Multi-resolution grid submarine topography modeling method and system

Abstract

The invention belongs to the technical field of submarine topography information processing, and discloses a submarine topography modeling method and system of a multi-resolution grid. The method comprises the following steps: generating a quadtree-based multi-resolution grid model; obtaining a multi-resolution grid model according to the terrain complexity evaluation index, and obtaining a node subdivision evaluation index for the terrain complexity; in the process of multi-resolution grid display rendering, cracks generated by splicing of adjacent grids with different resolutions are eliminated by using an 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, so that the space redundancy of data storage is remarkably reduced. According to the method, the spatial redundancy is reduced, the terrain features and the gradient are considered through the subdivision evaluation function, the complex features of the terrain can be well reserved, and the overall terrain precision is ensured.

Description

Technical Field

[0001] The present invention belongs to the technical field of seabed terrain information processing, and in particular relates to a multi-resolution grid seabed terrain modeling method and system. Background Art

[0002] Seafloor topography is one of the most complex landforms on Earth's surface. It encompasses vast abyssal plains, steep submarine ridges, and winding trenches, as well as complex volcanoes, rift valleys, and sediment accumulation zones. With the increasing frequency of marine surveys and development activities, the demand for high-precision, high-resolution seafloor topography data at full ocean depth is increasing. Therefore, establishing high-precision seafloor topography models is of great practical significance.

[0003] Currently, terrain modeling methods are primarily divided into regular grids and irregular triangulated networks (TINs). However, both methods have drawbacks. For complex terrain, increasing grid density to improve accuracy exponentially increases the amount of data required. For flat terrain, regular grids require a reduced grid density, otherwise redundant data is generated and system resources are wasted. TINs, with their complex data structure and computationally intensive nature, are difficult to implement for large-scale submarine terrain modeling.

[0004] When it comes to research on building large-scale, high-precision three-dimensional seabed terrain models, existing modeling methods are mainly based on LOD, which renders terrain models at different levels by adjusting the distance of the viewpoint and the scheduling of internal and external memory. However, for seabed terrain data, most areas are relatively flat, and only some areas have large terrain fluctuations. If a high grid density is used, the memory space occupied will increase, and a low grid density will lose terrain accuracy. Using LOD and Out-of-Core concepts based on the relationship between viewpoint and viewing distance requires a reasonable browsing rate to ensure the continuity of the terrain. Secondly, after determining the model of the bottom grid, additional data needs to be filled in to ensure the continuity of the LOD, further exacerbating spatial redundancy.

[0005] In summary, the technical problems and shortcomings of existing technologies are as follows: The main factor affecting submarine terrain modeling is that submarine terrain is one of the most complex terrain types on Earth's surface. It not only includes vast abyssal plains, steep submarine ridges, and winding trenches, but also contains complex volcanoes, rift valleys, and sediment accumulation areas. This complexity determines the need for high spatial resolution of submarine terrain. Traditional single-resolution grids cannot meet this requirement. Multi-resolution grid models based on LODs increase memory space usage if the grid density is high, while low grid density will reduce terrain accuracy. Summary of the Invention

[0006] To overcome the problems existing in the related art, the disclosed embodiments of the present invention provide a method and system for modeling submarine terrain with a multi-resolution grid. The purpose of the present invention is to propose a variable-resolution submarine terrain organization model based on a regular grid and a quadtree structure. First, high-resolution and low-resolution grid models are generated respectively. The terrain complexity in each grid is evaluated based on the low-resolution grid model. The grid data in the local complex terrain area is encrypted through the quadtree structure to meet the high-precision display of specific terrain, thereby generating a regular grid model with a mixture of multiple different resolutions to reduce the spatial redundancy caused by generating a high-resolution grid model where the terrain features of some areas vary greatly. This better realizes the three-dimensional modeling of submarine terrain.

[0007] The technical solution is as follows: A multi-resolution grid seabed terrain modeling method includes the following steps:

[0008] S1, grid interpolation is performed on the original discrete point cloud data to obtain data after regular grid visualization; segmentation is performed using the terrain complexity evaluation index until it can no longer be segmented or the grid cells corresponding to the high-resolution grid are segmented, and segmentation is stopped until it is no longer possible to segment or the grid cells corresponding to the high-resolution grid are segmented, generating a multi-resolution grid model based on a quadtree;

[0009] S2, obtaining a multi-resolution grid model based on the terrain complexity evaluation index, and obtaining a node subdivision evaluation index for the terrain complexity;

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

[0011] In step S1, the original discrete point cloud data is grid interpolated to obtain regular grid visualization data, including: using the inverse distance weighted method to grid the original discrete point cloud data, and determining the high-resolution grid resolution d based on the original discrete point cloud density and the flatness of the terrain. 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 is 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 segmentation and mesh encryption on the low-resolution grid are:

[0012] (X max -1)×2n =X min -1

[0013] (Y max -1)×2 n =Y min -1

[0014] Where n is the number of quadtree partitions.

[0015] In step S1, segmentation is performed using a terrain complexity evaluation index until no further segmentation is possible or the segmentation reaches a grid cell corresponding to a high-resolution grid, including:

[0016] Each low-resolution grid is traversed and searched, and each low-resolution grid is used as the root node of the quadtree. From each low-resolution grid, it is judged whether the evaluation index of node subdivision is met. If the conditions are not met, no subdivision is performed and the node is saved. Otherwise, the node is recursively split into four child nodes. The additional grid elevation values generated by the four child nodes are obtained from the corresponding high-resolution grid. Then, each child node is judged in turn until it can no longer be split or the grid unit corresponding to the high-resolution grid is split. The segmentation is stopped.

[0017] In step S2, node segmentation evaluation indicators for terrain complexity are obtained, including:

[0018] The terrain relief and slope are used as factors for node segmentation evaluation. By combining these two factors, it is determined whether the node should continue to be segmented. The definitions of terrain relief and slope are as follows:

[0019] Terrain relief:

[0020] Δh=h max -h min

[0021] Where Δh is the maximum elevation difference within the grid, h max is the maximum elevation value in the current grid cell, h min The minimum elevation value in the current grid cell

[0022] Terrain slope:

[0023] S max =max(S)

[0024] Where 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] Furthermore, for the calculation of the comprehensive slope value S of the current grid point, the grid slope of the grid point is calculated using the method of the surrounding eight neighborhoods. The elevation of the given grid point is Z5, and the elevation change rate of the grid point in the x and y directions is S.x and S y The calculation is shown in formula (1):

[0026]

[0027] Where Z1-Z9 are grid point elevation values, Δd is the grid size, and grids with less than eight neighbors are ignored and their values are defined as 0.

[0028]

[0029] Combining these two indicators, a function of grid complexity is defined as:

[0030]

[0031] Where w h and w s is the weight of height difference and slope, T h and T s is the threshold of height difference and slope, and C is a comprehensive evaluation index. When C exceeds the set threshold, the grid is subdivided until the low-resolution grid is subdivided to the grid accuracy corresponding to the high-resolution grid.

[0032] In step S1 , when visualizing a regular grid, the regular grid is divided into two triangular facets for rendering and display.

[0033] In step S3, the method of varying the resolution grid includes:

[0034] By increasing the number of triangles in the low-resolution mesh, the resolution of the low-resolution mesh is improved, the difference between the low-resolution mesh and the high-resolution mesh is reduced, and the holes that appear at the junction of the two are eliminated;

[0035] By additionally connecting the low-resolution grid points with the high-resolution grid points, the connecting lines are additionally edged at the junction of the low-resolution and high-resolution parts, thereby splitting the low-resolution triangular face into two and eliminating cracks.

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

[0037] The quadtree-based multi-resolution grid model generation module is used to perform grid interpolation on the original discrete point cloud data to obtain regular grid visualization data, and then divide it according to the terrain complexity evaluation index until it can no longer be divided or the grid cells corresponding to the high-resolution grid are divided. The multi-resolution grid model based on the quadtree is finally generated.

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

[0039] The crack elimination module is used to eliminate cracks that appear when adjacent grids of different resolutions are spliced together using the edge addition method during the multi-resolution grid display and rendering process based on the node segmentation evaluation index obtained 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, the functions of the multi-resolution grid seabed terrain modeling system are realized.

[0041] Furthermore, the system is carried on a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it can realize the functions of the above-mentioned multi-resolution grid seabed terrain modeling system.

[0042] In combination with all the above technical solutions, the beneficial effects of the present invention are as follows:

[0043] This invention uses a quadtree structure, combining high-resolution and low-resolution grids. The low-resolution grid is dynamically subdivided based on terrain complexity to achieve local densification, balancing high precision with low redundancy. Using terrain undulation and slope as evaluation factors, the grid is dynamically subdivided to avoid redundancy and ensure high precision in complex areas. Due to the different resolutions of adjacent areas of the multi-resolution grid, cracks appear at the junction of the high-resolution and low-resolution grids. By adding connecting edges between low-resolution and high-resolution grid points, the low-resolution triangles are split in two, eliminating the cracks.

[0044] The present invention reduces the number of high-precision grid points by combining a quadtree storage structure with a multi-resolution grid, thereby significantly reducing the spatial redundancy of data storage. While reducing spatial redundancy, the present invention can better preserve the complex features of the terrain and ensure overall terrain accuracy by taking into account terrain features and slopes through a subdivision evaluation function. The present invention reduces the number of triangles during rendering, reduces memory usage during the rendering process, and improves rendering efficiency, making it particularly suitable for the visualization of large-scale submarine terrain data. The method of the present invention can dynamically adjust the grid resolution according to the complexity of the terrain, ensuring high precision in locally complex areas while avoiding wasting computing resources in simple areas. BRIEF DESCRIPTION OF THE DRAWINGS

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

[0046] Figure 1 This is a flow chart of a multi-resolution grid seabed terrain modeling method provided by an embodiment of the present invention;

[0047] Figure 2 This is a schematic diagram of a method for gridding original discrete point cloud data using an inverse distance weighted method according to an embodiment of the present invention;

[0048] Figure 3 Schematic diagram of quadtree segmentation for each grid node provided by an embodiment of the present invention;

[0049] Figure 4 Schematic diagram of eight neighborhoods of a grid provided by an embodiment of the present invention;

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

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

[0052] Figure 7 Comparison chart of the multi-resolution terrain model of the present invention and different single-resolution terrain models. DETAILED DESCRIPTION

[0053] To make the above-mentioned objects, features, and advantages of the present invention more readily apparent, specific embodiments of the present invention are described in detail below with reference to the accompanying drawings. The following description sets forth numerous specific details to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways than those described herein, and those skilled in the art may make similar modifications without departing from the scope of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0054] The innovation of the present invention lies in: the innovation of the multi-resolution grid seabed terrain modeling method and system provided by the embodiment of the present invention lies in: based on the quadtree structure, the grid resolution is dynamically adjusted in combination with the terrain complexity (undulation and slope), high-resolution fine expression is adopted in complex areas, and low resolution is retained in flat areas to reduce data redundancy; the boundaries of grids of different resolutions are connected through 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 that of high-resolution models, providing an efficient and adaptive solution for large-scale seabed terrain modeling.

[0055] Example 1, as Figure 1As shown, the multi-resolution grid seabed terrain modeling method provided by the embodiment of the present invention includes:

[0056] S1, grid interpolation is performed on the original discrete point cloud data to obtain data after regular grid visualization; segmentation is performed using the terrain complexity evaluation index until it can no longer be segmented or the grid cells corresponding to the high-resolution grid are segmented, and segmentation is stopped until it is no longer possible to segment or the grid cells corresponding to the high-resolution grid are segmented, generating a multi-resolution grid model based on a quadtree;

[0057] S2, obtaining a multi-resolution grid model based on the terrain complexity evaluation index, and obtaining a node subdivision evaluation index for the terrain complexity;

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

[0059] Exemplarily, in step S1, finally generating a quadtree-based multi-resolution grid model includes:

[0060] Step 1.1, first grid interpolation is performed on the original discrete point cloud data to obtain data after regular grid visualization. Commonly used spatial interpolation methods are mainly inverse distance weighted method and kriging method. Kriging method is suitable for data with complex spatial structure and takes into account the correlation between data, but its calculation is too complicated. Although the inverse distance weighted method may have an excessive impact on the results due to outliers, it has high computational efficiency and produces a smooth surface suitable for terrain modeling. Therefore, this invention adopts the inverse distance weighted method to grid the original discrete point cloud data. The gridding process is as follows: Figure 2 Specifically including:

[0061] Determine 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 , where d max and d min Condition d should be met max =d min ×2 n , where n is any integer greater than 1, and the number of rows and columns of the low-resolution grid are X max 、Y max , the number of rows and columns of the high-resolution grid are X max 、Y max In order to meet the requirements of further encrypting the low-resolution grid by quadtree segmentation, 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] Where n is the number of quadtree splits. If the quadtree is split once, n is 1. If the quadtree is split once more, n is 2.

[0065] Therefore, it is necessary to first determine the number of rows and columns of the low-resolution grid, and then calculate the number of rows and columns of the high-resolution grid.

[0066] In step 1.2, when partitioning based on terrain complexity, each low-resolution grid is traversed and retrieved. Each low-resolution grid is used as the root node of a quadtree. Then, a judgment is made from each low-resolution grid. During the judgment process, a comprehensive evaluation metric, C, is used. When the calculated grid C exceeds a set threshold, the grid is subdivided until the low-resolution grid reaches the grid accuracy corresponding to the high-resolution grid. The evaluation metric for node subdivision is determined. If not, the node is not subdivided and saved. Otherwise, the node is recursively split into four child nodes. The additional grid elevation values generated for each of these four child nodes are obtained from the corresponding high-resolution grid. In the quadtree-based multi-resolution grid model, when a node in a low-resolution grid requires subdivision, it is split into four child nodes. The elevation values of these child 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, where each grid point stores an accurate elevation value. Low-resolution grids are used in areas with flatter terrain and larger grid spacing. When a node in the low-resolution grid meets the subdivision criteria, the system will split 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] Then judge each child node in turn until it can no longer be segmented or the grid unit corresponding to the high-resolution grid is segmented and the segmentation is stopped. The deeper the segmentation depth, the higher the resolution of the corresponding area and the higher the accuracy. Figure 3 Perform quadtree segmentation for each grid node, that is, node subdivision principle based on quadtree structure.

[0068] Exemplarily, in step S2, obtaining a node segmentation evaluation index for terrain complexity includes:

[0069] In order to avoid unnecessary subdivision and meet the accuracy requirements of local areas, this paper mainly uses terrain relief and terrain slope as factors for node subdivision evaluation. By combining these two factors, it is determined whether the node should continue to be subdivided. The definitions of terrain relief and terrain slope are as follows:

[0070] Terrain relief:

[0071] Δh=h max -h min

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

[0073] Terrain slope:

[0074] S max =max(S)

[0075] Where 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] Among them, for the calculation of S, the present invention adopts the method of eight surrounding neighborhoods to calculate the grid slope of the grid point, such as Figure 4 As shown in the eight-neighborhood grid diagram, the elevation of a given grid point is Z5, then the elevation change rate of the grid point in the x and y directions is S x and S y The calculation is shown in formula (1):

[0077]

[0078] Where Z1-Z9 are grid point elevation values, Δd is the grid size, and grids with less than eight neighbors are ignored and their values are defined as 0.

[0079]

[0080] Formulas (1) and (2) are both used to calculate the grid slope. Combining the terrain undulation and terrain slope index, a grid complexity function is defined as:

[0081]

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

[0083] For example, in step S3, when visualizing a regular grid, the regular grid is usually divided into two triangular facets for rendering. Compared with regular grids, triangular facets can more flexibly display irregular elevation changes. The current mainstream graphics rendering hardware is very efficient in processing triangles, and graphics APIs usually support the drawing of triangular facets. Therefore, the present invention chooses to divide the regular grid into two triangular facets for rendering and display.

[0084] In the process of multi-resolution grid display rendering, since adjacent grids may have different resolutions, terrain discontinuity will appear at the junction of two adjacent grids with different resolutions during the actual rendering process, which will appear as cracks in three dimensions. The main reason for the cracks is that the two adjacent grids have different resolutions. There will be several extra grid points at the junction of the high-resolution grid and the low-resolution grid. Since the grid point elevation values are interpolated from the original point cloud data, the elevation values of the extra grid points are not on the line connecting the two grid points at the junction of the low-resolution grid, so cracks will appear. In the rendering process, triangular holes will appear where grids with different resolutions meet, such as Figure 5 The principle of crack generation is shown.

[0085] To increase the number of vertices at the boundaries of the low-resolution grid, additional triangles are required. This method is more complex to implement and increases the complexity of the scene. However, it can readily handle any cracks that may arise, ensuring the accuracy of the terrain. It is suitable for describing more complex terrains and is particularly well-suited for applications in the variable-resolution grid proposed in this invention. The principle is to increase the number of triangles in the low-resolution grid, thereby increasing its resolution. This significantly reduces the difference between the low-resolution and high-resolution grids, effectively eliminating the potential voids that may appear at the junction of the two. Figure 6 This is the crack treatment effect diagram, which is achieved by additionally connecting the low-resolution grid points with the high-resolution grid points, such as Figure 6 The line connecting points AB and CD adds an extra edge to the part where the low-resolution and high-resolution triangles meet, splitting the low-resolution triangle into two to eliminate the cracks. The principle of adding edges is that when the resolutions of two adjacent meshes are different, the high-resolution mesh may produce holes in the low-resolution mesh area, such as Figure 5 To eliminate these holes, the present invention needs to perform edge processing on adjacent low-resolution grids to improve their resolution, thereby ensuring the continuity between grids and the integrity of the data.

[0086] The specific implementation steps of edge processing are:

[0087] 1. Detecting Resolution Differences Between Adjacent Meshes: First, the system traverses all adjacent meshes to detect if there are any significant resolution differences. If significant differences are detected, edge addition is required.

[0088] 2. Identify crack areas: At the boundaries between adjacent grids, the high-resolution grid has additional grid points. These grid points form areas of elevation discontinuity with the boundary points of the low-resolution grid, known as crack areas. The system marks these crack areas and determines the additional edges that need to be added.

[0089] 3. Adding extra edges: In crack areas, the system adds extra edges between the boundary points of the low-resolution mesh and the additional grid points of the high-resolution mesh. Specifically, the boundary points of the low-resolution mesh are connected to the adjacent grid points of the high-resolution mesh to form new triangles.

[0090] 4. Generate new triangular faces: By adding additional edges, the system will split the original regular grid into more triangular faces.

[0091] 5. Rendering: During the rendering process, the system will give priority to drawing these newly generated triangles to ensure smooth transitions between adjacent meshes and avoid visual cracks or discontinuities.

[0092] In embodiment 2, a multi-resolution grid seabed terrain modeling system provided by an embodiment of the present invention includes:

[0093] The quadtree-based multi-resolution grid model generation module is used to perform grid interpolation on the original discrete point cloud data to obtain regular grid visualization data, and then divide it according to the terrain complexity evaluation index until it can no longer be divided or the grid cells corresponding to the high-resolution grid are divided. The multi-resolution grid model based on the quadtree is finally generated.

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

[0095] The crack elimination module is used to eliminate cracks that appear when adjacent grids of different resolutions are spliced together using the edge addition method during the multi-resolution grid display and rendering process based on the node segmentation evaluation index obtained for terrain complexity.

[0096] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.

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

[0098] First, we used the checkpoint method to assess the accuracy of fixed-resolution and variable-resolution grids. This method evaluates the distance between the grid points and the original point cloud, calculating the root mean square error (RMSE). A smaller RMSE indicates higher accuracy. We also compared the number of grid points, the number of triangles used during rendering, and the memory usage of regular grids rendered using triangles for each resolution. The experimental results are shown in Table 1.

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

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

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

[0102] Secondly, the advantages and disadvantages of this method in expressing the terrain authenticity are observed and analyzed from the perspective of subjective vision. The present invention uses the VTK visualization toolbox to render fixed resolution grids at 0.1m and 1.6m resolutions and variable resolution grids at 0.1-1.6m resolutions and obtains the terrain dizziness map ( Figure 7 Comparison diagram of the multi-resolution terrain model of the present invention and different single-resolution terrain models). The higher the resolution of the grid, the better the rendering effect. However, the memory occupied by the rendering varies greatly, so it is necessary to select a grid with an appropriate resolution.

[0103] The above description is only a preferred specific implementation method of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with this technical field within the technical scope disclosed by the present invention and within the spirit and principles of the present invention should be covered by the scope of protection of the present invention.

Claims

1. A multi-resolution grid seabed terrain modeling method, characterized in that: The method comprises the following steps: S1, grid interpolation is performed on the original discrete point cloud data to obtain data after regular grid visualization; segmentation is performed using the terrain complexity evaluation index until it can no longer be segmented or the grid cells corresponding to the high-resolution grid are segmented, and segmentation is stopped until it is no longer possible to segment or the grid cells corresponding to the high-resolution grid are segmented, generating a multi-resolution grid model based on a quadtree; S2, obtaining a multi-resolution grid model based on the terrain complexity evaluation index, and obtaining a node subdivision evaluation index for the terrain complexity; S3, based on the obtained node subdivision evaluation index for terrain complexity, during the multi-resolution grid display and rendering process, the cracks that appear when adjacent grids of different resolutions are spliced are eliminated using the edge addition method.

2. The multi-resolution grid seabed terrain modeling method according to claim 1, characterized in that: In step S1, the original discrete point cloud data is grid interpolated to obtain regular grid visualization data, including: using the inverse distance weighted method to grid the original discrete point cloud data, and determining the high-resolution grid resolution d based on the original discrete point cloud density and the flatness of the terrain. 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 is 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 segmentation and mesh encryption on the low-resolution grid are: (X max -1)×2 n =X min -1 (AND max -1)×2 n =And min -1 Where n is the number of quadtree partitions.

3. The multi-resolution grid seabed terrain modeling method according to claim 1, characterized in that: In step S1, segmentation is performed using a terrain complexity evaluation index until no further segmentation is possible or the segmentation reaches a grid cell corresponding to a high-resolution grid, including: Each low-resolution grid is traversed and searched, and each low-resolution grid is used as the root node of the quadtree. From each low-resolution grid, it is judged whether the evaluation index of node subdivision is met. If the conditions are not met, no subdivision is performed and the node is saved. Otherwise, the node is recursively split into four child nodes. The additional grid elevation values generated by the four child nodes are obtained from the corresponding high-resolution grid. Then, each child node is judged in turn until it can no longer be split or the grid unit corresponding to the high-resolution grid is split. The segmentation is stopped.

4. The multi-resolution grid seabed terrain modeling method according to claim 1, characterized in that: In step S2, node segmentation evaluation indicators for terrain complexity are obtained, including: The terrain relief and slope are used as factors for node segmentation evaluation. By combining these two factors, it is determined whether the node should continue to be segmented. The definitions of terrain relief and slope are as follows: Terrain relief: Δh=h max -h min Where Δh is the maximum elevation difference within the grid, h max is the maximum elevation value in the current grid cell, h min The minimum elevation value in the current grid cell Terrain slope: S max =max(S) Where 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 multi-resolution grid seabed terrain modeling method according to claim 4, characterized in that: 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, and the elevation change rate of the grid point in the x and y directions is S. x and S y The calculation is shown in formula (1): Where Z1-Z9 are grid point elevation values, Δd is the grid size, and grids with less than eight neighbors are ignored and their values are defined as 0. Combining these two indicators, a function of grid complexity is defined as: Where w h and w s is the weight of height difference and slope, T h and T s is the threshold of height difference and slope, and C is a comprehensive evaluation index. When C exceeds the set threshold, the grid is subdivided until the low-resolution grid is subdivided to the grid accuracy corresponding to the high-resolution grid.

6. The multi-resolution grid seabed terrain modeling method according to claim 1, characterized in that: In step S1 , when visualizing a regular grid, the regular grid is divided into two triangular facets for rendering and display.

7. The multi-resolution grid seabed terrain modeling method according to claim 1, characterized in that: In step S3, the method of varying the resolution grid includes: By increasing the number of triangles in the low-resolution mesh, the resolution of the low-resolution mesh is improved, the difference between the low-resolution mesh and the high-resolution mesh is reduced, and the holes that appear at the junction of the two are eliminated; By additionally connecting the low-resolution grid points with the high-resolution grid points, the connecting lines are additionally edged at the junction of the low-resolution and high-resolution parts, thereby splitting the low-resolution triangular face into two and eliminating cracks.

8. A multi-resolution grid seabed terrain modeling system, characterized by: The system implements the multi-resolution grid seabed terrain modeling method according to any one of claims 1 to 7, and the system comprises: The quadtree-based multi-resolution grid model generation module is used to perform grid interpolation on the original discrete point cloud data to obtain regular grid visualization data, and then divide it according to the terrain complexity evaluation index until it can no longer be divided or the grid cells corresponding to the high-resolution grid are divided. The multi-resolution grid model based on the quadtree is finally generated. A node subdivision evaluation index acquisition module is used to obtain a multi-resolution grid model based on the terrain complexity evaluation index and obtain a node subdivision evaluation index for the terrain complexity; The crack elimination module is used to eliminate cracks that appear when adjacent grids of different resolutions are spliced together using the edge addition method during the multi-resolution grid display and rendering process based on the node segmentation evaluation index obtained for terrain complexity.

9. The multi-resolution grid seabed terrain modeling system according to claim 8, characterized in that: 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, the functions of the multi-resolution grid seabed terrain modeling system are realized.

10. The multi-resolution grid seabed terrain modeling system according to claim 8, characterized in that: The system is carried on a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it can realize the functions of the above-mentioned multi-resolution grid seabed terrain modeling system.

Images