A Smart Underwater Topographic Map Generation System and Method Based on Big Data

Abstract

This application relates to the field of surveying and mapping technology, and in particular to an intelligent underwater topographic map generation system and method based on big data. The method includes: acquiring surveying and mapping data; analyzing the surveying and mapping data to determine ground elevation points and non-ground elevation points; clustering the non-ground elevation points according to the distance between any two non-ground elevation points to obtain several point cloud clusters; acquiring preset structure information; matching each point cloud cluster according to the preset structure information to determine the attributes of each point cloud cluster; and generating an underwater topographic map based on the attributes and the ground elevation points. The method divides point cloud data into ground elevation points and non-ground elevation points, identifies different types of underwater objects, determines the attributes of each point cloud cluster, and generates an underwater topographic map containing ground features, topographic texture, and the spatial distribution of objects with different attributes based on the attributes of the point cloud clusters and the ground elevation points.

Description

Technical Field

[0001] This application relates to the field of surveying and mapping technology, and in particular to an intelligent underwater topographic map generation system and method based on big data. Background Technology

[0002] Underwater topographic maps are core foundational data for marine resource exploration, waterway maintenance, ecological protection, and underwater engineering construction. Traditional underwater topographic mapping techniques primarily rely on equipment such as sonar and multibeam echo sounders to acquire discrete elevation points, generating regular grids through interpolation algorithms (such as Kriging interpolation or inverse distance weighting) to create topographic maps. However, complex underwater environments (such as vegetation cover, shipwrecks, and rock distribution) result in a large number of non-topographic feature points mixed in with the elevation data. Semantic segmentation techniques based on machine learning have been introduced into underwater data processing, but they depend on a large number of labeled samples, making fully automated processing difficult. Summary of the Invention

[0003] This application provides an intelligent underwater topographic map generation system and method based on big data to solve the above problems.

[0004] In a first aspect, this application provides a method for intelligent generation of underwater topographic maps based on big data, the method comprising:

[0005] Acquire survey data, analyze the survey data, and determine ground elevation points and non-ground elevation points;

[0006] For the non-ground elevation points, based on the distance between any two non-ground elevation points, the non-ground elevation points are clustered to obtain several point cloud clusters.

[0007] Obtain preset structure information, and match each cloud cluster according to the preset structure information to determine the attributes of each cloud cluster.

[0008] An underwater topographic map is generated based on the attributes and the ground elevation points.

[0009] This solution ensures the quality and consistency of point cloud data. By analyzing elevation information, the point cloud data is divided into ground elevation points and non-ground elevation points. Based on the distance between any two non-ground elevation points, the non-ground layer point cloud data is clustered into several point cloud clusters, which helps in identifying different types of underwater objects. By matching preset structural information, the attributes of each point cloud cluster are determined, providing the necessary information for generating accurate underwater topographic maps. Based on the attributes of the point cloud clusters and ground elevation points, an underwater topographic map is generated that includes ground topography, terrain texture, and the spatial distribution of objects with different attributes.

[0010] Optionally, the analysis of the survey data to determine ground elevation points and non-ground elevation points includes:

[0011] Analyze the survey data to determine the elevation of each cloud data point;

[0012] For each point cloud data point, the point cloud data is divided into layers based on the elevation to obtain the layering result;

[0013] A preset height threshold is obtained, and the preset height threshold is compared with the average elevation of each layer in the layering result to obtain the comparison result.

[0014] Based on the comparison results, the ground layer and non-ground layer are determined;

[0015] The point cloud data of the ground layer is used as ground elevation points, and the point cloud data of the non-ground layer is used as non-ground elevation points.

[0016] This solution improves data accuracy and reliability through data cleaning. Elevation information for each point is extracted, providing necessary data support for data layering and analysis. Dividing the point cloud data into ground and non-ground layers helps distinguish underwater topography from other underwater objects, preparing for cluster analysis. Statistical analysis yields the mean and standard deviation of elevation for each layer. Threshold comparison allows for preliminary differentiation between ground and non-ground layers. Accurately identifying the ground layer helps generate more precise underwater topographic maps. Accurately identifying the non-ground layer helps identify and classify underwater objects. Labeling the point cloud data for both ground and non-ground layers facilitates data processing and analysis.

[0017] Optionally, for the non-ground elevation points, based on the distance between any two non-ground elevation points, the non-ground elevation points are clustered to obtain several point cloud clusters, including:

[0018] For each non-ground elevation point, calculate the distance between the current non-ground elevation point and any other non-ground elevation point in turn;

[0019] Based on the distance, calculate the average distance between the remaining non-ground elevation points and the current non-ground elevation point;

[0020] Based on the average distance, determine the local density index;

[0021] Based on the local density index, the non-ground elevation points are clustered to obtain several point cloud clusters.

[0022] This scheme analyzes the spatial relationships between points in a point cloud by calculating the distance between each non-ground elevation point and all other non-ground elevation points. Calculating the mean distance between each non-ground elevation point and all other non-ground elevation points helps assess the local density of that point, i.e., the distribution of surrounding points. Determining local density indices identifies which point cloud regions are dense and which are sparse, aiding in distinguishing different types of underwater objects such as dense aquatic plants and sparse rocks. Applying clustering algorithms divides the point cloud data into several clusters, with points within each cluster sharing similar spatial relationships and attributes, facilitating further analysis of the object type represented by each cluster. This results in several point cloud clusters, each representing a group of underwater objects of the same type.

[0023] Optionally, determining the local density index based on the mean distance includes:

[0024] Sort the distances by distance and plot the distance curves in ascending order;

[0025] Based on the distance curve, determine the steep inflection point;

[0026] The distance corresponding to the steep inflection point is used as the local density index of the current non-ground elevation point.

[0027] This method analyzes the spatial relationships between points in a point cloud by calculating the distance between each non-ground elevation point and all other non-ground elevation points. Sort the calculated distances in ascending order to visually observe the distance distribution and provide ordered data for drawing distance curves. Drawing distance curves visually displays the distance distribution and helps identify key features such as inflection points and distribution patterns. Analyzing the distance curves and identifying steep inflection points determines key points of change in the distance distribution, marking the transition from the ground layer to the non-ground layer. Using the distance corresponding to the steep inflection point as a local density index for the current non-ground elevation point accurately reflects the distribution density of points around that point.

[0028] Optionally, the step of matching each cloud cluster based on the preset structure information and determining the attributes of each cloud cluster includes:

[0029] Analyze the preset structural information to determine the geometric features of each structure;

[0030] The geometric features are matched with each point cloud cluster to determine the attributes of each point cloud cluster.

[0031] This solution establishes a standardized set of geometric feature templates by analyzing pre-defined structural information. These templates are used to identify and classify point cloud clusters. Extracting the geometric features of point cloud clusters helps distinguish different types of objects, such as those with different shapes, sizes, and densities. By matching the geometric features of point cloud clusters with pre-defined structural information, the similarity between the clusters and known structures can be identified. Based on the matching results, attributes can be determined for each point cloud cluster.

[0032] Optionally, matching the geometric features with each point cloud cluster to determine the attributes of each point cloud cluster includes:

[0033] For each point cloud cluster, the point cloud cluster is downsampled to obtain several simple geometric structures;

[0034] Analyze several simple geometric shapes to determine the possibilities of their combinations;

[0035] The geometric features are matched with each possible combination to determine the attributes of each point cloud cluster.

[0036] This scheme reduces the number of points in point cloud clusters through downsampling, lowering data complexity and improving processing efficiency and speed. Extracting simple geometric structures helps simplify the shape of point cloud clusters, making them easier to analyze and match, while also reducing the impact of noise and outliers. Analyzing the combination possibilities between simple geometric structures helps identify different object types present in the clusters. By matching geometric features and combination possibilities, the most likely object attributes for each point cloud cluster can be determined, thereby improving the accuracy and reliability of classification. Determining attributes for each point cloud cluster enables effective classification of underwater objects.

[0037] Optionally, for each point cloud cluster, the point cloud cluster is downsampled to obtain several simple geometric structures, including:

[0038] For each point cloud cluster, the density of the point cloud cluster is determined based on the distance.

[0039] Based on the density, the point cloud cluster is divided into several equal voxels;

[0040] Based on the division results, determine the voxel size;

[0041] Obtain the spatial coordinates of each point cloud data, and determine the voxel center of each equally divided voxel based on the spatial coordinates and the voxel size;

[0042] Based on the aforementioned physical properties, several simple geometric structures are obtained.

[0043] This scheme utilizes a method where calculating the density of point cloud clusters helps identify cluster density, making the downsampling process more reasonable. Dividing point cloud clusters into equally sized voxels simplifies the data structure, making processing more efficient while preserving the overall shape and characteristics of the clusters. Determining voxel sizes can adapt to different levels of detail in different regions based on the density and size of the point cloud clusters, ensuring that voxels can reasonably represent the characteristics of the clusters. Obtaining the spatial coordinates of each point cloud data point is fundamental for voxel partitioning and centroid calculation, providing necessary information for determining the position and shape of the voxels. Calculating the voxel centroids helps identify the point distribution within the voxels. Downsampling reduces the number of points in the point cloud clusters, lowers data complexity, and improves processing efficiency and speed. Extracting simple geometric structures clearly represents the basic shape and characteristics of the point cloud clusters.

[0044] Optionally, determining the voxel centroid of each equally divided voxel based on the spatial coordinates and the voxel size includes:

[0045] For each non-ground elevation point in each voxel segment, the index of the non-ground elevation point in the corresponding voxel segment is determined based on the spatial coordinates and the voxel size, using the following formula:

[0046] ;

[0047] in, This indicates the location of non-ground elevation points in each equally divided voxel. Axis index; This indicates the location of non-ground elevation points in each equally divided voxel. Axis index; This indicates the location of non-ground elevation points in each equally divided voxel. Axis index; The voxel size representing each segment of the voxel is in axis, axis, On-axis performance; The spatial coordinates of the non-ground elevation points in each equally divided voxel;

[0048] Based on the index, determine the non-ground elevation points corresponding to the corresponding equal voxels, and determine the center of mass of the voxels based on the spatial coordinates of the non-ground elevation points corresponding to the corresponding equal voxels, using the following formula:

[0049] ;

[0050] in, The spatial coordinates of the said body mass center; This indicates the number of non-ground elevation points in each equally divided voxel; This represents the spatial coordinates of non-ground elevation points within each equally divided voxel.

[0051] This scheme assigns non-ground elevation points to corresponding voxels by calculating an index. Determining the non-ground elevation points within each voxel based on the index helps organize and structure the data, making the point cloud data within each voxel more ordered. Counting the number of non-ground elevation points in each voxel helps calculate the voxel mass center, ensuring that the mass center accurately reflects the distribution of points within the voxel. Calculating the mass center yields the representative location of each voxel, which aids in identifying and analyzing the structure and characteristics of point cloud clusters.

[0052] Optionally, generating an underwater topographic map based on the attributes and the ground elevation points includes:

[0053] Based on the aforementioned attributes, determine the type of each cloud cluster;

[0054] Analyze the point cloud clusters to determine the overall structure of each point cloud cluster;

[0055] Based on the overall structure and type, determine the spatial distribution of each cloud cluster.

[0056] Based on the aforementioned ground elevation points, determine the ground trend and topographic texture.

[0057] An underwater topographic map is generated based on the ground surface trend, the terrain texture structure, and the spatial distribution.

[0058] This scheme, through attribute classification, can distinguish different types of point cloud clusters, such as land, aquatic plants, and rocks. Analyzing the overall structure helps identify the shape and distribution of point cloud clusters, providing a more accurate model for generating topographic maps. Determining the spatial distribution reveals the layers and complexity of underwater topography, contributing to the generation of more detailed and accurate topographic maps. By analyzing ground elevation points, the trend and texture of the underwater surface can be determined, providing crucial topographic information for map generation. Based on the surface trend, topographic texture, and spatial distribution of point cloud clusters, the generated underwater topographic maps will be more accurate and detailed, contributing to applications in marine exploration, environmental protection, and underwater ecological research.

[0059] Secondly, this application provides an intelligent underwater topographic map generation system based on big data, the system comprising:

[0060] The data analysis module is used to acquire and analyze survey data to determine ground elevation points and non-ground elevation points.

[0061] The clustering analysis module is used to cluster the non-ground elevation points according to the distance between any two non-ground elevation points to obtain several point cloud clusters.

[0062] The attribute analysis module is used to obtain preset structure information, match each point cloud cluster according to the preset structure information, and determine the attributes of each point cloud cluster.

[0063] The topographic map generation module is used to generate an underwater topographic map based on the attributes and the ground elevation points.

[0064] Optionally, when the data analysis module analyzes the surveying data to determine ground elevation points and non-ground elevation points, it is used for:

[0065] Analyze the survey data to determine the elevation of each cloud data point;

[0066] For each point cloud data point, the point cloud data is divided into layers based on the elevation to obtain the layering result;

[0067] A preset height threshold is obtained, and the preset height threshold is compared with the average elevation of each layer in the layering result to obtain the comparison result.

[0068] Based on the comparison results, the ground layer and non-ground layer are determined;

[0069] The point cloud data of the ground layer is used as ground elevation points, and the point cloud data of the non-ground layer is used as non-ground elevation points.

[0070] Optionally, when the clustering analysis module clusters the non-ground elevation points based on the distance between any two non-ground elevation points to obtain several point cloud clusters, it is used for:

[0071] For each non-ground elevation point, calculate the distance between the current non-ground elevation point and any other non-ground elevation point in turn;

[0072] Based on the distance, calculate the average distance between the remaining non-ground elevation points and the current non-ground elevation point;

[0073] Based on the average distance, determine the local density index;

[0074] Based on the local density index, the non-ground elevation points are clustered to obtain several point cloud clusters.

[0075] Optionally, when the clustering analysis module determines the local density index based on the mean distance, it is used for:

[0076] Sort the distances by distance and plot the distance curves in ascending order;

[0077] Based on the distance curve, determine the steep inflection point;

[0078] The distance corresponding to the steep inflection point is used as the local density index of the current non-ground elevation point.

[0079] Optionally, when the attribute analysis module matches each point cloud cluster according to the preset structure information and determines the attributes of each point cloud cluster, it is used for:

[0080] Analyze the preset structural information to determine the geometric features of each structure;

[0081] The geometric features are matched with each point cloud cluster to determine the attributes of each point cloud cluster.

[0082] Optionally, when the clustering analysis module matches the geometric features with each point cloud cluster to determine the attributes of each point cloud cluster, it is used for:

[0083] For each point cloud cluster, the point cloud cluster is downsampled to obtain several simple geometric structures;

[0084] Analyze several simple geometric shapes to determine the possibilities of their combinations;

[0085] The geometric features are matched with each possible combination to determine the attributes of each point cloud cluster.

[0086] Optionally, when the clustering analysis module downsamples each point cloud cluster to obtain several simple geometric structures, it is used for:

[0087] For each point cloud cluster, the density of the point cloud cluster is determined based on the distance.

[0088] Based on the density, the point cloud cluster is divided into several equal voxels;

[0089] Based on the division results, determine the voxel size;

[0090] Obtain the spatial coordinates of each point cloud data, and determine the voxel center of each equally divided voxel based on the spatial coordinates and the voxel size;

[0091] Based on the aforementioned physical properties, several simple geometric structures are obtained.

[0092] Optionally, when the clustering analysis module determines the voxel centroid of each equally divided voxel based on the spatial coordinates and the voxel size, it is used for:

[0093] For each non-ground elevation point in each voxel segment, the index of the non-ground elevation point in the corresponding voxel segment is determined based on the spatial coordinates and the voxel size, using the following formula:

[0094] ;

[0095] in, This indicates the location of non-ground elevation points in each equally divided voxel. Axis index; This indicates the location of non-ground elevation points in each equally divided voxel. Axis index; This indicates the location of non-ground elevation points in each equally divided voxel. Axis index; The voxel size representing each segment of the voxel is in axis, axis, On-axis performance; The spatial coordinates of the non-ground elevation points in each equally divided voxel;

[0096] Based on the index, determine the non-ground elevation points corresponding to the corresponding equal voxels, and determine the center of mass of the voxels based on the spatial coordinates of the non-ground elevation points corresponding to the corresponding equal voxels, using the following formula:

[0097] ;

[0098] in, The spatial coordinates of the said body mass center; This indicates the number of non-ground elevation points in each equally divided voxel; This represents the spatial coordinates of non-ground elevation points within each equally divided voxel.

[0099] Optionally, when the topographic map generation module generates an underwater topographic map based on the attributes and the ground elevation points, it is used for:

[0100] Based on the aforementioned attributes, determine the type of each cloud cluster;

[0101] Analyze the point cloud clusters to determine the overall structure of each point cloud cluster;

[0102] Based on the overall structure and type, determine the spatial distribution of each cloud cluster.

[0103] Based on the aforementioned ground elevation points, determine the ground trend and topographic texture.

[0104] An underwater topographic map is generated based on the ground surface trend, the terrain texture structure, and the spatial distribution. Attached Figure Description

[0105] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0106] Figure 1 This is a schematic diagram illustrating an application scenario provided in one embodiment of this application;

[0107] Figure 2 A flowchart illustrating an intelligent underwater topographic map generation method based on big data, provided as an embodiment of this application;

[0108] Figure 3 This is a schematic diagram of the structure of an intelligent underwater topographic map generation system based on big data, provided as an embodiment of this application. Detailed Implementation

[0109] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0110] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article, unless otherwise specified, generally indicates that the preceding and following related objects have an "or" relationship.

[0111] The embodiments of this application will now be described in further detail with reference to the accompanying drawings.

[0112] Underwater topographic maps are core foundational data for marine resource exploration, waterway maintenance, ecological protection, and underwater engineering construction. Traditional underwater topographic mapping techniques primarily rely on equipment such as sonar and multibeam echo sounders to acquire discrete elevation points, generating regular grids through interpolation algorithms (such as Kriging interpolation or inverse distance weighting) to create topographic maps. However, complex underwater environments (such as vegetation cover, shipwrecks, and rock distribution) result in a large number of non-topographic feature points mixed in with the elevation data. Semantic segmentation techniques based on machine learning have been introduced into underwater data processing, but they depend on a large number of labeled samples, making fully automated processing difficult. Based on this, this application provides an intelligent underwater topographic map generation system and method based on big data. The system acquires and analyzes surveying data to determine ground elevation points and non-ground elevation points. For non-ground elevation points, it clusters them according to the distance between any two non-ground elevation points, resulting in several point cloud clusters. It acquires preset structural information and matches each point cloud cluster according to this information to determine the attributes of each cluster. Based on the attributes and ground elevation points, an underwater topographic map is generated. This ensures the quality and consistency of the point cloud data. By analyzing elevation information, the point cloud data is divided into ground elevation points and non-ground elevation points. Clustering the non-ground layer point cloud data into several point cloud clusters based on the distance between any two non-ground elevation points helps identify different types of underwater objects. By matching preset structural information, the attributes of each point cloud cluster are determined, providing necessary information for generating accurate underwater topographic maps. Based on the attributes of the point cloud clusters and ground elevation points, an underwater topographic map is generated that includes ground topography, terrain texture, and the spatial distribution of objects with different attributes.

[0113] Figure 1 This application provides an illustration of an application scenario. In the process of generating underwater topographic maps, the method provided in this application is used to better distinguish underwater structures. Specifically, the method is applied to any server, where the server interacts with surveying equipment to acquire surveying data, ensuring the quality and consistency of the point cloud data. By analyzing elevation information, the point cloud data is divided into ground elevation points and non-ground elevation points. Based on the distance between any two non-ground elevation points, the non-ground layer point cloud data is clustered into several point cloud clusters, which helps to identify different types of underwater objects. By matching preset structural information, the attributes of each point cloud cluster are determined, providing necessary information for generating accurate underwater topographic maps. Based on the attributes of the point cloud clusters and ground elevation points, an underwater topographic map is generated that includes ground topography, terrain texture, and the spatial distribution of objects with different attributes.

[0114] For specific implementation details, please refer to the following examples.

[0115] Figure 2This is a flowchart illustrating an intelligent underwater topographic map generation method based on big data, provided as an embodiment of this application. The method of this embodiment can be applied to servers in the above scenarios. Figure 2 As shown, the method includes:

[0116] S201. Obtain survey data, analyze the survey data, and determine ground elevation points and non-ground elevation points;

[0117] Surveying data can be geospatial information such as elevation, latitude and longitude, and topographic features obtained through professional surveying equipment and measurement techniques.

[0118] Ground elevation points can be the elevation points that represent the underwater ground surface on underwater topographic maps.

[0119] Non-ground elevation points can be other elevation points in point cloud data besides ground elevation points.

[0120] Specifically, underwater elevation data is collected using surveying equipment such as sonar and laser ranging to generate point cloud datasets. The point cloud data is then cleaned, denoised, and formatted to ensure its quality and consistency. The point cloud data is analyzed to determine the elevation information of each point cloud cluster. Based on preset height thresholds, the point cloud data is divided into ground elevation points and non-ground elevation points.

[0121] S202. For non-ground elevation points, cluster the non-ground elevation points according to the distance between any two non-ground elevation points to obtain several point cloud clusters.

[0122] Point cloud clusters can be groups of non-ground elevation points with similar characteristics, formed by cluster analysis.

[0123] Specifically, for non-ground layer point cloud data, the distance between any two non-ground elevation points is calculated. A clustering algorithm is then used to cluster the non-ground layer point cloud data, forming several point cloud clusters.

[0124] S203. Obtain preset structure information, match each cloud cluster according to the preset structure information, and determine the attributes of each cloud cluster.

[0125] The preset structural information can be the geometric information of known underwater objects or terrain features, such as their shape, size, and distribution.

[0126] Specifically, the structural information of the geometric features of different types of underwater objects is acquired. The clustered point cloud groups are then matched with the preset structural information to determine the attributes of each point cloud cluster.

[0127] S204. Generate an underwater topographic map based on attributes and ground elevation points.

[0128] Underwater topographic maps can display elevation information, topographic texture, and distribution of underwater objects, showcasing underwater topographic features and structures.

[0129] Specifically, based on the attributes of point cloud clusters and ground elevation points, an underwater topographic map is generated, showing the ground trend, terrain texture, and spatial distribution of objects with different attributes.

[0130] This solution ensures the quality and consistency of point cloud data. By analyzing elevation information, the point cloud data is divided into ground elevation points and non-ground elevation points. Based on the distance between any two non-ground elevation points, the non-ground layer point cloud data is clustered into several point cloud clusters, which helps in identifying different types of underwater objects. By matching preset structural information, the attributes of each point cloud cluster are determined, providing the necessary information for generating accurate underwater topographic maps. Based on the attributes of the point cloud clusters and ground elevation points, an underwater topographic map is generated that includes ground topography, terrain texture, and the spatial distribution of objects with different attributes.

[0131] In some embodiments, the mapping data is analyzed to determine the elevation of each point cloud data; for each point cloud data, the point cloud data is layered according to the elevation to obtain a layering result; a preset height threshold is obtained, and the preset height threshold is compared with the average elevation of each layer in the layering result to obtain a comparison result; based on the comparison result, the ground layer and non-ground layer are determined; the point cloud data of the ground layer is used as ground elevation points, and the point cloud data of the non-ground layer is used as non-ground elevation points.

[0132] Point cloud data, which can be acquired by sonar equipment, laser scanners, or other measuring tools, is used to describe the detailed structure of underwater topography.

[0133] Elevation can be the vertical distance of underwater topography relative to sea level.

[0134] Layering results can be achieved by dividing point cloud data into several layers based on elevation during data processing.

[0135] The preset height threshold can be an elevation value set before data processing.

[0136] The mean elevation can be the average of the elevation values ​​of all points in a certain layer.

[0137] The comparison result can be obtained by comparing the preset height threshold with the average elevation in the layered results.

[0138] The ground layer can be the layer in the stratification results where the average elevation is lower than a preset height threshold.

[0139] The non-ground layer can be the layer in the stratification result whose average elevation is higher than the preset height threshold.

[0140] Specifically, the acquired mapping data is cleaned to remove outliers and noise, ensuring data quality. Point cloud data is analyzed to extract elevation information for each point. Based on the extracted elevation information, the point cloud data is layered. A height threshold can be set, classifying points above the threshold as non-ground layers and points below the threshold as ground layers. Statistical analysis is performed on the data for each layer, calculating the mean elevation, standard deviation, and other statistical measures for each layer. A preset height threshold is obtained and compared with the mean elevation of each layer. Based on the comparison results, the ground layer is determined. Typically, the ground layer has a lower mean elevation and a denser distribution of data points. The remaining layers are then identified as non-ground layers. Point cloud data from the ground layers are marked as ground elevation points, and point cloud data from the non-ground layers are marked as non-ground elevation points.

[0141] This solution improves data accuracy and reliability through data cleaning. Elevation information for each point is extracted, providing necessary data support for data layering and analysis. Dividing the point cloud data into ground and non-ground layers helps distinguish underwater topography from other underwater objects, preparing for cluster analysis. Statistical analysis yields the mean and standard deviation of elevation for each layer. Threshold comparison allows for preliminary differentiation between ground and non-ground layers. Accurately identifying the ground layer helps generate more precise underwater topographic maps. Accurately identifying the non-ground layer helps identify and classify underwater objects. Labeling the point cloud data for both ground and non-ground layers facilitates data processing and analysis.

[0142] In some embodiments, for each non-ground elevation point, the distance between the current non-ground elevation point and any other non-ground elevation point is calculated one by one; based on the distance, the average distance between the current non-ground elevation point and any other non-ground elevation point is calculated; based on the average distance, a local density index is determined; based on the local density index, the non-ground elevation points are clustered to obtain several point cloud clusters.

[0143] Distance can be the spatial interval between any non-ground elevation point and the current non-ground elevation point.

[0144] The mean distance can be the average distance between any non-ground elevation point and the current non-ground elevation point.

[0145] Local density indexes can be used to describe the distribution density of points around a given point.

[0146] Specifically, iterate through all non-ground elevation points, and for each non-ground elevation point, calculate its distance to all other non-ground elevation points. The distance can be calculated using the Euclidean distance formula: For the current non-ground elevation point, calculate the mean distance between it and all other non-ground elevation points. Based on the mean distance, determine the local density index of the current non-ground elevation point. The local density index can be a simple mean distance or a more complex metric. Apply hierarchical clustering and other clustering algorithms to cluster the non-ground elevation points. The clustering algorithm divides the points into several clusters based on the distances between points and the local density index. Based on the results of the clustering algorithm, several point cloud clusters are formed.

[0147] This scheme analyzes the spatial relationships between points in a point cloud by calculating the distance between each non-ground elevation point and all other non-ground elevation points. Calculating the mean distance between each non-ground elevation point and all other non-ground elevation points helps assess the local density of that point, i.e., the distribution of surrounding points. Determining local density indices identifies which point cloud regions are dense and which are sparse, aiding in distinguishing different types of underwater objects such as dense aquatic plants and sparse rocks. Applying clustering algorithms divides the point cloud data into several clusters, with points within each cluster sharing similar spatial relationships and attributes, facilitating further analysis of the object type represented by each cluster. This results in several point cloud clusters, each representing a group of underwater objects of the same type.

[0148] In some embodiments, the distances are sorted by distance and distance curves are plotted in ascending order; steep inflection points are determined based on the distance curves; and the distances corresponding to the steep inflection points are used as local density indicators for the current non-ground elevation points.

[0149] A distance curve is a graph that is drawn based on the distance between points in point cloud data processing.

[0150] A sharp inflection point can be a point on a distance curve where the curve suddenly changes from a steady change to a sharp rise.

[0151] Specifically, for each non-ground elevation point, calculate its distance to all other non-ground elevation points in the dataset. Sort the calculated distances in ascending order. Use the sorted distance data to plot a distance curve. The horizontal axis represents distance, and the vertical axis represents the frequency or number of distance occurrences. Analyze the distance curve to identify steep inflection points. These inflection points indicate a sudden increase in the rate of distance increase, marking the transition from the ground layer to the non-ground layer. Use the distance corresponding to the steep inflection point as a local density index for the current non-ground elevation point.

[0152] This method analyzes the spatial relationships between points in a point cloud by calculating the distance between each non-ground elevation point and all other non-ground elevation points. Sort the calculated distances in ascending order to visually observe the distance distribution and provide ordered data for drawing distance curves. Drawing distance curves visually displays the distance distribution and helps identify key features such as inflection points and distribution patterns. Analyzing the distance curves and identifying steep inflection points determines key points of change in the distance distribution, marking the transition from the ground layer to the non-ground layer. Using the distance corresponding to the steep inflection point as a local density index for the current non-ground elevation point accurately reflects the distribution density of points around that point.

[0153] In some embodiments, preset structural information is analyzed to determine the geometric features of each structure; the geometric features are matched with each point cloud cluster to determine the attributes of each point cloud cluster.

[0154] Geometric features can be used to distinguish and identify different underwater objects such as aquatic plants, rocks, and sediments.

[0155] Specifically, the pre-defined structural information of typical geometric features of known different types of underwater objects is analyzed. For each point cloud cluster, its geometric features are extracted. The distribution density of points in the cluster, the shape, size, and orientation of the cluster are calculated. The geometric features of each point cloud cluster are matched with the pre-defined structural information. The similarity between the geometric features of the cluster and the pre-defined structural features is compared. Based on the matching results, the attributes of each point cloud cluster are determined.

[0156] This solution establishes a standardized set of geometric feature templates by analyzing pre-defined structural information. These templates are used to identify and classify point cloud clusters. Extracting the geometric features of point cloud clusters helps distinguish different types of objects, such as those with different shapes, sizes, and densities. By matching the geometric features of point cloud clusters with pre-defined structural information, the similarity between the clusters and known structures can be identified. Based on the matching results, attributes can be determined for each point cloud cluster.

[0157] In some embodiments, for each point cloud cluster, the point cloud cluster is downsampled to obtain several simple geometric structures; the several simple geometric shapes are analyzed to determine the combination possibilities; the geometric features are matched with each combination possibility to determine the attributes of each point cloud cluster.

[0158] Simple geometric structures can be the basic geometric shapes extracted from point cloud data during point cloud data processing.

[0159] Combination possibility can be the possibility of combining simple geometric structures extracted in point cloud data processing according to certain rules and relationships to form more complex shapes.

[0160] Specifically, each point cloud cluster is downsampled to reduce the number of points in the point cloud. Meshization, clustering, or other downsampling techniques are used. Simple geometric structures such as planes, spheres, and cubes are extracted from the downsampled point cloud clusters. These geometric structures can represent the main shape features of the clusters. The combination possibilities between the extracted simple geometric structures are analyzed to determine which geometric structures might combine to form more complex objects. The extracted geometric features are matched with each combination possibility, and the similarity between each combination and the actual shape and structure of the point cloud cluster is evaluated. Based on the matching results, the most likely attributes are determined for each point cloud cluster.

[0161] This scheme reduces the number of points in point cloud clusters through downsampling, lowering data complexity and improving processing efficiency and speed. Extracting simple geometric structures helps simplify the shape of point cloud clusters, making them easier to analyze and match, while also reducing the impact of noise and outliers. Analyzing the combination possibilities between simple geometric structures helps identify different object types present in the clusters. By matching geometric features and combination possibilities, the most likely object attributes for each point cloud cluster can be determined, thereby improving the accuracy and reliability of classification. Determining attributes for each point cloud cluster enables effective classification of underwater objects.

[0162] In some embodiments, for each point cloud cluster, the density of the point cloud cluster is determined based on the distance; the point cloud cluster is divided into several equal voxels based on the density; the voxel size is determined based on the division result; the spatial coordinates of each point cloud data are obtained; the voxel mass center of each equal voxel is determined based on the spatial coordinates and voxel size; and several simple geometric structures are obtained based on the voxel mass center.

[0163] In point cloud data processing, density refers to the number of points per unit volume.

[0164] Equal voxels can be small cubic units that divide a cluster of point clouds in water.

[0165] The partitioning result can be the specific layout and structure of a point cloud cluster after it has been divided into equally divided voxels.

[0166] The voxel size can be the side length of an equally divided voxel.

[0167] Spatial coordinates can be the position of each point in the point cloud within the water.

[0168] The core of a voxel can be the average of the spatial coordinates of all points within a voxel.

[0169] Specifically, the density of a point cloud cluster is calculated based on the number and distribution of points within it. This can be achieved by counting the number of points per unit volume. Based on the cluster density, the point cloud cluster is divided into several equally spaced voxels. The size of each voxel is determined based on the size and density of the point cloud cluster. The voxel size should be small enough to capture the detailed structure within the cluster. The spatial coordinates of each point in the point cloud cluster are obtained to determine the voxel's position. The center of gravity of each voxel is calculated based on the spatial coordinates of points within each voxel. The point cloud cluster is downsampled by removing some points or merging neighboring points to reduce the number of points. Based on the downsampled point cloud data, simple geometric structures such as points, line segments, and planes are extracted.

[0170] This scheme utilizes a method where calculating the density of point cloud clusters helps identify cluster density, making the downsampling process more reasonable. Dividing point cloud clusters into equally sized voxels simplifies the data structure, making processing more efficient while preserving the overall shape and characteristics of the clusters. Determining voxel sizes can adapt to different levels of detail in different regions based on the density and size of the point cloud clusters, ensuring that voxels can reasonably represent the characteristics of the clusters. Obtaining the spatial coordinates of each point cloud data point is fundamental for voxel partitioning and centroid calculation, providing necessary information for determining the position and shape of the voxels. Calculating the voxel centroids helps identify the point distribution within the voxels. Downsampling reduces the number of points in the point cloud clusters, lowers data complexity, and improves processing efficiency and speed. Extracting simple geometric structures clearly represents the basic shape and characteristics of the point cloud clusters.

[0171] In some embodiments, for non-ground elevation points in each equal voxel, the index of the non-ground elevation point in the corresponding equal voxel is determined based on the spatial coordinates and voxel size, using the following formula:

[0172] ;

[0173] in, This indicates the location of non-ground elevation points in each equally divided voxel. Axis index; This indicates the location of non-ground elevation points in each equally divided voxel. Axis index; This indicates the location of non-ground elevation points in each equally divided voxel. Axis index; This indicates the voxel size of each equal voxel. axis, axis, On-axis performance; Represents the spatial coordinates of non-ground elevation points within each equally divided voxel;

[0174] Based on the index, determine the non-ground elevation points corresponding to the corresponding isomeric voxels, and determine the voxel center using the spatial coordinates of these non-ground elevation points, calculated using the following formula:

[0175] ;

[0176] in, Spatial coordinates representing the body's center of gravity; This indicates the number of non-ground elevation points in each equally divided voxel; This represents the spatial coordinates of non-ground elevation points within each equally divided voxel.

[0177] An index can be a number used in point cloud data processing to uniquely identify the position or sequence number of each element in the dataset.

[0178] The quantity can be the number of non-ground elevation points in each equally divided voxel.

[0179] Specifically, a spatial coordinate system is created based on the voxel size to organize the points in the point cloud data. For each non-ground elevation point, an index position is calculated based on its spatial coordinates and voxel size. Each non-ground elevation point is assigned to a corresponding voxel based on the calculated index. For each equally divided voxel, the number of non-ground elevation points it contains is counted. For each equally divided voxel, the voxel center is calculated using the spatial coordinates of all non-ground elevation points assigned to that voxel.

[0180] This scheme assigns non-ground elevation points to corresponding voxels by calculating an index. Determining the non-ground elevation points within each voxel based on the index helps organize and structure the data, making the point cloud data within each voxel more ordered. Counting the number of non-ground elevation points in each voxel helps calculate the voxel mass center, ensuring that the mass center accurately reflects the distribution of points within the voxel. Calculating the mass center yields the representative location of each voxel, which aids in identifying and analyzing the structure and characteristics of point cloud clusters.

[0181] In some embodiments, the type of each point cloud cluster is determined based on attributes; the point cloud clusters are analyzed to determine the overall structure of each point cloud cluster; the spatial distribution of each point cloud cluster is determined based on the overall structure and type; the ground trend and topographic texture are determined based on ground elevation points; and an underwater topographic map is generated based on the ground trend, topographic texture, and spatial distribution.

[0182] Attributes can be features such as color, intensity, number of times returned, and classification label for each point in point cloud data processing.

[0183] The type can be the category of objects or features represented by point cloud clusters, such as ground, aquatic plants, rocks, and sediments.

[0184] The overall structure can be the shape and layout of point cloud clusters.

[0185] Spatial distribution can be the distribution pattern and positional relationship of point cloud clusters in water.

[0186] The terrain can refer to the slope and undulation of the underwater surface.

[0187] Topographic texture can be the surface features of underwater land, such as sand, rocks, and sediments.

[0188] Specifically, attributes such as reflection intensity, color, and number of returns for each point are extracted from the point cloud data. These attributes help distinguish different types of point cloud clusters. Cluster analysis is performed on the point cloud data to divide it into different clusters based on the similarity between points. Each cluster may represent a different object type. Each cluster is analyzed to extract geometric features such as shape, size, and density to further determine the cluster type. Based on the extracted features and attributes, the types of land, aquatic plants, and rocks within each cluster are identified. The spatial distribution of each cluster is analyzed to determine its location and relationships in the water. Ground elevation points are analyzed to determine the trend and texture of the underwater surface. Based on the trend of ground elevation points, the texture of the terrain, and the spatial distribution of the clusters, an underwater topographic map is constructed.

[0189] This scheme, through attribute classification, can distinguish different types of point cloud clusters, such as land, aquatic plants, and rocks. Analyzing the overall structure helps identify the shape and distribution of point cloud clusters, providing a more accurate model for generating topographic maps. Determining the spatial distribution reveals the layers and complexity of underwater topography, contributing to the generation of more detailed and accurate topographic maps. By analyzing ground elevation points, the trend and texture of the underwater surface can be determined, providing crucial topographic information for map generation. Based on the surface trend, topographic texture, and spatial distribution of point cloud clusters, the generated underwater topographic maps will be more accurate and detailed, contributing to applications in marine exploration, environmental protection, and underwater ecological research.

[0190] Figure 3 A schematic diagram of the structure of an intelligent underwater topographic map generation system based on big data, provided in an embodiment of this application, is shown below. Figure 3 As shown, the underwater topographic map intelligent generation system 300 based on big data in this embodiment includes: a timed detection module 301, a data transmission module 302, an attribute analysis module 303, and a topographic map generation module 304.

[0191] Data analysis module 301 is used to acquire survey data, analyze the survey data, and determine ground elevation points and non-ground elevation points;

[0192] The clustering analysis module 302 is used to cluster the non-ground elevation points according to the distance between any two non-ground elevation points to obtain several point cloud clusters.

[0193] The attribute analysis module 303 is used to obtain preset structure information, match each cloud cluster according to the preset structure information, and determine the attributes of each cloud cluster.

[0194] The topographic map generation module 304 is used to generate an underwater topographic map based on the attributes and the ground elevation points.

[0195] Optionally, when the data analysis module 301 analyzes the surveying data and determines ground elevation points and non-ground elevation points, it is used for:

[0196] Analyze the survey data to determine the elevation of each cloud data point;

[0197] For each point cloud data point, the point cloud data is divided into layers based on the elevation to obtain the layering result;

[0198] A preset height threshold is obtained, and the preset height threshold is compared with the average elevation of each layer in the layering result to obtain the comparison result.

[0199] Based on the comparison results, the ground layer and non-ground layer are determined;

[0200] The point cloud data of the ground layer is used as ground elevation points, and the point cloud data of the non-ground layer is used as non-ground elevation points.

[0201] Optionally, when the clustering analysis module 302 clusters the non-ground elevation points based on the distance between any two non-ground elevation points to obtain several point cloud clusters, it is used for:

[0202] For each non-ground elevation point, calculate the distance between the current non-ground elevation point and any other non-ground elevation point in turn;

[0203] Based on the distance, calculate the average distance between the remaining non-ground elevation points and the current non-ground elevation point;

[0204] Based on the average distance, determine the local density index;

[0205] Based on the local density index, the non-ground elevation points are clustered to obtain several point cloud clusters.

[0206] Optionally, when the clustering analysis module 302 determines the local density index based on the mean distance, it is used to:

[0207] Sort the distances by distance and plot the distance curves in ascending order;

[0208] Based on the distance curve, determine the steep inflection point;

[0209] The distance corresponding to the steep inflection point is used as the local density index of the current non-ground elevation point.

[0210] Optionally, when the attribute analysis module 303 matches each point cloud cluster according to the preset structure information and determines the attributes of each point cloud cluster, it is used for:

[0211] Analyze the preset structural information to determine the geometric features of each structure;

[0212] The geometric features are matched with each point cloud cluster to determine the attributes of each point cloud cluster.

[0213] Optionally, when the clustering analysis module 302 matches the geometric features with each point cloud cluster to determine the attributes of each point cloud cluster, it is used for:

[0214] For each point cloud cluster, the point cloud cluster is downsampled to obtain several simple geometric structures;

[0215] Analyze several simple geometric shapes to determine the possibilities of their combinations;

[0216] The geometric features are matched with each possible combination to determine the attributes of each point cloud cluster.

[0217] Optionally, when the clustering analysis module 302 downsamples each point cloud cluster to obtain several simple geometric structures, it is used for:

[0218] For each point cloud cluster, the density of the point cloud cluster is determined based on the distance.

[0219] Based on the density, the point cloud cluster is divided into several equal voxels;

[0220] Based on the division results, determine the voxel size;

[0221] Obtain the spatial coordinates of each point cloud data, and determine the voxel center of each equally divided voxel based on the spatial coordinates and the voxel size;

[0222] Based on the aforementioned physical properties, several simple geometric structures are obtained.

[0223] Optionally, when the clustering analysis module 302 determines the voxel centroid of each equally divided voxel based on the spatial coordinates and the voxel size, it is used for:

[0224] For each non-ground elevation point in each voxel segment, the index of the non-ground elevation point in the corresponding voxel segment is determined based on the spatial coordinates and the voxel size, using the following formula:

[0225] ;

[0226] in, This indicates the location of non-ground elevation points in each equally divided voxel. Axis index; This indicates the location of non-ground elevation points in each equally divided voxel. Axis index; This indicates the location of non-ground elevation points in each equally divided voxel. Axis index; The voxel size representing each segment of the voxel is in axis, axis, On-axis performance; The spatial coordinates of the non-ground elevation points in each equally divided voxel;

[0227] Based on the index, determine the non-ground elevation points corresponding to the corresponding equal voxels, and determine the center of mass of the voxels based on the spatial coordinates of the non-ground elevation points corresponding to the corresponding equal voxels, using the following formula:

[0228] ;

[0229] in, The spatial coordinates of the said body mass center; This indicates the number of non-ground elevation points in each equally divided voxel; This represents the spatial coordinates of non-ground elevation points within each equally divided voxel.

[0230] Optionally, when the topographic map generation module 304 generates an underwater topographic map based on the attributes and the ground elevation points, it is used to:

[0231] Based on the aforementioned attributes, determine the type of each cloud cluster;

[0232] Analyze the point cloud clusters to determine the overall structure of each point cloud cluster;

[0233] Based on the overall structure and type, determine the spatial distribution of each cloud cluster.

[0234] Based on the aforementioned ground elevation points, determine the ground trend and topographic texture.

[0235] An underwater topographic map is generated based on the ground surface trend, the terrain texture structure, and the spatial distribution.

[0236] The system in this embodiment can be used to execute the methods of any of the above embodiments, and its implementation principle and technical effect are similar, so they will not be described again here.

Claims

1. A method for intelligent generation of underwater topographic maps based on big data, characterized in that, include: Acquire survey data, analyze the survey data, and determine ground elevation points and non-ground elevation points; For the non-ground elevation points, based on the distance between any two non-ground elevation points, the non-ground elevation points are clustered to obtain several point cloud clusters. Obtain preset structure information, and match each cloud cluster according to the preset structure information to determine the attributes of each cloud cluster. The preset structural information is the geometric information of the shape, size, and distribution characteristics of known underwater objects or terrain features; Based on the attributes and the ground elevation points, an underwater topographic map is generated, including: Based on the aforementioned attributes, determine the type of each cloud cluster; Analyze the point cloud clusters to determine the overall structure of each point cloud cluster; Based on the overall structure and type, determine the spatial distribution of each cloud cluster. Based on the aforementioned ground elevation points, determine the ground trend and topographic texture. An underwater topographic map is generated based on the ground surface trend, the terrain texture structure, and the spatial distribution.

2. The method according to claim 1, characterized in that, The analysis of the survey data to determine ground elevation points and non-ground elevation points includes: Analyze the survey data to determine the elevation of each cloud data point; For each point cloud data point, the point cloud data is divided into layers based on the elevation to obtain the layering result; A preset height threshold is obtained, and the preset height threshold is compared with the average elevation of each layer in the layering result to obtain the comparison result. Based on the comparison results, the ground layer and non-ground layer are determined; The point cloud data of the ground layer is used as ground elevation points, and the point cloud data of the non-ground layer is used as non-ground elevation points.

3. The method according to claim 2, characterized in that, For the non-ground elevation points, based on the distance between any two non-ground elevation points, the non-ground elevation points are clustered to obtain several point cloud clusters, including: For each non-ground elevation point, calculate the distance between the current non-ground elevation point and any other non-ground elevation point in turn; Based on the distance, calculate the average distance between the remaining non-ground elevation points and the current non-ground elevation point; Based on the average distance, determine the local density index; Local density index is an index that describes the distribution density of points around an elevation point. Based on the local density index, the non-ground elevation points are clustered to obtain several point cloud clusters.

4. The method according to claim 3, characterized in that, The step of determining the local density index based on the mean distance includes: Sort the distances by distance and plot the distance curves in ascending order; Based on the distance curve, determine the steep inflection point; The distance corresponding to the steep inflection point is used as the local density index of the current non-ground elevation point.

5. The method according to claim 2, characterized in that, The step of matching each cloud cluster based on the preset structure information and determining the attributes of each cloud cluster includes: Analyze the preset structural information to determine the geometric features of each structure; The geometric features are matched with each point cloud cluster to determine the attributes of each point cloud cluster.

6. The method according to claim 5, characterized in that, The step of matching the geometric features with each point cloud cluster to determine the attributes of each point cloud cluster includes: For each point cloud cluster, the point cloud cluster is downsampled to obtain several simple geometric structures; Simple geometric structures are basic geometric shapes extracted from point cloud data to represent the main shape features of point cloud clusters. Analyze several simple geometric shapes to determine the possibilities of their combinations; The geometric features are matched with each possible combination to determine the attributes of each point cloud cluster.

7. The method according to claim 6, characterized in that, For each point cloud cluster, the point cloud cluster is downsampled to obtain several simple geometric structures, including: For each point cloud cluster, the density of the point cloud cluster is determined based on the distance. Based on the density, the point cloud cluster is divided into several equal voxels; Based on the division results, determine the voxel size; Obtain the spatial coordinates of each point cloud data, and determine the voxel center of each equally divided voxel based on the spatial coordinates and the voxel size; Based on the aforementioned physical properties, several simple geometric structures are obtained.

8. The method according to claim 7, characterized in that, The step of determining the voxel centroid of each equally divided voxel based on the spatial coordinates and the voxel size includes: For each non-ground elevation point in each voxel segment, the index of the non-ground elevation point in the corresponding voxel segment is determined based on the spatial coordinates and the voxel size, using the following formula: ； in, This indicates the location of non-ground elevation points in each equally divided voxel. Axis index; This indicates the location of non-ground elevation points in each equally divided voxel. Axis index; This indicates the location of non-ground elevation points in each equally divided voxel. Axis index; The voxel size representing each segment of the voxel is in axis, axis, On-axis performance; The spatial coordinates of the non-ground elevation points in each equally divided voxel; Based on the index, determine the non-ground elevation points corresponding to the corresponding equal voxels, and determine the center of mass of the voxels based on the spatial coordinates of the non-ground elevation points corresponding to the corresponding equal voxels, using the following formula: ； in, The spatial coordinates of the said body mass center; This indicates the number of non-ground elevation points in each equally divided voxel; This represents the spatial coordinates of non-ground elevation points within each equally divided voxel.

9. A smart underwater topographic map generation system based on big data, characterized in that, The method applied to any one of claims 1-8 includes: The data analysis module is used to acquire and analyze survey data to determine ground elevation points and non-ground elevation points. The clustering analysis module is used to cluster the non-ground elevation points according to the distance between any two non-ground elevation points to obtain several point cloud clusters. The attribute analysis module is used to obtain preset structure information, match each point cloud cluster according to the preset structure information, and determine the attributes of each point cloud cluster. The topographic map generation module is used to generate an underwater topographic map based on the attributes and the ground elevation points.

Images