshold. S34: Use the connected region marking algorithm to mark the detected strong reflection area, give each ROI a unique identifier, and for each marked ROI, repeat the steps in S32 to calculate the Stokes vector and polarization angle A; S35: For each ROI, substitute the calculation formula in S31 The suppressed polarized radiation intensity is calculated, and the calculated suppressed polarized radiation intensity is used to replace the corresponding ROI area in the original image.

4. The method for acquiring and reconstructing a high dynamic range automotive circuit board image according to claim 1, wherein: In S4, the optimal exposure time is determined according to the reflective characteristics of different materials, and exposure parameters are set according to the functions of different areas, including: S41: For an object with k materials, each of them has k different polarization degrees, which are reflected as k peaks in the histogram of the polarization degree image. The peaks are divided into k regions and then back-mapped back to the polarization degree image, dividing the original image into k regions according to the polarization degree. S42: Calculate the Fresnel reflection coefficients of the object surface in each area for s-light and p-light based on the polarization phase of the projection light source and , , , calculate the incident angle and refraction angle , according to the law of refraction, the refractive index n of the material is obtained and the material type of the object surface is determined; S43: setting k exposures for objects of k types of materials, and synthesizing the optimal exposure time images of k areas into a measurement image.

5. The method for acquiring and reconstructing a high dynamic range automotive circuit board image according to claim 1, wherein: In the S5, a stripe coding strategy is used to mark overexposed areas, and a stripe pattern is set specifically to adapt to the unique geometry and material of automotive circuit boards, including: S51: After the polarization state is obtained, the area with a large vertical polarization component is processed to extract the soft mirror area, that is, the highlight area is segmented and the pixels in the highlight area are marked as 255; S52: Generate a projector version image including all matched saturated pixel clusters, wherein all pixels on and inside the boundary contours of the matched clusters are marked as 255 saturated, and the remaining pixels are marked as 0 unsaturated; S53: Mark the overexposed area by projecting a fringe pattern with a higher maximum input grayscale value, select a brightness value of 120 as the grayscale value of the projected fringe, and use a horizontal and vertical sinusoidal fringe projection method with a maximum input grayscale of 120 to match all projector pixels; S54: After the coefficient of the kth exposure point is determined, all saturated pixels are calculated to obtain the optimal intensity of the image captured by the camera, which is matched with the projection intensity of the projector to obtain the maximum input grayscale value, obtain the final adaptive projection fringe map, obtain the final adaptive projection fringe map, determine the intensity mapping coefficient of each cluster, and optimize the coefficient.

6. The method for acquiring and reconstructing an automotive circuit board image with a high dynamic range according to claim 1, wherein: In S7, the principal component analysis (PCA) method is used to fuse the regional flare suppression polarized radiation image with the original polarized radiation image to obtain a fused image without saturated pixels, including: S71: For the entire flare scene, select different ROI areas to perform flare polarization suppression, and obtain a series of regional flare suppression polarization radiation images. , where j is the ordinal number of ROI and n is the total number of ROIs; S72: Calculate the 2*2 covariance matrix of the flare suppression polarization radiation pattern and the original polarization radiation pattern, and calculate the eigenvector corresponding to the maximum eigenvalue of the eigenvalue diagonal matrix , , and are the eigenvectors corresponding to the largest eigenvalues; S73: According to and Calculate weighting coefficients and , , , calculate the fusion graph , the fused image is normalized and the pixel value range is divided into [0,255].

Images