SAR image scallop suppression method based on distance direction partitioning and Kalman filtering

Abstract

The present application relates to the technical field of SAR image processing, and particularly relates to a SAR image scallop suppression method based on distance direction blocking and Kalman filtering, comprising: obtaining a SAR image, and performing preprocessing to obtain a to-be-processed image; determining a scallop period based on the to-be-processed image; according to the scallop intensity distribution in the to-be-processed image, adaptively blocking the to-be-processed image along the distance direction to obtain a plurality of sub-images; based on the determined scallop period and each sub-image, obtaining scallop stripes of each sub-image through Kalman filtering; based on the obtained scallop stripes of all sub-images, obtaining a fused stripe image through splicing and fusion; and obtaining a target SAR image with scallop suppression by subtracting the fused stripe image from the to-be-processed image. The present application can effectively suppress the scallop effect in the SAR image obtained by ScanSAR technology, and improve the image quality.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to a SAR image scallop suppression method based on range-direction block division and Kalman filtering. Background Technology

[0002] Scanning Synthetic Aperture Radar (ScanSAR) is an important direction in the development of spaceborne synthetic aperture radar (SAR) technology. Its wide-swath mapping capability is beneficial for shortening the global observation cycle and monitoring rapidly changing large-scale surface phenomena, and it has been widely used. Within a single synthetic aperture time, ScanSAR scans multiple times along the range direction and receives data from multiple sub-bands, thus obtaining a mapping bandwidth several times that of traditional strip SAR, greatly expanding Earth observation capabilities.

[0003] Although ScanSAR boasts a wide mapping bandwidth, its unique discontinuous operating mode results in fine, periodically varying stripes along the azimuth direction in images. This radiation inhomogeneity is known as the scallop effect. The scallop effect not only severely impacts the visual quality of images but can also interfere with image interpretation, hindering post-processing such as feature extraction and image stitching, thus creating difficulties for practical applications. Summary of the Invention

[0004] To effectively suppress the scalloping effect in SAR images obtained by ScanSAR technology and improve image quality, this invention provides a SAR image scalloping suppression method, apparatus, electronic device, and storage medium based on range-direction block and Kalman filtering.

[0005] In a first aspect, embodiments of the present invention provide a SAR image scallop suppression method based on range-direction block partitioning and Kalman filtering, comprising:

[0006] Acquire SAR images and perform preprocessing to obtain the image to be processed;

[0007] Based on the image to be processed, the scallop cycle is determined;

[0008] Based on the scallop intensity distribution in the image to be processed, the image to be processed is adaptively divided into blocks along the distance direction to obtain multiple sub-images;

[0009] Based on the determined scallop period and each sub-image, the scallop stripes of each sub-image are obtained by Kalman filtering.

[0010] Based on the scallop stripes obtained from all sub-images, a fused stripe image is obtained by stitching and fusion.

[0011] By subtracting the fused stripe image from the image to be processed, a scallop-suppressed target SAR image is obtained.

[0012] Optionally, determining the scallop cycle includes:

[0013] Determine the size of the image to be processed, including the number of pixels m in the distance direction and the number of pixels n in the azimuth direction;

[0014] Calculate the average gray intensity of the image to be processed in the distance direction to obtain the average gray intensity sequence distributed along the azimuth direction;

[0015] Perform a Fast Fourier Transform on the obtained average grayscale sequence to obtain the corresponding frequency domain distribution;

[0016] Based on the obtained frequency domain distribution, the peak value of a defined region in the frequency domain is determined; the defined region is represented on the horizontal axis as floor(n / 500) to floor(n / 10), where floor() represents the zero-rounding function;

[0017] Based on the determined peak value's x-coordinate, the scallop period Δt = round(1 / t) is determined, where t represents the determined peak value's x-coordinate, and round() represents the rounding function.

[0018] Optionally, the step of adaptively dividing the image to be processed along the distance direction according to the scallop intensity distribution in the image to be processed to obtain multiple sub-images includes:

[0019] The image to be processed is uniformly divided into multiple small image blocks along the distance direction;

[0020] For each small image patch, calculate the corresponding average scallop intensity;

[0021] Determine the average scallop intensity difference between each small image patch and its adjacent small images, and take the side with the smaller average scallop intensity difference as the merging direction of the small images;

[0022] According to the determined merging direction, the small image blocks are merged to obtain multiple sub-images.

[0023] Optionally, the image to be processed is uniformly divided into multiple small image blocks along the distance direction, including:

[0024] The determination is made based on the number of pixels m in the distance direction of the image to be processed;

[0025] If the number of distance pixels m > 5000, then the number of uniform blocks is R = floor(m / 500);

[0026] If the number of distance pixels m satisfies 5000≥m≥2000, then the number of uniform blocks is R=floor(m / 300);

[0027] If the distance to the number of pixels m < 2000, then the number of uniform blocks is R = floor(m / 180); floor() represents the function to round to zero.

[0028] Optionally, for each small image patch, the corresponding average scallop intensity is calculated as follows:

[0029] Calculate the average gray intensity of the small image patch in the distance direction to obtain the average gray sequence distributed along the azimuth direction corresponding to the small image patch.

[0030] The average gray-level sequence distributed along the azimuth direction corresponding to the small image is uniformly divided into k segments, and the local scallop intensity of each segment is calculated to obtain k local scallop intensity values; k is a positive integer, determined according to the size of the small image along the azimuth direction, or according to the scallop period.

[0031] Based on the obtained k local scallop intensity values, the average scallop intensity corresponding to the small image patch is determined by averaging.

[0032] Optionally, a scalloped pattern of a sub-image is obtained by Kalman filtering, including:

[0033] Traverse each column of the sub-image along the azimuth direction, and calculate the mean and variance of all pixel values ​​in the filter window by matching a filter window for each column. These values ​​are used as the mean and variance of the current azimuth direction. The filter window is an integer multiple of the scallop period Δt.

[0034] The pixel sequence in each direction of the sub-image is read sequentially. The gray value in the pixel sequence is subtracted from the mean value corresponding to the current direction to obtain the observation value of the Kalman filter. The variance var corresponding to the current direction is used as the observation noise variance of the Kalman filter. Kalman filtering is performed. Through multiple iterations, the scallop intensity in the current direction is obtained, and finally the scallop stripes of the sub-image are obtained.

[0035] Optionally, obtaining the fused stripe image through stitching and fusion includes:

[0036] The scallop stripes of all the sub-images are stitched together;

[0037] For the pixels at the stitching point, a mean smoothing filter is applied along the distance direction to eliminate the boundary line caused by the stitching.

[0038] Secondly, embodiments of the present invention also provide a SAR image scallop suppression device based on range-direction block partitioning and Kalman filtering, comprising:

[0039] The image acquisition module is used to acquire SAR images and perform preprocessing to obtain the image to be processed;

[0040] The period determination module is used to determine the scallop period based on the image to be processed;

[0041] An adaptive segmentation module is used to adaptively segment the image to be processed along the distance direction according to the intensity distribution of scallops in the image to be processed, so as to obtain multiple sub-images;

[0042] The filtering module is used to obtain the scallop stripes of each sub-image by using Kalman filtering based on the determined scallop period and each sub-image;

[0043] The stitching module is used to obtain a fused stripe image by stitching and fusing the scallop stripes obtained from all sub-images;

[0044] A denoising module is used to obtain a scallop-suppressed target SAR image by subtracting the fused stripe image from the image to be processed.

[0045] Thirdly, embodiments of the present invention also provide an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, it implements the method described in any embodiment of this specification.

[0046] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to perform the methods described in any embodiment of this specification.

[0047] This invention provides a SAR image scallop suppression method, device, electronic device, and storage medium based on range-direction block partitioning and Kalman filtering. Targeting the characteristics of SAR images obtained using ScanSAR technology, this invention proposes scallop suppression based on range-direction block partitioning and Kalman filtering, thereby reducing image noise. This invention employs range-direction adaptive block partitioning, eliminating the need for manual adjustment of specific parameters and enabling automated image processing. The processing method is simple and exhibits excellent scallop suppression effects for SAR images of varying sizes, scenes, and scallop intensities, significantly improving image quality. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 This is a flowchart of a SAR image scallop suppression method based on range-direction block and Kalman filtering provided by an embodiment of the present invention;

[0050] Figure 2 A SAR image with scallop effect is shown;

[0051] Figure 3 It shows the result of Figure 2 The image shown is composed of the average grayscale curve formed by the average grayscale sequence distributed along the azimuth direction.

[0052] Figure 4 It shows Figure 3 The frequency domain distribution image corresponding to the average grayscale curve shown;

[0053] Figure 5 Another SAR image with scallop effect is shown;

[0054] Figure 6 It shows Figure 5 The adaptive block division result corresponding to the image shown;

[0055] Figure 7 This is a schematic diagram of a Kalman filtering process provided in an embodiment of the present invention;

[0056] Figure 8 for Figure 5 The image shown is obtained after scallop suppression;

[0057] Figure 9 for Figure 5 and Figure 8 A comparison of the average grayscale intensity distribution of the images shown;

[0058] Figure 10 for Figure 5 and Figure 8 A comparison of the average scallop intensity distribution in the images shown;

[0059] Figure 11 This is a hardware architecture diagram of an electronic device provided in an embodiment of the present invention;

[0060] Figure 12 This is a structural diagram of a SAR image scallop suppression device based on range-direction block division and Kalman filtering, provided by an embodiment of the present invention. Detailed Implementation

[0061] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0062] As mentioned earlier, the scallop effect not only severely affects the visual quality of images but can also interfere with image interpretation, hindering post-processing such as feature extraction and image stitching, thus causing difficulties for practical applications. Currently, although there are related technologies to suppress the scallop effect in SAR images, due to issues such as range-dependent spatial variability, the suppression effect is usually poor, with weak robustness and adaptability. Some SAR images still retain scallop stripes after processing.

[0063] To address the aforementioned issues, this invention proposes a SAR image scallop suppression method, apparatus, electronic device, and storage medium based on range-direction block partitioning and Kalman filtering. This method can adaptively partition the image in the range direction and process each sub-block separately, thereby reducing the impact of range-direction spatial variability on scallop suppression and improving the Kalman filtering effect.

[0064] The specific implementation of the above concept is described below.

[0065] Please refer to Figure 1 This invention provides a SAR image scallop suppression method based on range-direction block partitioning and Kalman filtering, the method comprising:

[0066] Step 100: Acquire SAR images and perform preprocessing to obtain the image to be processed;

[0067] Step 102: Based on the image to be processed, determine the scallop period, i.e., the pixel width of the scallop stripes;

[0068] Step 104: Based on the scallop intensity distribution in the image to be processed, the image to be processed is adaptively divided into blocks along the distance direction to obtain multiple sub-images;

[0069] Step 106: Based on the determined scallop period and each sub-image, the scallop stripes of each sub-image are obtained by Kalman filtering.

[0070] Step 108: Based on the scallop stripes obtained from all sub-images, a fused stripe image is obtained by stitching and fusing them.

[0071] Step 110: Subtract the fused stripe image from the to-be-processed image to obtain a scallop-suppressed target SAR image.

[0072] In the embodiment of the present invention, the image is adaptively blocked along the range direction, and Kalman filtering is performed on each sub-block (i.e., sub-image) respectively to determine the corresponding scallop stripe. This method can effectively reduce the influence of range-direction space variability on scallop suppression, and then determine the scallop stripe noise more accurately so as to remove it, finally improve the image quality, obtain a detection result closer to the real situation, and provide technical support for SAR image applications.

[0073] The following describes Figure 1 the execution manner of each step shown.

[0074] Optionally, for step 100, the performing preprocessing includes:

[0075] Step 100-0: Determine whether the range direction of the acquired SAR image is the vertical direction of the image; if yes, proceed to the subsequent steps, and if no, perform a transposition operation on the acquired SAR image to make the range direction the vertical direction;

[0076] Step 100-2: Perform unified quantization processing on the SAR image, and control the pixel value range to 0 to 255.

[0077] Further, for step 100-0, the manner for determining whether the range direction of the acquired SAR image is the vertical direction may adopt the following method:

[0078] Calculate the average value in the vertical direction and the horizontal direction of the SAR image respectively to obtain the average gray distribution of the SAR image in the horizontal direction and the vertical direction, and calculate the peak-valley difference thereof, that is, the difference between the maximum value and the minimum value, so as to obtain the peak-valley difference A1 in the horizontal direction and the peak-valley difference A2 in the vertical direction;

[0079] If A1 < A2, the range direction of the current SAR image is the horizontal direction, and a transposition operation is required;

[0080] If A1 > A2, the range direction of the current SAR image is the vertical direction, and no transposition operation is required.

[0081] Further, for step 100-2, it includes:

[0082] Read the maximum value I in the SAR image MAX and the minimum value I MIN , perform a unified quantization operation on all gray pixels in the SAR image, and the quantization formula is:

[0083]

[0084] In the formula, I0(x,r) represents the original SAR image, I1(x,r) represents the image obtained after uniform quantization, and x and r represent the position of the pixel in the horizontal and vertical directions, respectively, that is, the position in the azimuth and range directions.

[0085] Using the above embodiments, the acquired raw SAR images can be preprocessed for subsequent analysis and processing.

[0086] Optionally, for step 102, determining the scallop cycle includes:

[0087] Step 102-0: Determine the size of the image to be processed, including the number of pixels m in the distance direction and the number of pixels n in the azimuth direction;

[0088] Step 102-2: Calculate the average gray intensity of the image to be processed in the distance direction to obtain the average gray intensity sequence distributed along the azimuth direction. The average gray intensity sequence has n values.

[0089] Step 102-4: Perform a fast Fourier transform on the obtained average grayscale sequence to obtain the corresponding frequency domain distribution, which has a total of n points;

[0090] Step 102-6: Based on the obtained frequency domain distribution, determine the peak value of the defined region in the frequency domain; the defined region is represented on the horizontal axis as floor(n / 500)~floor(n / 10), where floor() represents the zero-rounding function;

[0091] Step 102-8: Based on the determined abscissa of the peak value, determine the scallop period Δt = round(1 / t), where t represents the determined abscissa of the peak value and round() represents the rounding function.

[0092] Using the above embodiments, the scallop cycle can be determined automatically and quickly without manual parameter adjustment. In a specific embodiment, such as... Figures 2 to 4 As shown, Figure 2 The image shown is a SAR image with scallop effect, measuring 10000*10000, where m=10000 and n=10000. Figure 3 The average grayscale curves composed of the corresponding average grayscale sequences are shown. Figure 4 The corresponding frequency domain distribution image is shown. Figure 2 Actual measurements of the SAR image shown indicate that the scallop period is 85 pixels wide. Figure 4 The peak abscissa of the frequency domain distribution region shown is 0.0118. Therefore, the calculated scallop period is round(1 / 0.0118) = round(84.75) = 85 pixels wide, which is consistent with the actual performance.

[0093] Optionally, step 104 further includes:

[0094] Step 104-0: Divide the image to be processed into multiple small image blocks evenly along the distance direction;

[0095] Step 104-2: For each small image patch, calculate the corresponding average scallop intensity.

[0096] Step 104-4: Determine the average scallop intensity difference between each small image patch and its adjacent small images, and take the side with the smaller average scallop intensity difference as the merging direction of the small image patch;

[0097] Step 104-6: Merge the small image blocks according to the determined merging direction to obtain multiple sub-images.

[0098] The above embodiments are based on the idea of ​​clustering, stitching together small image patches with similar average scallop intensity to more accurately determine scallop stripes using Kalman filtering. Directly applying Kalman filtering to the image can lead to poor results due to distance-space variability, resulting in limited image quality improvement. Compared to applying Kalman filtering to the entire image, the above embodiments effectively reduce the impact of distance-space variability on scallop suppression and significantly improve image quality. Dividing the image into uniform blocks and then applying Kalman filtering to each block relies too heavily on manually determining the number of blocks, resulting in poor adaptability. Furthermore, the specific block format may affect the final result; too few blocks will lead to little improvement, while too many blocks will increase computation and cause severe image "boundary lines." The above embodiments use adaptive block division, which not only has the advantages of not relying on manual intervention, low computational cost, and fast processing speed, but also has a better effect on scallop effect suppression.

[0099] Optionally, step 104-0 further includes:

[0100] The determination is made based on the number of pixels m in the distance direction of the image to be processed;

[0101] If the number of distance pixels m > 5000, then the number of uniform blocks is R = floor(m / 500);

[0102] If the number of distance pixels m satisfies 5000≥m≥2000, then the number of uniform blocks is R=floor(m / 300);

[0103] If the number of distance pixels m < 2000, then the number of uniform blocks is R = floor(m / 180).

[0104] Step 104-0 divides the image to be processed into uniform blocks. Too few blocks will make it difficult to reflect the advantages of adaptive slicing and clustering, while too many blocks will increase the amount of computation. The above embodiment divides the image into uniform blocks according to the distance of the image, which is beneficial to improve the computational efficiency and at the same time ensure the subsequent filtering effect.

[0105] Optionally, for step 104-2, the average scallop intensity for each small image patch is calculated as follows:

[0106] Calculate the average gray intensity of the small image patch in the distance direction to obtain the average gray sequence distributed along the azimuth direction corresponding to the small image patch, with a total of n values;

[0107] The average gray-level sequence corresponding to the small image patch along the azimuth direction is uniformly divided into k segments, and the local scallop intensity of each segment is calculated to obtain k local scallop intensity values; k is a positive integer, determined according to the size of the small image patch along the azimuth direction, such as setting k = floor(n / 50), or determined according to the scallop period, such as taking 0.5 to 1 scallop period; the expression for the local scallop intensity value is:

[0108]

[0109] Wherein, ΔP loc (j) represents the j-th local scallop intensity value, where j represents the j-th segment in the k-segment average grayscale sequence, with a value between 1 and k, and max[P(j)]. local Let P(j) represent the maximum value in the j-th segment of the average grayscale sequence. local This represents the minimum value in the j-th segment of the average grayscale sequence;

[0110] Based on the obtained k local scallop intensity values, the average scallop intensity corresponding to the small image patch is determined by averaging.

[0111] By performing the above processing on each small image patch, the average scallop intensity of all small patches can be obtained.

[0112] Regarding step 104-2, each small image selects the side with the smaller average scallop intensity difference from its neighboring small images as its merging direction. That is, it takes the neighboring small image with the smaller average scallop intensity difference as its merging target. After each small image is merged with its merging target, several sub-images with equal azimuth and unequal distance can be obtained.

[0113] Furthermore, regarding step 104, if fewer than four sub-images are obtained, the largest sub-image can be divided equally along the distance direction to obtain four sub-images. Alternatively, the image to be processed can be directly divided equally along the distance direction into four sub-images for subsequent Kalman filtering.

[0114] In a specific embodiment, such as Figure 5 and Figure 6 As shown, Figure 5 Another SAR image with scallop effect is shown. The image size is 10000*10000, m=10000, n=10000, so it will be evenly divided into R=floor(10000 / 500)=20 blocks. By calculating the average scallop intensity, the merging direction can be determined. Figure 6 The corresponding adaptive block division results are shown. Figure 6 In the diagram, vertical lines represent small image patches, horizontal coordinates indicate the patch numbers, the height of the vertical lines represents the average scallop intensity of the patch, and arrows indicate the merging direction of the patches determined by the difference in average scallop intensity. Figure 6 After merging in the indicated directions, six sub-images with equal azimuth and unequal range are obtained, thus completing adaptive block segmentation. Figure 6 The merged result is indicated by a dashed box.

[0115] Optionally, for step 106, a scalloped pattern of a sub-image is obtained by Kalman filtering, including:

[0116] Step 106-0: Traverse each column in the sub-image along the azimuth direction, and calculate the mean and variance of all pixel values ​​in the filter window by matching a filter window for each column. These values ​​are used as the mean and variance of the current azimuth direction. The filter window is an integer multiple of the scallop period Δt.

[0117] Step 106-2: Read the pixel sequence of each direction in the sub-image in sequence, subtract the mean value corresponding to the current direction from the gray value in the pixel sequence to obtain the observation value of the Kalman filter, and use the variance var corresponding to the current direction as the observation noise variance of the Kalman filter. Perform Kalman filtering operation, and through multiple iterations, obtain the scallop intensity (i.e., scallop noise intensity) of the current direction, and finally obtain the scallop stripes of the sub-image.

[0118] By performing the above steps on all sub-images, the scallop stripes of all sub-images can be obtained. Figure 7 The Kalman filter process is shown.

[0119] In the above embodiments, the filter window is an integer multiple of the scallop period, which helps improve the Kalman filtering effect and makes the calculated scallop stripes closer to reality. Generally, twice the pixel width, i.e., 2Δt, is used in the calculation. Of course, three or four times the pixel width can also be selected as the length of the filter window, but a filter window with twice the width is sufficient to achieve a good processing effect. Considering the computational complexity, choosing twice the pixel width as the width of the filter window is a better solution.

[0120] Further, for step 106-2, a Kalman filtering operation is performed, including:

[0121] Construct an additive noise model, the expression of which is:

[0122] S c (r,x)=S o (r,x)+o(x)

[0123] In the formula, r and x represent the positions in the range and azimuth directions, respectively, and S c (r,x) represents the sub-image pixels with scallop stripes, S o (r,x) represents the image pixel without scallop stripes, and o(x) represents the scallop intensity;

[0124] The prediction equations for the Kalman filter are constructed as follows:

[0125]

[0126] in, Let P represent the prior estimate for the i-th iteration. i - Let Q be the prior estimation error covariance for the i-th iteration, and let Q be the system noise variance.

[0127] The state update equation for the Kalman filter is:

[0128]

[0129] Among them, K i Let Kalman gain be the value of the i-th iteration. P is the posterior estimate for the i-th iteration, i.e., the filtered result of each step. i Let z be the posterior estimation error covariance of the i-th iteration, R be the observation noise variance, and z be the posterior estimation error covariance of the i-th iteration. i Let be the observation value in the i-th iteration.

[0130] After multiple iterations, the obtained posterior estimate is used as the scallop intensity in the current orientation. By determining the scallop intensity in each orientation of the sub-image, the scallop stripes of that sub-image can be finally obtained.

[0131] Optionally, the process of obtaining the fused stripe image by stitching and fusing in step 108 includes:

[0132] 108-0, will stitch together the scallop stripes of all sub-images;

[0133] 108-2, For the pixels at the splicing point, perform mean smoothing filtering along the distance direction to eliminate the boundary line caused by splicing.

[0134] The window length for mean smoothing filtering can be set as needed, for example, to two scallop periods. When the window length for mean smoothing filtering is two scallop periods, the target pixels are all pixels within one scallop stripe width on each side of the stitching. That is, for pixels within one period on each side of the stitching line, the following method is used for calculation.

[0135] The processing equation for smoothing filtering is:

[0136]

[0137] The window length of the mean smoothing filter is two scallop periods, where Δt represents one scallop period, i.e., the width of the scallop stripes, I(r) represents the pixel value of the stitched image, I'(r) represents the pixel value after smoothing filtering, and r represents the position of the image in the vertical direction.

[0138] Please see Figures 8 to 10 This invention compares the image effects before and after scallop suppression using the provided method. Figure 8 for Figure 5 The image shown is obtained after scallop suppression. Figure 9 for Figure 5 and Figure 8 The image shown is a comparison of the average grayscale intensity distribution. Figure 10 for Figure 5 and Figure 8 The image shown is a comparison of the average scallop intensity distribution. For ease of representation, Figure 9 and Figure 10 The original image is used to refer to... Figure 5 The image shown is referred to as the result image. Figure 8 The image shown is from Figure 5 , Figures 8 to 10 It is evident that the method provided by this invention can effectively handle the scallop effect in SAR images, significantly improve image quality, and exhibits excellent robustness.

[0139] like Figure 11 , Figure 12As shown, this embodiment of the invention provides a SAR image scallop suppression device based on range-direction block partitioning and Kalman filtering. The device embodiment can be implemented in software, hardware, or a combination of both. From a hardware perspective, such as... Figure 11 The diagram shown is a hardware architecture diagram of an electronic device for SAR image scallop suppression based on range-direction block and Kalman filtering, provided in an embodiment of the present invention. Except for... Figure 11 In addition to the processor, memory, network interface, and non-volatile memory shown, the electronic device in the embodiment may also include other hardware, such as a forwarding chip responsible for processing packets. Taking software implementation as an example, such as... Figure 12 As shown, a device in a logical sense is formed by the CPU of its host electronic device reading the corresponding computer program from non-volatile memory into memory and running it. This embodiment provides a SAR image scallop suppression device based on range-oriented block partitioning and Kalman filtering, comprising:

[0140] Image acquisition module 201 is used to acquire SAR images and perform preprocessing to obtain images to be processed;

[0141] The period determination module 202 is used to determine the scallop period based on the image to be processed;

[0142] The adaptive segmentation module 203 is used to adaptively segment the image to be processed along the distance direction according to the intensity distribution of scallops in the image to be processed, so as to obtain multiple sub-images.

[0143] The filtering module 204 is used to obtain the scallop stripes of each sub-image by Kalman filtering based on the determined scallop period and each sub-image.

[0144] The stitching module 205 is used to obtain a fused stripe image by stitching and fusing the scallop stripes obtained from all the sub-images.

[0145] The denoising module 206 is used to obtain a scallop-suppressed target SAR image by subtracting the fused stripe image from the image to be processed.

[0146] In this embodiment of the invention, the image acquisition module 201 can be used to execute step 100 in the above method embodiment, the period determination module 202 can be used to execute step 102 in the above method embodiment, the adaptive block segmentation module 203 can be used to execute step 104 in the above method embodiment, the filtering module 204 can be used to execute step 106 in the above method embodiment, the stitching module 205 can be used to execute step 108 in the above method embodiment, and the noise reduction module 206 can be used to execute step 110 in the above method embodiment.

[0147] It is understood that the structures illustrated in the embodiments of the present invention do not constitute a specific limitation on a SAR image scallop suppression device based on range-oriented block partitioning and Kalman filtering. In other embodiments of the present invention, a SAR image scallop suppression device based on range-oriented block partitioning and Kalman filtering may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0148] The information interaction and execution process between the modules in the above-mentioned device are based on the same concept as the method embodiment of the present invention, and the specific details can be found in the description in the method embodiment of the present invention, and will not be repeated here.

[0149] This invention also provides an electronic device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements a SAR image scallop suppression method based on range-direction block and Kalman filtering according to any embodiment of this invention.

[0150] This invention also provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program causes the processor to perform a SAR image scallop suppression method based on range-direction block and Kalman filtering according to any embodiment of this invention.

[0151] Specifically, a system or apparatus equipped with a storage medium may be provided, on which software program code implementing the functions of any of the embodiments described above is stored, and the computer (or CPU or MPU) of the system or apparatus may read and execute the program code stored in the storage medium.

[0152] In this case, the program code read from the storage medium can itself implement the function of any of the above embodiments, and therefore the program code and the storage medium storing the program code constitute part of the present invention.

[0153] Storage media embodiments for providing program code include floppy disks, hard disks, magneto-optical disks, optical disks (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), magnetic tapes, non-volatile memory cards, and ROMs. Alternatively, program code can be downloaded from a server computer via a communication network.

[0154] Furthermore, it should be clear that not only can the program code read by the computer be executed, but also the operating system or other components operating on the computer can be instructed based on the program code to perform some or all of the actual operations, thereby realizing the function of any of the embodiments described above.

[0155] Furthermore, it is understood that the program code read from the storage medium is written to the memory set in the expansion board inserted into the computer or to the memory set in the expansion module connected to the computer. Then, based on the instructions of the program code, the CPU or other components installed on the expansion board or expansion module execute some and all of the actual operations, thereby realizing the function of any of the above embodiments.

[0156] In summary, this invention provides a SAR image scallop suppression method, apparatus, electronic device, and storage medium based on range-direction block partitioning and Kalman filtering. This method effectively suppresses the scallop effect in images, reduces the impact of range-direction spatial variability, and significantly improves the image quality of SAR images. The technical solution provided by this invention has at least the following advantages:

[0157] (1) Adaptability: It can process images without adjusting specific parameters. The size of the distance block and the scallop period can be calculated by the program itself.

[0158] (2) Robustness: It has a very good scallop suppression effect for SAR images of different sizes, different scenes and different scallop strengths.

[0159] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0160] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as ROM, RAM, magnetic disk, or optical disk.

[0161] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A SAR image scallop suppression method based on range-direction block partitioning and Kalman filtering, characterized in that, include: Acquire SAR images and perform preprocessing to obtain the image to be processed; Based on the image to be processed, the scallop cycle is determined; Based on the scallop intensity distribution in the image to be processed, the image to be processed is adaptively divided into blocks along the distance direction to obtain multiple sub-images; Based on the determined scallop period and each sub-image, the scallop stripes of each sub-image are obtained by Kalman filtering. Based on the scallop stripes obtained from all sub-images, a fused stripe image is obtained by stitching and fusion. By subtracting the fused stripe image from the image to be processed, a scallop-suppressed target SAR image is obtained; The step of adaptively dividing the image to be processed into multiple sub-images along the distance direction based on the scallop intensity distribution in the image to be processed includes: The image to be processed is uniformly divided into multiple small image blocks along the distance direction; For each small image patch, calculate the corresponding average scallop intensity; Determine the average scallop intensity difference between each small image patch and its adjacent small images, and take the side with the smaller average scallop intensity difference as the merging direction of the small images; According to the determined merging direction, the small image blocks are merged to obtain multiple sub-images; Kalman filtering yields scalloped stripes in a sub-image, including: Traversing each column of the sub-image along the azimuth direction, by matching a filter window to each column, the mean and variance of all pixel values ​​within the filter window are calculated as the mean and variance of the current azimuth direction; the filter window is the scallop period Δ. t Integer multiples of; The pixel sequence in each direction of the sub-image is read sequentially. The gray value in the pixel sequence is subtracted from the mean value corresponding to the current direction to obtain the observation value of the Kalman filter. The variance var corresponding to the current direction is used as the observation noise variance of the Kalman filter. Kalman filtering is performed. Through multiple iterations, the scallop intensity in the current direction is obtained, and finally the scallop stripes of the sub-image are obtained.

2. The method according to claim 1, characterized in that, The determination of the scallop cycle includes: Determine the size of the image to be processed, including the number of pixels m in the distance direction and the number of pixels n in the azimuth direction; Calculate the average gray intensity of the image to be processed in the distance direction to obtain the average gray intensity sequence distributed along the azimuth direction; Perform a Fast Fourier Transform on the obtained average grayscale sequence to obtain the corresponding frequency domain distribution; Based on the obtained frequency domain distribution, the peak value of the defined region in the frequency domain is determined; the defined region is represented on the horizontal axis as floor(n / 500)~floor(n / 10), where floor() represents the zero-rounding function; Based on the determined peak value's x-axis, the scallop cycle Δ is determined. t =round(1 / t), where t represents the x-coordinate of the peak value, and round() represents the rounding function.

3. The method according to claim 1, characterized in that, The image to be processed is uniformly divided into multiple small image blocks along the distance direction, including: The determination is made based on the number of pixels m in the distance direction of the image to be processed; If the number of distance pixels m > 5000, then the number of uniform blocks is R = floor(m / 500); If the number of pixels m in the distance direction satisfies 5000≥m≥2000, then the number of uniform blocks is R=floor(m / 300); If the number of pixels m is less than 2000, then the number of uniform blocks is R = floor(m / 180); floor() represents the zero-rounding function.

4. The method according to claim 3, characterized in that, For each small image patch, the corresponding average scallop intensity is calculated as follows: Calculate the average gray intensity of the small image patch in the distance direction to obtain the average gray sequence distributed along the azimuth direction corresponding to the small image patch. The average gray-level sequence corresponding to the small image along the azimuth direction is uniformly divided into k The segments were divided, and the local scallop intensity of each segment was calculated to obtain... k Local scallop intensity values; k It is a positive integer, determined based on the size of the orientation of the small image patch, or based on the scallop cycle; Based on the obtained k The average scallop intensity corresponding to a small area of ​​the image is determined by averaging the local scallop intensity values.

5. The method according to claim 1, characterized in that, The process of stitching and fusing to obtain a fused stripe image includes: The scallop stripes of all the sub-images are stitched together; For the pixels at the stitching point, a mean smoothing filter is applied along the distance direction to eliminate the boundary line caused by the stitching.

6. A SAR image scallop suppression device based on range-direction block partitioning and Kalman filtering, characterized in that, include: The image acquisition module is used to acquire SAR images and perform preprocessing to obtain the image to be processed; The period determination module is used to determine the scallop period based on the image to be processed; An adaptive segmentation module is used to adaptively segment the image to be processed along the distance direction according to the intensity distribution of scallops in the image to be processed, so as to obtain multiple sub-images; The filtering module is used to obtain the scallop stripes of each sub-image by using Kalman filtering based on the determined scallop period and each sub-image; The stitching module is used to obtain a fused stripe image by stitching and fusing the scallop stripes obtained from all sub-images; A denoising module is used to obtain a scallop-suppressed target SAR image by subtracting the fused stripe image from the image to be processed; The step of adaptively dividing the image to be processed into multiple sub-images along the distance direction based on the scallop intensity distribution in the image to be processed includes: The image to be processed is uniformly divided into multiple small image blocks along the distance direction; For each small image patch, calculate the corresponding average scallop intensity; Determine the average scallop intensity difference between each small image patch and its adjacent small images, and take the side with the smaller average scallop intensity difference as the merging direction of the small images; According to the determined merging direction, the small image blocks are merged to obtain multiple sub-images; Kalman filtering yields scalloped stripes in a sub-image, including: Traversing each column of the sub-image along the azimuth direction, by matching a filter window to each column, the mean and variance of all pixel values ​​within the filter window are calculated as the mean and variance of the current azimuth direction; the filter window is the scallop period Δ. t Integer multiples of; The pixel sequence in each direction of the sub-image is read sequentially. The gray value in the pixel sequence is subtracted from the mean value corresponding to the current direction to obtain the observation value of the Kalman filter. The variance var corresponding to the current direction is used as the observation noise variance of the Kalman filter. Kalman filtering is performed. Through multiple iterations, the scallop intensity in the current direction is obtained, and finally the scallop stripes of the sub-image are obtained.

7. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the method as described in any one of claims 1-5.

8. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed in the computer, it causes the computer to perform the method of any one of claims 1-5.

Images