Gradient standard deviation based multi-beam line layout method, system and device

Abstract

The present application belongs to the technical field of seabed topography mapping, and particularly relates to a multi-beam survey line layout method, system and device based on gradient standard deviation. The present application optimizes partition design, first partitions through coverage times, then performs secondary partitioning after recognizing the internal topographic heterogeneity of the partition by using gradient standard deviation. This operation can recognize and process topography that is difficult to process by existing methods such as isobath ring-shaped areas by introducing gradient information, thereby avoiding unnecessary overlap of survey lines in such areas. In secondary partitioning, the weight is adaptively calculated based on the variance of coverage times and the variance of gradient amplitude in the partition, a two-dimensional feature vector is constructed for clustering, the influence of coverage redundancy and topographic complexity on the partition is automatically balanced, and the mapping coverage demand and seabed topographic features are taken into account.

Description

Technical Field

[0001] This invention belongs to the field of seabed topographic mapping technology, specifically relating to a method, system, and equipment for laying out multibeam survey lines based on gradient standard deviation. Background Technology

[0002] Patent CN118171586A discloses a method for laying out multibeam survey lines. The core idea is to calculate the grid coverage number matrix after dividing the sea area into grids, and then perform full-domain spatial clustering based on the coverage number to obtain first-level partitions. Survey lines are then laid out based on the partitioning results. This patented method only uses the coverage number as a single indicator for partitioning and does not consider the differences in seabed topography. The applicant's research found that in some areas after partitioning based on the coverage number model, especially in areas with ring-shaped isobaths and alternating steep and gentle topography, the water depth gradient varies greatly within the same partition. Unifying the spacing and orientation of survey lines will cause a large number of survey lines to overlap in the ring-shaped areas, resulting in surveying redundancy. Summary of the Invention

[0003] To address the problems in the background technology, this invention provides a method, system, and equipment for multibeam survey line layout based on gradient standard deviation.

[0004] The technical solution of the present invention is as follows: This invention provides a method for laying out multibeam survey lines based on gradient standard deviation, including: S1: Divide the sea area map of the task into grids and generate a two-dimensional water depth matrix; calculate the coverage number matrix based on the coverage number model, perform clustering and partitioning operations, and obtain multiple first-level partitions; S2: The central difference method is used to calculate the X and Y direction water depth gradient values ​​of all grids in the two-dimensional water depth matrix, and the X direction gradient matrix and Y direction gradient matrix are generated. For each primary zone, the larger of the standard deviations of the gradient in the X direction and the standard deviations of the gradient in the Y direction within the grid of the primary zone is selected as the terrain metric value for that primary zone. Collect all topographic measurement values ​​of the first-level partitions and sort them in ascending order. Calculate the interval between adjacent topographic measurement values ​​and take the midpoint value of the adjacent topographic measurement values ​​with the largest interval value as the gradient standard deviation threshold. For each primary partition, if the terrain metric value of the primary partition is greater than the gradient standard deviation threshold, then the primary partition is marked as to be partitioned into secondary partitions. S3: For each partition to be re-partitioned, normalize the coverage times and gradient magnitudes of all grids within the partition, construct a two-dimensional feature vector with adaptive weights or preset weights, cluster the two-dimensional feature vectors, and update the partitioning results. S4: The updated partitioning result is used as a first-level partition. S2 is executed until the preset conditions are met.

[0005] Based on the above-described multibeam survey line layout method based on gradient standard deviation, the adaptive weight in S3 is the proportion of the variance of the normalized gradient magnitude in the sum of the variance of the normalized coverage number and the variance of the normalized gradient magnitude.

[0006] Based on the multibeam survey line layout method based on gradient standard deviation described above, the gradient magnitude mentioned in S3 is obtained through the formula: To calculate; In the formula, Represents a grid The gradient magnitude; , They are grids Water depth gradient values ​​in the X and Y directions.

[0007] Based on the multibeam survey line layout method based on gradient standard deviation described above, the updating of the partitioning results in S3 includes boundary optimization of the secondary partitions obtained by clustering, as follows: For the second-level partitions obtained by clustering, the gradient magnitude is calculated grid by grid, and the contour lines with gradient magnitudes higher than a preset threshold are extracted as gradient contour lines; Traverse the boundary grid of the secondary partition, calculate the distance from the boundary grid to the nearest gradient contour line, and if the distance is less than the preset distance threshold, assign the boundary grid to the secondary partition with the largest area enclosed by the gradient contour line.

[0008] Furthermore, the multibeam survey line layout method based on gradient standard deviation described above also includes boundary optimization based on gradient local consistency, as follows: Take a preset local window to construct the gradient feature vector of the boundary grid, calculate the Euclidean distance between the gradient feature vector of the boundary grid and the average feature vector of each adjacent region, and divide the boundary grid into the partition with the smallest Euclidean distance.

[0009] Furthermore, based on the above-described multibeam survey line layout method based on gradient standard deviation, it also includes partitioning the boundary-optimized areas, calculating the ratio of the area of ​​each region to the total area of ​​the mission sea area map, and if the ratio is less than a preset ratio, merging the region into the adjacent partition with the smallest difference between the midpoint value of the gradient standard deviation of that region and the adjacent partition.

[0010] Based on the above-described multibeam survey line layout method based on gradient standard deviation, S1 specifically refers to: A three-dimensional coordinate system is established with the geometric center of the sea area to be tested as the origin. The sea area is gridded using a unit square grid to generate a two-dimensional water depth matrix. The coverage number model is combined to calculate the coverage number of each grid to generate a global coverage number matrix. Clustering and partitioning operations are performed on the global coverage number matrix to obtain multiple first-level partitions.

[0011] Based on the multibeam survey line layout method based on gradient standard deviation described above, after the preset conditions are met, it also includes: The spacing between survey lines in a zone is determined based on the maximum depth measurement width of the zone. Multibeam survey lines are then evenly distributed according to the total number of times the zone is covered, taking into account the direction of the isobaths, the direction of the tidal current, the channel axis, or the direction of the longest side of the zone.

[0012] The present invention also provides a multi-beam survey line layout system based on gradient standard deviation, comprising: The first-level partitioning module is used to divide the sea area map of the task into grids and generate a two-dimensional water depth matrix; it calculates the coverage number matrix based on the coverage number model, performs clustering partitioning operations, and obtains multiple first-level partitions. Gradient standard deviation discrimination module: The central difference method is used to calculate the water depth gradient values ​​in the X and Y directions of all grids in the two-dimensional water depth matrix, and generate the gradient matrix in the X direction and the gradient matrix in the Y direction. For each primary zone, the larger of the standard deviations of the gradient in the X direction and the standard deviations of the gradient in the Y direction within the grid of the primary zone is selected as the terrain metric value for that primary zone. Collect all topographic measurement values ​​of the first-level partitions and sort them in ascending order. Calculate the interval between adjacent topographic measurement values ​​and take the midpoint value of the adjacent topographic measurement values ​​with the largest interval value as the gradient standard deviation threshold. For each primary partition, if the terrain metric value of the primary partition is greater than the gradient standard deviation threshold, then the primary partition is marked as to be partitioned into secondary partitions. Second-level partitioning module: For each partition to be partitioned, the coverage count and gradient magnitude of all grids in the partition are normalized, and a two-dimensional feature vector is constructed according to adaptive weights or preset weights. The two-dimensional feature vector is then clustered to update the partitioning results. Iteration module: The updated partitioning result is used as a first-level partition and enters the gradient standard deviation discrimination module until the preset conditions are met.

[0013] The present invention also provides a multi-beam survey line layout device based on gradient standard deviation, including a processor and a memory, wherein the processor executes the computer program stored in the memory to implement the multi-beam survey line layout method based on gradient standard deviation.

[0014] Beneficial effects

[0015] The multibeam survey line layout method based on gradient standard deviation provided by this invention optimizes the zoning design. It first zones the lines based on the number of coverage iterations, then uses the gradient standard deviation to identify the topographic heterogeneity within each zone before performing a second zoning. This operation, by introducing gradient information, can identify and handle terrain features that are difficult to process using existing methods, such as isobath rings, avoiding unnecessary overlap of survey lines in these areas. During the second zoning, weights are adaptively calculated based on the variance of the number of coverage iterations and the variance of the gradient magnitude within each zone, constructing a two-dimensional feature vector for clustering. This automatically balances the impact of coverage redundancy and topographic complexity on the zoning, taking into account both surveying coverage requirements and seabed topographic features.

[0016] To address the issue that secondary partitioning results in partition boundaries crossing areas with gentle terrain or the same isobath area, this invention utilizes the obtained water depth gradient information to adjust the boundary grid assignment, making the partition boundaries fit as closely as possible to areas with drastic gradient changes, thereby improving the physical rationality of partitioning and the efficiency of survey line layout. Detailed Implementation

[0017] The following examples are intended to illustrate the present invention, and not to further limit the invention.

[0018] Example 1

[0019] This invention provides a method for laying out multibeam survey lines based on gradient standard deviation, comprising: S1: Divide the sea area map of the mission into grids and generate a two-dimensional water depth matrix; calculate the coverage number matrix based on the coverage number model, perform clustering and partitioning operations, and obtain multiple first-level partitions.

[0020] Specifically: A three-dimensional coordinate system is established with the geometric center of the sea area to be measured as the origin. The sea area is then meshed using a unit square grid to generate a two-dimensional water depth matrix. Combined with the coverage count model, the coverage count of each grid is calculated to generate a global coverage count matrix; clustering and partitioning operations are performed on the global coverage count matrix to obtain multiple first-level partitions.

[0021] When establishing a three-dimensional coordinate system, optionally, the vertical axis y-axis passes through the origin of the coordinate system and is parallel to the north-south direction, with due north being the positive direction of the y-axis; the horizontal axis x-axis passes through the origin of the coordinate system and is perpendicular to the y-axis, with due east being the positive direction of the x-axis; the vertical direction z-axis passes through the origin of the coordinate system and is positive downwards.

[0022] Two-dimensional water depth matrix , where m and n are the number of grid cells in the x-axis and y-axis directions, respectively.

[0023] Coverage Count Matrix Each element This indicates that, given the multibeam transducer opening angle θ, seawater depth, and seabed slope conditions... The theoretical number of times a unit surface element is covered by a multibeam strip when the grid point is the center is given below. For specific calculations, please refer to the published patent CN118171586A.

[0024] The coverage frequency matrix can reflect the coverage redundancy of different locations within the task area under a unified survey line layout scheme: areas with high coverage frequency are prone to excessive overlap, while areas with low coverage frequency face the risk of missed measurements.

[0025] S2: The central difference method is used to calculate the X and Y direction water depth gradient values ​​of all grids in the two-dimensional water depth matrix, and the X direction gradient matrix and Y direction gradient matrix are generated. For each primary zone, the larger of the standard deviations of the gradient in the X direction and the standard deviations of the gradient in the Y direction within the grid of the primary zone is selected as the terrain metric value for that primary zone. Collect all topographic measurement values ​​of the first-level partitions and sort them in ascending order. Calculate the interval between adjacent topographic measurement values ​​and take the midpoint value of the adjacent topographic measurement values ​​with the largest interval value as the gradient standard deviation threshold. For each primary partition, if the terrain metric value of the primary partition is greater than the gradient standard deviation threshold, then the primary partition is marked as to be partitioned into secondary partitions.

[0026] Specifically, the generation process of the gradient matrix in the X direction and the gradient matrix in the Y direction is as follows.

[0027] Two-dimensional water depth matrix The grid point function value is The grid spacing is , Calculate the water depth gradient values ​​in the X and Y directions, and denot them as follows: , .

[0028] This leads to the gradient matrix. and .

[0029] Preferably, in the specific implementation process, in addition to directly processing the first-level partition, the region whose area in the first-level partition is not less than a preset ratio can be selected as a candidate partition for gradient standard deviation discrimination.

[0030] Thus, for each candidate partition, the standard deviation of the gradient in the X direction for all grids within it is calculated. and the standard deviation of the gradient in the Y direction : , .

[0031] in This represents the number of valid pixels within the candidate partition. , This is the mean.

[0032] In the process of determining the gradient standard deviation threshold, if the number of candidate partitions is insufficient or the terrain measurement values ​​of all candidate partitions are equal and there is no positive interval, the gradient standard deviation threshold cannot be determined by the determination method provided by this invention. In this case, the threshold can be set according to actual needs.

[0033] Finally, by comparison, we can identify areas that do not require secondary zoning (areas with consistent terrain changes) and areas that require secondary zoning (areas with significant terrain differences).

[0034] This invention performs secondary partitioning based on primary partitioning. First, it distinguishes the differences in basic coverage redundancy by the number of coverages (partitioning method disclosed in patent CN118171586A). Then, it uses gradient standard deviation to identify the terrain heterogeneity within the partition and performs secondary splitting of partitions with obvious terrain undulations, thereby reducing the redundancy of overlapping survey lines in the ring contour line area.

[0035] S3: For each partition to be re-partitioned, normalize the coverage times and gradient magnitudes of all grids within the partition, construct a two-dimensional feature vector with adaptive weights or preset weights, cluster the two-dimensional feature vectors, and update the partitioning results.

[0036] The gradient magnitude is given by the formula: This can be calculated.

[0037] In the formula, Represents a grid The gradient magnitude; , They are grids Water depth gradient values ​​in the X and Y directions.

[0038] Adaptive weights or preset weights are used to adjust the relative importance of coverage times and gradient magnitudes in the secondary partitioning. Preferably, the adaptive weight is the proportion of the variance of the normalized gradient magnitude in the sum of the variance of the normalized coverage times and the variance of the normalized gradient magnitude.

[0039] This means that if the number of coverage times varies greatly within a partition, the partition should be determined more by the number of coverage times; if the gradient magnitude varies greatly, the partition should be determined more by the terrain complexity.

[0040] In addition, the preset weight can be 0.5, which is applicable to most general sea areas; if those skilled in the art have prior knowledge of the specific sea area characteristics, they can also manually set other values.

[0041] Finally, the K-means clustering algorithm is applied within the partition to divide the original partition into V sub-regions, and the entire partition result is updated.

[0042] This invention adaptively calculates weights based on the variance of coverage times and gradient magnitude within a partition, constructs a two-dimensional feature vector for clustering, and automatically balances the impact of coverage redundancy and terrain complexity on the partition, taking into account both surveying and mapping coverage requirements and seabed topographic features.

[0043] After the sub-regions are obtained by secondary clustering, the grid has been assigned new labels. However, the clustering algorithm does not consider the physical boundary characteristics of the seabed topography, which may cause the partition boundary to pass through the flat topography area or the same isobath area, which is not conducive to the subsequent survey line layout.

[0044] Furthermore, updating the partitioning results includes boundary optimization of the secondary partitions obtained from clustering, as follows: For the secondary partitions obtained by clustering, the gradient magnitude is calculated grid by grid, and the contour lines with gradient magnitudes higher than a preset threshold are extracted as gradient contour lines, i.e. candidate physical boundaries. Traverse the boundary grid of the secondary partition, calculate the distance from the boundary grid to the nearest gradient contour line, and if the distance is less than the preset distance threshold, assign the boundary grid to the secondary partition with the largest area enclosed by the gradient contour line.

[0045] The contour lines are based on the gradient magnitude matrix. Generate contour sets within the local region that needs optimization.

[0046] The preset threshold can be taken as the upper quartile of the gradient magnitude of all grids in the region, or it can be set by those skilled in the art based on their experience with the degree of topographic relief in the sea area.

[0047] In the specific implementation process, the boundary projection and adjustment based on gradient contour lines are performed as follows.

[0048] Suppose that the current secondary partitioning results in a set of subregions. Identify all grids located on the partition boundaries to form the boundary grid set. .right Grid in Calculate the distance from the nearest candidate physical boundary. If this distance is less than a preset distance threshold, then the mesh is... Reassign to the sub-region with the largest area within the area enclosed by the contour lines of that gradient; if the contour lines do not clearly define the area, then according to... The region with the smallest median gradient standard deviation among all sub-regions within the neighborhood is assigned to the sub-region.

[0049] After processing, morphological closing operations (dilation followed by erosion, with a 3×3 mesh as the structuring element) are used to fill any tiny gaps that may appear on the boundary and eliminate jagged isolated meshes. This step can bring the artificial cluster boundaries closer to the natural fault lines (high gradient zones) of the terrain, avoiding the partition boundaries from cutting off the ring areas of the isobaths in the gentle terrain.

[0050] If there are no closed or significant contour lines in the region that satisfy the condition that "the gradient magnitude is higher than the preset threshold", then proceed directly to the subsequent boundary optimization operation based on local gradient consistency.

[0051] Next, a gradient local consistency check is performed on the meshes still at the partition boundaries. Preferably, the boundary optimization based on gradient local consistency is performed as follows: Take a preset local window to construct the gradient feature vector of the boundary grid, calculate the Euclidean distance between the gradient feature vector of the boundary grid and the average feature vector of each adjacent region, and divide the boundary grid into the partition with the smallest Euclidean distance.

[0052] The specific implementation process is as follows.

[0053] For each boundary grid Take Centered window( (Choose 3 or 5), calculate the standard deviation of the gradient in the x-direction within this window. y-direction gradient standard deviation ,as well as Its own gradient magnitude , forming local gradient feature vectors .in Boundary grid The number of effective pixels within.

[0054] .

[0055] right All adjacent partitions Calculate the distance within the window of this partition. The most recent one that belongs to this partition The average eigenvector of the region The Euclidean distance metric is used. and Differences: .

[0056] If there exists a certain adjacent partition such that Less than Feature distance of the current partition , express The Euclidean distance between the average gradient features of the current partition will then be... Reassigned to make Minimum partition. Traverse all boundary grids, update labels after completing one round of redistribution, and repeat this process until no grid changes ownership or the preset iteration limit is reached (usually ≤5 times).

[0057] By fine-tuning local gradient similarity, unreasonable boundaries that may remain based on gradient contour lines are further eliminated and adjusted, ensuring that the terrain change patterns on both sides of the boundary have significant differences.

[0058] Boundary optimization may result in isolated sub-regions with very small areas.

[0059] Therefore, for the partitioned areas after boundary optimization, the ratio of each region's area to the total area of ​​the task's sea map is calculated. If the ratio is less than a preset ratio, the region is merged into the adjacent partition with the smallest difference between its gradient standard deviation midpoint value and that of the previous region. If multiple adjacent regions have the same gradient standard deviation midpoint value, they are merged into the region with the larger area.

[0060] The above operations utilize the obtained water depth gradient information to adjust the boundary grid assignments, so that the partition boundaries fit as closely as possible to areas of drastic gradient changes, thereby improving the physical rationality of the partitions and the efficiency of survey line layout.

[0061] S4: The updated partitioning result is used as a first-level partition. S2 is executed until the preset conditions are met.

[0062] The preset condition may be that the gradient standard deviation of all partitions does not exceed the gradient standard deviation threshold.

[0063] After the preset conditions are met, it also includes: The spacing between survey lines in a zone is determined based on the maximum depth measurement width of the zone. Multibeam survey lines are then evenly distributed according to the total number of times the zone is covered, taking into account the direction of the isobaths, the direction of the tidal current, the channel axis, or the direction of the longest side of the zone.

[0064] The multibeam survey line layout method based on gradient standard deviation provided by this invention optimizes the zoning design. It first zones the lines based on the number of coverage iterations, then uses the gradient standard deviation to identify the topographic heterogeneity within each zone before performing a second zoning. This operation, by introducing gradient information, can identify and handle terrain features that are difficult to process using existing methods, such as isobath rings, avoiding unnecessary overlap of survey lines in these areas. During the second zoning, weights are adaptively calculated based on the variance of the number of coverage iterations and the variance of the gradient magnitude within each zone, constructing a two-dimensional feature vector for clustering. This automatically balances the impact of coverage redundancy and topographic complexity on the zoning, taking into account both surveying coverage requirements and seabed topographic features.

[0065] The present invention also provides a multi-beam survey line layout system based on gradient standard deviation, comprising: The first-level partitioning module is used to divide the sea area map of the task into grids and generate a two-dimensional water depth matrix; it calculates the coverage number matrix based on the coverage number model, performs clustering partitioning operations, and obtains multiple first-level partitions. Gradient standard deviation discrimination module: The central difference method is used to calculate the water depth gradient values ​​in the X and Y directions of all grids in the two-dimensional water depth matrix, and generate the gradient matrix in the X direction and the gradient matrix in the Y direction. For each primary zone, the larger of the standard deviations of the gradient in the X direction and the standard deviations of the gradient in the Y direction within the grid of the primary zone is selected as the terrain metric value for that primary zone. Collect all topographic measurement values ​​of the first-level partitions and sort them in ascending order. Calculate the interval between adjacent topographic measurement values ​​and take the midpoint value of the adjacent topographic measurement values ​​with the largest interval value as the gradient standard deviation threshold. For each primary partition, if the terrain metric value of the primary partition is greater than the gradient standard deviation threshold, then the primary partition is marked as to be partitioned into secondary partitions. Second-level partitioning module: For each partition to be partitioned, the coverage count and gradient magnitude of all grids in the partition are normalized, and a two-dimensional feature vector is constructed according to adaptive weights or preset weights. The two-dimensional feature vector is then clustered to update the partitioning results. Iteration module: The updated partitioning result is used as a first-level partition and enters the gradient standard deviation discrimination module until the preset conditions are met.

[0066] The present invention also provides a multi-beam survey line layout device based on gradient standard deviation, including a processor and a memory, wherein the processor executes the computer program stored in the memory to implement the multi-beam survey line layout method based on gradient standard deviation.

[0067] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for laying out multibeam survey lines based on gradient standard deviation, characterized in that, include: S1: Divide the sea area map of the mission into grids and generate a two-dimensional water depth matrix; The coverage number matrix is ​​calculated based on the coverage number model, and clustering and partitioning operations are performed to obtain multiple first-level partitions. S2: The central difference method is used to calculate the X and Y direction water depth gradient values ​​of all grids in the two-dimensional water depth matrix, and the X direction gradient matrix and Y direction gradient matrix are generated. For each primary zone, the larger of the standard deviations of the gradient in the X direction and the standard deviations of the gradient in the Y direction within the grid of the primary zone is selected as the terrain metric value for that primary zone. Collect all topographic measurement values ​​of the first-level partitions and sort them in ascending order. Calculate the interval between adjacent topographic measurement values ​​and take the midpoint value of the adjacent topographic measurement values ​​with the largest interval value as the gradient standard deviation threshold. For each primary partition, if the terrain metric value of the primary partition is greater than the gradient standard deviation threshold, then the primary partition is marked as to be partitioned into secondary partitions. S3: For each partition to be re-partitioned, normalize the coverage times and gradient magnitudes of all grids within the partition, construct a two-dimensional feature vector with adaptive weights or preset weights, cluster the two-dimensional feature vectors, and update the partitioning results. S4: The updated partitioning result is used as a first-level partition. S2 is executed until the preset conditions are met.

2. The multibeam survey line layout method based on gradient standard deviation according to claim 1, characterized in that, The adaptive weights mentioned in S3 are the proportion of the variance of the normalized gradient magnitude in the sum of the variance of the normalized coverage number and the variance of the normalized gradient magnitude.

3. The multibeam survey line layout method based on gradient standard deviation according to claim 1, characterized in that, The gradient magnitude mentioned in S3 is obtained through the formula: To calculate; In the formula, Represents a grid The gradient magnitude; , They are grids Water depth gradient values ​​in the X and Y directions.

4. The multibeam survey line layout method based on gradient standard deviation according to claim 1, characterized in that, S3 describes updating the partitioning results, including boundary optimization of the secondary partitions obtained from clustering, as follows: For the second-level partitions obtained by clustering, the gradient magnitude is calculated grid by grid, and the contour lines with gradient magnitudes higher than a preset threshold are extracted as gradient contour lines; Traverse the boundary grid of the secondary partition, calculate the distance from the boundary grid to the nearest gradient contour line, and if the distance is less than the preset distance threshold, assign the boundary grid to the secondary partition with the largest area enclosed by the gradient contour line.

5. The multibeam survey line layout method based on gradient standard deviation according to claim 4, characterized in that, It also includes boundary optimization based on gradient local consistency, as follows: Take a preset local window to construct the gradient feature vector of the boundary grid, calculate the Euclidean distance between the gradient feature vector of the boundary grid and the average feature vector of each adjacent region, and divide the boundary grid into the partition with the smallest Euclidean distance.

6. The multibeam survey line layout method based on gradient standard deviation according to claim 5, characterized in that, It also includes optimizing the boundaries of the partitions, calculating the ratio of the area of ​​each region to the total area of ​​the mission sea map, and if the ratio is less than the preset ratio, merging the region into the adjacent partition with the smallest difference between the midpoint value of the gradient standard deviation of the region and the adjacent partition.

7. The multibeam survey line layout method based on gradient standard deviation according to claim 1, characterized in that, S1 specifically refers to: A three-dimensional coordinate system is established with the geometric center of the sea area to be tested as the origin. The sea area is gridded using a unit square grid to generate a two-dimensional water depth matrix. The coverage number model is combined to calculate the coverage number of each grid to generate a global coverage number matrix. Clustering and partitioning operations are performed on the global coverage number matrix to obtain multiple first-level partitions.

8. The multibeam survey line layout method based on gradient standard deviation according to claim 1, characterized in that, After the preset conditions are met, it also includes: The spacing between survey lines in a zone is determined based on the maximum depth measurement width of the zone. Multibeam survey lines are then evenly distributed according to the total number of times the zone is covered, taking into account the direction of the isobaths, the direction of the tidal current, the channel axis, or the direction of the longest side of the zone.

9. A multi-beam survey line layout system based on gradient standard deviation, characterized in that, include: The first-level partitioning module is used to divide the sea area map of the mission into grids and generate a two-dimensional water depth matrix. The coverage number matrix is ​​calculated based on the coverage number model, and clustering and partitioning operations are performed to obtain multiple first-level partitions. Gradient standard deviation discrimination module: The central difference method is used to calculate the water depth gradient values ​​in the X and Y directions of all grids in the two-dimensional water depth matrix, and generate the gradient matrix in the X direction and the gradient matrix in the Y direction. For each primary zone, the larger of the standard deviations of the gradient in the X direction and the standard deviations of the gradient in the Y direction within the grid of the primary zone is selected as the terrain metric value for that primary zone. Collect all topographic measurement values ​​of the first-level partitions and sort them in ascending order. Calculate the interval between adjacent topographic measurement values ​​and take the midpoint value of the adjacent topographic measurement values ​​with the largest interval value as the gradient standard deviation threshold. For each primary partition, if the terrain metric value of the primary partition is greater than the gradient standard deviation threshold, then the primary partition is marked as to be partitioned into secondary partitions. Second-level partitioning module: For each partition to be partitioned, the coverage count and gradient magnitude of all grids in the partition are normalized, and a two-dimensional feature vector is constructed according to adaptive weights or preset weights. The two-dimensional feature vector is then clustered to update the partitioning results. Iteration module: The updated partitioning result is used as a first-level partition and enters the gradient standard deviation discrimination module until the preset conditions are met.

10. A multi-beam survey line deployment device based on gradient standard deviation, characterized in that, It includes a processor and a memory, wherein the processor executes a computer program stored in the memory to implement the multibeam survey line layout method based on gradient standard deviation as described in any one of claims 1-8.