A method and system for modeling seafloor topography using a multi-resolution grid

Abstract

The application belongs to the technical field of seabed terrain information processing, and discloses a seabed terrain modeling method and system of a multi-resolution grid. The method comprises the following steps: generating a multi-resolution grid model based on a quadtree; obtaining the multi-resolution grid model according to a terrain complexity evaluation index, and acquiring a node subdivision evaluation index for the terrain complexity; in the process of multi-resolution grid display rendering, cracks appearing in the splicing of adjacent different resolution grids are eliminated by using the edge adding method. Through the combination of the quadtree storage structure and the multi-resolution grid, the number of high-precision grid points is reduced, thereby significantly reducing the spatial redundancy of data storage. While reducing the spatial redundancy, the application considers the terrain features and the slope through the subdivision evaluation function, can better retain the complex features of the terrain, and ensures the overall terrain accuracy.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of seabed topography information processing, and particularly relates to a seabed topography modeling method and system of a multi-resolution grid. BACKGROUND

[0002] Seabed topography is one of the most complex types of terrain on the earth's surface, including not only vast deep-sea plains, steep seabed ridges, and winding trenches, but also complex volcanoes, rift valleys, and sediment accumulation areas. As marine survey and development activities become increasingly frequent, there is a growing demand for high-precision, high-resolution seabed topography data in the entire deep-sea environment. Therefore, it is of great practical significance to establish a high-precision seabed topography model.

[0003] Currently, the methods of terrain modeling mainly include regular grid (Grid) and irregular triangle network (TIN). However, both have shortcomings. For complex topography, the grid density must be increased to improve accuracy, which will exponentially increase the amount of data. For flat terrain, the grid density needs to be reduced, otherwise there will be redundant data and waste of system resources. The data structure of the irregular triangle network (TIN) is complex and the calculation is large, making it difficult to achieve large-scale seabed topography modeling tasks.

[0004] For the construction of large-scale high-precision seabed three-dimensional terrain models, the existing modeling methods are mainly based on LOD, which renders different levels of terrain models through the relationship between the viewpoint distance and the in-out memory scheduling. However, for seabed topography data, most areas are relatively flat, and only some areas have large terrain undulations. If the grid density is high, it will increase the occupation of memory space, and if the grid density is low, it will lose the terrain accuracy. Using LOD and Out-of-Core ideas based on the relationship between the viewpoint and the viewing distance requires a reasonable browsing rate to ensure the continuity of the terrain, and additional data needs to be filled in to ensure the continuity of LOD after determining the model of the lowest grid, further exacerbating the spatial redundancy.

[0005] In summary, the technical problems and technical defects of the existing technology are that in the problem of seabed topography modeling, the main factor affecting modeling is that seabed topography is one of the most complex types of terrain on the earth's surface, including not only vast deep-sea plains, steep seabed ridges, and winding trenches, but also complex volcanoes, rift valleys, and sediment accumulation areas. This complexity determines the high demand for spatial resolution of seabed topography. Traditional single-resolution grids cannot meet this demand, and multi-resolution grid models based on LOD will increase the occupation of memory space if the grid density is high, and will lose terrain accuracy if the grid density is low. SUMMARY

[0006] To overcome the problems existing in related technologies, this invention discloses a method and system for multi-resolution grid-based seabed topography modeling. The purpose of this invention is to propose a variable-resolution seabed topography organization model based on regular grids and quadtree structures. First, high-resolution and low-resolution grid models are generated separately. Based on the low-resolution grid model, the topography complexity within each grid is evaluated. A quadtree structure is used to encrypt grid data in areas with complex local topography to meet the high-precision display requirements of specific topography. This generates multiple regular grid models with mixed resolutions, reducing spatial redundancy caused by generating high-resolution grid models in areas with significant topographic feature variations. This achieves better 3D seabed topography modeling.

[0007] The technical solution is as follows: A method for modeling seabed topography using multi-resolution grids, comprising the following steps:

[0008] S1. Perform grid interpolation on the original discrete point cloud data to obtain data after visualization of regular grids; divide the data according to the terrain complexity evaluation index until it can no longer be divided or the grid cell corresponding to the high-resolution grid is divided, and then stop the segmentation to generate a multi-resolution grid model based on quadtree.

[0009] S2, Based on the terrain complexity evaluation index, a multi-resolution grid model is obtained, and node subdivision evaluation indexes for terrain complexity are acquired.

[0010] S3, based on the obtained node subdivision evaluation index for terrain complexity, eliminates cracks that appear when adjacent grids of different resolutions are spliced ​​during the multi-resolution grid display and rendering process by using the edge addition method.

[0011] In step S1, grid interpolation is performed on the original discrete point cloud data to obtain data visualized as a regular grid. This includes: performing gridding processing on the original discrete point cloud data using the inverse distance weighting method, and determining the high-resolution grid resolution d based on the original discrete point cloud density and terrain flatness. max and low-resolution grid resolution d min d max and d min Satisfy condition d max =d min ×2 n n is any integer greater than 1; the number of rows and columns of the low-resolution grid are X min Y min The number of rows and columns of the high-resolution grid are X max Y max The conditions for performing quadtree partitioning and mesh refinement on low-resolution meshes are as follows:

[0012] (X max -1)×2n = X min -1

[0013] (Y max -1) x 2 n = Y min -1

[0014] In the formula, n is the number of quadtree partitioning.

[0015] In step S1, the division is performed by the terrain complexity evaluation index, and the division is stopped until it can no longer be divided or the grid cell corresponding to the high-resolution grid is divided, including:

[0016] Each low-resolution grid is searched and retrieved, each low-resolution grid is taken as a root node of a quadtree, it is judged from each low-resolution grid whether the node subdivision evaluation index is satisfied, if the condition is not satisfied, the subdivision is not performed and the node is saved, otherwise the node is recursively divided into four sub-nodes, the grid elevation values generated by the four sub-nodes are obtained from the corresponding high-resolution grid, and each sub-node is then judged, until it can no longer be divided or the grid cell corresponding to the high-resolution grid is divided.

[0017] In step S2, the node subdivision evaluation index for the terrain complexity is obtained, including:

[0018] The terrain relief degree and the terrain slope are taken as the factors for the node subdivision evaluation, whether the node continues to be subdivided is judged by combining the two factors, and the definitions of the terrain relief degree and the terrain slope are as follows:

[0019] Terrain relief degree:

[0020] Δh = h max - h min

[0021] In the formula, Δh is the maximum difference of the elevation in the grid, h max is the maximum elevation value in the current grid cell, and h min is the minimum elevation value in the current grid cell

[0022] Terrain slope:

[0023] S max = max(S

[0024] In the formula, S max is the maximum slope value of the current grid point, and S is the comprehensive slope value of the current grid point.

[0025] Further, for the calculation of the comprehensive slope value S of the current grid point, the grid slope of the grid point is calculated by the method of the surrounding eight neighborhoods, the elevation of the given grid point is Z5, the elevation change rate S of the grid point in the x and y directionsx and S y The calculation is shown in formula (1):

[0026]

[0027] In the formula, Z1-Z9 are grid point elevation values, and Δd is the grid size, and the calculation is ignored for the grid with no eight adjacent grids, and the value is defined as 0.

[0028]

[0029] In combination with the two indexes, a grid complexity function is defined as:

[0030]

[0031] In the formula, w h and w s are the weights of the height difference and the slope, T h and T s are the threshold values of the height difference and the slope, and C is a comprehensive evaluation index, when C exceeds a set threshold value, the grid is subdivided until the low-resolution grid is subdivided to the grid precision corresponding to the high-resolution grid.

[0032] In step S1, when the regular grid is visualized, the regular grid is divided into two triangular facets to be rendered and displayed.

[0033] In step S3, the method for changing the resolution of the grid comprises:

[0034] By increasing the number of triangular facets in the low-resolution grid, the resolution of the low-resolution grid is improved, the difference between the low-resolution grid and the high-resolution grid is reduced, and the gap between the two is eliminated.

[0035] By additionally connecting the low-resolution grid points and the high-resolution grid points, the connecting line additionally adds an edge to the joint part of the low-resolution grid and the high-resolution grid, and the triangular facet of the low-resolution grid is divided into two to eliminate the crack.

[0036] Another object of the present application is to provide a multi-resolution grid seabed terrain modeling system, which implements the multi-resolution grid seabed terrain modeling method, and the system comprises:

[0037] The multi-resolution grid model generation module based on the quadtree is used for grid interpolation on the original discrete point cloud data to obtain the data after regular grid visualization, and then the terrain complexity evaluation index is used for division until the division is stopped when the grid unit corresponding to the high-resolution grid is reached, and finally the multi-resolution grid model based on the quadtree is generated.

[0038] The node subdivision evaluation index acquisition module is used to obtain a multi-resolution grid model based on the terrain complexity evaluation index and to obtain node subdivision evaluation indexes for terrain complexity.

[0039] The crack elimination module is used to eliminate cracks that appear when adjacent grids of different resolutions are stitched together during the multi-resolution grid display and rendering process, based on the obtained node subdivision evaluation index for terrain complexity.

[0040] Furthermore, the system is mounted on a computer device, which includes at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor. When the processor executes the computer program, it implements the functions of the aforementioned multi-resolution grid seabed terrain modeling system.

[0041] Furthermore, the system is mounted on a computer-readable storage medium that stores a computer program, which, when executed by a processor, can realize the functions of the aforementioned multi-resolution grid-based seabed topography modeling system.

[0042] Combining all the above technical solutions, the beneficial effects of this invention are as follows:

[0043] This invention employs a quadtree structure, combining high-resolution and low-resolution grids. The low-resolution grid is dynamically subdivided based on terrain complexity to achieve localized densification, balancing high accuracy with low redundancy. The grid is dynamically subdivided using terrain undulation and slope as evaluation factors to avoid redundancy and ensure high accuracy in complex areas. Because adjacent areas of the multi-resolution grid have different resolutions, gaps appear at the junctions of high-resolution and low-resolution grids. These gaps are eliminated by adding connecting edges between low-resolution and high-resolution grid points, effectively splitting the low-resolution triangular facets in two.

[0044] This invention combines a quadtree storage structure with a multi-resolution grid to reduce the number of high-precision grid points, thereby significantly reducing spatial redundancy in data storage. While reducing spatial redundancy, this invention also considers terrain features and slope through a subdivision evaluation function, effectively preserving the complex features of the terrain and ensuring overall terrain accuracy. This invention reduces the number of triangles during rendering, lowering memory usage and improving rendering efficiency, making it particularly suitable for visualizing large-scale seabed terrain data. The method of this invention can dynamically adjust the grid resolution according to terrain complexity, ensuring high accuracy in locally complex areas while avoiding wasting computational resources in simpler areas. Attached Figure Description

[0045] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure;

[0046] Figure 1 This is a flowchart of the multi-resolution grid seabed topography modeling method provided in this embodiment of the invention;

[0047] Figure 2 This is a schematic diagram illustrating the principle of using the inverse distance weighting method to perform gridding processing on the original discrete point cloud data, as provided in this embodiment of the invention.

[0048] Figure 3 This is a schematic diagram of quadtree segmentation for each grid node provided in an embodiment of the present invention;

[0049] Figure 4 This is a schematic diagram of an eight-neighbor grid provided in an embodiment of the present invention;

[0050] Figure 5 This is a schematic diagram illustrating the crack generation principle provided in an embodiment of the present invention;

[0051] Figure 6 This is a diagram illustrating the crack treatment effect provided in an embodiment of the present invention;

[0052] Figure 7 Comparison diagram of the multi-resolution terrain model of this invention with different single-resolution terrain models. Detailed Implementation

[0053] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of the present invention. However, the present invention can be practiced in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0054] The innovation of this invention lies in the following: The multi-resolution grid-based seabed terrain modeling method and system provided in this embodiment are innovative in that: based on a quadtree structure, the grid resolution is dynamically adjusted according to the terrain complexity (undulation and slope), high-resolution fine representation is used in complex areas, and low resolution is retained in flat areas to reduce data redundancy; the boundaries of grids of different resolutions are connected by a crack elimination algorithm to ensure terrain continuity; experiments show that this method significantly reduces memory usage (by more than 80%) while maintaining rendering accuracy similar to high-resolution models, providing an efficient and adaptive solution for large-scale seabed terrain modeling.

[0055] Example 1, such as Figure 1As shown, the multi-resolution grid seabed topography modeling method provided in this embodiment of the invention includes:

[0056] S1. Perform grid interpolation on the original discrete point cloud data to obtain data after visualization of regular grids; divide the data according to the terrain complexity evaluation index until it can no longer be divided or the grid cell corresponding to the high-resolution grid is divided, and then stop the segmentation to generate a multi-resolution grid model based on quadtree.

[0057] S2, Based on the terrain complexity evaluation index, a multi-resolution grid model is obtained, and node subdivision evaluation indexes for terrain complexity are acquired.

[0058] S3, based on the obtained node subdivision evaluation index for terrain complexity, eliminates cracks that appear when adjacent grids of different resolutions are spliced ​​during the multi-resolution grid display and rendering process by using the edge addition method.

[0059] For example, in step S1, the final generation of a quadtree-based multi-resolution grid model includes:

[0060] Step 1.1: First, the original discrete point cloud data is interpolated to obtain data visualized as a regular grid. Commonly used spatial interpolation methods are the inverse distance weighting method and kriging. Kriging is suitable for data with complex spatial structures and considers the correlation between data, but its calculation is too complex. Although the inverse distance weighting method may have an excessive influence of outliers on the results, it has high computational efficiency and produces smooth surfaces, making it suitable for terrain modeling. Therefore, this invention adopts the inverse distance weighting method to grid the original discrete point cloud data. The gridding process is as follows: Figure 2 As shown. Specifically includes:

[0061] The high-resolution grid resolution d is determined based on the original discrete point cloud density and terrain flatness. max and low-resolution grid resolution d min , where d max and d min Condition d should be met max =d min ×2 n In the formula, n is any integer greater than 1, and let the number of rows and columns of the low-resolution grid be X. max Y max The number of rows and columns of the high-resolution grid are X max Y max To meet the requirement of further refining the low-resolution mesh through quadtree partitioning, the following conditions must be met:

[0062] (X max -1)×2 n =X min-1

[0063] (Y max -1)×2 n =Y min -1

[0064] In the formula, n is the number of quadtree partitioning, and n is 1 when the partitioning is once, and n is 2 when the partitioning is once again on this basis;

[0065] Therefore, the number of rows and columns of the low-resolution grid needs to be determined first, so as to calculate the number of rows and columns of the high-resolution grid.

[0066] In step 1.2, each low-resolution grid is traversed and searched, each low-resolution grid is taken as a root node of a quadtree, and then each low-resolution grid is judged. In the judgment process, the judgment is made according to the comprehensive evaluation index C. When the calculated grid C exceeds the set threshold, the grid is subdivided, and the subdivision is performed until the low-resolution grid is subdivided to the grid precision corresponding to the high-resolution grid. It is judged whether the evaluation index of node subdivision is satisfied. If the condition is not satisfied, the subdivision is not performed and the node is saved, otherwise the node is recursively divided into four sub-nodes, and the grid elevation values of the four sub-nodes are obtained from the corresponding high-resolution grid. In the quadtree-based multi-resolution grid model, when a node of a low-resolution grid needs to be subdivided, the node is divided into four sub-nodes. The elevation values of these sub-nodes are obtained from the corresponding high-resolution grid. The specific process is as follows: first, the original point cloud data is used to generate a high-resolution grid, and each grid point stores an accurate elevation value. The low-resolution grid is used for a relatively flat terrain area, and the grid spacing is larger. When the node of the low-resolution grid meets the subdivision condition, the system divides it into four sub-nodes. Each sub-node corresponds to a smaller area in the high-resolution grid, and its elevation value is directly obtained from the corresponding grid point in the high-resolution grid.

[0067] Each sub-node is then judged in turn until it can no longer be divided or is stopped from being divided into a grid unit corresponding to a high-resolution grid. The deeper the division, the higher the resolution of the corresponding area, and the higher the accuracy. Figure 3 Each grid node is quadtree-divided, that is, based on the node subdivision principle of the quadtree structure.

[0068] For example, in step S2, the node subdivision evaluation index for terrain complexity is obtained, including:

[0069] In order to avoid unnecessary subdivision and meet the requirement that the accuracy of the local area reaches the requirement, the terrain relief degree and the terrain slope are mainly used as the factors for evaluating the node subdivision in the present application. Whether the node continues to be subdivided is judged by combining the two factors. The definitions of the terrain relief degree and the terrain slope are as follows:

[0070] Terrain roughness:

[0071] Δh = h max -h min

[0072] In the formula, Δh is the maximum difference in elevation within the grid, h max is the maximum elevation value within the current grid cell, h min is the minimum elevation value within the current grid cell

[0073] Terrain slope:

[0074] S max = max(S)

[0075] In the formula, S max is the maximum slope value of the current grid point, and S is the comprehensive slope value of the current grid point.

[0076] In which, the calculation of S is calculated by the method of surrounding eight neighborhoods to calculate the grid slope of the grid point, as Figure 4 shown in the grid eight neighborhood diagram, the elevation of the given grid point is Z5, so the elevation change rate S x and S y of the grid point in the x, y direction

[0077]

[0078] In the formula, Z1-Z9 are all grid point elevation values, and Δd is the grid size. For the grid without eight neighborhoods, the calculation is ignored, and the value is defined as 0.

[0079]

[0080] Formula (1) and (2) are used to calculate the grid slope. Combined with the terrain roughness and terrain slope indexes, a grid complexity function is defined as:

[0081]

[0082] Formula (3) is used for the judgment condition of subsequent grid segmentation, in which w h and w s are the weights of the height difference and slope, T h and T s are the threshold values of the height difference and slope, and C is a comprehensive evaluation index. When C exceeds the set threshold value, the grid is subdivided until the low-resolution grid is subdivided to the grid precision corresponding to the high-resolution grid.

[0083] For example, in step S3, when the regular grid is visualized, it is usually rendered by dividing the regular grid into two triangular patches. Triangular patches can more flexibly display irregular elevation changes relative to the regular grid. The current mainstream graphics rendering hardware is very efficient in processing triangles, and the graphics API usually supports the drawing of triangular patches. Therefore, the present application selects to render and display by dividing the regular grid into two triangular patches.

[0084] In the process of multi-resolution grid display rendering, there may be different resolutions between adjacent grids. In the actual rendering process, the terrain will be discontinuous at the splicing place of two adjacent grids with different resolutions, which will be displayed as a crack in three dimensions. The main reason for the crack is that the two adjacent grids have different resolutions. The high-resolution grid will have several additional grid points at the joint with the low-resolution grid. Since the elevation value of the grid point is obtained by interpolation of the original point cloud data, the elevation value of the additional grid point is not on the line connecting the two grid points at the joint of the low-resolution grid. Therefore, a crack will be generated. In the rendering process, a triangular hole will appear at the joint of the different resolution grids, as shown in Figure 5 Crack generation principle.

[0085] To increase the vertices at the boundary of the low-resolution grid, additional triangular patches need to be added, which makes the implementation more complex and increases the complexity of the scene. However, it can handle possible cracks at any time, ensuring the accuracy of the terrain and being suitable for describing more complex terrain, and being more suitable for application in the application of the variable resolution grid proposed in the present application. The principle is to increase the number of triangular patches in the low-resolution grid to improve its resolution. In this way, the difference between the low-resolution grid and the high-resolution grid can be significantly reduced, thereby effectively eliminating the possible hole phenomenon at the joint of the two. Figure 6 That is, the crack processing effect diagram. By connecting the low-resolution grid points and the high-resolution grid points, such as Figure 6 the line connecting points A and B in the low-resolution grid and points C and D in the high-resolution grid, the low-resolution grid and the high-resolution grid are additionally processed at the joint. The low-resolution triangular patch is divided into two to eliminate the crack. The principle of adding edges is that when the resolution of adjacent two grids is different, the high-resolution grid may produce a hole in the area of the low-resolution grid, as shown on the right side. In order to eliminate these holes, the present application needs to add edges to the adjacent low-resolution grid to improve its resolution, thereby ensuring the continuity between the grids and the integrity of the data. Figure 5

[0086] The specific implementation steps of the edge adding process are as follows:

[0087] ​1. Detecting resolution differences between adjacent grids: First, the system will traverse all adjacent grids to detect if there is a resolution difference that is too large. If a difference that is too large is detected, edge addition processing is required.

[0088] 2. Determining crack areas: At the boundary of adjacent grids, the high-resolution grid will have some extra grid points, which will form an area of elevation discontinuity between the boundary points of the low-resolution grid, i.e. a crack area. The system will mark these crack areas and determine the additional edges that need to be added.

[0089] 3. Adding additional edges: In the crack area, the system will add additional edges between the boundary points of the low-resolution grid and the extra grid points of the high-resolution grid. Specifically, the boundary points of the low-resolution grid will be connected to the adjacent grid points of the high-resolution grid to form new triangular patches.

[0090] 4. Generating new triangular patches: Through the addition of additional edges, the system will divide the original regular grid into more triangular patches.

[0091] 5. Rendering process: In the rendering process, the system will preferentially draw these newly generated triangular patches to ensure smooth transitions between adjacent grids and avoid visual cracks or discontinuities.

[0092] In an embodiment 2, the multi-resolution grid-based seafloor terrain modeling system provided by the embodiment includes:

[0093] A multi-resolution grid model generation module based on a quadtree, configured to perform grid interpolation on original discrete point cloud data to obtain regular grid visualized data, and then divide the data based on a terrain complexity evaluation index until no further division is possible or the division reaches a grid unit corresponding to a high-resolution grid, and finally generate a multi-resolution grid model based on a quadtree;

[0094] A node subdivision evaluation index acquisition module, configured to obtain a multi-resolution grid model based on the terrain complexity evaluation index, and acquire a node subdivision evaluation index for terrain complexity;

[0095] A crack elimination module, configured to eliminate cracks that occur when adjacent grids of different resolutions are spliced, using an edge addition method based on the acquired node subdivision evaluation index for terrain complexity during multi-resolution grid display rendering.

[0096] In the above embodiments, the description of each embodiment focuses on different aspects. If a part is not described or recorded in detail in an embodiment, please refer to the relevant description of other embodiments.

[0097] In order to test the effect of the algorithm, the original point cloud data is generated into fixed resolution grids of 0.1m, 0.2m, 0.4m, 0.8m and 1.6m, and the method is used to generate grids with a resolution range of 0.1-1.6m, and the fixed resolution grids and the variable resolution grids are compared.

[0098] Firstly, the accuracy of the fixed resolution and variable resolution grids is evaluated by the idea of the check point method, that is, the distance between the position of the grid point and the original point cloud is evaluated, and the root mean square error RMSE is calculated, and the smaller the RMSE value is, the higher the accuracy is. Meanwhile, the number of grid points in each resolution grid, the number of triangular facets during rendering and the memory occupation during rendering of the regular grid are compared, and the experimental results are shown in Table 1.

[0099] Table 1 Comparison of statistical results of the multi-resolution grid and different single-resolution grids

[0100] Resolution RMSE / m Render memory footprint / MB 0.1m 0.0677 671 1.6m 0.3085 5 0.1 m - 1.6 m 0.0690 128

[0101] Through analysis of the data in Table 1, for the fixed resolution grid, the higher the generated grid resolution is, that is, the smaller the grid resolution value is, the smaller the corresponding RMSE value is, but the memory occupied during rendering will also increase exponentially. For the variable resolution grid, the resolution of the generated grid is between 0.1-1.6m, the corresponding RMSE is basically the same as that of the 0.1m resolution grid, and is much smaller than that of other resolution grids, and the number of grid points is greatly reduced compared with that of the 0.1m resolution grid, so the memory occupied during rendering is reduced by 80.93%, and is even less than that of the 0.2m resolution grid, greatly reducing the spatial redundancy of the seafloor terrain elevation model under the high resolution grid.

[0102] Secondly, the advantages and disadvantages of the method in the expression of terrain reality are observed and analyzed from the subjective visual angle. The fixed resolution grids of 0.1m and 1.6m and the variable resolution grids of 0.1-1.6m are rendered in turn by the VTK visualization tool box, and the terrain vertigo diagram is obtained. Figure 7 The higher the resolution of the grid is, the better the rendering effect is, but the memory occupied during rendering is quite different, so it is necessary to select a grid with a suitable resolution.

[0103] The above merely describes the preferred specific embodiments of the present application, but the protection scope of the present application is not limited thereto, and any modification, equivalent replacement and improvement within the technical range disclosed by the present application and within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method of modeling seafloor topography from a multiresolution grid, characterized by, The method comprises the following steps: S1, grid interpolation is performed on original discrete point cloud data to obtain regular grid visualized data; division is performed through a terrain complexity evaluation index until no further division is possible or division is stopped after division to a grid unit corresponding to a high-resolution grid, thereby generating a quadtree-based multi-resolution grid model; S2, the multi-resolution grid model is obtained according to the terrain complexity evaluation index, and a node subdivision evaluation index for terrain complexity is obtained; S3, based on the obtained node subdivision evaluation index for terrain complexity, in the process of multi-resolution grid display rendering, cracks occurring in the splicing of adjacent grids of different resolutions are eliminated by using an edge adding method; In step S2, the node subdivision evaluation index for terrain complexity is obtained, including: The terrain relief degree and the terrain slope are used as factors for node subdivision evaluation, and whether the node continues to be subdivided is determined by combining the two factors; In step S3, the number of triangular facets in the low-resolution grid is increased to improve the resolution of the low-resolution grid, so that the difference between the low-resolution grid and the high-resolution grid is reduced, and the voids occurring at the joint of the two are eliminated; By connecting the low-resolution grid point and the high-resolution grid point, the low-resolution and high-resolution joint part is additionally edge processed, and the low-resolution triangular facet is divided into two to eliminate the cracks.

2. The multi-resolution grid-based seafloor terrain modeling method of claim 1, wherein, In step S1, the original discrete point cloud data is grid interpolated to obtain regular grid visualized data, including: adopting an inverse distance weighting method to gridize the original discrete point cloud data, determining a high-resolution grid resolution d based on original discrete point cloud density and terrain flatness max and a low-resolution grid resolution d min , d max and d min satisfy the condition d max min × 2 n , n is any integer greater than 1; the row and column numbers of the low-resolution grid are X min , Y min , respectively, and the row and column numbers of the high-resolution grid are X max , Y max , respectively, and the condition for performing quadtree splitting and encryption grid on the low-resolution grid is:​ (X max -1)×2 n = X min -1 (Y max -1) x 2 n = Y min -1 In the formula, n is the number of quadtree divisions.

3. The multi-resolution grid-based seafloor terrain modeling method of claim 1, wherein, In step S1, division is performed through the terrain complexity evaluation index until no further division is possible or division is stopped after division to a grid unit corresponding to a high-resolution grid, including: Each low-resolution grid is searched and retrieved, each low-resolution grid is taken as a root node of a quadtree, whether the node subdivision evaluation index is met is determined from each low-resolution grid, if the condition is not met, the node is not subdivided and is saved, otherwise the node is recursively divided into four sub-nodes, the grid elevation values generated by the four sub-nodes are obtained from the corresponding high-resolution grid, and each sub-node is sequentially determined until no further division is possible or division is stopped after division to a grid unit corresponding to a high-resolution grid.

4. The multi-resolution grid-based seafloor terrain modeling method of claim 1, wherein, In step S2, the definitions of the terrain relief degree and the terrain slope are as follows: Terrain relief degree: Δh = h max -h min where Δh is the maximum difference in elevation within the grid, h max is the maximum elevation value within the current grid cell, h min is the minimum elevation value within the current grid cell Terrain slope: S max = max(S) In the formula, S max is the maximum slope value of the current grid point, and S is the comprehensive slope value of the current grid point.

5. The method of claim 4, wherein, For the current grid point comprehensive slope value S, the method of surrounding eight neighborhoods is adopted to calculate the grid slope of the grid point. The elevation of the given grid point is Z5, and the elevation change rate of the grid point in the x, y direction is S x and S y The calculation is shown in formula (1): In the formula, Z1-Z9 are grid point elevation values, and Δd is the grid size. For a grid without eight adjacent grids, the calculation is ignored, and the value is defined as 0; In combination with the two indexes, a grid complexity function is defined as: In the formula, w h and w s are weights of height difference and slope, T h and T s are threshold values of height difference and slope, and C is a comprehensive evaluation index. When C exceeds a set threshold value, the low-resolution grid is subdivided to the grid precision corresponding to the high-resolution grid.

6. The multi-resolution grid-based seafloor terrain modeling method of claim 1, wherein, In step S1, when the regular grid is visualized, the regular grid is divided into two triangular facets for rendering display.

7. A multi-resolution grid-based seafloor terrain modeling system, characterized by, The system implements the multi-resolution grid seabed terrain modeling method according to any one of claims 1-6, and the system comprises: A quadtree-based multi-resolution grid model generation module is configured to perform grid interpolation on original discrete point cloud data to obtain regular grid visualized data, then perform division through a terrain complexity evaluation index until no further division is possible or division is stopped after division to a grid unit corresponding to a high-resolution grid, and finally generate a quadtree-based multi-resolution grid model; The node subdivision evaluation index obtaining module is configured to obtain a multi-resolution grid model according to the terrain complexity evaluation index, and obtain a node subdivision evaluation index for terrain complexity. The crack elimination module is configured to eliminate cracks that occur when adjacent different-resolution grids are spliced, by using a boundary adding method, during multi-resolution grid display rendering, based on the obtained node subdivision evaluation index for terrain complexity.

8. The multi-resolution grid-based seafloor terrain modeling system of claim 7, wherein, The system is loaded on a computer device, which includes at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, and the processor implements the functions in the multi-resolution grid-based seafloor terrain modeling system when executing the computer program.

9. The multi-resolution grid-based seafloor terrain modeling system of claim 7, wherein, The system is loaded on a computer readable storage medium, which stores a computer program, and the computer program can implement the functions in the multi-resolution grid-based seafloor terrain modeling system when executed by a processor.

Images