Three-dimensional visual digital intelligent base platform based on multi-source data integration

Abstract

The invention relates to the technical field of digital intelligent base platforms, and particularly discloses a three-dimensional visual digital intelligent base platform based on multi-source data integration, and the platform comprises a data collection module which obtains the point cloud data of a target region, and builds an initial three-dimensional model; setting an updating period, regularly collecting new point cloud data and reconstructing a model, and numbering and storing the new point cloud data as historical modeling data; the data analysis module is used for dividing the initial model into three-dimensional grids and generating new grids after each update; calculating the similarity between each new grid and the initial grid, generating a change curve, and detecting the periodicity through Fourier analysis and other methods; the updating module is used for predicting an optimal model of a new grid in combination with the current similarity and historical data if the similarity curve has periodicity; integrating all the new grid data, and outputting an updated three-dimensional model; according to the method, a model updating strategy is optimized through periodic analysis, redundant calculation is reduced, and modeling efficiency and accuracy are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of digital intelligence base platforms, and specifically to a three-dimensional visualization digital intelligence base platform based on multi-source data integration. Background Art

[0002] The Digital Intelligence Platform integrates multiple advanced technologies to provide comprehensive, intelligent services for government and enterprise users. Leveraging professional GIS software and 3D modeling tools, it transforms collected data into high-precision 3D models. These models accurately reproduce the shape, position, size, and other information of objects, providing a foundation for subsequent analysis and application.

[0003] Under today's existing technology, key data such as the target area's topography, buildings, roads, and water systems is not static but changes significantly over time. For example, as urban development progresses, new buildings are erected, older structures are demolished and rebuilt, road networks are continuously expanded and optimized, and river systems can change direction and form due to natural factors or human intervention. These dynamic changes keep the data foundation of the 3D digital intelligence platform in constant flux.

[0004] In order to ensure that the generated three-dimensional digital intelligence base platform can accurately and truly reflect the latest conditions in the real world, it is usually necessary to re-collect relevant data in a timely manner and update the platform accordingly. However, this process of re-collecting data is not only tedious and complicated, but also involves huge cost investment. From a manpower perspective, it is necessary to organize a professional surveying and mapping team to go deep into the site and use advanced measuring instruments and equipment to obtain the latest geographic information data point by point, line by line, and surface by surface. The manpower cost required during this period is very high. From a material perspective, the purchase, maintenance and replacement of measuring equipment, the deployment and use of data collection vehicles, and the investment in a large number of storage devices all constitute a considerable expense. In addition, the time cost cannot be ignored. Large-scale data updates often take a long time to complete. During this period, some application scenarios with high timeliness requirements may be missed, which in turn affects the practicality of the platform and the timeliness of decision support. Summary of the Invention

[0005] The purpose of the present invention is to provide a three-dimensional visualization digital intelligence base platform based on multi-source data integration to solve the above technical problems.

[0006] The purpose of the present invention can be achieved through the following technical solutions: The data acquisition module of the 3D visualization digital intelligence platform based on multi-source data integration: acquires point cloud data within the target area and establishes an initial 3D model based on the point cloud data; sets an update cycle to obtain an update node, and re-acquires new point cloud data within the target area each time an update node is reached; re-establishes the 3D model based on the new point cloud data, numbers them sequentially, and obtains historical modeling data; Data analysis module: divide the initial three-dimensional model into three-dimensional rectangular grids to obtain new grids and initial grids, and obtain new grid data and initial grid data; sequentially obtain the similarity between each new grid data and the initial grid data, obtain a similarity change curve, and determine whether the similarity change curve of the new grid has periodicity; Update module: obtain the current point cloud data at the latest update node. If the similarity change curve of the new grid is periodic, obtain the current similarity between the current point cloud data of the new grid and the initial grid data; obtain the historical modeling of the new grid based on the current similarity and the historical modeling data; and obtain the updated three-dimensional modeling from the historical modeling of all new grids.

[0007] As a further solution of the present invention: the historical modeling data includes three-dimensional models of various numbers, and the point cloud data is three-dimensional point cloud data obtained by splicing and fusing point clouds of all viewing angles or positions of the target area.

[0008] As a further solution of the present invention, the process of splicing and fusing point cloud data of all viewing angles or positions includes: All point clouds in the point cloud data are obtained, and the corner points, edges, and planes of all point clouds are obtained and recorded as point cloud features; feature descriptors are obtained according to the point cloud features, and the correspondence between each point cloud is obtained based on a feature matching algorithm, and a transformation matrix is determined, and the point cloud data is converted into the same three-dimensional coordinate system to obtain the final point cloud data.

[0009] As a further solution of the present invention, the process of establishing an initial three-dimensional model based on the point cloud data includes: Based on the spherical Voronoi diagram method, all point clouds are regarded as one site, a spherical Voronoi diagram is constructed, and a grid model is generated according to the Voronoi diagram; and based on the modeling method of the spring-mass model, the point cloud data is regarded as a spring-mass system, and the equilibrium state of each point cloud is obtained by simulating physical mechanics, and finally a three-dimensional model with a smooth surface is obtained.

[0010] As a further solution of the present invention: the process of obtaining the new grid data and the initial grid data includes: The initial three-dimensional model is divided into a three-dimensional rectangular grid to obtain a plurality of three-dimensional rectangular grids, and the three-dimensional rectangular grids in the initial three-dimensional model are recorded as initial grids, and the three-dimensional rectangular grids in the three-dimensional model are recorded as new grids; the new point cloud data in the new grids are recorded as new grid data, and the point cloud data in the initial grids are recorded as initial grid data.

[0011] As a further solution of the present invention: the process of obtaining the similarity includes: Based on the feature extraction algorithm, feature points of the initial grid data are extracted, recorded as initial grid feature points, and a matching relationship between the initial grid feature points is established; feature points of the new grid data are extracted, recorded as new grid feature points, and the number M of new grid feature points that satisfy the matching relationship is obtained to obtain a similarity P=M / min(N1, N2), where min(N1, N2) represents the minimum value between the total number N1 of the initial grid feature points and the total number N2 of the new grid feature points.

[0012] As a further solution of the present invention, the process of determining whether the similarity change curve of the new square is periodic includes: Obtain a curve expression f(x) of the similarity change curve, where x represents a number. If there are any two positive numbers a and T such that f(x) satisfies: , then the similarity change curve has periodicity; otherwise, the similarity change curve does not have periodicity.

[0013] As a further solution of the present invention: the process of obtaining the historical modeling of the new grid includes: Obtain all coordinate points on the similarity change curve of the new square, obtain the similarity corresponding to each coordinate point, obtain the absolute value of the difference between the similarity of each coordinate point and the current similarity, and select the similarity with the smallest absolute value of the difference, which is recorded as the adjacent similarity; obtain the number corresponding to the adjacent similarity, and obtain a three-dimensional model of the number, model the area within the new square on the three-dimensional model, and record it as the historical model of the new square.

[0014] Beneficial effects of the present invention: This method automatically matches historical modeling data by analyzing the periodicity of similarity change curves, reducing redundant data collection and full model reconstruction. Local remodeling is triggered only for areas with significant or non-periodic changes (such as new construction sites), avoiding resource waste in global data processing. Feature matching and coordinate system normalization (e.g., spherical Voronoi diagram + spring-mass model) ensure seamless integration of spliced 3D point cloud data, eliminating misalignment issues associated with multi-view acquisition. 3D rectangular meshing combined with feature point matching (SIFT / SURF algorithms) enables sub-grid change detection, calculates similarity, and sensitively captures subtle changes. If the similarity curve meets periodicity, the optimal historical model fragment (e.g., the model corresponding to adjacent similarities) is automatically invoked, avoiding repeated calculations and improving response speed. New modeling is immediately triggered for areas with non-periodic changes (e.g., sudden demolition) to ensure data timeliness. Through a closed-loop "dynamic perception-intelligent analysis-precise update" approach, this method achieves low-cost, high-precision, and adaptive operation and maintenance of a 3D digital intelligence platform, addressing the two core pain points of resource waste and data lag in traditional static modeling. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The present invention will be further described below with reference to the accompanying drawings.

[0016] Figure 1 It is a flow chart of the three-dimensional visualization digital intelligence base platform based on multi-source data integration of the present invention. DETAILED DESCRIPTION

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0018] See also Figure 1 As shown, the present invention is a three-dimensional visualization digital intelligence base platform based on multi-source data integration, including: Data acquisition module: acquires point cloud data within the target area and establishes an initial 3D model based on the point cloud data; sets an update cycle, reaches an update node every other update cycle, reacquires point cloud data within the target area each time it reaches the update node, records it as new point cloud data, and re-establishes a 3D model based on the new point cloud data. The re-established 3D models are numbered in sequence; and historical modeling data is obtained based on the numbered 3D models. It is understood that by scanning the target area to obtain point cloud data of its surface, this data can reflect the three-dimensional shape and structure of the target; using this data, an initial three-dimensional model of the target can be constructed; in order to adapt to the changes of the target area over time, a fixed update cycle is set; whenever this cycle is reached, an update operation will be triggered to re-collect the point cloud data of the target area; the three-dimensional model will be rebuilt using the newly collected point cloud data, and each reconstructed model will be numbered to record the change history of the model; all numbered three-dimensional models will be arranged in sequence to form historical modeling data for subsequent analysis and comparison; In a preferred embodiment of the present invention, the point cloud data is three-dimensional point cloud data obtained by splicing and fusing point clouds of all viewing angles or positions of the target area; In a preferred embodiment of the present invention, the process of stitching and fusing point cloud data of all viewing angles or positions includes: Obtain all point clouds in the point cloud data, and obtain corner points, edges, and planes of all point clouds, which are recorded as point cloud features; obtain feature descriptors based on the point cloud features, and obtain the correspondence between each point cloud based on a feature matching algorithm, determine a transformation matrix, and transform the point cloud data into the same three-dimensional coordinate system to obtain final point cloud data; In a preferred embodiment of the present invention, the process of establishing an initial three-dimensional model based on the point cloud data includes: Based on the spherical Voronoi diagram method, all point clouds are regarded as one site, a spherical Voronoi diagram is constructed, and a mesh model is generated based on the Voronoi diagram; and based on the modeling method of the spring-mass model, the point cloud data is regarded as a spring-mass system, and the equilibrium state of each point cloud is obtained by simulating physical mechanics, and finally a three-dimensional model with a smooth surface is obtained; It should be noted that the spherical Voronoi diagram is a spatial partitioning method based on geometric principles. It treats point cloud data as stations and generates a grid model by constructing a spherical Voronoi diagram. Each station (point cloud) corresponds to a region, which is composed of the set of points closest to the station. This method can efficiently divide the three-dimensional space into multiple non-overlapping regions, providing a basic framework for subsequent modeling. The spring-mass model is a modeling method that treats point cloud data as a spring-mass system to simulate the behavior of springs in physics. In this system, point clouds are treated as mass points, connected by virtual springs. By simulating physical and mechanical processes such as force balance and energy minimization, the entire system reaches a stable state, thereby obtaining a smooth 3D model. As a preferred embodiment of the present invention, the process of setting the update period includes: The update cycle of the 3D modeling of the target area needs to be determined based on factors such as the speed of urban development, technical means, application requirements, and resource budget. For high-speed development areas, the recommended update cycle is 3-6 months; for medium-speed change areas, the recommended update cycle is 1-2 years; and for low-speed stable areas, the recommended update cycle is 3-5 years. It should be noted that the target area may change over time (due to urban development, changes in the natural environment, etc.). Regularly updating the 3D model ensures that the model always accurately reflects the latest actual situation. By comparing 3D models at different time points, it is possible to analyze the changing trends of the target area and even predict future development trends. In areas such as urban planning and resource management, timely and accurate 3D models can provide decision makers with important visual information, helping them make more scientific and reasonable decisions. Data analysis module: divides the initial three-dimensional model into three-dimensional rectangular grids to obtain a plurality of three-dimensional rectangular grids, records the three-dimensional rectangular grids in the initial three-dimensional model as initial grids, and records the three-dimensional rectangular grids in the three-dimensional model as new grids; records the new point cloud data in the new grids as new grid data, and records the point cloud data in the initial grids as initial grid data; Sequentially obtaining similarities between the data of each new square and the data of the initial square to obtain similarities of the new squares in the three-dimensional model with each number, and establishing similarity change curves based on the similarities of the new squares with each number; performing periodicity analysis on the similarity change curves to determine whether the similarity change curves of the new squares have periodicity; It is understandable that dividing a 3D model into multiple small, rectangular grids (voxels) makes it easier to perform local analysis of the model. For each new grid, the similarity of the point cloud data contained in it is calculated with respect to the initial grid. This usually involves comparing the shape, density, or other relevant features of the two point cloud datasets. Over time, a curve is generated showing the change in similarity between the new and old grids. This curve is analyzed to determine whether there are repetitive patterns or cycles, that is, to determine whether the change in similarity shows regular fluctuations. As a preferred embodiment of the present invention, the process of obtaining the similarity of the new square includes: Extracting feature points of the initial grid data based on a feature extraction algorithm (such as SIFT or SURF), recording them as initial grid feature points, and establishing a matching relationship between the initial grid feature points; extracting feature points of the new grid data, recording them as new grid feature points, and obtaining the number M of new grid feature points that satisfy the matching relationship, and obtaining a similarity P = M / min(N1, N2), where min(N1, N2) represents the minimum value between the total number N1 of the initial grid feature points and the total number N2 of the new grid feature points; It should be noted that the matching relationship refers to establishing a correspondence between feature points in two or more point cloud datasets. This correspondence is based on the attributes of the feature points (such as position, normal vector, descriptor, etc.) to determine whether they represent the same point in the same space or points with similar features. The feature extraction algorithms (such as SIFT and SURF) are used to detect and describe local features from images or point cloud data. SIFT (Scale Invariant Feature Transform) and SURF (Speeded Up Robust Features) are two widely used algorithms that can identify stable feature points under different viewing angles and lighting conditions. The matching relationship between the feature points in the initial grid and the new grid is determined by comparing them. This usually involves calculating a similarity metric for the feature points, such as the Euclidean distance, and selecting the most similar pair as a match. The similarity is calculated based on the ratio of the number of matched feature points to the total number of feature points. It is worth noting that a pair of matched feature points is recorded as two matched feature points; As a preferred embodiment of the present invention, the process of establishing the similarity change curve includes: A two-dimensional coordinate system is established with the number as the horizontal coordinate and the similarity as the vertical coordinate; each numbered new square and its similarity are converted into coordinate points of corresponding positions on the two-dimensional coordinate system, and each coordinate point is connected with a smooth curve to obtain a similarity change curve; In a preferred embodiment of the present invention, the process of periodically analyzing the similarity change curve includes: Obtain a curve expression f(x) of the similarity change curve, where x represents a number. If there are any two positive numbers a and T such that f(x) satisfies: , then the similarity change curve has periodicity; otherwise, the similarity change curve does not have periodicity; Update module: When the latest update node is reached, the new point cloud data at the latest update node is obtained and recorded as the current point cloud data; if the similarity change curve of the new grid is periodic, the similarity between the current point cloud data of the new grid and the initial grid data is obtained and recorded as the current similarity; based on the current similarity and the historical modeling data, the historical modeling of the new grid is obtained; and the updated 3D modeling is obtained from the historical modeling of all new grids; It is understood that at the scheduled update time, the system will capture the latest state of the target area and generate the current point cloud data, which reflects the latest geometric information of the scene. If the similarity changes of the new grid are detected to be periodic, it means that the change pattern of the target area is predictable. At this time, by comparing the similarity of the current point cloud with the initial grid data, a quantitative change indicator (current similarity) can be obtained. Combined with the historical modeling data, the historical model of the new grid can be more accurately restored or predicted. The historical modeling results of all new grids are used to comprehensively construct or update the 3D model of the entire target area. As a preferred embodiment of the present invention, if the similarity change curve of the new square does not have similarity, the new square is re-modeled according to the current point cloud data and directly recorded as the historical modeling of the new square; It is understandable that if the similarity change curve does not show obvious periodicity or regularity, that is, there is no similarity, it means that the change of the new square is irregular, or there is insufficient historical data to identify a stable change pattern. In this case, since it is impossible to rely on historical trends to predict the future state, the new square is directly modeled using the latest point cloud data. This method assumes that the currently observed state is the most accurate representation and is suitable for rapidly changing or unpredictable scenarios. This strategy ensures that the model is updated promptly, even in the absence of clear patterns of change, to always reflect the latest state of the target area. This is particularly important for handling sudden events or short-term fluctuations in dynamic environments. When complex computations are detected (such as for periodic analysis), directly using the latest data for modeling reduces unnecessary computational overhead, thereby improving the overall efficiency of the system.

[0019] In a preferred embodiment of the present invention, the process of obtaining the historical modeling of the new grid includes: Obtain all coordinate points on the similarity change curve of the new square, obtain the similarity corresponding to each coordinate point, obtain the absolute value of the difference between the similarity of each coordinate point and the current similarity, and select the similarity with the smallest absolute value of the difference, recording it as the adjacent similarity; obtain the number corresponding to the adjacent similarity, and obtain a three-dimensional model of the number, and model the area within the new square on the three-dimensional model, recording it as the historical model of the new square; It should be noted that by regularly updating point cloud data and combining it with historical information, it is ensured that the model reflects the latest status and that historical trends are used to improve the accuracy of modeling. Instead of rebuilding the entire model with each update, only the areas that have changed are updated as needed, thereby saving computing resources and time. It can effectively handle gradual or sudden changes in the target area, making the final three-dimensional model closer to the actual situation.

[0020] The above is a detailed description of an embodiment of the present invention. However, the content described is only a preferred embodiment of the present invention and should not be considered to limit the scope of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A 3D visualization digital intelligence platform based on multi-source data integration, characterized by: include: Data acquisition module: acquires point cloud data within the target area and builds an initial three-dimensional model based on the point cloud data; An update cycle is set to obtain an update node, and new point cloud data within the target area is reacquired each time the update node is reached; a three-dimensional model is rebuilt based on the new point cloud data, and the data is numbered in sequence, and historical modeling data is obtained; Data analysis module: divide the initial three-dimensional model into three-dimensional rectangular grids to obtain new grids and initial grids, and obtain new grid data and initial grid data; sequentially obtain the similarity between each new grid data and the initial grid data, obtain a similarity change curve, and determine whether the similarity change curve of the new grid has periodicity; Update module: obtains the current point cloud data at the latest update node, and if the similarity change curve of the new grid is periodic, obtains the current similarity between the current point cloud data of the new grid and the initial grid data; obtains the historical modeling of the new grid based on the current similarity and the historical modeling data; The updated three-dimensional model is obtained from the historical modeling of all new squares.

2. The three-dimensional visualization digital intelligence base platform based on multi-source data integration according to claim 1 is characterized in that: The historical modeling data includes three-dimensional models of various numbers, and the point cloud data is three-dimensional point cloud data obtained by splicing and fusing point clouds of all viewing angles or positions of the target area.

3. The three-dimensional visualization digital intelligence base platform based on multi-source data integration according to claim 2 is characterized in that: The process of stitching and fusing point cloud data from all viewpoints or positions includes: All point clouds in the point cloud data are obtained, and the corner points, edges, and planes of all point clouds are obtained and recorded as point cloud features; feature descriptors are obtained according to the point cloud features, and the correspondence between each point cloud is obtained based on a feature matching algorithm, and a transformation matrix is determined, and the point cloud data is converted into the same three-dimensional coordinate system to obtain the final point cloud data.

4. The three-dimensional visualization digital intelligence base platform based on multi-source data integration according to claim 3 is characterized in that: The process of establishing an initial three-dimensional model based on the point cloud data includes: Based on the spherical Voronoi diagram method, all point clouds are regarded as one site, a spherical Voronoi diagram is constructed, and a grid model is generated according to the Voronoi diagram; and based on the modeling method of the spring-mass model, the point cloud data is regarded as a spring-mass system, and the equilibrium state of each point cloud is obtained by simulating physical mechanics, and finally a three-dimensional model with a smooth surface is obtained.

5. The three-dimensional visualization digital intelligence base platform based on multi-source data integration according to claim 1 is characterized in that: The process of obtaining the new grid data and the initial grid data includes: The initial three-dimensional model is divided into a three-dimensional rectangular grid to obtain a plurality of three-dimensional rectangular grids, and the three-dimensional rectangular grids in the initial three-dimensional model are recorded as initial grids, and the three-dimensional rectangular grids in the three-dimensional model are recorded as new grids; the new point cloud data in the new grids are recorded as new grid data, and the point cloud data in the initial grids are recorded as initial grid data.

6. The three-dimensional visualization digital intelligence base platform based on multi-source data integration according to claim 1 is characterized in that: The process of obtaining the similarity includes: Based on the feature extraction algorithm, feature points of the initial grid data are extracted, recorded as initial grid feature points, and a matching relationship between the initial grid feature points is established; feature points of the new grid data are extracted, recorded as new grid feature points, and the number M of new grid feature points that satisfy the matching relationship is obtained to obtain a similarity P=M / min(N1, N2), where min(N1, N2) represents the minimum value between the total number N1 of the initial grid feature points and the total number N2 of the new grid feature points.

7. The three-dimensional visualization digital intelligence base platform based on multi-source data integration according to claim 1 is characterized in that: The process of determining whether the similarity change curve of the new square is periodic includes: Obtain a curve expression f(x) of the similarity change curve, where x represents a number. If there are any two positive numbers a and T such that f(x) satisfies: , then the similarity change curve has periodicity; otherwise, the similarity change curve does not have periodicity.

8. The three-dimensional visualization digital intelligence base platform based on multi-source data integration according to claim 1 is characterized in that: The process of obtaining the historical modeling of the new grid includes: Obtain all coordinate points on the similarity change curve of the new square, obtain the similarity corresponding to each coordinate point, obtain the absolute value of the difference between the similarity of each coordinate point and the current similarity, and select the similarity with the smallest absolute value of the difference, which is recorded as the adjacent similarity; obtain the number corresponding to the adjacent similarity, and obtain a three-dimensional model of the number, model the area within the new square on the three-dimensional model, and record it as the historical model of the new square.

Images