Lee filtering method and Lee filtering system based on GPU

Abstract

The invention discloses a Lee filtering method and a Lee filtering system based on a GPU, and relates to the technical field of signal processing, the Lee filtering method comprises the following steps: S1, defining a memory pool in the GPU, and dividing the memory pool in a global memory of the GPU; s2, calculating and applying for a memory pool with a matched size according to the size of a sonar image to be filtered; s3, loading image data to be filtered to the memory pool; s4, calculating a variance sum mean value of each sliding window; s5, calculating a weighting coefficient of the sliding window according to the variance and the mean value of the sliding window; s6, calculating a Lee filtering algorithm according to the weighting coefficient of the sliding window so as to realize data filtering of the sliding window; and S7, splicing the data of each sliding window into filtered sonar image data. In this way, the Lee filtering method can be carried out based on the physical characteristics of a GPU platform, design and development calculation efficiency is high, memory consumption is low, and calculation of large-scale data volume is supported.

Description

Technical Field

[0001] The present invention relates to the technical field of signal processing, and in particular to a Lee filtering method and a Lee filtering system based on a GPU. Background Art

[0002] The Lee filter algorithm is an adaptive filtering denoising algorithm based on statistical principles. It is widely used in the fields of signal processing and image analysis because it can effectively reduce the noise in sonar images and improve the quality and readability of sonar images, especially in processing noisy signal remote sensing imaging images. However, as an adaptive filtering algorithm, the three main steps of the Lee filter algorithm: calculation of local statistical characteristics of the sliding window, calculation of filter coefficients, and filtering operations all need to be performed on each pixel in the image. Therefore, when processing high-resolution images and large-scale image data, the computational complexity of the Lee filter algorithm will increase significantly as the number of local images sampled by the sliding window increases, resulting in high computational complexity, large computational memory requirements, and low computational efficiency. Summary of the invention

[0003] An object of the present invention is to provide a GPU-based Lee filtering method and Lee filtering system, wherein the Lee filtering method can be performed based on the physical characteristics of the GPU platform, and the design and development has high computing efficiency, low memory consumption, and supports the calculation of large-scale data volumes.

[0004] An object of the present invention is to provide a GPU-based Lee filtering method and Lee filtering system, wherein the Lee filtering method can be implemented based on the PTX instruction set, and the design and development has high computing efficiency, low memory consumption, and supports the calculation of large-scale data volumes.

[0005] In order to achieve at least one of the above purposes, the technical solution adopted by the present invention is to provide a GPU-based Lee filtering method, wherein the Lee filtering method comprises the following steps: S1, define a memory pool in the GPU and divide the memory pool in the global memory of the GPU; S2, calculates and applies for a memory pool of matching size according to the size of the sonar image to be filtered; S3, loading the image data to be filtered into the memory pool; S4, calculate the variance and mean of each sliding window; S5, calculating the weight coefficient of the sliding window according to the variance and mean of the sliding window; S6, calculating the Lee filtering algorithm according to the weighted coefficient of the sliding window to implement data filtering of the sliding window; S7, splicing the data of each sliding window into filtered sonar image data.

[0006] As a further improvement of the technical solution of the present invention, step S4 includes the following steps: S4.1, divide the data range of each sliding window according to the GPU thread index; S4.2, using the data loading instruction in the PTX instruction set to load each sliding window into each shared memory of the GPU; S4.3, using the combined calculation and summation formula in the PTX instruction set to calculate the mean value in each shared memory; S4.4, using the PTX instruction set, square the difference between the pixel data in the shared memory and the mean to obtain the variance.

[0007] As a further improvement of the technical solution of the present invention, the calculation formula for calculating the weighted coefficient of the sliding window according to the variance and mean of the sliding window in step S5 satisfies: , where the parameters is the total number of pixels in the sliding window, and the parameter is the variance of the sliding window, and the parameter is the mean value of all pixels in the sliding window.

[0008] As a further improvement of the technical solution of the present invention, the calculation formula of the Lee filter algorithm calculated according to the weighted coefficient of the sliding window in step S6 satisfies: , where the parameters It is a sliding window The mean square error of the direction, parameter It is a sliding window The mean square error in the direction.

[0009] As a further improvement of the technical solution of the present invention, in step S4, thread indexes in width and height directions are constructed according to the width and height dimensions of the sonar image to mark each sliding window, and step S7 further includes the steps of: S7.1, judging whether the sliding window belongs to an edge position according to the thread indexes in the width and height directions marked by each sliding window, if the sliding window does not belong to an edge position, partially overlapping the data of adjacent edges of two adjacent sliding windows, taking the average of the overlapped data and loading it into a new shared memory, if the sliding window belongs to an edge position, directly loading the data of the sliding window into the new shared memory; S7.2, according to the thread indexes in the width and height directions marked by each sliding window, the data in the new shared memory are spliced ​​along the width and height directions of the sonar image, and the splicing results are loaded into the memory pool to obtain the filtered sonar image data.

[0010] According to another aspect of the present invention, the present invention further provides a GPU-based Lee filtering system, comprising: A memory pool partitioning unit, used to define a memory pool in the GPU and partition the memory pool in the global memory of the GPU; A memory pool application unit is used to calculate and apply for a memory pool of matching size according to the size of the sonar image to be filtered; A data loading unit, used for loading the image data to be filtered into the memory pool; A first calculation unit, used to calculate the variance and mean of each sliding window; A second calculation unit, used for calculating a weight coefficient of the sliding window according to the variance and mean of the sliding window; A third calculation unit, used for calculating a Lee filter algorithm according to a weighted coefficient of the sliding window, so as to implement data filtering of the sliding window; The splicing unit is used to splice the data of each sliding window into filtered sonar image data.

[0011] As a further improvement of the technical solution of the present invention, the process of the first calculation unit calculating the variance and mean of each sliding window is: Divide the data range of each sliding window according to the GPU thread index; Use the data loading instructions in the PTX instruction set to load each sliding window into each shared memory of the GPU; Use the combined calculation and summation formula in the PTX instruction set to calculate the mean value in each shared memory; The PTX instruction set is used to square the difference between the pixel data in the shared memory and the mean to obtain the variance.

[0012] As a further improvement of the technical solution of the present invention, the calculation formula for calculating the weighted coefficient of the sliding window by the second calculation unit according to the variance and mean of the sliding window satisfies: , where the parameters is the total number of pixels in the sliding window, and the parameter is the variance of the sliding window, and the parameter is the mean value of all pixels in the sliding window.

[0013] As a further improvement of the technical solution of the present invention, the calculation formula of the Lee filtering algorithm calculated by the third calculation unit according to the weighted coefficient of the sliding window satisfies: , where the parameters It is a sliding window The mean square error of the direction, parameter It is a sliding window The mean square error of the direction.

[0014] As a further improvement of the technical solution of the present invention, the stitching process of the stitching unit stitching the data of each sliding window into the filtered sonar image data is: According to the thread index in the width and height directions marked by each sliding window, it is determined whether the sliding window belongs to the edge position. If the sliding window does not belong to the edge position, the data of the adjacent edges of two adjacent sliding windows are partially overlapped, and the overlapped data is averaged and loaded into the new shared memory. If the sliding window belongs to the edge position, the data of the sliding window is directly loaded into the new shared memory; According to the thread index in the width and height directions marked by each sliding window, the data in the new shared memory is spliced ​​along the width and height directions of the sonar image, and the splicing result is loaded into the memory pool to obtain the filtered sonar image data.

[0015] Compared with the prior art, the present invention has at least the following beneficial effects: First, the Lee filtering method of the present invention can be carried out based on the physical characteristics of the GPU platform, and the design and development has high computational efficiency, low memory consumption, and supports the calculation of large-scale data volumes; Second, the Lee filtering method of the present invention can be implemented based on the PTX instruction set, and the design and development has high computational efficiency, low memory consumption, and supports the computation of large-scale data volumes; Third, the Lee filtering method of the present invention can be performed based on the PTX instruction set, has higher versatility, and can achieve higher computing performance on GPUs of different architectures, which is beneficial to improving computing speed and memory utilization, and can also shorten the compilation process, making the compilation speed faster; Fourthly, when stitching image data, it can be done in a shared memory, so that the Lee filtering method of the present invention has a faster filtering speed and a shorter filtering time.

[0016] Other beneficial effects of the Lee filtering method and Lee filtering system of the present invention will be further disclosed in the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flow chart of a GPU-based Lee filtering method according to a preferred embodiment of the present invention.

[0018] Figure 2 It is a block diagram of a GPU-based Lee filtering system according to a preferred embodiment of the present invention.

[0019] Figure 3 It is a block diagram of a computing device according to a preferred embodiment of the present invention.

[0020] In the figure: 200, Lee filter system; 210, memory pool division unit; 220, memory pool application unit; 230, data loading unit; 240, first calculation unit; 250, second calculation unit; 260, third calculation unit; 270, splicing unit; 300, computing device; 310, processor; 320, memory; 330, input device; 340, output device. DETAILED DESCRIPTION

[0021] Before describing in detail any embodiment of the present invention, it should be understood that the present invention is not limited in its application to the construction and arrangement details of the components described below or illustrated in the following figures. The present invention can have other embodiments and can be practiced or carried out in various ways. In addition, it should be understood that the words and terms used here are for descriptive purposes and should not be considered restrictive. The use of "including" or "having" and its variations herein is intended to cover the items and their equivalents and additional items displayed below. Unless otherwise specified or limited, the terms "install", "connect", "support" and "couple" and their variations are widely used and cover direct installation and indirect installation, connection, support and connection. In addition, "connect" and "couple" are not limited to physical or mechanical connections or connections.

[0022] Furthermore, on the first aspect, in the disclosure of the present invention, the orientations or positional relationships indicated by terms such as “longitudinal”, “transverse”, “upper”, “lower”, “front”, “back”, “left”, “right”, “vertical”, “horizontal”, “top”, “bottom”, “inside” and “outside” are based on the orientations or positional relationships shown in the accompanying drawings, which are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore the above terms cannot be understood as limitations on the present invention; on the second aspect, the term “one” should be understood as “at least one” or “one or more”, that is, in one embodiment, the number of an element may be one, while in another embodiment, the number of the element may be multiple, and the term “one” cannot be understood as a limitation on the quantity.

[0023] With reference to the accompanying drawings of the present invention Figure 1 A GPU-based Lee filtering method according to a preferred embodiment of the present invention will be disclosed and explained in the following description, wherein the Lee filtering method is used to filter sonar images.

[0024] Specifically, the Lee filtering method includes the following steps: S1, defining a memory pool in the GPU, and dividing the memory pool in the global memory of the GPU; S2, calculating and applying for a memory pool of matching size according to the size of the sonar image to be filtered; S3, loading the image data to be filtered into the memory pool; S4, calculating the variance and mean of each sliding window; S5, calculating the weighting coefficient of the sliding window according to the variance and mean of the sliding window; S6, calculating the Lee filtering algorithm according to the weighting coefficient of the sliding window to filter the data of the sliding window; S7, splicing the data of each sliding window into the filtered sonar image data. Through such steps, the Lee filtering method of the present invention relies on the physical characteristics of the GPU platform, so that the design and development of the Lee filtering method of the present invention is more efficient, has lower memory consumption, and supports large-scale data volume. For example, in a specific example of the Lee filtering method of the present invention, the amount of data supported by the Lee filtering method can reach 5GB for a single sonar image. In addition, the Lee filtering method of the present invention can significantly reduce the overhead caused by frequent memory application and release during the calculation process by designing a memory pool in the GPU, and can ensure fast access during the calculation process.

[0025] In a specific example of the Lee filtering method of the present invention, in step S2, the Lee filtering method calculates and applies for a memory pool according to twice the size of the sonar image to be filtered to meet the needs of subsequent calculation processes.

[0026] Further, the step S4 includes the following steps: S4.1, dividing the data range of each sliding window according to the GPU thread index; S4.2, using the data loading instruction in the PTX instruction set to load each sliding window into each shared memory of the GPU; S4.3, using the combined calculation and summation formula in the PTX instruction set to respectively calculate the mean in each shared memory; S4.4, using the PTX instruction set to square the difference between the pixel data and the mean in the shared memory to obtain the variance. Through such steps, the Lee filtering method of the present invention is implemented by relying on the PTX instruction set, and the design and development has high computing efficiency, low memory consumption, and supports the calculation of large-scale data volumes, which can reach a single sonar image of 5GB. In addition, by relying on the PTX instruction set to implement, the Lee filtering method of the present invention has higher versatility, can play a higher computing performance on GPUs of different architectures, and is conducive to improving the computing speed and memory utilization. Moreover, by relying on the PTX instruction set to implement, the Lee filtering method of the present invention has a shorter compilation process and a faster compilation speed.

[0027] In a specific example of the Lee filtering method of the present invention, the calculation formula for calculating the weighted coefficient of the sliding window according to the variance and mean of the sliding window in step S5 satisfies: , where the parameters is the total number of pixels in the sliding window, and the parameter is the variance of the sliding window, and the parameter is the mean value of all pixels in the sliding window.

[0028] In a specific example of the Lee filtering method of the present invention, the calculation formula of the Lee filtering algorithm calculated according to the weighted coefficient of the sliding window in step S6 satisfies: , where the parameters It is a sliding window The mean square error of the direction, parameter It is a sliding window The mean square error of the direction.

[0029] Further, in step 4, a thread index in the width and height direction is constructed according to the size of the sonar image in the width and height direction, which is used to mark each sliding window. In step S7, the specific process of splicing the data of each sliding window into the filtered sonar image data is as follows: first, according to the thread index in the width and height direction marked by each sliding window, it is determined whether the sliding window belongs to the edge position; if the sliding window does not belong to the edge position, the data of the adjacent edges of two adjacent sliding windows are partially overlapped, and the overlapped data is loaded into a new shared memory after taking the average; if the sliding window belongs to the edge position, the data of the sliding window is directly loaded into the new shared memory; secondly, according to the thread index in the width and height direction marked by each sliding window, the data of the new shared memory is spliced ​​along the width and height direction of the sonar image, and the splicing result is loaded into the memory pool to obtain the filtered sonar image data.

[0030] That is to say, the step of splicing the data of each sliding window into the filtered sonar image data executed in step S7 is performed in the shared memory, which makes the Lee filtering method of the present invention faster and takes less time. In addition, since the Lee filtering method is spliced ​​by thread index marking, large-scale data can be spliced ​​at the same time, so the Lee filtering method of the present invention can filter higher resolution sonar images. At the same time, the step of splicing the data of each sliding window into the filtered sonar image data executed in step S7 is implemented by the PTX instruction set directly operating the shared memory, so the compilation time is shorter.

[0031] Attached Figure 2A block diagram of a GPU-based Lee filtering system 200 of a preferred embodiment of the present invention is shown, wherein the Lee filtering system 200 includes a memory pool partitioning unit 210, a memory pool application unit 220, a data loading unit 230, a first computing unit 240, a second computing unit 250, a third computing unit 260 and a splicing unit 270.

[0032] The memory pool division unit 210 is used to define a memory pool in the GPU and divide the memory pool in the global memory of the GPU; the memory pool application unit 220 is used to calculate and apply for a memory pool of matching size according to the size of the sonar image to be filtered; the data loading unit 230 is used to load the image data to be filtered into the memory pool; the first calculation unit 240 is used to calculate the variance and mean of each sliding window; the second calculation unit 250 is used to calculate the weighting coefficient of the sliding window according to the variance and mean of the sliding window; the third calculation unit 260 is used to calculate the Lee filtering algorithm according to the weighting coefficient of the sliding window to achieve data filtering of the sliding window; the splicing unit 270 is used to splice the data of each sliding window into the filtered sonar image data. In this way, the Lee filtering system 200 of the present invention relies on the physical characteristics of the GPU platform, so that the design and development of the Lee filtering system 200 of the present invention has higher calculation efficiency, lower memory consumption, and supports large-scale data volume, which can reach 5GB for a single sonar image.

[0033] In a specific example of the Lee filter system 200 of the present invention, the process of the first calculation unit 240 calculating the variance and mean of each sliding window is: dividing the data range of each sliding window according to the GPU thread index; using the data loading instruction in the PTX instruction set to load each sliding window into each shared memory of the GPU; using the combined calculation and summation formula in the PTX instruction set to respectively calculate the mean in each shared memory; using the PTX instruction set to square the difference between the pixel data and the mean in the shared memory to obtain the variance. In this way, the Lee filter system 200 of the present invention is implemented by relying on the PTX instruction set, and the design and development has high calculation efficiency, low memory consumption, and supports the calculation of large-scale data volumes, which can reach a single sonar image of 5GB. In addition, by relying on the PTX instruction set to implement, the Lee filter system 200 of the present invention has higher versatility, can play a higher computing performance on GPUs of different architectures, and is conducive to improving the computing speed and memory utilization. And, by relying on the PTX instruction set to implement, the Lee filter system 200 of the present invention has a shorter compilation process and a faster compilation speed.

[0034] In a specific example of the Lee filter system 200 of the present invention, the stitching process of the stitching unit 270 stitching the data of each sliding window into the filtered sonar image data is: according to the thread index in the width and height directions marked by each sliding window, determine whether the sliding window belongs to the edge position; if the sliding window does not belong to the edge position, locally overlap the data of adjacent edges of two adjacent sliding windows, and take the average of the overlapped data and load it into a new shared memory; if the sliding window belongs to the edge position, directly load the data of the sliding window into the new shared memory; according to the thread index in the width and height directions marked by each sliding window, stitch the data of the new shared memory along the width and height directions of the sonar image, and load the stitching result into the memory pool to obtain the filtered sonar image data.

[0035] That is, the splicing unit 270 splices data in the shared memory, which makes the Lee filter system 200 of the present invention faster and takes less time. In addition, since the Lee filter system 200 is spliced ​​by thread index marking, large-scale data can be spliced ​​at the same time, so the Lee filter system 200 of the present invention can filter higher-resolution sonar images. At the same time, the splicing unit 270 is implemented by directly operating the shared memory using the PTX instruction set, so the compilation time is shorter.

[0036] like Figure 3 As shown, a computing device 300 according to a preferred embodiment of the present invention includes one or more processors 310 and one or more memories 320 connected to the processors 310 .

[0037] The processor 310 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the computing device 300 to perform desired functions.

[0038] The memory 320 may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may include, for example, a read-only memory (ROM), a hard disk, a flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 310 may execute the program instructions to implement the optimization method of the present invention described above and / or other desired functions.

[0039] Specifically, when the processor 310 executes the program instructions stored in the memory 320, the computing device 300 executes the Lee filtering method, which includes the following steps: S1, defining a memory pool in the GPU and dividing the memory pool in the global memory of the GPU; S2, calculating and applying for a memory pool of matching size according to the size of the sonar image to be filtered; S3, loading the image data to be filtered into the memory pool; S4, calculating the variance and mean of each sliding window; S5, calculating the weighting coefficient of the sliding window according to the variance and mean of the sliding window; S6, calculating the Lee filtering algorithm according to the weighting coefficient of the sliding window to filter the data of the sliding window; S7, splicing the data of each sliding window into the filtered sonar image data. Through such steps, the computing device 300 of the present invention relies on the physical characteristics of the GPU platform, so that the design and development of the computing device 300 of the present invention has higher computing efficiency, lower memory consumption, and supports large-scale data volume. For example, in a specific example of the computing device 300 of the present invention, the amount of data supported by the computing device 300 can reach 5 GB for a single sonar image.

[0040] In a specific example of the computing device 300 of the present invention, the computing device 300 may also include an input device 330 and an output device 340, and these components are interconnected through a bus system and / or other forms of connection mechanisms (not shown). The input device 330 may include, for example, a keyboard, a mouse, etc. The output device 340 may output various information to the outside. The output device 340 may include, for example, a display, a speaker, a printer, a communication network and a remote output device connected thereto, etc.

[0041] It should be understood by those skilled in the art that the embodiments of the present invention described above and shown in the accompanying drawings are only examples and do not limit the present invention. The purpose of the present invention has been fully and effectively achieved. The functional and structural principles of the present invention have been demonstrated and explained in the embodiments, and the embodiments of the present invention may be modified or altered in any way without departing from the principles.

Claims

1. The GPU-based Lee filtering method is characterized by: The Lee filtering method comprises the following steps: S1, define a memory pool in the GPU and divide the memory pool in the global memory of the GPU; S2, calculates and applies for a memory pool of matching size according to the size of the sonar image to be filtered; S3, loading the image data to be filtered into the memory pool; S4, calculate the variance and mean of each sliding window; S5, calculating the weight coefficient of the sliding window according to the variance and mean of the sliding window; S6, calculating the Lee filtering algorithm according to the weighted coefficient of the sliding window to implement data filtering of the sliding window; S7, splicing the data of each sliding window into filtered sonar image data.

2. The Lee filtering method according to claim 1, characterized in that: The step S4 comprises the following steps: S4.1, divide the data range of each sliding window according to the GPU thread index; S4.2, using the data loading instruction in the PTX instruction set to load each sliding window into each shared memory of the GPU; S4.3, using the combined calculation and summation formula in the PTX instruction set to calculate the mean value in each shared memory; S4.4, using the PTX instruction set, square the difference between the pixel data in the shared memory and the mean to obtain the variance.

3. The Lee filtering method according to claim 2, characterized in that: The calculation formula for calculating the weighted coefficient of the sliding window according to the variance and mean of the sliding window in step S5 satisfies: , where the parameters is the total number of pixels in the sliding window, and the parameter is the variance of the sliding window, and the parameter is the mean value of all pixels in the sliding window.

4. The Lee filtering method according to claim 3, characterized in that: The calculation formula of the Lee filter algorithm calculated according to the weighted coefficient of the sliding window in step S6 satisfies: , where the parameters It is a sliding window The mean square error of the direction, parameter It is a sliding window The mean square error in the direction.

5. The Lee filtering method according to claim 4, characterized in that: In the step S4, thread indexes in the width and height directions are constructed according to the sizes of the sonar image in the width and height directions, for marking each sliding window, and the step S7 further includes the steps of: S7.1, judging whether the sliding window belongs to an edge position according to the thread indexes in the width and height directions marked by each sliding window, if the sliding window does not belong to an edge position, partially overlapping the data of adjacent edges of two adjacent sliding windows, taking the average of the overlapped data and loading it into a new shared memory, if the sliding window belongs to an edge position, directly loading the data of the sliding window into the new shared memory; S7.2, according to the thread indexes in the width and height directions marked by each sliding window, the data in the new shared memory are spliced ​​along the width and height directions of the sonar image, and the splicing results are loaded into the memory pool to obtain the filtered sonar image data.

6. GPU-based Lee filtering system, characterized by: include: A memory pool partitioning unit, used to define a memory pool in the GPU and partition the memory pool in the global memory of the GPU; A memory pool application unit is used to calculate and apply for a memory pool of matching size according to the size of the sonar image to be filtered; A data loading unit, used for loading the image data to be filtered into the memory pool; A first calculation unit, used to calculate the variance and mean of each sliding window; A second calculation unit, used for calculating a weight coefficient of the sliding window according to the variance and mean of the sliding window; A third calculation unit, used for calculating a Lee filter algorithm according to a weighted coefficient of the sliding window, so as to implement data filtering of the sliding window; The splicing unit is used to splice the data of each sliding window into filtered sonar image data.

7. The Lee filter system according to claim 6, characterized in that: The process of the first calculation unit calculating the variance and mean of each sliding window is: Divide the data range of each sliding window according to the GPU thread index; Use the data loading instructions in the PTX instruction set to load each sliding window into each shared memory of the GPU; Use the combined calculation and summation formula in the PTX instruction set to calculate the mean value in each shared memory; The PTX instruction set is used to square the difference between the pixel data in the shared memory and the mean to obtain the variance.

8. The Lee filter system according to claim 7, characterized in that: The calculation formula for calculating the weighted coefficient of the sliding window by the second calculation unit according to the variance and mean of the sliding window satisfies: , where the parameters is the total number of pixels in the sliding window, and the parameter is the variance of the sliding window, and the parameter is the mean value of all pixels in the sliding window.

9. The Lee filter system according to claim 8, characterized in that: The calculation formula of the Lee filter algorithm calculated by the third calculation unit according to the weighted coefficient of the sliding window satisfies: , where the parameters It is a sliding window The mean square error of the direction, parameter It is a sliding window The mean square error in the direction.

10. The Lee filter system according to claim 9, characterized in that: The stitching process of the stitching unit stitching the data of each sliding window into the filtered sonar image data is: According to the thread index in the width and height directions marked by each sliding window, it is determined whether the sliding window belongs to the edge position. If the sliding window does not belong to the edge position, the data of the adjacent edges of two adjacent sliding windows are partially overlapped, and the overlapped data is averaged and loaded into the new shared memory. If the sliding window belongs to the edge position, the data of the sliding window is directly loaded into the new shared memory; According to the thread index in the width and height directions marked by each sliding window, the data in the new shared memory is spliced ​​along the width and height directions of the sonar image, and the splicing result is loaded into the memory pool to obtain the filtered sonar image data.

Images