Focal plane polarization demosaicing method based on local gradient and channel correlation

The focal plane polarization demosaicing method addresses poor demosaicing effects by utilizing local gradient and channel correlation to optimize interpolation and reduce field of view errors, enhancing accuracy and efficiency.

CN119067880AActive Publication Date: 2024-12-03BEIJING INST OF TECH
View PDF 4 Cites 0 Cited by

Patent Information

Application Number
CN202410920983.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-10
Publication Date
2024-12-03
Estimated Expiration
2044-07-10

AI Technical Summary

Technical Problem

Current polarization demosaicing methods for split-focus plane polarization imaging systems assume uniform distribution of intensity, linear polarization degree, and polarization angle, leading to poor demosaicing effects and increased computational costs, and they often require expensive hardware and long processing times.

Method used

A focal plane polarization demosaicing method based on local gradient and channel correlation, using local gradient to optimize bilinear interpolation, designing polarization channel correlation weights with symmetry characteristics, and applying them in a differential model to reduce instantaneous field of view errors.

Benefits of technology

Improves demosaicing accuracy by reflecting actual scene characteristics, reducing computational costs, and minimizing field of view errors in polarization mosaic images.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119067880A_ABST
    Figure CN119067880A_ABST
Patent Text Reader

Abstract

The invention discloses a focal plane polarization demosaicing method based on local gradient and channel correlation, and belongs to the field of polarization imaging and image processing. The implementation method comprises the following steps: for a polarization mosaic image, obtaining a sparse image through down-sampling; and calculating the local gradient of the current pixel by using adjacent pixels in other polarization directions, and designing a smoothness weight to optimize a bilinear interpolation process, thereby generating an initial demosaiced image. Calculating a relation between the minimum cost function and a normalized cross correlation coefficient through the normalized cross correlation coefficient and guided filtering, and designing a polarization channel correlation weight with a symmetry characteristic; and applying the weight in a polarization channel difference model to further optimize a demosaicing result. According to the method, the distribution characteristics of the intensity, the linear polarization degree and the polarization angle can be reflected more accurately in an actual application scene, the polarization demosaicing effect is improved, and the instantaneous view field error of the focal plane polarization mosaic image is reduced.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to a focal plane polarization demosaicing method, in particular to a focal plane polarization demosaicing method based on local gradient and channel correlation, and belongs to the field of polarization imaging and image processing. Background Art

[0002] Polarization is one of the main physical properties of electromagnetic waves. Introducing polarization information into conventional imaging technology can effectively broaden the dimension of information acquisition. Polarization imaging technology has broad application prospects in military security, biomedicine, materials science, environmental monitoring and other fields, and provides important technical support for applications such as target detection and identification, disease diagnosis and drug development, material characterization and defect detection.

[0003] The split-focus plane polarization imaging system integrates a micro-polarizer array with a super-pixel structure on the detector surface, and simultaneously obtains scene radiation in different polarization directions. It has no moving parts, a compact structure, and is firm and reliable, and has become a research hotspot in polarization imaging technology in recent years. However, since each pixel in the super-pixel structure can only obtain radiation in a single polarization direction, the original image collected by the split-focus plane polarization imaging system is mosaic, and there is an instantaneous field of view error, which affects the reconstruction of the polarization image and the calculation of the polarization information.

[0004] The accuracy of current polarization image demosaicing methods continues to improve, but the computational cost also increases. Complex polarization image demosaicing requires expensive hardware and long processing time, which runs counter to the original intention of using a split-focal plane polarization imaging system. Methods based on polarization channel correlation have received increasing attention due to their interpolation accuracy and processing efficiency. However, such methods usually rely on certain prior knowledge, but this prior knowledge is often difficult to obtain or lacks universality. Therefore, how to reasonably utilize polarization channel correlation and design a universal polarization demosaicing method is a key issue worthy of study. Summary of the invention

[0005] The current polarization demosaicing method based on polarization channel correlation generally assumes that the intensity, linear polarization degree and polarization angle of the scene are uniformly distributed, ignoring the complexity of the scene, so that the polarization demosaicing effect in practical applications is poor. The purpose of the present invention is to propose a polarization demosaicing method based on local gradient and channel correlation, which uses local gradient to optimize bilinear interpolation to generate an initial demosaiced image; through normalized cross-correlation coefficients and guided filtering, polarization channel correlation weights with symmetry characteristics are designed; the weights are applied in the polarization channel difference model to further optimize the demosaicing results and reduce the instantaneous field of view error of the polarization mosaic image of the focal plane.

[0006] The purpose of the present invention is achieved through the following technical solutions.

[0007] The present invention discloses a focal plane polarization demosaicing method based on local gradient and channel correlation, which uses pixels in adjacent other polarization directions to calculate the local gradient of the current pixel, and designs a smoothness weight to optimize the bilinear interpolation process, thereby generating an initial demosaiced image. Through normalized cross-correlation coefficients and guided filtering, the relationship between the minimized cost function and the normalized cross-correlation coefficient is calculated, and a polarization channel correlation weight with symmetry characteristics is designed. The weight is applied in a polarization channel differential model to further optimize the demosaicing result. Compared with the existing focal plane polarization demosaicing method based on polarization channel correlation, the present invention can more accurately reflect the distribution characteristics of intensity, linear polarization degree and polarization angle in actual application scenarios, improve the polarization demosaicing effect, and reduce the instantaneous field of view error of the focal plane polarization mosaic image.

[0008] The present invention discloses a focal plane polarization demosaicing method based on local gradient and channel correlation, comprising the following steps:

[0009] Step 1: For the polarization mosaic image, obtain a sparse image by downsampling.

[0010] For the polarization mosaic image, a sparse image is obtained by downsampling according to equation (1).

[0011] I α =I⊙M α (1)

[0012] Among them, I represents the original mosaic image, I α represents the sparse image of the α polarization channel, M α represents the binary downsampling matrix, and ⊙ represents the Hadamard product.

[0013]

[0014] Step 2: For the polarized mosaic image, the local gradient of the current pixel is calculated using the adjacent pixels in other polarization directions, and the smoothness weight is designed to optimize the bilinear interpolation process to generate the initial demosaiced image.

[0015] Step 2.1: Calculate the gradient of the current pixel in the horizontal, vertical and diagonal directions using adjacent pixels in other polarization directions according to formula (3), and perform linear weighting to determine the local gradient of the current pixel.

[0016]

[0017] Where LG(i,j) represents the local gradient at (i,j).

[0018] Step 2.2: Design smoothness weights to optimize bilinear interpolation and avoid blurring caused by the bilinear interpolation method in edge and detail areas.

[0019]

[0020] Among them, W LG (i,j) represents the smoothness weight at (i,j), ranging from [0,1], LG max Represents the maximum gradient in the image.

[0021] Step 2.3: After obtaining the smoothness weight, optimize the bilinear interpolation.

[0022]

[0023] in, represents the initial demosaiced image of the α polarization channel generated by the local gradient method, F represents the bilinear interpolation filter, Represents a convolution operation.

[0024] Step 3: According to the initial demosaiced image obtained in step 2, the normalized cross-correlation coefficients of different polarization channels are calculated, the relationship between the minimized cost function and the normalized cross-correlation coefficient is calculated using guided filtering, and a polarization channel correlation weight with symmetry characteristics is designed.

[0025] Step 3.1: There is a correlation between different channels of the polarization image, and the correlation is used to further optimize the demosaicing result. According to formula (6), the initial demosaicing images of different polarization channels are subjected to normalized cross-correlation processing to obtain a normalized cross-correlation coefficient that measures the similarity of the two images.

[0026]

[0027] in, Represents the normalized cross-correlation coefficient that measures the similarity of the initial demosaiced images of the α and β polarization channels. The range is [-1,1], where -1 indicates a complete negative correlation, 1 indicates a complete positive correlation, and 0 indicates no linear correlation. The larger the absolute value, the higher the similarity. and denote the average intensity values ​​of the initial demosaiced images for the α and β polarization channels, respectively, represents the covariance of the initial demosaiced image for the α and β polarization channels, and Denotes the standard deviation of the initial demosaiced images for the α and β polarization channels, respectively.

[0028] Step 3.2: In guided filtering, take the input image I inThe output image is I out is the guide image I guide The linear transformation of

[0029]

[0030] Among them, a(i,j) and b(i,j) are the linear coefficients within the filter window ω, and Respectively represent the output image I out and the guide image I guide The intensity value within the filter window ω.

[0031] According to formula (8), a(i,j) and b(i,j) in each filter window are obtained by minimizing the cost function E(a(i,j),b(i,j)).

[0032]

[0033] Among them, λ is a regularization term to prevent the solution a(i,j) from being too large, and is also the smoothing threshold of the guided filter. The purpose of using guided filtering in polarized image demosaicing is to obtain more accurate interpolation results without smoothing the input image, so λ is 0.

[0034] By linearly solving formula (8), we obtain the closed-form solutions of a(i,j) and b(i,j).

[0035]

[0036] in, Represents the input image I in and the guide image I guide The covariance within the filter window ω is, Represents the guidance image I guide The variance within the filter window ω is, and Represents the input image I in and the guide image I guide The mean value within the filter window ω.

[0037] Substituting formula (9) and (10) into formula (8), we get the minimum cost function and The relationship within the filter window ω.

[0038]

[0039] in, Represents the input image I in Variance within the filter window ω.

[0040] Initial demosaiced images of the α and β polarization channels and As input image and guidance image, the minimum cost function and The relationship within the filter window ω is

[0041]

[0042] Step 3.3: Based on the symmetry of the similarity measure of the initial demosaiced images of the α and β polarization channels, the similarity measure of the initial demosaiced images of the α and β polarization channels within the filter window ω is calculated. Defined as

[0043]

[0044] Step 3.4: Based on similarity measurement The α and β polarization channel correlation weights Defined as

[0045]

[0046] Step 4: Using the initial demosaiced image obtained in step 2, calculate the difference images of different polarization channels.

[0047] Use step 2 to get the initial demosaiced image Then, the difference image Δ between the α and β polarization channel images at the α polarization channel sampling point is calculated according to formula (15): α,β .

[0048]

[0049] Step 5: After obtaining the polarization channel difference image according to step 4, optimize the bilinear interpolation according to the local gradient method of step 2 to obtain the polarization channel difference estimation image.

[0050] Use step 4 to obtain the difference image Δ of the α and β polarization channels α,β Then, according to the local gradient method and smoothness weight optimization bilinear interpolation process of equations (3)-(5) in step 2, the differential estimation images of the α and β polarization channels are obtained.

[0051] Step 6: Add the initial demosaiced image obtained in step 2 and the differential estimation images of different polarization channels obtained in step 5 to obtain a polarization channel estimation image.

[0052] According to formula (16), the initial demosaiced image of the β polarization channel obtained in step 2 is The difference estimation images of the α and β polarization channels obtained in step 5 are Add together to obtain the α polarization channel image estimated using the β polarization channel

[0053]

[0054] Step 7: Using the polarization channel correlation weights obtained in step 3, weighted averaging is performed on the polarization channel estimation images obtained in step 6 to obtain a final demosaiced image.

[0055] According to equation (17), the correlation weights of the α and β polarization channels obtained in step 3 are The α polarization channel image estimated from the β polarization channel obtained in step 6 Perform weighted averaging to further optimize the demosaicing result, reduce the instantaneous field of view error, and obtain the final demosaicing image I α .

[0056]

[0057] Beneficial effects:

[0058] 1. Existing polarization demosaicing methods for focal plane based on polarization channel correlation generally assume that the intensity, linear polarization degree and polarization angle of the scene are uniformly distributed, ignoring the complexity of the scene, resulting in poor polarization demosaicing effects in practical applications. The polarization demosaicing method for focal plane based on local gradient and channel correlation disclosed in the present invention uses local gradient to optimize bilinear interpolation to generate an initial demosaiced image; polarization channel correlation weights with symmetry characteristics are designed by normalizing the cross-correlation coefficient and guided filtering; the weights are introduced into the polarization channel differential model to further optimize the demosaicing results and reduce the instantaneous field of view error of the focal plane polarization mosaic image.

[0059] 2. The present invention discloses a defocused plane polarization demosaicing method based on local gradient and channel correlation, which calculates the normalized cross-correlation coefficients of different polarization channels according to the obtained initial demosaiced image; uses guided filtering to calculate the relationship between the minimized cost function and the normalized cross-correlation coefficient; designs a polarization channel correlation weight with symmetry characteristics; introduces the weight into the polarization channel difference model to further optimize the demosaicing result and reduce the instantaneous field of view error of the defocused plane polarization mosaic image.

[0060] 3. The focal plane polarization demosaicing method based on local gradient and channel correlation disclosed in the present invention obtains sparse images of different polarization channels, and then uses the pixels of other adjacent polarization directions to calculate the local gradient of the current pixel; by designing the smoothness weight to optimize the bilinear interpolation process, an initial demosaiced image is generated, thereby improving the initial demosaicing effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 This is a flow chart of the focal plane polarization demosaicing method based on local gradient and channel correlation of the present invention.

[0062] Figure 2 This is a polarization demosaicing result of an actual scene image according to the present invention.

[0063] Among them: 1-original polarization mosaic image, 2-single polarization channel sparse image, 3-initial demosaic image, 4-polarization channel correlation weight, 5-smoothness weight, 6-0° polarization channel sparse image, 7-polarization channel difference image, 8-polarization channel difference estimation image, 9-0° polarization channel image estimated by other polarization channels, 10-final 0° polarization channel demosaic image, 11-sub-focal plane polarization mosaic image of the actual scene, 12-demosaic image of four polarization channels, 13-Stokes vector image, 14-linear polarization degree and polarization angle image. DETAILED DESCRIPTION

[0064] In order to better illustrate the purpose and advantages of the present invention, the invention is further described below with reference to the accompanying drawings and examples.

[0065] Embodiment 1:

[0066] like Figure 1 As shown, the focal plane polarization demosaicing method based on local gradient and channel correlation of this embodiment has the following specific implementation steps:

[0067] Step 1: Taking a four-channel (0°, 45°, 90° and 135°) polarization mosaic image as an example, for the original polarization mosaic image 1, a single polarization channel sparse image 2 is obtained by downsampling according to formulas (1) and (2).

[0068] Step 2: For the original polarized mosaic image 1, according to formulas (3)-(5), the local gradient of the current pixel is calculated using the pixels of other adjacent polarization directions, and the smoothness weight 5 is designed to optimize the bilinear interpolation process to obtain the initial demosaiced image 3.

[0069] Step 3: Using the initial demosaiced image obtained in step 2, according to formulas (6)-(14), calculate the normalized cross-correlation coefficients of different polarization channels, use guided filtering to calculate the relationship between the minimized cost function and the normalized cross-correlation coefficient, and design a polarization channel correlation weight 4 with symmetry characteristics.

[0070] Step 4: Taking the 0° polarization channel sparse image 6 as an example, using the initial demosaiced image 3 obtained in step 2, according to formula (15), calculate the differential images 7 of different polarization channels.

[0071] Step 5: Using the polarization channel difference image 7 obtained in step 4, optimize the bilinear interpolation according to the local gradient method in step 2 to obtain the polarization channel difference estimation image 8.

[0072] Step 6: According to formula (16), the initial demosaiced image 3 obtained in step 2 and the differential estimated images 8 of different polarization channels obtained in step 5 are added to obtain the estimated 0° polarization channel image 9 of other polarization channels (45°, 90° and 135°).

[0073] Step 7: Using the polarization channel correlation weight 4 obtained in step 3, according to formula (17), weighted average is performed on the polarization channel estimation image 9 obtained in step 6 to obtain the final demosaiced image 10.

[0074] like Figure 2 As shown, it is a polarization demosaicing result of an actual scene image disclosed in this embodiment. The focal plane polarization imaging system obtains a polarization mosaic image 11 with pixel interval undersampling in four polarization directions; since a single pixel only responds to the intensity of the current polarization direction, it is necessary to generate four interpolated single-channel polarization images 12 through a demosaicing method; and calculate the Stokes vector image 13 and the linear polarization degree and polarization angle image 14 of the target scene. The polarization mosaic image acquisition device is the PHX050S visible light focal plane polarization imaging system of LUCID Company of Canada, with a resolution of 2448×2048, an image bit number of 8 bits, and polarization directions of 0°, 45°, 90° and 135°. 100 images of the same static scene are continuously acquired and averaged to reduce the influence of temporal noise.

[0075] The focal plane polarization demosaicing method based on local gradient and channel correlation disclosed in this embodiment acts on the process of generating a demosaiced image 12 of four polarization channels from an original polarization mosaic image 11. After obtaining the demosaiced image 12 of four polarization channels, the Stokes vector image 13, linear polarization degree and polarization angle image 14 of the scene are calculated according to equations (18) and (19).

[0076] S0=(I0+I 45 +I 90 +I 135 ) / 2(18a)

[0077] S1=I0-I 90 (18b)

[0078] S2=I 45 -I 135 (18c)

[0079] Among them, I0, I 45 ,I 90 and I135 Represents the radiation intensity images of four polarization directions (0°, 45°, 90° and 135°). S0 represents the total radiation intensity of the scene, S1 represents the horizontal or vertical linear polarization component, and S2 represents the ±45° linear polarization component. Since circular polarization components are rare in nature, the fourth component of the Stokes vector is usually 0.

[0080]

[0081] Wherein, DoLP and AoP represent the degree of linear polarization and angle of polarization images respectively.

[0082] Since S1 and S2 are obtained by image difference in different polarization directions, the accuracy of the demosaicing method directly affects the calculation of linear polarization degree and polarization angle. Therefore, the key to the processing of the focal plane polarization image lies in the effectiveness of the demosaicing effect. By demosaicing the focal plane polarization image, the instantaneous field error of the focal plane polarization mosaic image can be reduced.

[0083] The specific description above further illustrates the purpose, technical solutions and beneficial effects of the invention in detail. It should be understood that the above is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A focal plane polarization demosaicing method based on local gradient and channel correlation, characterized by: The following steps are included: Step 1: For the polarization mosaic image, a sparse image is obtained by downsampling; Step 2: For the polarized mosaic image, the local gradient of the current pixel is calculated using the pixels of other adjacent polarization directions, and the smoothness weight is designed to optimize the bilinear interpolation process, thereby generating an initial demosaiced image; Step 3: According to the initial demosaiced image obtained in step 2, the normalized cross-correlation coefficients of different polarization channels are calculated, the relationship between the minimization cost function and the normalized cross-correlation coefficient is calculated by guided filtering, and the polarization channel correlation weight with symmetry characteristics is designed; Step 4: Using the initial demosaiced image obtained in step 2, calculate the difference images of different polarization channels; Step 5: After obtaining the polarization channel difference image according to step 4, optimize the bilinear interpolation according to the local gradient method of step 2 to obtain the polarization channel difference estimation image; Step 6: Add the initial demosaiced image obtained in step 2 and the differential estimation images of different polarization channels obtained in step 5 to obtain a polarization channel estimation image; Step 7: Using the polarization channel correlation weights obtained in step 3, weighted averaging is performed on the polarization channel estimation images obtained in step 6 to obtain a final demosaiced image.

2. The method for de-mosaicing polarization in a focal plane based on local gradient and channel correlation as claimed in claim 1, characterized in that: Step 1 is implemented as follows: For the polarization mosaic image, a sparse image is obtained by downsampling according to equation (1); I α =I⊙M α (1) Among them, I represents the original mosaic image, I α represents the sparse image of the α polarization channel, M α represents the binary downsampling matrix, ⊙ represents the Hadamard product; 3. The method for de-mosaicing polarization in a focal plane based on local gradient and channel correlation as claimed in claim 1, characterized in that: Step 2 is implemented as follows: Step 2.1, according to formula (3), use the pixels of other adjacent polarization directions to calculate the gradient of the current pixel in the horizontal, vertical and diagonal directions, and perform linear weighting to determine the local gradient of the current pixel; Among them, LG(i,j) represents the local gradient at (i,j); Step 2.2, design smoothness weights to optimize bilinear interpolation and avoid blurring caused by the bilinear interpolation method in edge and detail areas; Among them, W LG (i,j) represents the smoothness weight at (i,j), ranging from [0,1], LG max Represents the maximum gradient in the image; Step 2.3, after obtaining the smoothness weight, optimize the bilinear interpolation; in, represents the initial demosaiced image of the α polarization channel generated by the local gradient method, F represents the bilinear interpolation filter, Represents a convolution operation.

4. The method for de-mosaicing polarization in a focal plane based on local gradient and channel correlation as claimed in claim 1, characterized in that: Step 3 is implemented as follows: Step 3.1, there is a correlation between different channels of the polarization image, and the correlation is used to further optimize the demosaicing result; according to formula (6), the initial demosaicing images of different polarization channels are subjected to normalized cross-correlation processing to obtain a normalized cross-correlation coefficient for measuring the similarity of the two images; in, Represents the normalized cross-correlation coefficient that measures the similarity of the initial demosaiced images of the α and β polarization channels. The range is [-1,1], where -1 represents a complete negative correlation, 1 represents a complete positive correlation, and 0 represents no linear correlation. The larger the absolute value, the higher the similarity. and denote the average intensity values ​​of the initial demosaiced images for the α and β polarization channels, respectively, represents the covariance of the initial demosaiced image for the α and β polarization channels, and represent the standard deviation of the initial demosaiced images for the α and β polarization channels, respectively; Step 3.2: In guided filtering, take the input image I in The output image is I out is the guide image I guide The linear transformation of Among them, a(i,j) and b(i,j) are the linear coefficients within the filter window ω, and Respectively represent the output image I out and the guide image I guide The intensity value within the filter window ω; According to formula (8), a(i,j) and b(i,j) in each filter window are obtained by minimizing the cost function E(a(i,j),b(i,j)); Among them, λ is a regularization term to prevent the solved a(i,j) from being too large, and is also the smoothing threshold of the guided filter. The purpose of using guided filtering in polarization image demosaicing is to obtain more accurate interpolation results without smoothing the input image, so λ is 0. By linearly solving formula (8), we obtain the closed-form solutions of a(i,j) and b(i,j); in, Represents the input image I in and the guide image I guide The covariance within the filter window ω is, Represents the guidance image I guide The variance within the filter window ω is, and Represents the input image I in and the guide image I guide The mean value within the filter window ω; Substituting formula (9) and (10) into formula (8), we get the minimum cost function and The relationship within the filter window ω; in, Represents the input image I in Variance within the filter window ω; Initial demosaiced images of the α and β polarization channels and As input image and guidance image, the minimum cost function and The relationship within the filter window ω is Step 3.3: Based on the symmetry of the similarity measure of the initial demosaiced images of the α and β polarization channels, the similarity measure of the initial demosaiced images of the α and β polarization channels within the filter window ω is calculated. Defined as Step 3.4: Based on similarity measurement The α and β polarization channel correlation weights Defined as 5. The method for de-mosaicing polarization in a focal plane based on local gradient and channel correlation as claimed in claim 1, characterized in that: Step 4 is implemented as follows: Use step 2 to get the initial demosaiced image Then, the difference image Δ between the α and β polarization channel images at the α polarization channel sampling point is calculated according to formula (15): α,β ; 6. The method for de-focusing polarization in the focal plane based on local gradient and channel correlation as claimed in claim 1, characterized in that: Step 5 is implemented as follows: Use step 4 to obtain the difference image Δ of the α and β polarization channels α,β Then, according to the local gradient method and smoothness weight optimization bilinear interpolation process of equations (3)-(5) in step 2, the differential estimation images of the α and β polarization channels are obtained.

7. The method for de-focusing polarization in the focal plane based on local gradient and channel correlation as claimed in claim 1, characterized in that: Step 6 is implemented as follows: According to formula (16), the initial demosaiced image of the β polarization channel obtained in step 2 is The difference estimation images of the α and β polarization channels obtained in step 5 are Add together to obtain the α polarization channel image estimated using the β polarization channel 8. The method for de-focusing polarization in the focal plane based on local gradient and channel correlation as claimed in claim 1, characterized in that: Step 7 is implemented as follows: According to equation (17), the correlation weights of the α and β polarization channels obtained in step 3 are The α polarization channel image estimated from the β polarization channel obtained in step 6 Perform weighted averaging to further optimize the demosaicing result, reduce the instantaneous field of view error, and obtain the final demosaicing image I α ;

Citation Information

Patent Citations

  • Image processing device, image processing method, and imaging device

    CN109804621A

  • Rapid polarization demosaicing method based on polarization prior

    CN113706378A

  • Image demosaicing method and device

    CN118233761A

  • Image demosaicing method and device

    CN118233769A