A 2D forward-looking sonar image denoising method

Abstract

A 2D forward-looking sonar image denoising method comprises image acquisition (1), image preprocessing (2), threshold segmentation (3), noise segmentation (4), target enhancement (5), and region of interest extraction (6). The method integrates the traditional image processing method with adaptive threshold segmentation, Hough transform and cluster statistical method, effectively solving the problem of excessive image noise caused by payload body noise, inter-sector interference and uneven target echo characteristics during the imaging process of the forward-looking sonar, achieving better removal of background noise while retaining the effective area of ​​obstacles and targets, obtaining high-quality regions of interest, and laying a foundation for subsequent further target recognition and high-precision positioning work.

Description

Technical Field

[0001] The present invention relates to a 2D forward-looking sonar image denoising method, in particular to an image noise removal method for a 2D forward-looking sonar payload image. Background Art

[0002] Forward-looking sonar is an important payload carried by underwater equipment. It can detect underwater obstacles and moving and static targets in a large range, and is the basis for autonomous obstacle avoidance and target identification and tracking. However, factors such as water noise, interference between payload sectors, and uneven target echo intensity introduce strong noise into the sonar image. Up to now, noise interference is still an important challenge for sonar image target detection. There is currently no target detection method designed specifically for sonar images. Most of the related research methods are based on optical image processing methods and optimize the imaging characteristics of acoustic images to achieve the purpose of filtering out noise and obtaining the target area. The images formed by different payload equipment in complex marine environments are diverse. Due to insufficient samples, existing research cannot meet the needs of underwater equipment for target detection. Summary of the invention

[0003] The technical problem solved by the present invention is: to overcome the shortcomings of the prior art and provide a 2D forward-looking sonar image denoising method, which can filter out some typical noises caused by problems such as water noise, interference between payload sectors and uneven target echo intensity, and meet the mission system requirements through testing.

[0004] The technical solution of the present invention is: a 2D forward-looking sonar image denoising method, comprising the following steps:

[0005] Sonar images are obtained from the forward-looking sonar in real time, and the threshold method is used for preliminary segmentation. Then, the noise is segmented according to the noise characteristics to obtain the noise map. Then, according to the target characteristics, the target display is enhanced. Finally, the initial segmentation, noise segmentation, and target enhancement map are fused to obtain the final region of interest.

[0006] Furthermore, the threshold method adopts a global adaptive threshold segmentation algorithm based on histogram statistical characteristics.

[0007] Furthermore, the global adaptive threshold segmentation algorithm based on histogram statistical features comprises the following steps:

[0008] First, the grayscale values ​​of the image from 0 to 255, a total of 256, are counted to create a grayscale histogram;

[0009] Then, start calculating the threshold;

[0010] Finally, the image is traversed with the threshold, and pixels above the threshold are set to 255, otherwise they are set to 0.

[0011] Furthermore, the calculation of the threshold value comprises the following steps:

[0012] When calculating the threshold, count the number of pixels from high brightness 255 to low brightness 0, record the number of pixels as α, and keep recording the gray value δ at this time;

[0013] The reason for going from 255 to 0 is that in sonar images, targets often have strong reflective characteristics, corresponding to the bright spots in the image, with high grayscale values, and are foregrounds; water bodies have weak reflective characteristics, corresponding to the low bright spots in the image, with low grayscale values, and are backgrounds;

[0014] Then, when the counted pixel number α is greater than 500, the gray value δ at this time is found, which is the segmentation threshold.

[0015] Furthermore, the method for segmenting noise according to noise characteristics is an arc extraction method based on improved Hough transform; when extracting arcs in the arc extraction method based on improved Hough transform, a horizontal scanning method is used to scan the image.

[0016] Furthermore, in the horizontal scanning method used for scanning the image, the method for obtaining the angular resolution is: Among them, deg is the angular resolution, r is the pixel distance from the point to the sonar head, and value is the value of the angular resolution deg.

[0017] Furthermore, the method for enhancing target display is a fusion of radial compensation method and clustering method; wherein

[0018] The radial compensation method utilizes the "tailing" and "shadow" characteristics of the target imaging, scans from left to right along the sector radial direction, and retains the pixels that meet the width;

[0019] The clustering method analyzes the region of interest of the image and uses the distance as an evaluation index of similarity.

[0020] Furthermore, in the radial compensation method, during scanning, the resolution is 0.001 radians.

[0021] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the 2D forward-looking sonar image denoising method are implemented.

[0022] A 2D forward-looking sonar image denoising device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the 2D forward-looking sonar image denoising method are implemented.

[0023] The advantages of the present invention compared with the prior art are:

[0024] (1) The present invention integrates traditional image processing methods, adaptive threshold segmentation, Hough transform and cluster statistics methods to effectively solve the problem of excessive image noise caused by payload body noise, sector interference and target echo characteristics during the imaging process of forward-looking sonar. It achieves the goal of removing background noise while retaining the effective area of ​​obstacles and targets, thereby obtaining a high-quality area of ​​interest.

[0025] (2) The present invention can be seamlessly connected with subsequent obstacle and target detection technologies, and has been tested on an underwater mobile platform. The algorithm has good real-time performance and meets the needs of unmanned platforms such as AUV and ROV. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 is a flow chart of the method of the present invention;

[0027] Figure 2 Schematic diagram of the processing process taking a certain image as an example;

[0028] Figure 3 Extract contrast images for regions of interest in typical sonar images. DETAILED DESCRIPTION

[0029] In order to better understand the above technical scheme, the technical scheme of the present application is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical scheme of the present application, rather than limitations on the technical scheme of the present application. In the absence of conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other.

[0030] The following is a further detailed description of a 2D forward-looking sonar image denoising method provided by an embodiment of the present application in conjunction with the accompanying drawings. The specific implementation method may include (eg Figures 1 to 3 As shown in the figure): image acquisition (1), image preprocessing (2), initial segmentation (3), noise segmentation (4), target enhancement (5), and region of interest extraction (6).

[0031] The image preprocessing described in step (2) includes one or any combination of image median filtering, bilateral filtering and morphological opening operation.

[0032] The initial segmentation in step (3) adopts a global adaptive threshold segmentation algorithm based on histogram statistical features.

[0033] The processing described in step (4) adopts an arc extraction method based on improved Hough transform.

[0034] The processing described in step (5) first uses a fused radial compensation method and clustering method to enhance the target to reduce target breakage caused by echo unevenness, and then uses a morphological filtering method to further optimize the target area.

[0035] In the solution provided in the embodiment of the present application, Figure 1 As shown, the steps of the present invention include image acquisition (1), image preprocessing (2), initial segmentation (3), noise segmentation (4), target enhancement (5), and region of interest extraction (6). That is:

[0036] Firstly, the sonar image is obtained from the forward-looking sonar in real time, and the threshold method is used for preliminary segmentation. Then, the noise is segmented according to the noise characteristics to obtain the noise map. Then, according to the target characteristics, the target display is enhanced. Finally, the initial segmentation, noise segmentation and target enhancement map are fused to obtain the final region of interest.

[0037] It should be noted that the threshold method here adopts a global adaptive threshold segmentation algorithm based on histogram statistical features.

[0038] First, we count the 256 gray levels of the image from 0 to 255 and create a gray level histogram.

[0039] Then, the threshold calculation begins.

[0040] When calculating the threshold, the number of pixels is counted from high brightness 255 to low brightness 0, the number of pixels is recorded as α, and the grayscale value δ at this time is continuously recorded.

[0041] The reason for going from 255 to 0 is that in sonar images, targets often have strong reflective characteristics, corresponding to the bright spots in the image, with high grayscale values, and are foregrounds; water bodies have weak reflective characteristics, corresponding to the low bright spots in the image, with low grayscale values, and are backgrounds.

[0042] Then, when the counted pixel number α is greater than 500, the gray value δ at this time is found, which is the segmentation threshold.

[0043] It should be noted that, during the statistical process, if the gray value δ is less than 100 and the number of pixels α is still less than 500, the statistical process is stopped and 100 is used as the segmentation threshold.

[0044] Finally, the image is traversed with the threshold, and pixels above the threshold are set to 255, otherwise they are set to 0.

[0045] Figure 2 Figure 2 in the middle is the image after threshold segmentation.

[0046] Noise segmentation mainly removes the main body noise of the sonar equipment. Affected by the inconsistency of the equipment sector and the strong reflection effect of the target, ring noise will appear in the sonar image along with the high-intensity target, affecting the segmentation of the real target. In this regard, by analyzing the characteristics of the noise, an arc extraction method based on the improved Hough transform is adopted.

[0047] It should be noted that when extracting arcs, the image is scanned using the lateral scanning method. Among them, the angular resolution decreases as the distance from the sonar head increases. The angular resolution is obtained according to the following formula:

[0048]

[0049] Figure 2 Figure 3 in the middle is the extracted arc noise.

[0050] Directly removing arc noise based on threshold segmentation will lose target characteristics, and during threshold segmentation, due to the inconsistency of target reflection intensity, the target will be segmented into multiple small blocks, such as Figure 2 The middle one is shown in No. 4. Therefore, target enhancement is required.

[0051] A fusion of radial compensation and clustering methods is used to enhance the target.

[0052] The radial compensation method analyzes and utilizes the "tailing" and "shadow" characteristics in the target imaging, scans along the sector radially from left to right, and retains the pixels that meet the width.

[0053] It should be noted that when scanning, a resolution of 0.001 radians will obtain a better result of comprehensive time cost and compensation effect. When the target pixel width is 7 pixels, a better enhancement result can be obtained.

[0054] The clustering method is to analyze the interesting regions of the image and use the distance as the evaluation index of similarity.

[0055] It should be noted that if the pixel value of pixel point P1 is (x1, y1) and the pixel value of pixel point P2 is (x2, y2), then the distance Where x is the column where the pixel is located, and y is the row where the pixel is located.

[0056] It should be further explained that when the block spacing d of the region of interest is less than 4 pixels, it is considered to be the same segmented target, and the two regions of interest are connected. Finally, the morphological opening and closing operations are used in turn to filter and enhance the region of interest, and finally the final target region is obtained. Figure 2 Figure 5 in the middle is the region of interest that was finally extracted. Figure 2 Figure 6 in the middle shows the target in the original image.

[0057] The present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and when the computer instructions are executed on a computer, the computer executes Figure 1 The method described.

[0058] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) that contain computer-usable program code.

[0059] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0060] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0061] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0062] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

[0063] The contents not described in detail in the specification of the present invention belong to the common knowledge of those skilled in the art.

Claims

1. A 2D forward-looking sonar image denoising method, characterized in that: The steps include: The sonar image is obtained from the forward-looking sonar in real time, and the threshold method is used for preliminary segmentation. Then the noise is segmented according to the noise characteristics to obtain the noise map. Then the target display is enhanced according to the target characteristics. Finally, the initial segmentation, noise segmentation, and target enhancement map are fused to obtain the final region of interest. The threshold method adopts a global adaptive threshold segmentation algorithm based on histogram statistical characteristics; The global adaptive threshold segmentation algorithm based on histogram statistical features comprises the following steps: First, the grayscale values ​​of the image from 0 to 255, a total of 256, are counted to create a grayscale histogram; Then, start calculating the threshold; Finally, the image is traversed with the threshold, and pixels above the threshold are set to 255, otherwise they are set to 0; The calculation of the threshold comprises the following steps: When calculating the threshold, count the number of pixels from high brightness 255 to low brightness 0, record the number of pixels as α, and keep recording the gray value δ at this time; The reason for going from 255 to 0 is that in sonar images, targets often have strong reflective characteristics, corresponding to the bright spots in the image, with high grayscale values, and are foregrounds; water bodies have weak reflective characteristics, corresponding to the low bright spots in the image, with low grayscale values, and are backgrounds; Then, when the counted pixel number α is greater than 500, find the gray value δ at this time, which is the segmentation threshold; The method for segmenting noise according to noise characteristics is an arc extraction method based on improved Hough transform; when extracting arcs in the arc extraction method based on improved Hough transform, a horizontal scanning method is used for scanning the image; The image is scanned using a lateral scanning method, and the angular resolution is obtained by: Where, deg is the angular resolution, r is the pixel distance from the point to the sonar head, and value is the value of the angular resolution deg; The method for enhancing target display is a fused radial compensation method and clustering method; wherein The radial compensation method utilizes the "tailing" and "shadow" characteristics in the target imaging, scans from left to right along the sector radial direction, and retains the pixel points that meet the width; The clustering method is to analyze the region of interest of the image and use the distance as the evaluation index of similarity; In the radial compensation method, the resolution is 0.001 radian during scanning.

2. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method as claimed in claim 1 are implemented.

3. A 2D forward-looking sonar image denoising device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method as claimed in claim 1 are implemented.

Images