An underwater three-dimensional terrain and multi-beam image sonar data simulation system and method

Abstract

The present invention discloses an underwater three-dimensional terrain and multi-beam bathymetric sonar simulation method and system, belonging to the field of underwater detection technology, and can be widely applied in the fields of three-dimensional technology and sonar images. The present invention obtains a two-dimensional sonar image of the underwater terrain through the synthetic aperture sonar in the acquisition unit, and obtains a measured multi-beam sonar image data set through the multi-beam sonar in the acquisition unit; a data simulation unit simulates the navigation data of the underwater vehicle and the detection method of the multi-beam image sonar; the first model construction unit mainly constructs a target simulation model and an underwater three-dimensional point cloud terrain, the second model construction unit mainly constructs a sonar image simulation model and an original simulated sonar image, and the third model unit mainly constructs a sonar image style model and a final sonar image. The above units ensure that the system can complete three-dimensional scene simulation and multi-beam image sonar data simulation in a complex and changeable underwater environment, and generate a complete and clear sonar image.

Description

Technical Field

[0001] The present invention relates to an underwater three-dimensional terrain and multi-beam image sonar data simulation system and method, belonging to the technical field of underwater detection. Technical Background

[0002] In recent years, due to the important role played by underwater scene three-dimensional reconstruction in engineering projects such as marine resource development and underwater emergency rescue, it has gradually become the research focus in the field of underwater detection. Currently, the main underwater detection methods are acoustic detection and optical detection modes. Although the optical images detected by using optical vision perception technology can cover various color information and structural information within the detection range, due to the very strong absorption and scattering effects of dissolved and suspended substances in water on light, light shows exponential attenuation when propagating underwater. In some turbid waters, the visibility is even less than ten meters, severely limiting the perception range. Based on sonar acoustic vision perception technology, because the wavelength of sound waves is relatively long and the propagation distance underwater is relatively far, it is widely used in the field of underwater detection.

[0003] Using the form of an underwater vehicle equipped with a multi-beam sonar is one of the most common detection modes in current underwater detection. However, due to factors such as the high cost of underwater vehicles and sonars and the complex and changeable underwater environment, it is difficult to obtain underwater data, which severely limits the progress of research in aspects such as underwater detection. Therefore, it is of great significance to simulate underwater three-dimensional terrain and sonar images. Summary of the Invention

[0004] To solve the above technical problems, the present invention provides a simulation of an underwater three-dimensional scene and a multi-beam image sonar data simulation system and method, which can complete three-dimensional scene simulation and multi-beam image sonar data simulation in a complex and changeable underwater environment and generate complete and clear sonar images.

[0005] The technical solution of the present invention is realized as follows:

[0006] An underwater three-dimensional terrain and multi-beam image sonar data simulation system, the system includes: an acquisition unit, a first model construction unit, a second model construction unit, a third model construction unit, and a data simulation unit;

[0007] The acquisition unit, the synthetic aperture sonar in the acquisition unit acquires a high-definition underwater terrain two-dimensional sonar image and converts it into a grayscale image, and the multi-beam sonar in the acquisition unit acquires an underwater terrain measured multi-beam sonar image dataset;

[0008] The first model construction unit constructs an underwater three-dimensional point cloud terrain based on the two-dimensional sonar image;

[0009] A data simulation unit that simulates the navigation data of an underwater vehicle and the detection method of a multibeam image sonar to obtain a simulation result;

[0010] A second model construction unit that constructs a sonar image simulation model based on the simulation result and the underwater three-dimensional point cloud terrain to obtain an original simulated sonar image;

[0011] A third model construction unit that constructs a sonar image style model based on the measured multibeam sonar image dataset of the underwater terrain, and brings the original simulated sonar image into the sonar image style model to construct a final sonar image.

[0012] Further, the method for obtaining the original simulated sonar image is specifically as follows:

[0013] According to the simulation, obtain the movement trajectory of the underwater vehicle underwater, and bring the detection method of the simulated multibeam image sonar into the movement trajectory; perform three-dimensional transformation according to the simulated orientation angle, find the nearest neighbor points of the detection rays of the simulated multibeam image sonar and the simulated three-dimensional point cloud terrain, and obtain the distance between the detection points of each ray and the simulated multibeam image sonar; determine the width of the simulated two-dimensional multibeam image sonar image according to the detection horizontal opening angle of the simulated multibeam image sonar, and determine the height of the simulated two-dimensional multibeam image sonar image and the distance represented by one pixel point in the column pixels according to the detection distance of the simulated multibeam image sonar; bring the distance between the target points detected by the simulation and the simulated multibeam image sonar into the two-dimensional image to generate an initial simulated sonar image.

[0014] Further, the detection method of the simulated multibeam image sonar is specifically as follows:

[0015] Determine the longitudinal opening angle and the horizontal opening angle of the simulated multibeam sonar, and emit a ray every 1° with the simulated multibeam sonar as the origin in the horizontal opening angle, and emit a ray every 0.1 degree in the longitudinal opening angle.

[0016] Further, the navigation data includes: timestamp data, navigation azimuth data, attitude angle data of the underwater vehicle, depth data of the underwater vehicle, and navigation speed data.

[0017] An underwater three-dimensional terrain and multibeam image sonar data simulation method, including the following steps:

[0018] Step 1, the synthetic aperture sonar in the acquisition unit acquires a high-definition two-dimensional sonar image of the underwater terrain and converts it to the gray space as the first gray image;

[0019] Step 2. The first model construction unit establishes a target simulation model according to the KLSG dataset and the first gray image, and obtains an underwater two-dimensional terrain image;

[0020] Step 3. The first model construction unit constructs an underwater three-dimensional point cloud terrain based on the underwater two-dimensional terrain image;

[0021] Step 4. The data simulation unit simulates the navigation data of the underwater vehicle and the multi-beam image sonar detection method to obtain a simulation result;

[0022] Step 5. The second model construction unit constructs a sonar image simulation model by combining the simulation result with the underwater three-dimensional point cloud terrain to obtain an original simulated sonar image;

[0023] Step 6. The multi-beam sonar in the acquisition unit acquires a set of measured multi-beam sonar image datasets of the underwater terrain. The third model construction unit constructs a sonar image style model based on the measured multi-beam sonar image datasets of the underwater terrain, and brings the original simulated sonar image into the sonar image style model to obtain the final sonar image.

[0024] Further, the first model construction unit establishes a target simulation model based on the KLSG dataset and the first grayscale image, and obtains an underwater two-dimensional terrain image. The specific method is as follows:

[0025] Randomly select an image from the KLSG dataset as the image to be processed, convert it to the grayscale space as the second grayscale image; segment the second grayscale image and perform binarization processing to obtain a binarized foreground image; use the binarized foreground image as a mask and combine it with the image to be processed to obtain a foreground image of the target area; bring the binarized foreground image into the CANNY algorithm to obtain the edge information of the binarized foreground image; brighten the foreground image and randomly add it to the first grayscale image to obtain an initial fusion image, and perform Gaussian filtering along the image edge in the initial fusion image according to the edge information of the above binarized foreground image to generate the final underwater terrain two-dimensional image.

[0026] Further, the method for constructing an underwater three-dimensional point cloud terrain based on the underwater two-dimensional terrain image is as follows:

[0027] According to the simulation requirements, set the number of point clouds corresponding to the image pixel points, use the abscissa and ordinate of the pixel points of the image as the abscissa and ordinate of the three-dimensional point cloud respectively, and extract the grayscale value of the two-dimensional terrain image as the height of the three-dimensional point cloud to establish an underwater three-dimensional point cloud terrain. The specific mapping relationship is shown in the following formula:

[0028]

[0029] Where: X map is the abscissa of the three-dimensional point cloud; Y map is the ordinate of the three-dimensional point cloud; Z map is the height of the three-dimensional point cloud; X imgis the abscissa of the image, Y img is the ordinate of the image, is the grayscale value corresponding to the x and y coordinates in the image; n and m are the adjustment coefficients of the three-dimensional terrain;

[0030] After generating the initial three-dimensional point cloud, interpolation is performed between adjacent point clouds using the method of linear interpolation, as shown in the following formula:

[0031]

[0032] Where: is the abscissa of the j-th point cloud; is the ordinate of the j-th point cloud, is the height of the j-th point cloud, and l is the number of differences between adjacent point clouds.

[0033] Furthermore, the multi-beam sonar in the acquisition unit acquires a set of measured multi-beam sonar image data sets of the underwater terrain. The third model construction unit constructs a sonar image style model based on the measured multi-beam sonar image data sets of the underwater terrain and obtains the final sonar image. The specific method is as follows:

[0034] Segment the images in a set of measured multi-beam sonar image data sets of the underwater terrain acquired by the multi-beam sonar in the acquisition unit, separate the foreground and background of the images, generate an image background data set, and construct a sonar image style model based on the image background data set; bring the image background data set into the sonar image style model for training; perform Gaussian blur on the original simulated sonar image, and bring the Gaussian-blurred original sonar image into the sonar image style model to generate the final simulated sonar image.

[0035] Beneficial effects:

[0036] 1. The present invention sets up an acquisition unit, a first model construction unit, a second model construction unit, a third model construction unit, and a data simulation unit. The five units work together. Only by acquiring a high-definition two-dimensional sonar image of the underwater terrain once through a synthetic aperture sonar, multiple simulations can be completed and the clarity of the final sonar image can be improved by combining the image style transfer technology, reducing the high cost of acquiring high-definition two-dimensional sonar images of the underwater terrain multiple times using a synthetic aperture sonar.

[0037] 2. When the second model construction unit constructs the original simulated sonar image, it combines the simulation results of the data simulation unit, ensuring the accuracy of the initial simulated sonar image.

[0038] 3. For the longitudinal opening angle and transverse opening angle of the true multi-beam sonar, a ray is emitted every 1° in the transverse opening angle with the simulated multi-beam sonar as the origin, and a ray is emitted every 0.1 degree in the longitudinal opening angle, ensuring the clarity and accuracy of the subsequent image generation.

[0039] 4. According to the simulation requirements, set the number of point clouds corresponding to the image pixel points. Use the abscissa and ordinate of the pixel points of the image to correspond to the abscissa and ordinate of the three-dimensional point cloud respectively, and extract the gray value of the two-dimensional terrain image as the height of the three-dimensional point cloud to establish an underwater three-dimensional point cloud terrain, improving the accuracy of converting from a two-dimensional terrain image to a three-dimensional point cloud terrain and ensuring clear and accurate later imaging. Brief Description of the Drawings

[0040] Figure 1 is a flowchart of an underwater three-dimensional terrain and multi-beam image sonar data simulation method;

[0041] Figure 2 is a schematic diagram of a simulated three-dimensional point cloud terrain;

[0042] Figure 3 is a schematic diagram of the detection process of a simulated multi-beam image sonar in a simulated three-dimensional terrain;

[0043] Figure 4 is a simulated multi-beam sonar image. Detailed Embodiment

[0044] The present invention will be further described below in conjunction with the drawings in the specification and specific embodiments.

[0045] In this embodiment, in combination with the actual situation, an underwater three-dimensional terrain and multi-beam image sonar data simulation system and method (such as Figure 1 ) are provided, including the following content:

[0046] An underwater three-dimensional terrain and multi-beam image sonar data simulation system, which includes: an acquisition unit, a first model construction unit, a second model construction unit, a third model construction unit, and a data simulation unit; among them, the functions of each unit are as follows:

[0047] The acquisition unit, the synthetic aperture sonar in the acquisition unit acquires a high-definition underwater terrain two-dimensional sonar image and converts it into a gray-scale image, and the multi-beam sonar in the acquisition unit acquires a measured multi-beam sonar image dataset;

[0048] The first model construction unit constructs a target simulation model and constructs an underwater three-dimensional point cloud terrain;

[0049] The second model construction unit constructs a sonar image simulation model;

[0050] The third model construction unit constructs a sonar image style model;

[0051] The data simulation unit simulates the navigation data of an underwater vehicle and simulates the detection method of a multi-beam image sonar.

[0052] An underwater three-dimensional terrain and multi-beam image sonar data simulation method includes the following specific methods:

[0053] Step 1. The synthetic aperture sonar in the acquisition unit acquires a high-definition two-dimensional sonar image of the underwater terrain and converts it into the gray space as the first gray image;

[0054] It should be noted that the acquired high-definition two-dimensional sonar image of the underwater terrain is generally a synthetic aperture sonar detection image. The acquired synthetic aperture sonar image is converted to the gray space, and the conversion formula is as follows:

[0055] Gray = 0.229*R + 0.587*G + 0.114*B (1)

[0056] Step 2. The first model construction unit establishes a target simulation model according to the KLSG dataset and the first gray image, and obtains an underwater two-dimensional terrain image;

[0057] In this step, first, a to-be-processed image is randomly selected from the KLSG image dataset as the target image. After the image is segmented by the K-Means algorithm and then binaryzation analysis is performed, according to the characteristics of the dataset image, the one with a small area is selected as the foreground image, that is, the target ship or the target aircraft;

[0058] Process the above binaryzation image: On the one hand, use it as a mask to extract the target of the target image; on the other hand, bring the binaryzation image into the CANNY algorithm to obtain the edge information of the target image.

[0059] Randomly superimpose it on the first gray image according to the size of the target image. According to the edge information of the target image, perform Gaussian filtering on the target edge of the superimposed image to blur the target edge and the background. The size of the filtering kernel is 3×3, and finally generate a simulated two-dimensional terrain image;

[0060] Step 3. The first model construction unit constructs an underwater three-dimensional point cloud terrain according to the underwater two-dimensional terrain image (such as Figure 2 );

[0061] The specific method of establishing an underwater three-dimensional point cloud terrain using the simulated two-dimensional terrain image is: map the horizontal and vertical coordinates of the image into the three-dimensional space, and map the gray value of the image into the height of the three-dimensional point cloud. The specific mapping relationship is as shown in formula (2):

[0062]

[0063] Where: X map is the abscissa of the three-dimensional point cloud; Y map is the ordinate of the three-dimensional point cloud; Z map is the height of the three-dimensional point cloud; X imgis the abscissa of the image, Y img is the ordinate of the image, is the gray value corresponding to the x and y coordinates in the image; n and m are the adjustment coefficients of the three-dimensional terrain. In this example, the value of n is 100 and the value of m is 0.01. The specific values are determined according to the actual simulation requirements;

[0064] After generating the initial three-dimensional point cloud, linear interpolation is used to interpolate between adjacent point clouds. In this example, the number of interpolations between adjacent point clouds is 99, as shown in formula (3) specifically:

[0065]

[0066] Where: is the abscissa of the j-th point cloud; is the ordinate of the j-th point cloud, is the height of the j-th point cloud.

[0067] Step 4. The data simulation unit simulates the navigation data of the underwater vehicle and the multi-beam image sonar detection method to obtain the simulation results;

[0068] In this embodiment, the referred navigation data includes timestamp data T, navigation azimuth data Cog, underwater vehicle attitude angle data (heading, pith, roll), underwater vehicle depth data set Depth, and navigation speed data v;

[0069] The referred multi-beam image sonar detection method is: determine the longitudinal opening angle and the transverse opening angle of the simulated multi-beam sonar. In this embodiment, the set transverse opening angle is 80° and the longitudinal opening angle is 1°, that is, determine the three-dimensional space detected by the multi-beam sonar. A ray is emitted every 0.15° with the simulated multi-beam sonar as the origin in the transverse opening angle, and a ray is emitted every 0.1 degree in the longitudinal opening angle.

[0070] Step 5. The second model construction unit combines the simulation results with the underwater three-dimensional point cloud terrain to construct a sonar image simulation model, and obtains the original simulated sonar image;

[0071] According to the simulated movement trajectory of the underwater vehicle underwater, the simulated multi-beam image sonar detection method is brought into the movement trajectory. The schematic diagram of the detection process of the simulated multi-beam image sonar in the simulated three-dimensional terrain is as Figure 3 shown;

[0072] Preprocess the simulation results. Among them, the timestamp data T is the recording time of each item of data, and the time interval set T for each data simulation can be calculated using T 1 , as shown in formulas (4) and (5):

[0073] T = {t1, t2, t3, …, t n-2 , t n-1 , t n} (4)

[0074] T 1 = {t2 - t1, t3 - t2, …, t n-1 - t n-2 , t n - t n-1} (5)

[0075] The navigation speed data V is the instantaneous speed of the simulation. The average speed set V of the data acquisition time interval can be approximately calculated based on V, as shown in Equations (6) and (7): 1 , as shown in Equations (6) and (7):

[0076] V = {v1, v2, v3, …, v n-2 , v n-1 , v n} (6)

[0077]

[0078] The difference data set Depth of each depth and the previous depth can be calculated based on the depth data set Depth, as shown in Equations (8) and (9): 1 , as shown in Equations (8) and (9):

[0079] Depth = {d1, d2, d3, …, d n-2 , d n-1 , d n} (8)

[0080] Depth 1 = {d2 - d1, d3 - d2, …, d n-1 - d n-2 , d n - d n-1} (9)

[0081] Using the data time interval data set T 1 and the interval average speed set V 1 The movement distance data set D of the vehicle within the time interval can be calculated. The calculation formula is shown in Equation (10):

[0082]

[0083] A rotation matrix R is established based on the carrier attitude angle data. The rotation matrix is shown in Equation (11):

[0084]

[0085] Where: h, p, and r respectively represent the elements in the Heading, Pitch, and Roll datasets.

[0086] Construct a sonar image simulation model according to the above algorithm, and bring the simulated multi-beam sonar image detection method into the constructed sonar image simulation model; it should be noted that since the rays emitted by the simulation may not intersect a certain point in the three-dimensional point cloud map, therefore, in this embodiment, only the point cloud in the three-dimensional point cloud map that is closest to the sonar detection line emitted by the simulation is calculated as the target point cloud.

[0087] Record the distance between the target point cloud and the vehicle, and set the detection distance of the sonar detection and the mapping relationship in the sonar image. In this embodiment, the simulated sonar detection distance is 10m, and the height of the simulated sonar image is 1000 pixel points. Generate the original simulated sonar image according to the ratio of the distance between the target point cloud and the vehicle to the sonar detection distance, and randomly add patch noise to it.

[0088] Step 6. The multi-beam sonar in the acquisition unit acquires a set of measured multi-beam sonar image datasets of the underwater terrain. The third model construction unit constructs a sonar image style model according to the measured multi-beam sonar image datasets of the underwater terrain and obtains the final sonar image;

[0089] In this embodiment, the acquired measured multi-beam sonar image datasets of the underwater terrain are two-dimensional image sonars of the BlueView MB2250-W-DL model detected according to the real underwater terrain;

[0090] In this embodiment, the specific process of constructing the sonar image style model is as follows: Segment the images in the measured dataset (such as Figure 4 ) to separate the foreground and background of the image, remove the foreground objects in the image, generate a background image dataset, and construct a sonar image style model according to the dataset; Bring the dataset into the sonar image style model constructed by the fast image style transfer algorithm optimized based on the offline model for training. In this embodiment, the single-style single-transfer algorithm is used;

[0091] Perform Gaussian blur on the simulated original sonar image; Bring the Gaussian-blurred original sonar image into the sonar image style model to generate the final simulated sonar image.

[0092] For the step numbers in the above method embodiments, they are only set for the convenience of elaboration and explanation, and no limitation is imposed on the order between the steps. The execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0093] The above is a specific description of the preferred embodiment of the present invention. However, the present invention is not limited to the described embodiment. Those skilled in the art can make various equivalent deformations or substitutions without departing from the spirit of the present invention, and these equivalent deformations or substitutions are all included within the scope defined by the claims of this application.

Claims

1. An underwater three-dimensional terrain and multi-beam image sonar data simulation system, characterized in that The system includes: an acquisition unit, a first model construction unit, a second model construction unit, a third model construction unit, and a data simulation unit; The acquisition unit, wherein the synthetic aperture sonar in the acquisition unit acquires a high-definition two-dimensional sonar image of the underwater terrain and converts it into a grayscale image, and the multi-beam sonar in the acquisition unit acquires a measured multi-beam sonar image dataset of the underwater terrain; The first model construction unit constructs an underwater three-dimensional point cloud terrain based on the two-dimensional sonar image; The data simulation unit simulates the navigation data of the underwater vehicle and the detection method of the multi-beam image sonar; The second model construction unit constructs a sonar image simulation model based on the simulation result and the underwater three-dimensional point cloud terrain to obtain an original simulated sonar image; The third model construction unit constructs a sonar image style model based on the measured multi-beam sonar image dataset of the underwater terrain, and brings the original simulated sonar image into the sonar image style model to construct a final sonar image.

2. The system according to claim 1, wherein The method for obtaining the original simulated sonar image is specifically as follows: According to the simulation, obtain the movement trajectory of the underwater vehicle underwater, and bring the simulated multi-beam image sonar detection method into the movement trajectory; perform three-dimensional transformation according to the simulated orientation angle, find the nearest neighbor points between the detection rays of the simulated multi-beam image sonar and the simulated three-dimensional point cloud terrain, and obtain the distance between the detection points of each ray and the simulated multi-beam image sonar; determine the width of the simulated two-dimensional multi-beam image sonar image according to the detection lateral opening angle of the simulated multi-beam image sonar, and determine the height of the simulated two-dimensional multi-beam image sonar image and the distance represented by one pixel in the column pixels according to the detection distance of the simulated multi-beam image sonar; bring the distance between the simulated detection target points and the simulated multi-beam image sonar into the two-dimensional image to generate an initial simulated sonar image.

3. The system according to claim 1, wherein The method for simulating the detection method of the multi-beam image sonar is specifically as follows: Determine the longitudinal opening angle and the lateral opening angle of the simulated multi-beam sonar, emit a ray every 1° with the simulated multi-beam sonar as the origin in the lateral opening angle, and emit a ray every 0.1 degree in the longitudinal opening angle.

4. The system according to claim 1 or 2, wherein The navigation data includes: timestamp data, navigation azimuth data, attitude angle data of the underwater vehicle, depth data of the underwater vehicle, and navigation speed data.

5. A method for simulating underwater three-dimensional terrain and multibeam image sonar data, characterized in that, It includes the following steps: Step 1, the synthetic aperture sonar in the acquisition unit acquires a high-definition two-dimensional sonar image of the underwater terrain and converts it to the grayscale space as the first grayscale image; Step 2. The first model construction unit establishes a target simulation model according to the KLSG dataset and the first grayscale image, and obtains an underwater two-dimensional terrain image; Step 3. The first model construction unit constructs an underwater three-dimensional point cloud terrain according to the underwater two-dimensional terrain image; Step 4. The data simulation unit simulates the navigation data of the underwater vehicle and the detection method of the multi-beam image sonar to obtain a simulation result; Step 5. The second model construction unit combines the simulation result with the underwater three-dimensional point cloud terrain to construct a sonar image simulation model to obtain an original simulated sonar image; Step 6. The multi-beam sonar in the acquisition unit acquires a set of measured multi-beam sonar image datasets of underwater terrain. The third model construction unit constructs a sonar image style model based on the measured multi-beam sonar image datasets of underwater terrain, and brings the original simulated sonar image into the sonar image style model to obtain the final sonar image.

6. The method according to claim 5, characterized in that, The first model construction unit constructs a target simulation model based on the KLSG dataset and the first grayscale image, and obtains an underwater two-dimensional terrain image. The specific method is as follows: Randomly select an image from the KLSG dataset as the image to be processed, convert it to the grayscale space as the second grayscale image; segment the second grayscale image and perform binarization processing to obtain a binarized foreground image; use the binarized foreground image as a mask and combine it with the image to be processed to obtain the foreground image of the target area; bring the binarized foreground image into the CANNY algorithm to obtain the edge information of the binarized foreground image. Brighten the foreground image and randomly add it to the first grayscale image to obtain an initial fused image, and perform Gaussian filtering along the image edge in the initial fused image according to the edge information of the above binarized foreground image to generate the final underwater terrain two-dimensional image.

7. The method according to claim 5, wherein Based on the underwater two-dimensional terrain image, construct an underwater three-dimensional point cloud terrain. The specific method is as follows: According to the simulation requirements, set the number of point clouds corresponding to the image pixel points. Respectively use the abscissa and ordinate of the pixel points of the image as the abscissa and ordinate of the three-dimensional point cloud, and extract the grayscale value of the two-dimensional terrain image as the height of the three-dimensional point cloud to establish an underwater three-dimensional point cloud terrain. The specific mapping relationship is shown in the following formula: Where: X map is the abscissa of the three-dimensional point cloud; Y map is the ordinate of the three-dimensional point cloud; Z map is the height of the three-dimensional point cloud; X img is the abscissa of the image, Y img is the ordinate of the image, is the gray value corresponding to the x and y coordinates in the image; n and m are the adjustment coefficients of the three-dimensional terrain; After generating the initial three-dimensional point cloud, use the method of linear interpolation to interpolate between adjacent point clouds. Specifically, it is shown in the following formula: Wherein: is the abscissa of the j-th point cloud; is the ordinate of the j-th point cloud, is the height of the j-th point cloud, and l is the number of differences between adjacent point clouds.

8. The method according to claim 5, wherein The multi-beam sonar in the acquisition unit acquires a set of measured multi-beam sonar image datasets of underwater terrain. The third model construction unit constructs a sonar image style model based on the measured multi-beam sonar image datasets of underwater terrain to obtain the final sonar image. The specific method is as follows: Segment the images in a set of measured multi-beam sonar image datasets of underwater terrain acquired by the multi-beam sonar in the acquisition unit, separate the image foreground and the image background, generate an image background dataset, and construct a sonar image style model according to the image background dataset; bring the image background dataset into the sonar image style model for training; perform Gaussian blur on the original simulated sonar image, and bring the Gaussian-blurred original sonar image into the sonar image style model to generate the final simulated sonar image.

Images