Internal wave fringe positioning method, system and device based on SAR image

Abstract

The present application discloses a method, system, and device for locating internal wave streaks based on SAR images, relating to the field of marine remote sensing information technology. The method comprises: dividing and preprocessing a SAR image to obtain multiple SAR image blocks; calculating shape parameters and scale parameters based on the pixel values of each SAR image block; determining whether each SAR image block satisfies a K distribution based on the shape parameters and scale parameters; determining different filtering algorithms based on the shape parameters and scale parameters, and filtering the SAR image blocks that satisfy the K distribution to obtain multiple denoised SAR image blocks; processing each denoised SAR image block based on a multi-scale ridgeline positioning algorithm to obtain candidate locations of internal wave streaks in each denoised SAR image block; and segmenting and locating the internal wave streaks using a union-find set and an octree based on the candidate locations of the internal wave streaks in each denoised SAR image block. The present application can effectively reduce coherent speckle noise and accurately locate internal wave streaks.

Description

Technical Field

[0001] The present application relates to the field of marine remote sensing information technology, and in particular to a method, system and device for internal wave fringe positioning based on SAR images. Background Art

[0002] In applications of using Synthetic Aperture Radar (SAR) data to acquire ocean internal wave and surface characteristics, the speckle noise in SAR imaging data causes inaccurate and discontinuous pixel position detection, interfering with the precise acquisition of internal wave streak location information. Previous studies have employed statistical filtering methods such as Lee and Frost to suppress noise and highlight edges, and have employed conventional edge detection methods for edge detection, but these efforts have not fundamentally addressed the problem. The underlying issue lies in the conflict between the uncertainty of inter-pixel differences caused by speckle noise, a phenomenon inherent in SAR imaging, and the certainty required by conventional edge detection. With the advancement of deep learning research, some researchers have also attempted to address this issue using neural networks. However, the desired streak features in different subspaces are still affected by the inherent speckle noise in SAR images. Summary of the Invention

[0003] The purpose of this application is to provide a method, system and device for locating internal wave fringes based on SAR images, which can effectively reduce coherent speckle noise and accurately locate internal wave fringes.

[0004] To achieve the above objectives, this application provides the following solutions.

[0005] In a first aspect, the present application provides a method for internal wave fringe positioning based on SAR images, comprising:

[0006] The SAR image is divided into blocks and preprocessed to obtain a plurality of SAR image blocks; shape parameters and scale parameters are calculated based on the pixel values of each SAR image block; whether each SAR image block satisfies K distribution is determined based on the shape parameters and the scale parameters; different filtering algorithms are determined based on the shape parameters and the scale parameters, and SAR image blocks that satisfy K distribution are filtered to obtain a plurality of denoised SAR image blocks; each denoised SAR image block is processed based on a multi-scale ridge positioning algorithm to obtain candidate positions of internal wave fringes in each denoised SAR image block; based on the candidate positions of internal wave fringes in each denoised SAR image block, the internal wave fringes are segmented and located using a union-find set and an octree.

[0007] In a second aspect, the present application provides an internal wave fringe positioning system based on SAR images, comprising:

[0008] The invention relates to a block segmentation and preprocessing module for performing block segmentation and preprocessing on the SAR image to obtain a plurality of SAR image blocks; a calculation module for calculating shape parameters and scale parameters based on the pixel values of each SAR image block; a judgment module for judging whether each SAR image block satisfies the K distribution based on the shape parameters and the scale parameters; a filtering module for determining different filtering algorithms based on the shape parameters and the scale parameters, and performing filtering processing on the SAR image blocks that satisfy the K distribution to obtain a plurality of denoised SAR image blocks; an initial positioning module for processing each denoised SAR image block based on a multi-scale ridge positioning algorithm to obtain a position candidate point of the internal wave stripes in each denoised SAR image block; and a segmentation and positioning module for segmenting and positioning the internal wave stripes using a union-find set and an octree based on the position candidate points of the internal wave stripes in each denoised SAR image block.

[0009] In a third aspect, the present application provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-mentioned internal wave fringe positioning method based on SAR images.

[0010] According to the specific embodiments provided in this application, this application has the following technical effects.

[0011] (1) By using the K distribution characteristics of SAR image noise to associate multiple filtering algorithms, it can better adapt to the changes in SAR image noise and better suppress image noise.

[0012] (2) The denoised SAR image blocks are processed based on the multi-scale ridge positioning algorithm, which can avoid the uncertainty of local window processing and effectively improve the noise resistance performance.

[0013] (3) The internal wave stripe segmentation processing technology based on union-find set and octree can efficiently realize the segmentation processing of internal wave stripes. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0015] Figure 1 This is a diagram of the application environment of an internal wave fringe positioning method based on SAR images in one embodiment of the present application.

[0016] Figure 2A flowchart of an internal wave fringe positioning method based on SAR images is provided in one embodiment of the present application.

[0017] Figure 3 Schematic diagram of SAR image processing using different filtering algorithms.

[0018] Figure 4 Schematic diagram of the selected strip data.

[0019] Figure 5 Schematic diagram of the distribution of extreme points of one-dimensional strip data with widths of 1, 11, and 31 respectively.

[0020] Figure 6 Schematic diagram of the analysis results when the window size is 11 at different resolutions.

[0021] Figure 7 Schematic diagram of the internal wave stripes in the northeast (EN) diagonal direction of the denoised SAR image block.

[0022] Figure 8 for Figure 7 Schematic diagram of one-dimensional stripe data of internal wave stripes.

[0023] Figure 9 Schematic diagram of the extreme points obtained using the half-height reference line method.

[0024] Figure 10 Schematic diagram of the overall advancement of the strip in four directions.

[0025] Figure 11 Schematic diagram of the process of multi-scale ridge positioning algorithm.

[0026] Figure 12 This is a schematic diagram showing an example of internal wave positioning and extraction results in SAR images.

[0027] Figure 13 This is a schematic diagram of the second example of internal wave positioning and extraction results in SAR images.

[0028] Figure 14 This is a schematic diagram of example three of the internal wave positioning and extraction results of SAR images.

[0029] Figure 15 This is a fourth example diagram of the results of internal wave positioning and extraction from SAR images.

[0030] Figure 16 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0031] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0032] The purpose of this application is to utilize the K-distribution characteristics of SAR image noise and associate multiple statistical filtering methods to effectively suppress image noise. Furthermore, by combining the streak characteristics of internal waves in SAR images, the edges of the sea surface streaks are enhanced, making subsequent internal wave streak detection more effective. Furthermore, a multi-scale ridgeline location algorithm is proposed and utilized to accurately determine the possible locations of internal waves, and the correlation of internal wave streaks is used to obtain the entire internal wave streak. Finally, a union-find and octree algorithm is used to automatically segment and number the internal wave streaks. This application can provide effective methods and technical support for subsequent applications such as automatic inversion of internal waves.

[0033] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0034] The internal wave fringe positioning method based on SAR image provided in the embodiment of the present application can be applied to Figure 1In the application environment shown, the terminal 102 communicates with the server 104 via a network. The data storage system can store data that the server 104 needs to process. The data storage system can be set up separately, integrated with the server 104, or placed on a cloud or other server. The terminal 102 can send the SAR image to be processed to the server 104. After receiving the SAR image to be processed, the server 104 divides and preprocesses the SAR image to obtain multiple SAR image blocks; calculates shape parameters and scale parameters based on the pixel values of each SAR image block; determines whether each SAR image block satisfies the K distribution based on the shape parameters and scale parameters; determines different filtering algorithms based on the shape parameters and scale parameters, and performs filtering processing on the SAR image blocks that satisfy the K distribution to obtain multiple denoised SAR image blocks; processes each denoised SAR image block based on a multi-scale ridgeline positioning algorithm to obtain candidate locations of internal wave fringes in each denoised SAR image block; and segmentes and locates the internal wave fringes based on the candidate locations of the internal wave fringes in each denoised SAR image block using a union-find set and an octree. The server 104 may feed back the obtained internal wave fringes to the terminal 102. Furthermore, in some embodiments, the internal wave fringe positioning method based on SAR images may also be implemented independently by the server 104 or the terminal 102. For example, the terminal 102 may directly perform internal wave fringe positioning on the SAR image to be processed, or the server 104 may obtain the SAR image to be processed from a data storage system and perform internal wave fringe positioning on the SAR image to be processed.

[0035] Terminal 102 may include, but is not limited to, various desktop computers, laptops, smartphones, tablet computers, IoT devices, and portable wearable devices. IoT devices may include smart speakers, smart TVs, smart air conditioners, and smart car devices. Portable wearable devices may include smart watches, smart bracelets, and head-mounted devices. Server 104 may be implemented as a standalone server or a server cluster consisting of multiple servers, or may be a cloud server.

[0036] In an exemplary embodiment, Figure 2 As shown, a method for internal wave fringe positioning based on SAR images is provided. The method is executed by a computer device, specifically a computer device such as a terminal or a server, or a terminal and a server. In the embodiment of the present application, the method is applied to Figure 1 The server 104 in FIG. 1 is used as an example to illustrate the method, which includes the following steps S1 to S6.

[0037] S1: Divide and preprocess the SAR image to obtain multiple SAR image blocks.

[0038] S2: Calculate shape parameters and scale parameters based on the pixel values of each SAR image block.

[0039] Estimate the shape parameter ν and scale parameter c of the K distribution based on the image processed in step S1. The shape parameter ν is typically used to describe the intensity of coherent speckle noise in SAR images, namely, the ratio of the variance to the mean of multiplicative noise. The scale parameter c reflects the heterogeneity of the backscatter intensity of the ground objects. Specifically, the shape parameter ν and scale parameter c are calculated using the mean and variance.

[0040]

[0041]

[0042] in, is the mean pixel value, is the pixel value variance.

[0043] S3: Determining whether each SAR image block satisfies a K-distribution based on the shape parameter and the scale parameter. Specifically, generating theoretical distribution data based on the shape parameter and the scale parameter; and determining whether each SAR image block satisfies a K-distribution based on a difference between the theoretical distribution data and a pixel value of each SAR image block.

[0044] The SAR imaging process differs from conventional optical imaging. The fading caused by the phase accumulation of azimuthally scattered energy is the primary cause of speckle formation. While speckle noise is inherent, it also obeys certain laws, such as the theoretical K distribution, and reflects the characteristics of actual terrain features. However, due to errors in the sensor and data imaging process, the actual generated image does not fully adhere to this distribution law. The purpose of this application's K distribution test analysis is to adaptively select a filtering algorithm based on the test results.

[0045] The specific method to determine whether the data obeys the null hypothesis of K distribution includes: using chi-square test, dividing the data value range into 10 intervals, comparing the actual distribution with the theoretical distribution, and calculating the chi-square critical statistic based on the principle of confidence level greater than 95%. When the chi-square critical value is greater than the statistic, it can be considered to satisfy the K distribution. The cumulative distribution function of K distribution is defined as:

[0046]

[0047] in, is the gamma function, is the modified Bessel function of the second kind, and x is the input pixel value.

[0048] S4: Determine different filtering algorithms based on the shape parameters and scale parameters, and perform filtering on the SAR image blocks that satisfy the K distribution to obtain multiple denoised images. SAR image blocks that do not satisfy the K distribution are discarded.

[0049] Specifically, when the shape parameter is greater than the shape parameter threshold (such as (0.0625), the enhanced Lee filtering algorithm is used to filter the SAR image blocks that meet the K distribution; when the shape parameter is less than the shape parameter threshold and the scale parameter is greater than the scale parameter threshold (such as 4.88e -4 ), the Gamma-MAP filtering algorithm is used to filter the SAR image blocks that meet the K distribution; when the scale parameter is less than the scale parameter threshold, the mean filtering algorithm or the local statistical filtering algorithm is used to filter the SAR image blocks that meet the K distribution.

[0050] When the ν value is large, a small window Lee filter is used to reduce detail loss; when the ν value is small and the c value is large, a large window Gamma-MAP filter is used.

[0051] This embodiment effectively suppresses speckle noise by applying different window sizes and filtering methods based on the statistical parameters (ν and c) and the K distribution assumption. For the internal wave region, high c-value areas are filtered using an enhanced Lee filter or Gamma-MAP filter, while low c-value areas are typically filtered using traditional mean filtering or local statistical filtering. This process is implemented dynamically.

[0052] (1) Enhanced Lee filtering algorithm.

[0053] Enhanced Lee filtering is an improved version of the Lee filtering method, using different filters based on different regions. It divides an image into three regions: uniform regions, where speckle noise can be simply smoothed out using a mean filter; non-uniform regions, where texture information should be preserved when removing noise, and the Lee filtering method can be applied; and regions containing discrete point targets, where the filter should preserve the original values as much as possible.

[0054] The expression of the enhanced Lee filtering algorithm is:

[0055]

[0056] in, , is the weight function of Lee filtering; , is the local coefficient of variation; is the local standard deviation, is the original grayscale value of each pixel in the smoothing window; All are threshold values; ; ; is the imaging view number. is the pixel gray value, is the gray value of the pixel after smoothing; is the average grayscale value of pixels in the window; is the variance of the pixel values in the smoothing window.

[0057] The enhanced Lee filter uses the following criteria: 1) When , the mean value is taken in the filter sub-window instead of the value of the central pixel; 2) , the filtering algorithm is used to calculate the filtering value of the central pixel in the filtering sub-window; 3) , retain the central pixel value.

[0058] (2) Gamma-MAP filtering algorithm.

[0059] This filtering method is a maximum a posteriori probability (MAP) filtering method obtained based on the prior distribution of the data and the likelihood function. The MAP filter assumes that the coherent speckle is a multiplicative Gamma distribution. More information can be obtained without prior knowledge of the probability density function of . The Gamma-MAP filtering process adopts the following criteria:

[0060] 1) When hour, 2) When hour, ;3) When hour, (L≠1), (L=1).

[0061] in, ; ; ; Intermediate variables ; is the size of the smoothing window.

[0062] Figure 3 The schematic diagram of SAR image processing using different filtering algorithms is shown in Figure 2. Figure 3 (a) in the figure is the original image. Figure 3 (b) in the figure is the result after processing by the enhanced Lee filter algorithm. Figure 3 Figure (c) shows the result after applying Gamma-MAP filtering and traditional Gaussian low-pass filtering. The results show that this method can significantly improve the accuracy of SAR image denoising and target detection, thus providing good data conditions for subsequent steps.

[0063] S5: Processing each denoised SAR image block based on a multi-scale ridge positioning algorithm to obtain candidate positions of internal wave fringes in each denoised SAR image block.

[0064] While this holistic multi-scale ridgeline precision positioning technique incorporates the concept of valley tracking in maps, it differs significantly from earlier methods. Here, we apply this holistic multi-scale ridgeline precision positioning technique to the detection of structures similar to "mountain ridges," and achieve the marking of internal wave fringes through holistic filtering, de-slope, and extrema processing.

[0065] This method offers four significant improvements over earlier valley tracking methods in maps. First, rather than processing multiple directions pixel by pixel within a sliding window, nor does it directly determine whether each pixel has a "ridge"-like structure with high center and low sides, it instead processes one-dimensional strip data in four different directions to mark extreme value locations. Second, after acquiring one-dimensional strip data in a particular direction, it selects a one-dimensional sliding window size based on the typical width of internal waves (1km-2km) and performs sliding averaging and de-sloping on the one-dimensional strip data, thus avoiding the uncertainty of local window processing. Third, when acquiring one-dimensional strip data, instead of simply taking the strip width of a pixel, it selects data of a width d and averages them. This processing method further reduces image noise, resulting in better noise immunity. Figure 5 The distribution of extreme points for widths of 1, 11, and 31 are shown. It can be seen that a larger strip width effectively suppresses speckle noise. Fourthly, multi-scale methods, especially super-resolution processing, can detect more extreme points at high resolution, thereby improving the connectivity of internal wave fringes. At low resolution, more reliable extreme points can be detected, thereby reducing the missed detection rate. Figure 6 The following are the analysis results when the window size is 11 at different resolutions. It can be seen that more extreme points may appear at high resolution, while the corresponding extreme points may not be detected at low resolution.

[0066] like Figure 11 As shown, the specific processing steps are:

[0067] S51: Acquire multi-scale one-dimensional strip data of each denoised SAR image block. Specifically, it includes: performing multi-scale upsampling on each denoised SAR image block to obtain multi-scale sampled image blocks; selecting strip data of set width in multiple directions of the sampled image blocks of each scale (such as Figure 4 As shown); the strip data are averaged in the vertical direction to obtain multi-scale one-dimensional strip data.

[0068] The image data blocks are upsampled on the scales [-1 0 1 2], where -1 corresponds to super-resolution interpolation sampling, and 1 and 2 correspond to downsampling. For each sampled image, strips of data with a width of d are selected in the four directions [EW SN EN ES] and the obtained data are averaged in the vertical direction to obtain one-dimensional strip data.

[0069] Taking the EN direction as an example, the detailed description of the processing steps is as follows: 1) Assuming that the size of the image block at the current scale is N*N, starting from the d / 2th row, read the data from the 1st to the dth row in the image to form a data matrix of size d*N. 2) Average each column of the data matrix (i.e., average in the east-west vertical direction) to obtain the 1*N one-dimensional strip data corresponding to the d / 2th row, and then process the data by S52 and S53. The processing method for the other rows of the image block (the d / 2+1th to the Nd / 2th row) is similar. The strip data acquisition and processing methods for other directions (i.e., EW, SN, and ES) are similar. Figure 7 This is an example of reading strip data in the northeast (EN) diagonal direction of the SAR image block after denoising. There are three internal wave fringes below the two lines in the upper right corner. The one-dimensional strip data of the three internal wave fringes are as follows: Figure 8 shown.

[0070] S52: Perform sliding average and de-slope processing on the multi-scale one-dimensional strip data.

[0071] The size of the sliding window is selected according to the internal wave width, and the one-dimensional strip data are subjected to sliding averaging and de-tilting.

[0072] To determine the specific locations of internal waves, the one-dimensional strip data was processed as follows: 1) A larger sliding window (win_size = 51) was selected for the one-dimensional strip data. The first sliding average was calculated using the sliding average method. This result was subtracted from the first sliding average to obtain the one-dimensional data without the trend line, thus achieving de-trending. 2) A smaller sliding window (win_size = 11) was then selected for the one-dimensional data without the trend line, and the second sliding average was calculated using the sliding average method.

[0073] S53: Detecting extreme points of the one-dimensional strip data using a half-height reference line, where the extreme points are candidate positions of the internal wave fringes.

[0074] The half-height reference line method can be used to set the size of the mutation during the evaluation process, thereby ignoring some noise and obtaining the extreme points and their corresponding positions. Figure 9As shown in Figure 2, the corresponding positions of the extreme values (red marked points) are obtained when the mutation grayscale value is set to 2, where the two points on the right can correspond to the internal wave stripes in the image block.

[0075] The position candidate points are represented by a label matrix of the same size as the current scale image block, which is an N*N matrix. Figure 7 The third internal wave stripe in the EN direction ( Figure 7 Taking the longest internal wave stripe in the middle blue stripe as an example, the corresponding position of the extreme point is determined after using the half-height reference line (such as Figure 9 Middle horizontal coordinate x ), according to the coordinates of the row number of the starting point of the current scale image block, and Figure 9 The horizontal coordinates in the matrix are used to obtain the two-dimensional coordinates of the candidate position points. In the N*N marking matrix, the candidate position points are marked as 1 and the others are marked as 0. The element position of the candidate position point in the N*N marking matrix is ( i y - a , a ), the coordinate position in the sampled image block is ( a , i y - a ),in, a is the first i The horizontal coordinate of the strip corresponding to the row, i y is the first i The vertical coordinate of the strip to which the row corresponds. a The calculation formula is as follows.

[0076] .

[0077] in, Indicates taking an integer as the result of the calculation. The four detection directions at the current scale can be represented by a label matrix.

[0078] like Figure 10 As shown in FIG, the method advances row by row, column by column, and diagonally along the direction perpendicular to the strip to complete the detection and marking of extreme points of all one-dimensional strip data at this scale.

[0079] S6: Based on the candidate positions of the internal wave fringes in each denoised SAR image block, the internal wave fringes are segmented and located using a union-find set and an octree.

[0080] The union-find method is computationally efficient in image segmentation. This technique is used in this application to efficiently segment different internal wave fringes. The octree technique can effectively handle inappropriate connections within the detected fringes. The main processing steps are as follows.

[0081] S61: Fusing the candidate position points of the internal wave fringes in each denoised SAR image block to obtain fused candidate position points.

[0082] The internal wave ridge line marking results (internal wave stripe position candidate points) at different scales are fused, and the candidate pixels marked in the multi-scale image are retained as possible marking points in the internal wave ridge line feature stripes.

[0083] S62: According to the eight-neighborhood connectivity rule, the fused candidate position points are segmented using a union-find set to obtain multiple connectivity graphs.

[0084] The fused location candidate points are segmented into different connected graphs using the retrieval and merging processing in the union-find method.

[0085] S63: Removing loop structures from each connected graph based on the octree and performing auxiliary processing to obtain segmented and located internal wave fringes. This specifically includes: representing each connected graph using an octree; determining whether a loop structure exists in the connected graph based on the relationship between child nodes and parent nodes in the octree, and removing the loop structure if present; screening the connected graph after removing the loop structure, and removing isolated points and redundant edges to obtain segmented and located internal wave fringes.

[0086] Specifically, each connected graph is represented using an octree. Based on the relationship between child nodes and parent nodes in the tree and by marking whether they have been visited, possible loop structures in the connected graph are determined. Coarse-grained loop structures are then removed based on the hierarchy. Each connected graph is then retraversed to remove isolated points, redundant edges, and connected graphs containing only short lines. Each internal wave characteristic stripe is numbered and output.

[0087] This application addresses the streak characteristics of internal waves in SAR images and proposes a complete set of methods for accurately locating internal wave streaks. The proposed K-distribution test processing and filtering technology is the first key point of this application. This technology can better adapt to changes in SAR data noise and improve the applicability of the technical method. The proposed overall multi-scale ridge positioning technology is the second key point of this application. This technology avoids the uncertainty of local window processing and can effectively improve noise resistance by increasing the strip width. The proposed internal wave streak segmentation processing technology based on union-find set and octree is the third key point of this application. This technology can efficiently implement the segmentation processing of internal wave streaks.

[0088] According to the above process, multiple sets of ocean internal wave SAR images were processed, including SAR images with larger scale internal waves and images with smaller scale internal waves. Figure 12The results of detecting the internal wave ridge features of a SAR image are given. In this image, the scale of the ocean internal waves is relatively small, and the wavelength of the internal waves is about 200m-500m. Figure 12 (a) is the result of preprocessing the original SAR image. Figure 12 (b) is the internal wave ridge result obtained when the downsampling scale interval parameter is 2. Figure 12 (c) is the result obtained when the scale interval is 4, where the numbers are the detected internal wave ridges that are discontinuous in space. Figure 12 (d) is the result obtained after the detection results at two scales are fused and isolated points and small line segments are automatically removed. It can be seen that the multi-scale processing method better retains the internal wave stripe features that cannot be effectively detected at a single scale due to noise. To illustrate the adaptability of this method, Figure 13-15 The processing results of internal wave SAR images (blocks) at different times and in different areas are given respectively. Figure 14 and Figure 15 Two-scene SAR images have a larger spatial scale. Figure 13 (a) is the result of preprocessing the original SAR image. Figure 13 (b) is the internal wave ridge result obtained when the downsampling scale interval parameter is 4. Figure 13 (c) in the figure is the result obtained when the scale interval is 6. Figure 13 (d) in the figure is the result obtained by fusion of the detection results at two scales and automatically removing isolated points and small line segments. Figure 14 (a) and Figure 15 (a) is the result of preprocessing the original SAR image. Figure 14 (b) and Figure 15 (b) in the figure is the result of internal wave ridge positioning and extraction.

[0089] Based on the same inventive concept, embodiments of the present application also provide a SAR image-based internal wave fringe positioning system. The solution provided by this system is similar to the solution described in the aforementioned method. Therefore, the specific limitations of one or more SAR image-based internal wave fringe positioning system embodiments provided below can be found in the aforementioned limitations of the SAR image-based internal wave fringe positioning method and are not further elaborated here.

[0090] In an exemplary embodiment, an internal wave fringe positioning system based on SAR images is provided, comprising the following modules.

[0091] The block division and preprocessing module is used to divide and preprocess the SAR image into blocks to obtain multiple SAR image blocks.

[0092] The calculation module is used to calculate the shape parameters and scale parameters based on the pixel values of each SAR image block.

[0093] A judgment module is used to judge whether each SAR image block satisfies K distribution based on the shape parameter and the scale parameter.

[0094] The filtering module is used to determine different filtering algorithms based on shape parameters and scale parameters, and perform filtering processing on the SAR image blocks that meet the K distribution to obtain multiple denoised SAR image blocks.

[0095] The initial positioning module is used to process each denoised SAR image block based on a multi-scale ridge positioning algorithm to obtain candidate positions of internal wave fringes in each denoised SAR image block.

[0096] The segmentation and positioning module is used to segment and locate the internal wave fringes based on the position candidate points of the internal wave fringes in each denoised SAR image block using a union-find set and an octree.

[0097] In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in the above-mentioned method embodiments. The computer device can be a server or a terminal, and its internal structure can be as shown in FIG. Figure 16 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O) and a communication interface. The processor, memory and input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store data to be processed. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for internal wave fringe positioning based on SAR images is implemented.

[0098] Those skilled in the art will understand that Figure 16The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present application and does not constitute a limitation on the computer device to which the solution of the present application is applied. A specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement. In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps of the above-mentioned method embodiments when executing the computer program.

[0099] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0100] In an exemplary embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0101] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0102] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments. In particular, any reference to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0103] The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processors involved in the various embodiments provided herein may include, but are not limited to, general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic units, data processing logic units based on quantum computing, and the like.

[0104] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0105] This document uses specific examples to illustrate the principles and implementation methods of this application. The description of the above examples is only intended to help understand the method and core concept of this application. At the same time, for those skilled in the art, based on the concept of this application, there may be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for internal wave fringe positioning based on SAR images, characterized in that: include: Divide and preprocess the SAR image to obtain multiple SAR image blocks; Calculate shape parameters and scale parameters based on the pixel values of each SAR image block; Determining whether each SAR image block satisfies a K distribution based on the shape parameter and the scale parameter; Based on the shape parameters and scale parameters, different filtering algorithms are determined to filter the SAR image blocks that meet the K distribution to obtain multiple denoised SAR image blocks; SAR image blocks that do not meet the K distribution are discarded; Different filtering algorithms are determined based on shape parameters and scale parameters to filter the SAR image blocks that meet the K distribution, including: When the shape parameter is greater than the shape parameter threshold and the scale parameter is greater than the scale parameter threshold, the enhanced Lee filtering algorithm is used to filter the SAR image blocks that meet the K distribution; When the shape parameter is less than the shape parameter threshold and the scale parameter is greater than the scale parameter threshold, the Gamma-MAP filtering algorithm is used to filter the SAR image blocks that meet the K distribution; When the scale parameter is less than the scale parameter threshold, the mean filtering algorithm or the local statistical filtering algorithm is used to filter the SAR image blocks that meet the K distribution; Each denoised SAR image block is processed based on a multi-scale ridgeline positioning algorithm to obtain candidate locations of internal wave fringes in each denoised SAR image block; Based on the candidate positions of internal wave fringes in each denoised SAR image block, the internal wave fringes are segmented and located using union-find set and octree.

2. The internal wave fringe positioning method based on SAR images according to claim 1, characterized in that: The shape and scale parameters are calculated based on the pixel values of each SAR image block, including: Calculate the mean and variance of the pixel values of each SAR image block; calculating a shape parameter based on the mean and the variance; A scale parameter is calculated based on the shape parameter and the mean.

3. The internal wave fringe positioning method based on SAR images according to claim 1, characterized in that: Determining whether each SAR image block satisfies the K distribution based on the shape parameter and the scale parameter specifically includes: generating theoretical distribution data based on the shape parameter and the scale parameter; Whether each SAR image block satisfies the K distribution is determined based on the difference between the theoretical distribution data and the pixel value of each SAR image block.

4. The internal wave fringe positioning method based on SAR images according to claim 1, characterized in that: Each denoised SAR image block is processed based on a multi-scale ridgeline positioning algorithm to obtain candidate locations of internal wave fringes in each denoised SAR image block, specifically including: Obtain multi-scale one-dimensional strip data of each denoised SAR image block; Perform sliding average and de-slope processing on multi-scale one-dimensional strip data; The extreme points of the processed one-dimensional strip data are detected using the half-height reference line, and the detected extreme points are the candidate positions of the internal wave fringes.

5. The internal wave fringe positioning method based on SAR images according to claim 1, characterized in that: Obtain multi-scale one-dimensional strip data for each denoised SAR image block, specifically including: Perform multi-scale upsampling on each denoised SAR image block to obtain a multi-scale sampling image block; Selecting strip data of set width in multiple directions of the sampled image blocks at each scale; The strip data are averaged in the vertical direction to obtain multi-scale one-dimensional strip data.

6. The internal wave fringe positioning method based on SAR images according to claim 1, characterized in that: Based on the candidate positions of the internal wave fringes in each denoised SAR image block, the internal wave fringes are segmented and located using a union-find set and octree. Specifically, the following steps are performed: The candidate positions of the internal wave stripes in each denoised SAR image block are fused to obtain fused candidate positions; According to the eight-neighborhood connectivity rule, the fused location candidate points are segmented using the union-find set to obtain multiple connected graphs; Based on the octree, the ring structure in each connected graph is removed and auxiliary processing is performed to obtain the internal wave stripes after segmentation and positioning.

7. The internal wave fringe positioning method based on SAR images according to claim 1, characterized in that: Based on the octree, the ring structure in each connected graph is removed and auxiliary processing is performed to obtain the internal wave stripes after segmentation and positioning, including: Each connected graph is represented by an octree; Based on the relationship between the child nodes and the parent nodes in the octree, determine whether there is a ring structure in the connected graph, and remove it if there is a ring structure; The connected graph after removing the ring structure is screened, and isolated points and redundant edges are removed to obtain the internal wave stripes after segmentation and positioning.

8. An internal wave fringe positioning system based on SAR images, characterized in that: include: A block division and preprocessing module is used to divide and preprocess the SAR image to obtain multiple SAR image blocks; A calculation module, configured to calculate shape parameters and scale parameters based on pixel values of each SAR image block; a judgment module, configured to judge whether each SAR image block satisfies a K distribution based on the shape parameter and the scale parameter; A filtering module is used to determine different filtering algorithms based on shape parameters and scale parameters, filter the SAR image blocks that meet the K distribution, and obtain multiple denoised SAR image blocks; SAR image blocks that do not meet the K distribution are discarded; Different filtering algorithms are determined based on shape parameters and scale parameters to filter the SAR image blocks that meet the K distribution, including: When the shape parameter is greater than the shape parameter threshold and the scale parameter is greater than the scale parameter threshold, the enhanced Lee filtering algorithm is used to filter the SAR image blocks that meet the K distribution; When the shape parameter is less than the shape parameter threshold and the scale parameter is greater than the scale parameter threshold, the Gamma-MAP filtering algorithm is used to filter the SAR image blocks that meet the K distribution; When the scale parameter is less than the scale parameter threshold, the mean filtering algorithm or the local statistical filtering algorithm is used to filter the SAR image blocks that meet the K distribution; The initial positioning module is used to process each denoised SAR image block based on a multi-scale ridge positioning algorithm to obtain the position candidate points of the internal wave fringes in each denoised SAR image block; The segmentation and positioning module is used to segment and locate the internal wave fringes based on the position candidate points of the internal wave fringes in each denoised SAR image block using a union-find set and an octree.

9. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the internal wave fringe positioning method based on SAR images according to any one of claims 1 to 7.

Images