Marine radar image denoising method and device based on multistage filtering algorithm

Abstract

The invention provides a marine radar image denoising method and device based on a multistage filtering algorithm, and the method comprises the steps: carrying out the preprocessing of a to-be-processed marine radar image, and obtaining a first image; performing edge gradient detection on the first image based on the gradient direction corresponding to each pixel point in the first image and the target identified in the first image, and performing noise detection on the first image based on the gray value corresponding to each pixel point in the first image after the edge gradient detection is passed; after the noise detection is passed, processing the first image by adopting bilateral filtering to obtain a second image; and image enhancement is carried out after image sharpening is carried out on the second image, and a denoised marine radar image is obtained. According to the method, multiple filtering technologies such as Gaussian filtering, median filtering and bilateral filtering are combined, multiple kinds of noise in the marine radar image are effectively removed, especially image edge and target information can be kept in a complex environment, and the target detection and navigation precision is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and in particular to a method and device for denoising a marine radar image based on a multi-stage filtering algorithm. Background Art

[0002] Marine radar images are often affected by various noises, including sea surface noise, weather interference, co-frequency interference, speckle noise, etc. These noises not only affect the image quality, causing image blur or distortion, but also seriously interfere with radar target detection and tracking, thus affecting the navigation safety and collision avoidance accuracy of ships.

[0003] Existing marine radar image denoising technologies mostly rely on hardware solutions, such as digital signal processors (DSP) or field programmable gate arrays (FPGA) and other devices for real-time image processing. Although hardware solutions can provide a certain degree of noise filtering effect, their design and debugging are complex, with poor flexibility and difficulty in adapting to changing noise environments. Moreover, hardware solutions can usually only process specific types of noise, and it is difficult to provide comprehensive and effective denoising effects for complex noise types (such as sea surface noise and weather interference). In addition, hardware filtering requires high costs and is not suitable for marine radar systems with frequent environmental changes.

[0004] Compared with hardware filtering, software-based image filtering methods have higher flexibility and customizability. By using software-implemented filtering algorithms, parameters can be flexibly adjusted according to the characteristics of different noises to improve the denoising effect and avoid high hardware costs. However, existing software filtering methods usually rely on a single filtering technology (such as Gaussian filtering or median filtering, etc.), and it is difficult to balance the needs of denoising and detail preservation when faced with a variety of complex noise types in marine radar images. Therefore, the existing single filtering method often cannot effectively handle the various noises in radar images, and it is easy to cause image details to be lost or target information to be unclear. Summary of the invention

[0005] The object of the present invention is to provide a method and device for denoising a marine radar image based on a multi-stage filtering algorithm to improve the above-mentioned problem.

[0006] In order to achieve the above objectives, the present application provides the following technical solutions:

[0007] On the one hand, an embodiment of the present application provides a method for denoising a marine radar image based on a multi-stage filtering algorithm, the method comprising:

[0008] Acquire the ship radar image to be processed;

[0009] The to-be-processed marine radar image is preprocessed to obtain a first image; edge gradient detection is performed on the first image based on the gradient direction corresponding to each pixel point in the first image and the target identified in the first image; after the edge gradient detection passes, noise detection is performed on the first image based on the grayscale value corresponding to each pixel point in the first image; after the noise detection passes, the first image is processed using bilateral filtering to obtain a second image; image sharpening is performed on the second image and then image enhancement is performed to obtain a denoised marine radar image.

[0010] In a second aspect, an embodiment of the present application provides a marine radar image denoising device based on a multi-stage filtering algorithm, the device comprising:

[0011] An acquisition module, used for acquiring the marine radar image to be processed;

[0012] The denoising module is used to preprocess the to-be-processed marine radar image to obtain a first image; perform edge gradient detection on the first image based on the gradient direction corresponding to each pixel point in the first image and the target identified in the first image, and perform noise detection on the first image based on the grayscale value corresponding to each pixel point in the first image after the edge gradient detection passes; use bilateral filtering to process the first image after the noise detection passes to obtain a second image; perform image sharpening and then image enhancement on the second image to obtain a denoised marine radar image.

[0013] In a third aspect, an embodiment of the present application provides a marine radar image denoising device based on a multi-stage filtering algorithm, the device comprising a memory and a processor. The memory is used to store a computer program; the processor is used to implement the steps of the marine radar image denoising method based on the multi-stage filtering algorithm when executing the computer program.

[0014] In a fourth aspect, an embodiment of the present application provides a readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the above-mentioned method for denoising a marine radar image based on a multi-stage filtering algorithm are implemented.

[0015] The beneficial effects of the present invention are:

[0016] 1. The present invention combines Gaussian filtering, median filtering and bilateral filtering to effectively remove various noises in marine radar images, especially in complex environments, it can maintain image edges and target information, and significantly improve target detection and navigation accuracy.

[0017] 2. The present invention can gradually remove different types of noise while retaining edge details and target information in the image through a combination of multiple filtering techniques. The algorithm optimizes the quality of radar images through the following filtering steps: Gaussian filtering: smoothing the image, removing high-frequency noise, and retaining relatively clear edge information. Median filtering: removing speckle noise and co-frequency interference, and further smoothing the image without losing edge details. Bilateral filtering: combining spatial weights and grayscale difference weights to enhance the solidity of the target area while retaining edge details. The multi-stage filtering algorithm of the present invention can effectively cope with a variety of complex noises, optimize radar image quality, and improve the accuracy of target detection. Compared with traditional single filtering methods, this algorithm has higher noise removal and detail retention capabilities, and is particularly suitable for complex noise environments commonly found in marine radar images.

[0018] 3. The present invention adopts a multi-level filtering framework to gradually remove various noises through a combination of different filtering methods, and effectively retains the details and edge information of the image during the denoising process; the present invention processes different noise types in steps, and adopts different strategies such as Gaussian filtering, median filtering, bilateral filtering, etc. for different types of noise to ensure that the image processing effect is maximized; the present invention dynamically adjusts the filtering parameters, and automatically adjusts the filtering parameters according to the noise intensity and the characteristics of the target area to achieve the best denoising effect.

[0019] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or be understood by implementing the embodiments of the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0021] Figure 1 It is a schematic flow chart of a method for denoising a marine radar image based on a multi-stage filtering algorithm described in an embodiment of the present invention;

[0022] Figure 2 Schematic diagram of the structure of a marine radar image denoising device based on a multi-stage filtering algorithm described in an embodiment of the present invention;

[0023] Figure 3 It is a schematic diagram of the structure of a marine radar image denoising device based on a multi-stage filtering algorithm described in an embodiment of the present invention. DETAILED DESCRIPTION

[0024] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings herein can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0025] It should be noted that similar reference numerals or letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings. At the same time, in the description of the present invention, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0026] Example 1

[0027] like Figure 1 As shown, this embodiment provides a method for denoising a marine radar image based on a multi-stage filtering algorithm, and the method includes step S1 and step S2.

[0028] Step S1, obtaining a ship radar image to be processed;

[0029] Step S2, preprocessing the to-be-processed marine radar image to obtain a first image; performing edge gradient detection on the first image based on the gradient direction corresponding to each pixel point in the first image and the target identified in the first image, and performing noise detection on the first image based on the grayscale value corresponding to each pixel point in the first image after the edge gradient detection passes; processing the first image by using bilateral filtering after the noise detection passes to obtain a second image; performing image sharpening and then image enhancement on the second image to obtain a denoised marine radar image.

[0030] In step S2, the to-be-processed marine radar image is preprocessed to obtain a first image, and the specific implementation steps include step S21;

[0031] Step S21, graying the to-be-processed marine radar image to obtain a gray image, and normalizing the gray image to obtain a first image.

[0032] This step is specifically as follows: normalizing the grayscale image, scaling the grayscale image pixel values ​​to a range of 0 to 255, and obtaining a first image. Before multi-level filtering, the marine radar image is first preprocessed to ensure the stability and accuracy of the image in the subsequent processing stage. The purpose of converting to a grayscale image is to simplify processing; scaling the pixel values ​​to a range of 0 to 255 is to ensure consistency of processing.

[0033] In step S2, based on the gradient direction corresponding to each pixel point in the first image and the target identified in the first image, the specific implementation steps of performing edge gradient detection on the first image include steps S22 and S23;

[0034] Step S22, using the Sobel operator to calculate the gradient intensity corresponding to each pixel in the first image; comparing the gradient intensity corresponding to all the pixels with a preset gradient intensity threshold, screening out the pixels with a gradient intensity greater than the threshold, and recording them as first pixels; calculating the gradient direction corresponding to each pixel, forming a square frame of a preset size with each first pixel as the center, and taking the standard deviation of the gradient directions of all the pixels contained in each square frame as the standard deviation corresponding to each first pixel; taking the first pixel whose standard deviation is greater than the standard deviation threshold as a rough point;

[0035] In this step, the gradient direction corresponding to each pixel is calculated as follows:

[0036] (1) The Sobel operator is used to calculate the gradients of the pixels in the first image in the horizontal and vertical directions, specifically:

[0037] G x (x, y) = I (x + 1, y) - I (x - 1, y) (horizontal gradient)

[0038] G y (x, y) = I (x, y + 1) - I (x, y - 1) (vertical gradient)

[0039] In the formula, I(x+1, y) is the grayscale value of the first image at the pixel point (x+1, y), I(x-1, y) is the grayscale value of the first image at the pixel point (x-1, y), I(x, y+1) is the grayscale value of the first image at the pixel point (x, y+1), and I(x, y-1) is the grayscale value of the first image at the pixel point (x, y-1); Gx(x, y) is the horizontal gradient of the first image at the pixel point (x, y), and Gy(x, y) is the vertical gradient of the first image at the pixel point (x, y);

[0040] When calculating the gradient, the edge pixels of the image may cross the boundary. Therefore, the edge pixels need to be processed separately. The processing method can be:

[0041] Horizontal gradient: The horizontal gradient of the pixel on the left boundary is replaced by the horizontal gradient of its right neighboring pixel, and the horizontal gradient of the pixel on the right boundary is replaced by the horizontal gradient of its left neighboring pixel.

[0042] Vertical gradient: The horizontal gradient of the pixel on the upper boundary is replaced by the horizontal gradient of its lower neighboring pixel, and the horizontal gradient of the pixel on the lower right boundary is replaced by the horizontal gradient of its upper neighboring pixel.

[0043] (2) Calculate the gradient strength:

[0044]

[0045] G(x,y) is the intensity of the first image at the pixel (x,y);

[0046] (3) Calculate the gradient direction:

[0047]

[0048] θ(x,y) is the gradient direction of the first image at the pixel point (x,y), in radians, indicating the direction of the edge.

[0049] In addition, in this step, the preset size of the square frame is 3*3 or 5*5; that is, the side length of the square frame is 3 pixels or 5 pixels;

[0050] Step S23, using OpenCV to identify the contours of the target in the first image and the coordinates of the pixel points contained in each contour; counting the number of pixel points contained in each contour and recording it as a first value, counting the number of rough points contained in each contour according to the coordinates of the pixel points contained in each contour and recording it as a second value, and taking the ratio of the second value to the first value as the roughness ratio of each target; calculating the mean of the roughness ratio of each target to obtain the overall roughness ratio, comparing the overall roughness ratio with a preset overall roughness ratio threshold, when the overall roughness ratio is greater than the preset overall roughness ratio threshold, determining that the detection result is not smooth, and the edge gradient detection fails, and using Gaussian filtering to process the first image, and then performing edge gradient detection after processing; otherwise, determining that the detection result is smooth, and the edge gradient detection passes.

[0051] In this step, the connected component labeling (CCL) algorithm in OpenCV can be used to obtain the contour of the target in the first image and the coordinates of the pixels contained in each contour. This is a conventional method, and this step will not be described in detail. At the same time, in this step, if the detection result is not smooth, that is, the roughness is obvious (the gradient direction jumps too much or too large), it means that the image noise is high or the edge is not smooth. At this time, Gaussian filtering can be applied to smooth the image by setting a suitable filter kernel size (such as 5x5 or 7x7). Gaussian filtering performs weighted averaging of the pixels around each pixel through a Gaussian function to remove high-frequency noise in the image.

[0052] In step S2, the specific implementation steps of performing noise detection on the first image based on the grayscale value corresponding to each pixel in the first image include steps S24 and S25;

[0053] Step S24, selecting a local window with each pixel as the center on the first image, and taking the mean of the grayscale values ​​of all the pixels in the window as the local mean corresponding to each pixel; calculating the local variance corresponding to each pixel based on the local mean corresponding to each pixel, the grayscale values ​​of all the pixels in the window and the size of the window, and calculating the local standard deviation based on the local variance;

[0054] The specific steps of this step are:

[0055] (1) Local mean calculation:

[0056]

[0057] In the formula, I(x+i,y+i) is the grayscale value of the first image at the pixel point (x+i,y+i); w is the window size, which can be 3×3 or 5×5, for example; μ x,y is the local mean corresponding to the central pixel (x, y).

[0058] (2) Local variance calculation: The local variance measures the fluctuation of brightness around each pixel. A larger variance usually means that there is a large grayscale change in the local area, which may be a manifestation of noise.

[0059]

[0060] In the formula, is the local variance corresponding to the central pixel (x, y); I(x+i, y+j) is the gray value of the first image at the pixel (x+i, y+j); μ x,y is the local mean corresponding to the central pixel (x, y).

[0061] Step S25, multiply the local standard deviation corresponding to each pixel by a preset coefficient to obtain a local standard deviation threshold, compare the absolute value of the difference between the gray value corresponding to each pixel and the corresponding local mean with the local standard deviation threshold, if the absolute value is greater than the local standard deviation threshold, determine that the pixel is a noise point; at the same time, compare the local variance corresponding to each pixel with the preset local variance threshold, if the local variance is greater than the local variance threshold, determine that the pixel is a noise point; based on the number of noise points and the local variance, determine whether the first image passes the noise detection.

[0062] The preset coefficient in this step is a parameter to control the sensitivity, usually between 1 and 2; this step can determine whether each pixel is likely to be a noise point by calculating the local mean and variance. Noise points usually show large differences from surrounding pixels, or have a high variance in their local area. Therefore, based on these statistical characteristics, some thresholds can be set to determine noise points.

[0063] In step S25, potential noise points are identified by calculating the local mean and variance around each pixel. Speckle noise usually shows large grayscale fluctuations in a local area, while real image details usually change more slowly. Therefore, through this method, we can detect pixels with abnormal changes, that is, possible noise points.

[0064] Meanwhile, the specific implementation steps of judging whether the first image passes the noise detection based on the number of noise points and the local variance in step S25 include step S251 and step S252;

[0065] Step S251, calculating the mean of the local variances corresponding to all the pixels in the first image to obtain the global variance mean; counting the number of noise points, and taking the ratio between the number of noise points and the total number of pixels in the first image as the noise point ratio;

[0066] Step S252, setting conditions, condition one is comparing the global variance mean with a preset global variance mean threshold, and the global variance mean is greater than the global variance mean threshold; condition two is comparing the noise point ratio with a preset noise point ratio threshold, and the noise point ratio is greater than the noise point ratio threshold; when any one of the conditions is met, the detection result is determined to be the presence of noise, the noise detection fails, and the first image is processed by median filtering, and noise detection is performed after processing; otherwise, the detection result is determined to be the absence of noise, and the noise detection passes.

[0067] In this step, the noise point ratio threshold can be 5% or 10%; at the same time, under the interference of complex weather or waves, speckle noise will be generated, and after median filtering, the speckle noise in the image can be effectively removed.

[0068] In step S2, the specific implementation steps of using bilateral filtering to process the first image include step S26;

[0069] Step S26, determining key parameters in bilateral filtering, the key parameters including neighborhood size, color standard deviation and space standard deviation; after determination, using bilateral filtering to process the first image to obtain a second image.

[0070] In this step, the key parameters in bilateral filtering are determined. The key parameters include neighborhood size, color standard deviation and spatial standard deviation. The specific implementation steps include step S261;

[0071] Step S261, using OpenCV to identify the contour of the target in the first image after noise detection and the coordinates of the pixels contained in each contour, based on the contour of the target in the first image after noise detection and the coordinates of the pixels contained in each contour, record the area formed by each contour as the target area, record the number of pixels contained in each contour as the area of ​​the target area, calculate the standard deviation of the grayscale values ​​of the pixels contained in each contour and record the standard deviation calculation result as the grayscale standard deviation of the target area; the calculation formula of the neighborhood size d is:

[0072]

[0073]

[0074] Among them, K is the adjustment coefficient, which is 0.1; N is the number of target areas; Area i is the area of ​​target region i;

[0075] Color standard deviation σ C The calculation formula is:

[0076]

[0077]

[0078] Among them, α is the coefficient, which is 1; β is the offset, which is 10; is the average value of the grayscale standard deviation of N target areas; σ i is the grayscale standard deviation of the i-th target area;

[0079] Spatial standard deviation σ S The calculation formula is:

[0080]

[0081] In this step, the method of using OpenCV for recognition is the same as that in step S23;

[0082] In step S2, after completing the multi-stage filtering, the last step is to post-process and optimize the image to improve the image quality:

[0083] Image sharpening: A sharpening filter (Laplacian filter) can be used to slightly enhance the image and enhance the details and edges in the image.

[0084] Image enhancement: Contrast enhancement (Histogram Equalization) can be used later to further improve the visual quality of the image.

[0085] This embodiment has the following advantages:

[0086] 1. Multi-level filtering framework:

[0087] Existing solutions: Most existing radar image denoising technologies rely on a single filtering method, such as Gaussian filtering, median filtering, etc. These methods are usually optimized for a certain type of noise, cannot handle multiple types of noise at the same time, and often lose image details, especially edge and target information, while denoising.

[0088] The present invention proposes a multi-level filtering framework, which combines three filtering techniques, namely Gaussian filtering, median filtering and bilateral filtering, to remove various noises in the image step by step. Each filtering technique is optimized for a specific type of noise and can effectively retain the edge details and target information of the image during the denoising process. This multi-level processing scheme can more comprehensively and finely handle the noise problem in radar images, avoiding problems such as insufficient noise removal or image blur in traditional methods.

[0089] 2. Step-by-step processing for different noise types:

[0090] Existing solutions: Although traditional hardware filtering solutions (such as those based on digital signal processors (DSPs) or field programmable gate arrays (FPGAs)) can provide efficient real-time processing, they lack flexibility and are difficult to quickly adapt to different types of noise changes. In addition, their hardware design, debugging and upgrade processes are complex and costly.

[0091] Technology of the Invention: The multi-stage filtering method of this patent has strong flexibility and adjustability. Users can dynamically adjust the filtering algorithm and parameters according to different noise environments and actual needs. This enables the technology to efficiently denoise in various complex marine environments, greatly improving the adaptability and flexibility of the algorithm.

[0092] 3. Step-by-step processing for different noise types:

[0093] Existing solutions: A single filtering method (such as Gaussian filtering) can usually only handle one type of noise. However, for complex marine radar images, the types and properties of noise are diverse and intertwined, and a single method is difficult to achieve the best effect.

[0094] The present invention adopts different filtering strategies for different noise types through step-by-step processing. For example:

[0095] Gaussian filtering: removes high-frequency noise and smoothes the image, retaining clearer edges.

[0096] Median filtering: removes speckle noise and co-frequency interference while avoiding excessive image blur.

[0097] Bilateral filtering: By combining spatial weights and grayscale difference weights, the solidity of the target area is enhanced, especially the smoothing of the darker areas inside the target in the image, avoiding excessive smoothing of the background area while maintaining the edges.

[0098] 4. Dynamically adjust parameters:

[0099] Existing solutions: Existing filtering methods usually use fixed filtering parameters and cannot be adjusted in real time according to the specific noise conditions of the image. This results in poor results in environments with large noise changes.

[0100] The present invention optimizes the filtering process in real time by dynamically adjusting the filtering parameters (such as the spatial standard deviation and color standard deviation in bilateral filtering) according to the standard deviation of the target area and the different characteristics of the noise. This technology can ensure that the filtering effect can be optimized in different scenarios and avoid the problems of over-smoothing or insufficient denoising.

[0101] Example 2

[0102] like Figure 2 As shown, this embodiment provides a marine radar image denoising device based on a multi-stage filtering algorithm, and the device includes an acquisition module 1 and a denoising module 2.

[0103] An acquisition module 1 is used to acquire a ship radar image to be processed;

[0104] De-noising module 2 is used to pre-process the to-be-processed marine radar image to obtain a first image; perform edge gradient detection on the first image based on the gradient direction corresponding to each pixel point in the first image and the target identified in the first image, and perform noise detection on the first image based on the grayscale value corresponding to each pixel point in the first image after the edge gradient detection passes; use bilateral filtering to process the first image after the noise detection passes to obtain a second image; perform image sharpening and then image enhancement on the second image to obtain a denoised marine radar image.

[0105] It should be noted that, regarding the device in the above embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment of the method, and will not be elaborated here.

[0106] Example 3

[0107] Corresponding to the above method embodiments, the embodiments of the present disclosure also provide a ship radar image denoising device based on a multi-stage filtering algorithm. The ship radar image denoising device based on a multi-stage filtering algorithm described below and the ship radar image denoising method based on a multi-stage filtering algorithm described above can be referred to each other.

[0108] Figure 3 FIG. 3 is a block diagram of a ship radar image denoising device 300 based on a multi-stage filtering algorithm according to an exemplary embodiment. Figure 3 As shown, the ship radar image denoising device 300 based on the multi-stage filtering algorithm may include: a processor 301, a memory 302. The ship radar image denoising device 300 based on the multi-stage filtering algorithm may also include one or more of a multimedia component 303, an I / O interface 304, and a communication component 305.

[0109] The processor 301 is used to control the overall operation of the marine radar image denoising device 300 based on the multi-stage filtering algorithm to complete all or part of the steps in the marine radar image denoising method based on the multi-stage filtering algorithm. The memory 302 is used to store various types of data to support the operation of the marine radar image denoising device 300 based on the multi-stage filtering algorithm, and these data may include, for example, instructions for any application or method operating on the marine radar image denoising device 300 based on the multi-stage filtering algorithm, and application-related data, such as contact data, sent and received messages, pictures, audio, video, etc. The memory 302 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, disk or optical disk. The multimedia component 303 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signal may be further stored in the memory 302 or sent via the communication component 305. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 304 provides an interface between the processor 301 and other interface modules, and the above-mentioned other interface modules can be keyboards, mice, buttons, etc. These buttons can be virtual buttons or physical buttons. The communication component 305 is used for wired or wireless communication between the marine radar image denoising device 300 based on the multi-stage filtering algorithm and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G or 4G, or a combination of one or more of them, so the corresponding communication component 305 can include: Wi-Fi module, Bluetooth module, NFC module.

[0110] In an exemplary embodiment, the ship radar image denoising device 300 based on the multi-stage filtering algorithm can be implemented by one or more application specific integrated circuits (Application Specific Integrated Circuit, referred to as ASIC), digital signal processors (Digital Signal Processor, referred to as DSP), digital signal processing devices (Digital Signal Processing Device, referred to as DSPD), programmable logic devices (Programmable Logic Device, referred to as PLD), field programmable gate arrays (Field Programmable Gate Array, referred to as FPGA), controllers, microcontrollers, microprocessors or other electronic components, and is used to execute the above-mentioned ship radar image denoising method based on the multi-stage filtering algorithm.

[0111] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, and when the program instructions are executed by a processor, the steps of the above-mentioned ship radar image denoising method based on a multi-stage filtering algorithm are implemented. For example, the computer-readable storage medium can be the above-mentioned memory 302 including program instructions, and the above-mentioned program instructions can be executed by the processor 301 of the ship radar image denoising device 300 based on a multi-stage filtering algorithm to complete the above-mentioned ship radar image denoising method based on a multi-stage filtering algorithm.

[0112] Example 4

[0113] Corresponding to the above method embodiment, the embodiment of the present disclosure further provides a readable storage medium. The readable storage medium described below and the ship radar image denoising method based on the multi-stage filtering algorithm described above can be referred to each other.

[0114] A readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the marine radar image denoising method based on a multi-stage filtering algorithm in the above method embodiment.

[0115] The readable storage medium may specifically be a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or other readable storage medium that can store program codes.

[0116] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for denoising a marine radar image based on a multi-stage filtering algorithm, characterized in that: include: Acquire the ship radar image to be processed; Preprocessing the to-be-processed marine radar image to obtain a first image; Based on the gradient direction corresponding to each pixel point in the first image and the target identified in the first image, edge gradient detection is performed on the first image, and after the edge gradient detection passes, noise detection is performed on the first image based on the grayscale value corresponding to each pixel point in the first image; After the noise detection is passed, the first image is processed by using bilateral filtering to obtain a second image; The second image is subjected to image sharpening and then image enhancement to obtain a denoised marine radar image.

2. The marine radar image denoising method based on a multi-stage filtering algorithm according to claim 1 is characterized in that: Preprocessing the to-be-processed marine radar image to obtain a first image includes: The to-be-processed marine radar image is gray-scaled to obtain a gray-scale image, and the gray-scale image is normalized to obtain a first image.

3. The marine radar image denoising method based on a multi-stage filtering algorithm according to claim 1 is characterized in that: Based on the gradient direction corresponding to each pixel point in the first image and the target identified in the first image, edge gradient detection is performed on the first image, including: The Sobel operator is used to calculate the gradient intensity corresponding to each pixel in the first image; the gradient intensity corresponding to all the pixels is compared with a preset gradient intensity threshold, and the pixels with a gradient intensity greater than the gradient intensity threshold are screened out and recorded as first pixels; the gradient direction corresponding to each pixel is calculated, and a square frame of a preset size is formed with each first pixel as the center, and the standard deviation of the gradient direction of all the pixels contained in each square frame is used as the standard deviation corresponding to each first pixel; the first pixel whose standard deviation is greater than the standard deviation threshold is used as a rough point; OpenCV is used to identify the contours of the target in the first image and the coordinates of the pixel points contained in each contour; the number of pixel points contained in each contour is counted and recorded as a first value, the number of rough points contained in each contour is counted according to the coordinates of the pixel points contained in each contour and recorded as a second value, and the ratio of the second value to the first value is used as the roughness ratio of each target; the roughness ratio of each target is averaged to obtain the overall roughness ratio, and the overall roughness ratio is compared with a preset overall roughness ratio threshold. When the overall roughness ratio is greater than the preset overall roughness ratio threshold, the detection result is determined to be not smooth and the edge gradient detection fails, and the first image is processed by Gaussian filtering, and the edge gradient detection is performed after the processing; otherwise, the detection result is determined to be smooth and the edge gradient detection passes.

4. The method for denoising a marine radar image based on a multi-stage filtering algorithm according to claim 1, characterized in that: Performing noise detection on the first image based on the grayscale value corresponding to each pixel in the first image includes: Select a local window with each pixel as the center on the first image, and take the mean of the grayscale values ​​of all pixels in the window as the local mean corresponding to each pixel; calculate the local variance corresponding to each pixel based on the local mean corresponding to each pixel, the grayscale values ​​of all pixels in the window and the size of the window, and calculate the local standard deviation based on the local variance; The local standard deviation corresponding to each pixel point is multiplied by a preset coefficient to obtain a local standard deviation threshold. The absolute value of the difference between the grayscale value corresponding to each pixel point and the corresponding local mean is compared with the local standard deviation threshold. If the absolute value is greater than the local standard deviation threshold, the pixel point is determined to be a noise point. At the same time, the local variance corresponding to each pixel point is compared with the preset local variance threshold. If the local variance is greater than the local variance threshold, the pixel point is also determined to be a noise point. Based on the number of noise points and the local variance, it is determined whether the first image passes the noise detection.

5. The method for denoising a marine radar image based on a multi-stage filtering algorithm according to claim 4, characterized in that: Judging whether the first image passes the noise detection based on the number of noise points and the local variance includes: Calculate the mean of the local variances corresponding to all the pixels in the first image to obtain the global variance mean; count the number of noise points, and take the ratio between the number of noise points and the total number of pixels in the first image as the noise point ratio; Conditions are set, where condition one is that the global variance mean is compared with a preset global variance mean threshold, and the global variance mean is greater than the global variance mean threshold; condition two is that the noise point ratio is compared with a preset noise point ratio threshold, and the noise point ratio is greater than the noise point ratio threshold; when any one of the conditions is met, the detection result is determined to be that noise exists, the noise detection fails, and the first image is processed by median filtering, and noise detection is performed after processing; otherwise, the detection result is determined to be that noise does not exist, and the noise detection passes.

6. The method for denoising a marine radar image based on a multi-stage filtering algorithm according to claim 1, characterized in that: The first image is processed using bilateral filtering, including: The key parameters in the bilateral filter are determined, and the key parameters include the neighborhood size, the color standard deviation and the space standard deviation. After the determination, the first image is processed by using the bilateral filter to obtain the second image.

7. The method for denoising a marine radar image based on a multi-stage filtering algorithm according to claim 1, characterized in that: Determine the key parameters in bilateral filtering, which include neighborhood size, color standard deviation, and spatial standard deviation, including: OpenCV is used to identify the contours of the target in the first image after noise detection and the coordinates of the pixels contained in each contour. Based on the contours of the target in the first image after noise detection and the coordinates of the pixels contained in each contour, the area formed by each contour is recorded as the target area, the number of pixels contained in each contour is recorded as the area of ​​the target area, the standard deviation of the grayscale values ​​of the pixels contained in each contour is calculated and the standard deviation calculation result is recorded as the grayscale standard deviation of the target area; the calculation formula of the neighborhood size d is: Among them, K is the adjustment coefficient, which is 0.1; N is the number of target areas; Area i is the area of ​​target region i; Color standard deviation σ C The calculation formula is: Among them, α is the coefficient, which is 1; β is the offset, which is 10; is the average value of the grayscale standard deviation of N target areas; σ i is the grayscale standard deviation of the i-th target area; Spatial standard deviation σ S The calculation formula is:

8. A marine radar image denoising device based on a multi-stage filtering algorithm, characterized in that: include: An acquisition module, used for acquiring the marine radar image to be processed; A denoising module, used for preprocessing the to-be-processed marine radar image to obtain a first image; Based on the gradient direction corresponding to each pixel point in the first image and the target identified in the first image, edge gradient detection is performed on the first image, and after the edge gradient detection passes, noise detection is performed on the first image based on the grayscale value corresponding to each pixel point in the first image; After the noise detection is passed, the first image is processed by using bilateral filtering to obtain a second image; The second image is subjected to image sharpening and then image enhancement to obtain a denoised marine radar image.

9. A marine radar image denoising device based on a multi-stage filtering algorithm, characterized in that: include: Memory for storing computer programs; A processor, configured to implement the steps of the marine radar image denoising method based on a multi-stage filtering algorithm as claimed in any one of claims 1 to 7 when executing the computer program.

10. A storage medium, characterized in that: The storage medium stores a computer program, which, when executed by a processor, implements the steps of the method for denoising a marine radar image based on a multi-stage filtering algorithm as claimed in any one of claims 1 to 7.

Images