Interferogram image spectrum recovery-oriented slanted stripe noise correction method and image signal-to-noise ratio optimization method

Abstract

The application discloses a slant stripe noise correction method and an image signal-to-noise ratio optimization method for interferometric image spectrum recovery, and solves the technical problem that the prior art is difficult to ensure the smoothness of the signal-to-noise ratio between the bands of the recovered image while removing the slant stripe noise of the recovered image. The method can not only effectively remove the slant stripe noise of some bands of the recovered image, but also ensure the smoothness of the signal-to-noise ratio between the bands of the recovered image, thereby significantly improving the quality of the recovered image. The application removes the slant stripe noise in the frequency domain by using Fourier transform, and adjusts the pixel value distribution of the image after noise removal by using a segmented pixel interval linear stretching method, so that the pixel value distribution is close to the distribution characteristics of normal bands, thereby effectively removing the slant stripe noise of some bands of the recovered image, ensuring the smoothness of the signal-to-noise ratio between the bands of the recovered image, and significantly improving the quality of the recovered image. The application is suitable for fine processing of hyperspectral imaging.

Description

Technical Field

[0001] This invention relates to a method for spectral restoration of interferometric images, specifically to a method for correcting diagonal stripe noise and optimizing the signal-to-noise ratio of images for spectral restoration of interferometric images. Background Technology

[0002] During the spectral reconstruction of interferometric images using a large-aperture static interferometric spectrometer, diagonal stripes appear in some bands of the reconstructed image due to periodic noise. These stripes not only affect the visual appearance of the image, but more seriously, they significantly reduce the quality of the reconstructed image, especially in applications requiring high image accuracy.

[0003] While some methods exist for removing diagonal stripe noise in existing technologies, these methods often neglect the smoothness of the signal-to-noise ratio (SNR) across different bands of the restored image. This results in anomalies in the SNR of the restored image in bands containing diagonal stripe noise. Therefore, a new method is urgently needed that can effectively remove diagonal stripe noise from the restored image while ensuring the smoothness of the SNR across different bands. Summary of the Invention

[0004] The purpose of this invention is to provide a method for correcting diagonal stripe noise and optimizing the signal-to-noise ratio (SNR) of interferometric image spectral restoration. This addresses the technical problem of existing technologies, which struggle to ensure the smoothness of the SNR across different bands of the restored image while removing diagonal stripe noise. This method not only effectively removes diagonal stripe noise from certain bands of the restored image but also ensures the smoothness of the SNR across different bands, thereby significantly improving the quality of the restored image.

[0005] This invention uses Fourier transform to remove diagonal stripe noise in the frequency domain and employs a linear stretching method to segment pixel intervals to adjust the pixel value distribution of the denoised image, making it close to the distribution characteristics of normal bands. This effectively removes diagonal stripe noise from some bands of the restored image, ensuring the smoothness of the signal-to-noise ratio between different bands of the restored image, thereby significantly improving the quality of the restored image. It is suitable for fine processing of hyperspectral imaging.

[0006] To achieve the above objectives and complete the above inventive concept, the present invention adopts the following technical solution:

[0007] A method for correcting oblique stripe noise in interferometric image spectral restoration, characterized by the following steps:

[0008] Step 1: Spectral Image Reconstruction

[0009] Reconstructing spectral images from interferometric images;

[0010] Step 2: Remove diagonal stripe noise:

[0011] 2.1 Select the image S containing diagonal stripe noise from the spectral image. a For image S a Calculate the mean value for each row along the row direction to obtain the mean vector, and then use the mean vector to obtain the preprocessed image S. a ′;

[0012] 2.2. Preprocess the image S a Transforming from the spatial domain to the frequency domain yields the frequency domain image P. a ; Calculate the frequency domain image P a Calculate the amplitude at each frequency point and then calculate the mean m of all amplitudes. p The amplitude will exceed Qm p The frequency points are set to 0 to remove diagonal stripe noise, resulting in a denoised image P. a ′; where Q is a preset multiple and Q≥1;

[0013] 2.3. Denoising the image P a The frequency domain is transformed into the spatial domain to obtain the spatial domain image S. a "; For spatial domain image S a "Perform brightness restoration to obtain the brightness restored image S." b This completes the correction of oblique stripe noise for spectral restoration of interferometric images.

[0014] Furthermore, step 1 specifically includes:

[0015] Based on spectral restoration technology, through interferometric images Reconstructed spectral images The process is as follows:

[0016]

[0017] Where Recover() represents the spectral recovery method, σ represents periodic noise, and H I W represents the height of the interferometric image. I F represents the width of the interference pattern. I H represents the number of frames in the interferometric image. S W represents the height of the spectral image. S C represents the width of the spectral image. S F represents the number of bands in a spectral image. I =H S .

[0018] Furthermore, step 2.1 specifically includes:

[0019] From spectral images Image S with diagonal stripe noise is selected from the image. a For image S aCalculate the mean of each row along the row direction. Obtain the mean vector Then obtain the preprocessed image S a ′:

[0020]

[0021] Where, p (i,j) Image S a The pixel value in the i-th row and j-th column.

[0022] Furthermore, step 2.2 specifically includes:

[0023] Using Fourier transform to preprocess the image S a Transforming from the spatial domain to the frequency domain yields the frequency domain image P. a =fft(S a ′), where fft(·) denotes the Fourier transform method;

[0024] Calculate the frequency domain image P a Calculate the amplitude at each frequency point and then calculate the average amplitude. Where, p k This represents the k-th frequency point, N represents the number of frequency points, and |·| represents the absolute value operation;

[0025] The magnitude will exceed Qm p The frequency points are set to 0 to remove diagonal stripe noise, resulting in a denoised image. Where Q is a preset multiple.

[0026] Furthermore, in step 2.2: Q = 4.

[0027] Furthermore, step 2.3 specifically includes:

[0028] Using inverse Fourier transform to denoise the image P a The frequency domain is transformed into the spatial domain to obtain the spatial domain image S. a "=ifft(P a ′), where ifft(·) denotes the inverse Fourier transform method;

[0029] For spatial domain image S a Perform brightness restoration to obtain a brightness-restored image.

[0030] Complete the correction of diagonal stripe noise for spectral restoration of interferometric images.

[0031] An image signal-to-noise ratio optimization method for interferometric image spectral restoration, based on the aforementioned diagonal stripe noise correction method for interferometric image spectral restoration, is characterized by including the following steps:

[0032] Step 1: Spectral Image Reconstruction

[0033] Reconstructing spectral images from interferometric images;

[0034] Step 2: Remove diagonal stripe noise:

[0035] 2.1 Select the image S containing diagonal stripe noise from the spectral image. a For image S a Calculate the mean value for each row along the row direction to obtain the mean vector, and then use the mean vector to obtain the preprocessed image S. a ′;

[0036] 2.2. Preprocess the image S a Transforming from the spatial domain to the frequency domain yields the frequency domain image P. a ; Calculate the frequency domain image P a Calculate the amplitude at each frequency point and then calculate the mean m of all amplitudes. p The amplitude will exceed Qm p The frequency points are set to 0 to remove diagonal stripe noise, resulting in a denoised image P. a ′; where Q is a preset multiple and Q≥1;

[0037] 2.3. Denoising the image P a The frequency domain is transformed into the spatial domain to obtain the spatial domain image S. a "; For spatial domain image S a "Perform brightness restoration to obtain the brightness restored image S." b ;

[0038] Step 3: Obtain the image pixel distribution;

[0039] spectral image Image pixel value P without diagonal stripe noise j Arrange the pixels from smallest to largest to obtain pixel distribution d1; restore the brightness of the image S. b pixel value P′ j Arrange the pixels from smallest to largest to obtain pixel distribution d2;

[0040] Step 4: Divide the pixel range;

[0041] Pixel distribution d1 and pixel distribution d2 are divided into Y pixel intervals respectively. The Y pixel intervals of pixel distribution d1 correspond one-to-one with the Y pixel intervals of pixel distribution d2, and the number of pixels in the corresponding pixel intervals is equal, Y≥2.

[0042] Step 5: Obtain the maximum and minimum pixel values;

[0043] Obtain the maximum pixel value P within each pixel interval of the pixel distribution d1. maxand minimum pixel value P min And obtain the maximum pixel value P′ in each pixel interval of the pixel distribution d2. max and minimum pixel value P′ min ;

[0044] Step 6: Linear stretching;

[0045] For each pixel value P in each pixel interval of pixel distribution d1 j One-to-one matching to each pixel value P′ within the corresponding interval of pixel distribution d2. j To obtain the pixel values ​​after linear stretching

[0046]

[0047] Thus, the brightness recovery image S b Linear stretching of pixel values ​​yields the restored image S. b The signal-to-noise ratio optimization for interferometric image spectral restoration was completed.

[0048] Furthermore, step 4 specifically involves:

[0049] Pixel distributions d1 and d2 are each divided into three pixel intervals. The three pixel intervals of pixel distribution d1 correspond one-to-one with the three pixel intervals of pixel distribution d2, and the number of pixels in the corresponding pixel intervals is equal.

[0050] Furthermore, in step 4, pixel distribution d1 and pixel distribution d2 are divided into three pixel intervals: a small value interval, a medium value interval, and a large value interval.

[0051] Furthermore, in step 4, the number of pixels in the small value range and the large value range is n = 1000.

[0052] The beneficial effects of this invention are:

[0053] 1. The oblique stripe noise correction method and image signal-to-noise ratio optimization method for spectral restoration of interferometric images provided by this invention can not only improve the quality of the restored image, but also enhance the reliability and applicability of hyperspectral imaging data, meeting the needs of the high-precision imaging field.

[0054] 2. This invention employs frequency domain analysis technology and removes oblique stripe noise through Fourier transform, effectively solving the problem of oblique stripes caused by periodic noise during image restoration in large-aperture static interferometric spectrometers.

[0055] 3. Compared with existing technologies that only focus on noise removal, this invention pays special attention to the smoothness of the signal-to-noise ratio between different bands of the restored image. By using a linear stretching strategy to segment pixel intervals, the pixel distribution of the image after noise removal is adjusted, so that the signal-to-noise ratio of the restored image maintains a smooth transition between different bands, avoiding abnormal fluctuations in the signal-to-noise ratio and further improving the overall image quality.

[0056] 4. This invention can make targeted adjustments based on the image characteristics of different spectral bands, making it suitable for various hyperspectral imaging application scenarios.

[0057] 5. By introducing frequency domain processing and linear stretching technology for segmented pixel intervals, this invention further optimizes the image processing flow on the basis of noise correction, simplifies the operation steps, improves the overall processing efficiency, and helps to achieve fast and efficient processing of hyperspectral imaging systems. Attached Figure Description

[0058] Figure 1 This is a flowchart illustrating an embodiment of the oblique stripe noise correction method and image signal-to-noise ratio optimization method for interferometric image spectral restoration of the present invention;

[0059] Figure 2 This is a schematic diagram of the restoration of an interference image to a spectral image in an embodiment of the present invention;

[0060] Figure 3 This is a schematic diagram of an image with diagonal stripe noise in an embodiment of the present invention;

[0061] Figure 4 These are image pixel distribution curves in embodiments of the present invention; wherein, (a) is a pixel distribution curve of an image without diagonal stripe noise, and (b) is a pixel distribution curve of an image with diagonal stripe noise removed;

[0062] Figure 5 This is a schematic diagram of adjusting the segmented pixel range to remove diagonal stripe noise in an embodiment of the present invention;

[0063] Figure 6 This is a pixel distribution curve of the image after segmenting pixel intervals and adjusting to remove diagonal stripe noise in an embodiment of the present invention.

[0064] Figure 7 This is a schematic diagram of the signal-to-noise ratio of an image in an embodiment of the present invention, wherein (a) is a schematic diagram of the signal-to-noise ratio of an image with diagonal stripe noise, (b) is a schematic diagram of the signal-to-noise ratio of an image with diagonal stripe noise removed, and (c) is a schematic diagram of the signal-to-noise ratio of an image with diagonal stripe noise removed by segmenting pixel intervals. Detailed Implementation

[0065] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings and embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0066] This invention provides a method for correcting diagonal stripe noise in interferometric image spectral reconstruction, used to address the problem of diagonal stripe noise appearing in certain bands during the reconstruction of interferometric images. Figure 1 As shown, the specific steps are as follows:

[0067] Step 1: Spectral image reconstruction;

[0068] like Figure 2 As shown, based on spectral restoration technology, focal plane interferometric images acquired using an integrating sphere are used to analyze the interferometric images. Reconstructed spectral images The process is as follows:

[0069]

[0070] Here, Recover() represents the spectral recovery method, including baseline removal, apodization, phase transformation, and Fourier transform. σ represents periodic noise, and H... I W represents the height of the interferometric image. I F represents the width of the interference pattern. I H represents the number of frames in the interferometric image. S W represents the height of the spectral image. S C represents the width of the spectral image. S F represents the number of bands in a spectral image. I =H S Interference image and spectral images All data are three-dimensional.

[0071] Step 2: Remove diagonal stripe noise;

[0072] 2.1 From spectral images Image S with diagonal stripe noise is selected from the image. a Image as Figure 3 As shown, for image S a Calculate the mean of each row along the row direction. Obtain the mean vector Through the mean vector Obtain the preprocessed image S a ′:

[0073]

[0074] Where, p (i,j) Image S a The pixel value in the i-th row and j-th column.

[0075] 2.2. Using Fourier transform to process the preprocessed image S a Transforming from the spatial domain to the frequency domain yields the frequency domain image P. a =fft(S a ′), where fft(·) denotes the Fourier transform method;

[0076] Calculate the frequency domain image P a Calculate the amplitude at each frequency point and then calculate the average amplitude. Where, p k This represents the k-th frequency point, N represents the number of frequency points, and |·| represents the absolute value operation;

[0077] The magnitude will exceed Qm p The frequency points are set to 0 to remove diagonal stripe noise, resulting in a denoised image. Where Q is a preset multiple and Q≥1, and in this embodiment, Q=4 is preferred.

[0078] 2.3. Using inverse Fourier transform to denoise the image P a The frequency domain is transformed into the spatial domain to obtain the spatial domain image S. a "=ifft(P a ′), where ifft(·) denotes the inverse Fourier transform method;

[0079] For spatial domain image S a Perform brightness restoration to obtain a brightness-restored image.

[0080] Complete the correction of diagonal stripe noise for spectral restoration of interferometric images.

[0081] This invention provides an image signal-to-noise ratio optimization method for interferometric image spectral reconstruction, which addresses the problems of diagonal stripe noise in some bands and uneven signal-to-noise ratio between bands in the reconstructed image. Figure 1 As shown, the specific steps are as follows:

[0082] Step 1: Spectral image reconstruction;

[0083] like Figure 2 As shown, based on spectral restoration technology, focal plane interferometric images acquired using an integrating sphere are used to analyze the interferometric images. Reconstructed spectral images The process is as follows:

[0084]

[0085] Here, Recover() represents the spectral recovery method, which includes baseline removal, apodization, phase transformation, and Fourier transform. σ represents periodic noise.

[0086] Step 2: Remove diagonal stripe noise;

[0087] 2.1 From spectral images Image S with diagonal stripe noise is selected from the image. a Image as Figure 3 As shown, for image S a Calculate the mean of each row along the row direction. Obtain the mean vector Through the mean vector Obtain the preprocessed image S a ′:

[0088]

[0089] Where, p (i,j) Image S a The pixel value in the i-th row and j-th column.

[0090] 2.2. Using Fourier transform to process the preprocessed image S a Transforming from the spatial domain to the frequency domain yields the frequency domain image P. a =fft(S a ′), where fft(·) denotes the Fourier transform method;

[0091] Calculate the frequency domain image P a Calculate the amplitude at each frequency point and then calculate the average amplitude. Where, p k This represents the k-th frequency point, N represents the number of frequency points, and |·| represents the absolute value operation;

[0092] The amplitude exceeds 4 times the average amplitude m p The frequency points are set to 0 to remove diagonal stripe noise, resulting in a denoised image.

[0093] 2.3. Using inverse Fourier transform to denoise the image P a The frequency domain is transformed into the spatial domain to obtain the spatial domain image S. a "=ifft(P a ′), where ifft(·) denotes the inverse Fourier transform method;

[0094] For spatial domain image S a Perform brightness restoration to obtain a brightness-restored image. Brightness restoration image S bThis refers to removing diagonal stripe noise from the image;

[0095] Step 3: Obtain the image pixel distribution;

[0096] spectral image Image pixel value P without diagonal stripe noise j Arrange the pixels from smallest to largest to obtain pixel distribution d1; restore the brightness of the image S. b pixel value P′ j Arranged from smallest to largest, the pixel distribution d2 is obtained, as follows: Figure 4 As shown;

[0097] Step 4: Divide the pixel range;

[0098] Pixel distributions d1 and d2 are each divided into Y pixel intervals. The Y pixel intervals of pixel distribution d1 correspond one-to-one with the Y pixel intervals of pixel distribution d2, and the number of pixels in the corresponding pixel intervals is equal, Y≥2. In this embodiment, Y=3 is preferred, that is, pixel distributions d1 and d2 are divided into three pixel intervals, namely, the small value interval, the middle value interval, and the large value interval. The number of pixels in the small value interval and the large value interval is n=1000. That is, the first 1000 pixel values ​​in the image arranged from smallest to largest are defined as the small value interval, the last 1000 pixel values ​​are defined as the large value interval, and the remaining pixel values ​​are the middle value interval.

[0099] Step 5: Obtain the maximum and minimum pixel values;

[0100] Obtain the maximum pixel value P within each pixel interval of the pixel distribution d1. max and minimum pixel value P min And obtain the maximum pixel value P′ in each pixel interval of the pixel distribution d2. max and minimum pixel value P′ min ;

[0101] Step 6: Linear stretching;

[0102] For each pixel value P in each pixel interval of pixel distribution d1 j One-to-one matching to each pixel value P′ within the corresponding interval of pixel distribution d2. j To obtain the pixel values ​​after linear stretching

[0103]

[0104] Thus, the brightness recovery image S b Perform linear stretching of pixel values, such as Figure 5 As shown, the restored image S is obtained. b This involves optimizing the signal-to-noise ratio of the image for spectral restoration of interferometric images. The restored image S...b The pixel distribution of ′ is as follows Figure 6 As shown.

[0105] Step 7: Calculate the signal-to-noise ratio to evaluate image quality.

[0106] like Figure 7 As shown, the image S with diagonal stripe noise is calculated. a Image S with diagonal stripe noise removed b Image S segmentation and pixel interval adjustment to remove diagonal stripe noise b The signal-to-noise ratio (SNR) of the bands was used to assess the improvement in image quality. By improving the smoothness of the SNR between different bands, the quality of the restored image was effectively improved.

[0107] Signal-to-noise ratio (SNR) is typically defined as the ratio of signal strength to noise strength, which can be approximated by the mean and standard deviation of the image. The formula is as follows:

[0108]

[0109] Among them, P signal It is the signal power, P noise This is noise power. μ signal σ represents the average pixel intensity of the image. noise This represents the standard deviation of pixel intensity in an image.

[0110] Solve for μ signal and σ noise The formula is as follows:

[0111] Average pixel intensity μ signal :

[0112]

[0113] Among them, S b ′(i,j) represents adjusting the pixel value at position (i,j) of the image after removing diagonal stripe noise.

[0114] Pixel intensity standard deviation σ noise :

[0115]

[0116] This invention effectively improves the quality of the restored image by enhancing the continuity and smoothness of the signal-to-noise ratio (SNR) across different bands. The SNR is typically defined as the ratio of signal strength to noise intensity, which can be approximated by the mean and standard deviation of the image.

[0117] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions within the technical scope disclosed in the present invention should be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for correcting oblique stripe noise in spectral restoration of interferometric images, characterized in that, Includes the following steps: Step 1: Spectral Image Reconstruction Through interferometric images Reconstructed spectral images Step 2: Remove diagonal stripe noise: 2.1 From spectral images Image S with diagonal stripe noise is selected from the image. a For image S a Calculate the mean of each row along the row direction. Obtain the mean vector The preprocessed image S is obtained through the mean vector. a ′: Where, p (i，j) Image S a The pixel value in the i-th row and j-th column; 2.

2. The preprocessed image Sa is transformed from the spatial domain to the frequency domain using Fourier transform to obtain the frequency domain image P. a =fft(S a ′), where fft(·) denotes the Fourier transform method; Calculate the frequency domain image P a Calculate the amplitude at each frequency point and then calculate the average amplitude. Where, p k This represents the k-th frequency point, N represents the number of frequency points, and |·| represents the absolute value operation; The magnitude will exceed Qm p The frequency points are set to 0 to remove diagonal stripe noise, resulting in a denoised image. Where Q is a preset multiple and Q≥1; 2.

3. Using inverse Fourier transform to denoise the image P a The frequency domain is transformed into the spatial domain to obtain the spatial domain image S. a "=ifft(P a ′), where ifft(·) denotes the inverse Fourier transform method; For spatial domain image S a Perform brightness restoration to obtain a brightness-restored image. Complete the correction of diagonal stripe noise for spectral restoration of interferometric images.

2. The method for correcting oblique stripe noise for spectral restoration of interferometric images according to claim 1, characterized in that, Step 1 is as follows: Based on spectral restoration technology, through interferometric images Reconstructed spectral images The process is as follows: Where Recover() represents the spectral recovery method, σ represents periodic noise, and H I W represents the height of the interferometric image. I F represents the width of the interference pattern. I H represents the number of frames in the interferometric image. S W represents the height of the spectral image. S C represents the width of the spectral image. S F represents the number of bands in a spectral image. I =H S .

3. The method for correcting oblique stripe noise for spectral restoration of interferometric images according to claim 2, characterized in that, In step 2.2: Q = 4.

4. A method for optimizing the signal-to-noise ratio of an image for spectral restoration of interferometric images, based on the oblique stripe noise correction method for spectral restoration of interferometric images according to any one of claims 1-3, characterized in that, Includes the following steps: Step 1: Spectral Image Reconstruction Reconstructing spectral images from interferometric images; Step 2: Remove diagonal stripe noise: 2.1 Select the image S containing diagonal stripe noise from the spectral image. a For image S a Calculate the mean value for each row along the row direction to obtain the mean vector, and then use the mean vector to obtain the preprocessed image S. a ′; 2.

2. Preprocess the image S a Transforming from the spatial domain to the frequency domain yields the frequency domain image P. a ; Calculate the frequency domain image P a Calculate the amplitude at each frequency point and then calculate the mean m of all amplitudes. p ; The magnitude will exceed Qm p The frequency points are set to 0 to remove diagonal stripe noise, resulting in a denoised image P. a ′; where Q is a preset multiple and Q≥1; 2.

3. Denoising the image P a The frequency domain is transformed into the spatial domain to obtain the spatial domain image S. a "; For spatial domain image S a "Perform brightness restoration to obtain the brightness restored image S." b ; Step 3: Obtain the image pixel distribution; spectral image Image pixel value P without diagonal stripe noise j Arrange the pixels from smallest to largest to obtain pixel distribution d1; restore the brightness of the image S. b pixel value P j Arrange the pixels from smallest to largest to obtain pixel distribution d2; Step 4: Divide the pixel range; Pixel distribution d1 and pixel distribution d2 are divided into Y pixel intervals respectively. The Y pixel intervals of pixel distribution d1 correspond one-to-one with the Y pixel intervals of pixel distribution d2, and the number of pixels in the corresponding pixel intervals is equal, Y≥2. Step 5: Obtain the maximum and minimum pixel values; Obtain the maximum pixel value P within each pixel interval of the pixel distribution d1. max and minimum pixel value P min And obtain the maximum pixel value P′ in each pixel interval of the pixel distribution d2. max and minimum pixel value P′ min ; Step 6: Linear stretching; For each pixel value P in each pixel interval of pixel distribution d1 j One-to-one matching to each pixel value P′ within the corresponding interval of pixel distribution d2. j To obtain the pixel values ​​after linear stretching Thus, the brightness recovery image S b Linear stretching of pixel values ​​yields the restored image S. b The signal-to-noise ratio optimization for interferometric image spectral restoration was completed.

5. The image signal-to-noise ratio optimization method for interferometric image spectral restoration according to claim 4, characterized in that, Step 4 is as follows: Pixel distributions d1 and d2 are each divided into three pixel intervals. The three pixel intervals of pixel distribution d1 correspond one-to-one with the three pixel intervals of pixel distribution d2, and the number of pixels in the corresponding pixel intervals is equal.

6. The image signal-to-noise ratio optimization method for interferometric image spectral restoration according to claim 5, characterized in that: In step 4, pixel distribution d1 and pixel distribution d2 are divided into three pixel intervals: small value interval, medium value interval, and large value interval.

7. The image signal-to-noise ratio optimization method for interferometric image spectral restoration according to claim 6, characterized in that: In step 4, the number of pixels in the small value range and the large value range is n = 1000.

Images