A data management system based on unmanned aerial vehicle surveying and mapping

Abstract

This invention discloses a data management system based on UAV mapping, specifically relating to the field of data management technology. It involves collecting UAV mapping data and performing spatial coordinate registration and unified encoding; dividing spatial regions based on terrain elevation variation coefficient and obstacle projection coverage to obtain high-density and low-density verification zones; calculating the feature matching degree of overlapping areas in adjacent frames using a local feature matching algorithm for the high-density verification zone, and outputting first correction data; performing data compensation using spatial interpolation and cubic spline interpolation for the low-density verification zone to generate second correction data; generating a full-scene reconstruction data frame based on the first and second correction data; allocating regional weight factors based on the full-scene reconstruction data frame, performing differential error correction encoding, and outputting UAV mapping results data, effectively improving the spatial adaptability of data verification and enhancing the overall accuracy and completeness of UAV mapping data.

Description

Technical Field

[0001] This invention relates to the field of data management technology, and more specifically, to a data management system based on unmanned aerial vehicle (UAV) mapping. Background Technology

[0002] Unmanned aerial vehicle (UAV) mapping technology is widely used in the field of spatial data acquisition and management in complex terrains and urban environments.

[0003] Existing UAV mapping data management methods typically employ a unified global verification mechanism, which fails to adequately consider the local spatial differences in the mapping scenario. This makes it difficult to effectively guarantee the data quality in key areas, and local errors accumulate and amplify during processing, seriously affecting the integrity and reliability of the mapping data and failing to meet the requirements for high-precision spatial mapping results. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a data management method and system based on unmanned aerial vehicle (UAV) mapping to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A data management system based on UAV mapping includes:

[0007] Data acquisition module: Acquires the set of raw mapping data frames obtained by the UAV platform and outputs the processed mapping data set;

[0008] Region division module: Based on the spatial characteristics of the surveying and mapping dataset, the surveying and mapping area is divided into high-density verification area and low-density verification area by means of terrain elevation variation coefficient and obstacle projection coverage;

[0009] Feature matching module: Uses a local feature matching algorithm to calculate the matching degree of SIFT feature points in the overlapping area of ​​adjacent frames for the high-density verification region;

[0010] Redundancy check module: When the matching degree is lower than the first threshold, the sliding window CRC check is started and the first correction data is output;

[0011] Data compensation module: Performs data compensation on low-density verification areas, performs spatial interpolation compensation on areas occluded by obstacles, and reconstructs the point cloud topology structure by cubic spline interpolation on areas with abrupt terrain changes, and outputs second correction data;

[0012] Data alignment module: Based on the first and second correction data, it generates a full-scene reconstruction data frame by aligning the boundary data of the high-density verification area and the low-density verification area;

[0013] The differential error correction module performs integrity management on the full-scene reconstruction data frames, dynamically allocates regional weight factors, performs differential error correction coding, and outputs UAV mapping results data.

[0014] In a preferred embodiment, the raw mapping data frame set acquired by the UAV platform is collected, and the processed mapping data set is output, specifically as follows:

[0015] Simultaneously collect 3D point cloud data and 2D image data within the target area;

[0016] The collected 3D point cloud data and 2D image data are time-stamped and tagged according to the recorded flight trajectory information;

[0017] Based on the timestamp synchronization marker, the collected 3D point cloud data and 2D image data are respectively subjected to spatial coordinate registration and data fusion processing to generate a standard format mapping data frame set containing complete spatial coordinate system information and image feature information;

[0018] The standard format surveying data frame set is uniformly encoded according to the spatial data format specification, and the uniformly encoded surveying data set is output.

[0019] In a preferred embodiment, based on the spatial characteristics of the mapping dataset, the mapping area is divided into high-density verification zones and low-density verification zones using the terrain elevation variation coefficient and obstacle projection coverage. Specifically:

[0020] The mapping data frame set is divided into spatial grids, and multiple grid units are generated according to a preset grid size.

[0021] Calculate the terrain elevation variation coefficient for the point cloud data within each grid cell according to the height value distribution;

[0022] The optical image data within each grid cell is mapped to the point cloud data through a projection model to calculate the obstacle projection coverage.

[0023] The terrain elevation variation coefficient and obstacle projection coverage are compared using thresholds. Grid cells with both variation coefficient and coverage higher than the predetermined threshold are marked as high-density verification areas, while grid cells with both variation coefficient and coverage lower than the predetermined threshold are marked as low-density verification areas.

[0024] In a preferred embodiment, a local feature matching algorithm is used to calculate the matching degree of SIFT feature points in the overlapping areas of adjacent frames for the high-density verification region, specifically as follows:

[0025] Two-dimensional image data of the overlapping area between adjacent frames is extracted from the mapping data frames within the grid cells marked as high-density verification areas;

[0026] Detect and extract spatial scale-invariant feature points from two-dimensional image data in overlapping regions;

[0027] The feature vectors of the extracted spatial scale-invariant feature transformation feature points are calculated, and the feature vectors between adjacent frames are matched.

[0028] The matching degree is calculated based on the ratio between the number of successfully matched feature vectors and the total number of extracted feature vectors.

[0029] In a preferred embodiment, when the matching degree is lower than a first threshold, a sliding window CRC check is initiated, and first correction data is output, specifically as follows:

[0030] Set the first threshold as the minimum allowable value for feature vector matching degree;

[0031] For mapping data frames within grid cells of high-density verification zones where the feature vector matching degree is lower than the first threshold, a sliding window cyclic redundancy check is initiated.

[0032] The sliding window method is used to sequentially extract continuous data segments from the mapping data frame to be verified, and a cyclic redundancy check operation is performed on each continuous data segment.

[0033] Output the first correction data of the high-density check area grid cells generated after cyclic redundancy check.

[0034] In a preferred embodiment, data compensation is performed on the low-density verification area by spatial interpolation compensation for obstacle-occluded areas and cubic spline interpolation reconstruction of the point cloud topology for areas with abrupt terrain changes, and the second correction data is output as follows:

[0035] Identify obstacle-occluded areas and terrain abrupt change areas within the low-density verification area;

[0036] When performing data compensation for point cloud data within the area obstructed by obstacles, a spatial interpolation algorithm based on neighboring point clouds is used for interpolation compensation.

[0037] When performing data compensation on point cloud data in areas of abrupt terrain change, a cubic spline curve interpolation algorithm is used to reconstruct the point cloud topology.

[0038] Output the second correction data generated by the low-density check area grid cells.

[0039] In a preferred embodiment, based on the first correction data and the second correction data, a full-scene reconstruction data frame is generated by aligning the boundary data of the high-density verification region and the low-density verification region, specifically as follows:

[0040] Based on the spatial coordinates of the grid cells, the boundary regions of adjacent grid cells between the high-density and low-density verification regions are determined.

[0041] Based on the overlapping point cloud data characteristics of adjacent grid cells within the boundary region, a spatial coordinate registration algorithm is used to align the boundary data of the first correction data of the grid cells in the high-density verification area and the second correction data of the grid cells in the low-density verification area.

[0042] Based on the aligned boundary data, all correction data of the high-density verification area and the low-density verification area are merged to generate a full-scene reconstruction data frame covering the entire surveying area.

[0043] In a preferred embodiment, integrity management is performed on the full-scene reconstruction data frame, regional weight factors are dynamically allocated, differential error correction coding is performed, and UAV mapping results data are output, specifically as follows:

[0044] Data integrity assessment is performed on the reconstructed data frames of the entire scene, and the regional error weighting factor corresponding to each grid cell in the high-density and low-density verification areas is calculated.

[0045] Based on the regional error weighting factor, the regional error weighting factor corresponding to each grid cell is assigned to the data point cloud within the corresponding grid cell.

[0046] Differential error correction coding is used to perform differential coding on the data point cloud after assigning regional error weight factors, generating UAV mapping results data.

[0047] The technical effects and advantages of the data management system based on UAV mapping proposed in this invention are as follows:

[0048] By employing spatial feature analysis and adaptive regional partitioning, refined management of UAV mapping data was achieved, overcoming local data anomalies caused by terrain changes, object occlusion, and uneven flight paths. In high-density areas, local feature matching and sliding window CRC verification were used to improve the accuracy of error detection and correction in key areas. In low-density areas, spatial interpolation and cubic spline interpolation reconstruction methods were used to effectively compensate for missing data. Dynamic differential error correction coding based on regional weight factors improved the overall integrity and reliability of the mapping results data, ensuring the accuracy and stability of mapping data application. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the structure of a data management system based on UAV mapping according to the present invention. Detailed Implementation

[0050] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention. Example 1

[0051] Figure 1 This invention provides a data management system based on UAV mapping, comprising:

[0052] Data acquisition module: Acquires the set of raw mapping data frames obtained by the UAV platform and outputs the processed mapping data set;

[0053] Region division module: Based on the spatial characteristics of the surveying and mapping dataset, the surveying and mapping area is divided into high-density verification area and low-density verification area by means of terrain elevation variation coefficient and obstacle projection coverage;

[0054] Feature matching module: Uses a local feature matching algorithm to calculate the matching degree of SIFT feature points in the overlapping area of ​​adjacent frames for the high-density verification region;

[0055] Redundancy check module: When the matching degree is lower than the first threshold, the sliding window CRC check is started and the first correction data is output;

[0056] Data compensation module: Performs data compensation on low-density verification areas, performs spatial interpolation compensation on areas occluded by obstacles, and reconstructs the point cloud topology structure by cubic spline interpolation on areas with abrupt terrain changes, and outputs second correction data;

[0057] Data alignment module: Based on the first and second correction data, it generates a full-scene reconstruction data frame by aligning the boundary data of the high-density verification area and the low-density verification area;

[0058] The differential error correction module performs integrity management on the full-scene reconstruction data frames, dynamically allocates regional weight factors, performs differential error correction coding, and outputs UAV mapping results data.

[0059] The system collects the raw mapping data frames acquired by the UAV platform and outputs the processed mapping data set, including:

[0060] Simultaneously collect 3D point cloud data and 2D image data within the target area;

[0061] In UAV mapping, the target area refers to a specific spatial region to be mapped, such as mountains, urban building complexes, or complex transportation networks. For this target area, the UAV platform is equipped with both a 3D LiDAR sensor and a high-definition optical image sensor to ensure that, during flight, it can simultaneously record the 3D spatial structure information and surface texture image information of spatial targets within the target area. For example, when conducting a mapping task in a mountainous area, the UAV, at a specific flight altitude and path, simultaneously maps the mountain's surface topography, vegetation cover, and the height and morphological structure of features, thereby acquiring 3D point cloud data and 2D image data within the target area. The 3D point cloud data includes feature location information obtained through LiDAR scanning, containing the 3D spatial coordinates (X, Y, and Z coordinates) of each scanned point in a Cartesian coordinate system. The 2D image data is high-resolution image data captured by the high-definition optical image sensor, recording the surface texture, color, spectral characteristics, and shape contour information of features within the target area.

[0062] The collected 3D point cloud data and 2D image data are time-stamped and tagged according to the recorded flight trajectory information;

[0063] The flight trajectory information recorded in real time by the UAV flight platform serves as a unified reference, enabling timestamp synchronization between different types of data. The flight trajectory information includes the spatial location coordinates (such as latitude and longitude coordinates, elevation data) recorded in real time by the UAV, along with corresponding time information. For example, using the Global Positioning System (GPS), each spatial location point corresponds to a unique GPS satellite time stamp. A one-to-one correspondence is established between the 3D point cloud data collected by LiDAR and the 2D image data captured by high-definition optical image sensors and the timestamps of the corresponding location points in the flight trajectory information. Providing timestamp synchronization ensures accurate correspondence and consistent correlation between different data types.

[0064] Based on the timestamp synchronization marker, the collected 3D point cloud data and 2D image data are respectively subjected to spatial coordinate registration and data fusion processing to generate a standard format mapping data frame set containing complete spatial coordinate system information and image feature information;

[0065] The specific implementation method of spatial coordinate registration is as follows: Using 3D point cloud data synchronized with timestamps as the basic coordinate reference data, and 2D image data as the data to be registered, precise spatial coordinate registration is achieved by utilizing the mapping relationship between 3D spatial coordinates and 2D image projection coordinates. For example, based on the principle of perspective projection transformation, the pixel coordinates in the 2D image data are matched with the corresponding 3D spatial coordinates in the 3D point cloud data, thereby determining the spatial position of the 2D image data in the 3D spatial coordinate system and achieving precise spatial coordinate registration between the data. Data fusion processing includes fusing the spatially registered 3D point cloud data and 2D image data according to their spatial position correspondence. For example, color and texture information in the 2D image data are directly mapped to the corresponding 3D point cloud data points, resulting in complete fused data that contains both spatial coordinate information and image feature information. The data set formed after the above data fusion is stored according to the common standard format in the spatial data industry, thereby generating a unified and standardized set of surveying and mapping data frames.

[0066] The standard format surveying data frame set is uniformly encoded according to the spatial data format specification, and the uniformly encoded surveying data set is output.

[0067] Unified encoding processing refers to the data encoding conversion of a set of standard-formatted surveying and mapping data frames according to a pre-defined spatial data format specification. For example, using the GeoTIFF or LAS standard format specifications published by the Open Geospatial Consortium (OPCW), commonly used in existing spatial data management, the data content of each surveying and mapping data frame, such as spatial coordinate information, color information, texture information, and timestamp information, is encoded into a standard format specification for storage. This ensures the consistency and interoperability of the output surveying and mapping data set. Surveying and mapping data sets that have undergone unified encoding processing facilitate regional division, data verification, data compensation, and overall reconstruction.

[0068] Based on the spatial characteristics of the surveying dataset, the surveyed area is divided into high-density verification zones and low-density verification zones using the terrain elevation variation coefficient and obstacle projection coverage, including:

[0069] The mapping data frame set is divided into spatial grids, and multiple grid units are generated according to a preset grid size.

[0070] The surveying data frame set consists of standard-format surveying data generated after unified encoding processing, containing complete three-dimensional spatial coordinate information and two-dimensional image feature information of the surveyed target area. Spatial grid division refers to the spatial partitioning of the entire surveyed target area, dividing the overall spatial range into multiple grid units of uniform size and regular shape, each grid unit being called a grid cell. The size of the grid cell is predetermined, for example, the horizontal width and vertical length of the grid cell are 10 meters, and the spatial height is set according to the actual situation of the surveyed object. For example, for the surveying of urban building complexes, the grid height is generally set to 5 meters. By setting a spatial coordinate origin, based on a unified spatial rectangular coordinate system, the entire surveying data frame set is spatially divided using an equidistant method, so that each spatial grid cell corresponds to a fixed spatial position interval, thereby generating a series of grid cell sets with definite spatial coordinates and clear boundaries.

[0071] Calculate the terrain elevation variation coefficient for the point cloud data within each grid cell according to the height value distribution;

[0072] The 3D point cloud data contained in each independent grid cell is analyzed and calculated to obtain the terrain elevation variation of each grid cell. The method for calculating the terrain elevation variation coefficient is as follows: First, the 3D coordinate height value of each point cloud data point in the grid cell is extracted, that is, the Z coordinate value in the spatial coordinate system. Then, all Z coordinate values ​​in each grid cell are statistically analyzed to calculate the standard deviation and mean of the Z coordinate values ​​in the grid cell. The terrain elevation variation coefficient is the ratio between the standard deviation of the height values ​​in the grid cell and the mean height value. For example, if the standard deviation of the height values ​​in the point cloud data in the grid cell is 0.8 meters and the mean height is 10 meters, then the terrain elevation variation coefficient in the grid cell is 0.08. The terrain elevation variation coefficient can accurately reflect the terrain undulation and surface elevation changes in each grid cell. If the terrain elevation variation coefficient is high, it indicates that the terrain changes drastically in the area, such as obvious height changes in building clusters, mountains, and deep valleys. A low terrain elevation variation coefficient indicates that the terrain is relatively flat.

[0073] The optical image data within each grid cell is mapped to the point cloud data through a projection model to calculate the obstacle projection coverage.

[0074] A projection model is used to map 2D image data to 3D point cloud data. This projection model includes a perspective projection model, which uses the mapping between 3D spatial coordinates and 2D image plane pixel coordinates to accurately overlay 2D image data within a grid cell onto the corresponding 3D point cloud data. Specifically, the projection model parameters for image sensor capture are first established, including focal length, principal point coordinates, camera position, and orientation. Based on the projection model, each pixel of the 2D image data is projected onto its corresponding 3D spatial position. Using the overlaid image information, the projected area of ​​obstacles (such as trees, buildings, and bridges) within each grid cell is calculated, and the proportion of this projected area to the entire grid cell's horizontal area is determined—the obstacle projection coverage rate. For example, if a grid cell has a horizontal area of ​​100 square meters, and the calculated projected coverage area of ​​obstacles after mapping from the 2D image data is 30 square meters, then the obstacle projection coverage rate is 30%.

[0075] The terrain elevation variation coefficient and obstacle projection coverage are compared with thresholds. Grid cells with both variation coefficient and coverage higher than the predetermined threshold are marked as high-density verification areas, and grid cells with both variation coefficient and coverage lower than the predetermined threshold are marked as low-density verification areas.

[0076] A predetermined threshold is set for both the coefficient of variation of terrain elevation and the obstacle projection coverage. For example, the predetermined threshold for the coefficient of variation of terrain elevation is 0.05, and the predetermined threshold for the obstacle projection coverage is 20%. The calculation results are evaluated for each grid cell. If the coefficient of variation of terrain elevation of a grid cell is greater than 0.05, and the obstacle projection coverage of the grid cell is also greater than 20%, then the grid cell is marked as a high-density verification area. If the coefficient of variation of terrain elevation of a grid cell is less than or equal to 0.05, and the obstacle projection coverage is also less than or equal to 20%, then the grid cell is marked as a low-density verification area. For example, densely built-up urban areas or areas with significant mountain undulations are marked as high-density verification areas, while flat grasslands or open water areas are marked as low-density verification areas. This method completes the accurate division and marking of grid cells throughout the entire surveying area.

[0077] For the high-density verification region, a local feature matching algorithm is used to calculate the matching degree of SIFT feature points in the overlapping areas of adjacent frames, including:

[0078] Two-dimensional image data of the overlapping area between adjacent frames is extracted from the mapping data frames within the grid cells marked as high-density verification areas;

[0079] Because UAV mapping platforms continuously capture multiple mapping data frames during mapping tasks, there is a certain degree of overlap between any two adjacent data frames to ensure complete coverage of the mapping area. For example, in mapping densely built-up urban areas, the UAV takes images every 5 meters horizontally along a pre-planned flight path, with each image overlapping the next by approximately 2 meters. The first step is to extract the two-dimensional image data of the overlapping area corresponding to the spatial location in adjacent mapping data frames. The extraction method for the two-dimensional image data is as follows: using the precise spatial coordinate information recorded in the mapping data frames, the common overlap range between two adjacent data frames is determined; the two-dimensional image data corresponding to the common overlap range is then cropped and extracted from each data frame to obtain a set of overlapping two-dimensional image data between two adjacent frames. Through cropping, the two-dimensional image overlap area can be obtained for each pair of adjacent frames.

[0080] Detect and extract spatial scale-invariant feature points from two-dimensional image data in overlapping regions;

[0081] Spatial scale-invariant feature transform (SMT) feature points are a special type of image feature point that possesses stable feature information under different scales, viewpoints, and rotation conditions, and are often used to achieve accurate registration and matching between images. First, SMT detection is performed individually on the extracted 2D image data of each overlapping region. This includes Gaussian pyramid scale space construction, difference image calculation, and spatial extremum detection. For example, for overlapping region data of a 1024×1024 pixel 2D image, after Gaussian blurring and downsampling, a multi-level pyramid structure is constructed. Each level detects and identifies feature points through difference Gaussian calculation. The 2D image data processed by SMT will generate a large number of feature points, precisely corresponding to salient features on the surface of ground objects in the image, such as building corners, bridge intersections, and road intersections. For example, 800 scale-invariant feature points may be detected within the overlapping region. Each feature point accurately records its coordinate position and scale feature information in the 2D image data; this feature point information is then stored as a feature point set.

[0082] The feature vectors of the extracted spatial scale-invariant feature transformation feature points are calculated, and the feature vectors between adjacent frames are matched.

[0083] To achieve accurate matching between adjacent frames, it is necessary to calculate the feature vector corresponding to each feature point. The feature vector can accurately represent the texture and orientation features of the local region corresponding to each feature point. The calculation method is as follows: for the two-dimensional image data region at the location of each feature point, a local image region of size 16×16 pixels is extracted with the feature point as the center, and then divided into 4×4 small regions. The local gradient direction histogram is calculated in each small region, thereby obtaining a 128-dimensional feature vector for each feature point. The above calculation process is performed one by one for all feature points in the overlapping region, and finally the feature vector set of all feature points in the overlapping region is obtained. Based on the feature vector set of the corresponding overlapping region between two adjacent frames, feature vector matching calculation is performed one by one. The matching method is to calculate the Euclidean distance between two feature vectors. For example, if the Euclidean distance between two 128-dimensional feature vectors is calculated to be 0.2, and the set matching judgment threshold is 0.3, then the two feature vectors are judged to be successfully matched. All feature vectors are matched in the above manner to obtain the result set of feature vector matching.

[0084] The matching degree is calculated based on the ratio between the number of successfully matched feature vectors and the total number of extracted feature vectors.

[0085] The matching degree is the percentage of feature points that successfully match feature vectors out of the total number of extracted feature points. For example, if 800 feature points are detected and extracted in an overlapping region, and 800 feature vectors are calculated, and after matching calculations between adjacent data frames, if 720 feature vectors are successfully matched, then the feature vector matching degree of the overlapping region is 720 divided by 800, which is 90%. A higher matching degree indicates a more accurate spatial correspondence and data quality between adjacent frames within the overlapping region; a lower matching degree indicates anomalies in the image data between adjacent frames.

[0086] When the matching degree is lower than the first threshold, a sliding window CRC check is initiated, and the first correction data is output, including:

[0087] Set the first threshold as the minimum allowable value for feature vector matching degree;

[0088] Feature vector matching degree reflects the accuracy and effectiveness of image data registration between surveying data frames. To ensure the effective implementation of surveying data frame verification, a first threshold is introduced as the minimum allowable value for feature vector matching degree to determine whether error correction and verification processing needs to be initiated for surveying data frames within a grid cell. The process for setting the first threshold is as follows: based on the analysis of a large amount of historical surveying data and empirical values, for example, in surveying data of densely built-up urban areas, historical data statistical analysis shows that grid cells with a feature vector matching degree of 85% or higher have high data reliability, while grid cells with a feature vector matching degree below 85% may have data registration deviations or local data errors. Therefore, 85% is set as the minimum allowable value for feature vector matching degree, i.e., the first threshold. The method for determining the first threshold can also be appropriately adjusted according to the characteristics and requirements of the surveying target area. For example, in mountainous terrain surveying tasks, due to the complex terrain structure and large variations in overlapping areas, it may be necessary to appropriately reduce the first threshold to 80%.

[0089] For mapping data frames within grid cells of high-density verification zones where the feature vector matching degree is lower than the first threshold, a sliding window cyclic redundancy check is initiated.

[0090] Cyclic Redundancy Check (CRBC) is a verification method that uses polynomials to calculate check codes and detect data errors. Combining CRBC with a sliding window approach improves the accuracy and effectiveness of data error location. Specifically, taking a grid cell in a high-density check area as an example, if the calculated feature vector matching degree between adjacent frames is 75%, which is lower than the set first threshold of 85%, then a sliding window CRBC is automatically initiated. Initiating the check means automatically calling the sliding window CRBC method. The size and step size of the sliding window are pre-set. For example, the sliding window length is determined to be 256 data units, and the sliding window step size is 128 data units, gradually sliding and capturing continuous data segments to ensure that each CRBC covers a sufficient data range while having appropriate overlap, facilitating the detection and confirmation of data error locations.

[0091] The sliding window method is used to sequentially extract continuous data segments from the mapping data frame to be verified, and a cyclic redundancy check operation is performed on each continuous data segment.

[0092] A sliding window method is used to perform cyclic redundancy check (CRC) calculations on each segment of the surveying data frame to accurately locate potential data errors within the frame. The specific implementation process of the sliding window method is as follows: First, starting from the initial data unit of the surveying data frame, continuous data segments are extracted, each containing 256 data units. The first CRC calculation is performed, for example using the CRC-32 check method, to obtain a 32-bit CRC value for each continuous data segment. The window then slides forward 128 data units to extract the next continuous data segment, and the same CRC-32 check calculation is performed until all data within the entire surveying data frame has undergone sliding window extraction and CRC calculation. For example, in a surveying data frame containing 1024 data units within a grid cell, the sliding window first extracts data units 1 to 256, then 129 to 384, and so on, achieving comprehensive verification of all data within the data frame. The sliding window cyclic redundancy check method can be used to determine whether there are data errors in each data segment. For example, if the CRC value calculated for a certain data segment is inconsistent with the pre-stored CRC reference value, it is determined that there is a data error in the data segment.

[0093] Output the first correction data of the high-density check zone grid cells generated after cyclic redundancy check;

[0094] For each grid cell mapping data frame, data segments without errors detected through cyclic redundancy check (CRC) are considered valid data. For data segments with abnormal CRC values, their specific locations and anomaly types are recorded. The erroneous data segments are then reacquired and replaced by re-retrieving the original data from the mapping data repository or through a data retransmission mechanism to ensure the integrity and accuracy of the data frame. After completing the verification and replacement of all data segments, all consecutive data segments that have been verified and confirmed to be error-free or corrected are recombined to form a complete set of corrected mapping data frames. This set of corrected mapping data frames is the first set of corrected data. The first set of corrected data is stored in a standard format data file, such as the aforementioned LAS or GeoTIFF file format.

[0095] Data compensation is performed on the low-density verification area. Spatial interpolation compensation is applied to areas obscured by obstacles, and cubic spline interpolation is used to reconstruct the point cloud topology in areas with abrupt terrain changes. The second correction data is then output, including:

[0096] Identify obstacle-occluded areas and terrain abrupt change areas within the low-density verification area;

[0097] First, the data frames within the low-density verification area are examined to identify local areas with obstructions or abrupt changes in terrain. Obstruction-occluded areas refer to regions in the surveying area where buildings, trees, tall structures (such as towers, utility poles, and communication base stations), or terrain features (such as cliffs and valleys) obstruct the line of sight of the UAV's lidar sensor or high-definition image sensor, resulting in incomplete surveying data, missing data, or data density below the average level. For example, in urban surveying, narrow roads between high-rise and low-rise buildings may result in data omissions due to building obstruction; in mountainous terrain surveying, the line of sight in the back slope of mountains and canyons may be obstructed by steep mountains, resulting in lower data density than adjacent areas and blank areas. Abrupt terrain areas refer to areas with drastic changes in surface undulation or elevation, such as riverbank edges, steep slope edges, and excavation boundaries. These typically manifest as sparse or even missing 3D point cloud data, with significant elevation changes between adjacent data, making it impossible to accurately reconstruct the complete terrain through simple linear interpolation or smoothing. Therefore, by analyzing the data of all grid cells in the low-density verification area, such as calculating the missing rate, average density, and degree of elevation change of the adjacent area of ​​the point cloud data in each cell, obstacle-occupied areas and terrain change areas can be identified and marked, providing basic information on location and type for data compensation.

[0098] When performing data compensation for point cloud data within the area obstructed by obstacles, a spatial interpolation algorithm based on neighboring point clouds is used for interpolation compensation.

[0099] Data compensation is performed on identified obstacle-occluded areas to correct data loss caused by occlusion. The spatial interpolation algorithm based on neighboring point clouds refers to using effective point cloud data surrounding the obstacle-occluded area to perform spatial interpolation to supplement missing point cloud data. This includes: firstly, defining the interpolation area range for each identified obstacle-occluded area, for example, the horizontal spatial range of missing data in the obstacle-occluded area is 20m × 20m; extracting effective point cloud data within a 5m to 10m range around the obstacle as input data for interpolation calculation; and then using spatial interpolation methods to compensate for the data in the obstacle-occluded area. For example, the coordinates and elevation values ​​of each missing point are calculated using the inverse distance weighted interpolation method. Specifically, for a point at a missing location, the nearest several effective point clouds are extracted, such as all effective point clouds within 10 meters of the missing location. A weighted average is then performed using the inverse distance of each effective point cloud data point as a weight, thereby calculating the data value at the missing location. For example, if a missing point is located within an obstacle-occluded area, and there are 5 valid point cloud data points within a 5-meter radius of that point, each with different coordinate and elevation values, the elevation of the missing point can be calculated using the inverse distance weighted interpolation method: (multiply the elevation value of each valid point by the inverse of its distance, sum the results), and then divide by the sum of all the inverse distances to obtain the interpolated data. Through interpolation calculations, obstacle-occluded areas that were originally sparse or even missing data can obtain a more complete and accurate data compensation effect, thereby improving the overall consistency and integrity of the mapping data frame.

[0100] When performing data compensation on point cloud data in areas of abrupt terrain change, a cubic spline curve interpolation algorithm is used to reconstruct the point cloud topology.

[0101] For areas with abrupt topographic changes—that is, regions with drastic changes in surface elevation and sparse point cloud data—a cubic spline interpolation algorithm is employed for data compensation and topological reconstruction. The cubic spline interpolation algorithm is a method that achieves high-precision, smooth interpolation using piecewise cubic polynomial functions, suitable for compensating and reconstructing terrain data with significant curvature changes. Specifically: First, based on the limited effective point cloud data identified within the abrupt topographic change area, effective elevation points are determined to define the boundaries of the area. These effective elevation points are typically located at the boundaries of areas with denser data. Using these effective elevation points as boundary control points for cubic spline interpolation, interpolation parameters between each point are calculated. By establishing a continuous and smooth cubic polynomial curve, the trend and characteristics of surface topographic changes are described. Based on the generated cubic spline curve equation, point-by-point interpolation and reconstruction are performed on the missing point cloud data within the abrupt topographic change area to restore the elevation and coordinate data within the area. For example, in the riverbank edge area, where there is only sparse point cloud data, by extracting several effective points on the bank as control points, using a cubic spline interpolation algorithm to generate curve equations, and then reconstructing complete topographic point cloud data point by point within the riverbank area based on the curve equations, the accurate and continuous riverbank topographic contour and topology can be restored, ensuring the integrity and accuracy of the surveying data.

[0102] Output the second correction data generated by the low-density verification grid cells;

[0103] The second calibration data refers to the complete set of mapping data frames after interpolation compensation for obstacle-occluded areas and cubic spline curve interpolation reconstruction for terrain abrupt changes. It is stored in a standardized data format, such as LAS or GeoTIFF, forming the second calibration data file. The second calibration data file records the coordinates, elevation values, and data compensation information of data points within each grid cell of the low-density calibration area, describing the location of data points and the complete representation of the surface topology after interpolation compensation.

[0104] Based on the first and second correction data, a full-scene reconstruction data frame is generated by aligning the boundary data of the high-density verification region and the low-density verification region, including:

[0105] Based on the spatial coordinates of the grid cells, the boundary regions of adjacent grid cells between the high-density and low-density verification regions are determined.

[0106] High-density verification zones refer to areas with a topographic elevation variation coefficient greater than 0.05 and an obstacle projection coverage rate higher than 20%, typically corresponding to densely built-up urban areas or complex mountainous terrain areas. Low-density verification zones, on the other hand, are areas with lower values ​​for both indicators, typically corresponding to flat grasslands or open water areas. The spatial coordinates of a grid cell refer to the range of X, Y, and Z coordinate values ​​determined by the grid cell in a unified spatial rectangular coordinate system. Specifically, firstly, using the grid cell coordinates of the divided and marked high-density and low-density verification zones, grid cells located at the boundaries of the high-density and low-density verification zones are extracted. These grid cells are directly adjacent in spatial location, and their spatial coordinate ranges overlap or border each other. For example, in a city surveying task, if multiple consecutive grid cells in a certain area are spatially adjacent, with some grid cells marked as high-density verification zones and others as low-density verification zones, then the boundary areas between these grid cells form boundary regions. To accurately determine the boundary region, all adjacent grid cells between the high-density and low-density verification areas are compared and analyzed in a spatial rectangular coordinate system. The coordinate range of the boundary grid cells in space is extracted and determined, and the boundary region is marked.

[0107] Based on the overlapping point cloud data characteristics of adjacent grid cells within the boundary region, a spatial coordinate registration algorithm is used to align the boundary data of the first correction data of the grid cells in the high-density verification area and the second correction data of the grid cells in the low-density verification area.

[0108] Both the first and second correction data are stored in a unified standard format, such as LAS or GeoTIFF, recording the spatial coordinate information of the point cloud data within each grid cell. Although both the first and second correction data have been verified and compensated within their respective regions, slight deviations or overlaps in data position may still exist in the boundary region where the high-density verification area meets the low-density verification area. Therefore, spatial coordinate registration processing is required for the data in the boundary region.

[0109] The spatial coordinate registration algorithm is as follows: First, overlapping point cloud data from adjacent grid cells within the boundary region is extracted. Overlapping data points are those that are close to each other in spatial coordinates and represent the same ground surface. A correspondence is established by calculating the distance difference and positional relationship between data points in two grid cells within the overlapping region. Taking an urban area as an example, the boundary grid cells of the high-density verification zone and the low-density verification zone record point cloud data of adjacent building edges or road intersections. These data are strictly corresponding in spatial location. Through feature point correspondence analysis and calculation, such as using the iterative nearest point algorithm, the distance difference and rotation relationship between corresponding feature points within the boundary region are determined. Based on the corresponding feature point information, translation and rotation matrices are calculated for registering the first and second correction data. For example, if a 0.5-meter translation error and a 1-degree rotation error are found between a boundary data frame in the high-density verification zone and the boundary data frame in the low-density verification zone, the registration calculation corrects these translation and rotation errors, aligning the overlapping data points of the two boundary grid cells and completing the spatial coordinate registration.

[0110] Based on the aligned boundary data, all correction data of the high-density verification area and the low-density verification area are merged to generate a full-scene reconstruction data frame covering the entire surveying area.

[0111] The specific method for merging is as follows: First, based on the boundary area data that has been spatially registered, determine the splicing position between the data frames of the high-density verification area and the low-density verification area, eliminating any spatial gaps or overlaps that may exist between the data of the high-density verification area and the low-density verification area; then, seamlessly splice and merge the first correction data frames of all high-density verification areas with the second correction data frames of low-density verification areas according to the positional relationship of the boundary areas, forming complete and continuous surveying and mapping area data; for example, in the area where urban and suburban areas meet, the urban area belongs to the high-density verification area and the suburban area belongs to the low-density verification area. After the boundary area data is spliced ​​after registration, it can completely and accurately reflect the terrain and land feature characteristics at the boundary between urban and suburban areas; the merged data is re-encoded and stored to generate a unified full-scene reconstruction data frame covering the entire surveying and mapping area;

[0112] The full-scene reconstruction data frame is a unified standard format data file, such as LAS or GeoTIFF format data. The file contains the three-dimensional coordinates and elevation values ​​of all land features and landforms in the entire surveying target area, as well as the corresponding two-dimensional image texture, color and morphological feature information, forming accurate and unified surveying and mapping result data.

[0113] Integrity management is performed on the full-scene reconstruction data frames, regional weight factors are dynamically allocated, differential error correction coding is performed, and UAV mapping results data are output, including:

[0114] Data integrity assessment is performed on the reconstructed data frames of the entire scene, and the regional error weighting factor corresponding to each grid cell in the high-density and low-density verification areas is calculated.

[0115] To ensure the integrity and reliability of the reconstructed data frames across the entire scenario, an integrity assessment is required. This assessment aims to identify and quantify potential residual data errors or uncertainties within the dataset, providing error reference information for data encoding and data storage management.

[0116] Specifically, data integrity is assessed for each of the high-density and low-density verification areas in the full-scene reconstruction data frame. The data integrity assessment method is as follows: First, the point cloud data density and consistency within the data frame are quantitatively analyzed for each grid cell, including calculating the point cloud density within the grid cell, i.e., the number of valid data points per unit space. Second, the spatial continuity and consistency of the point cloud within the grid cell are evaluated, including calculating the average spatial distance and standard deviation between adjacent point clouds. For example, for a high-density verification area grid cell in an urban area, the point cloud density may reach 200 data points per square meter within a 10m × 10m space. By calculating the standard deviation of the distance between each data point and its adjacent data points, the spatial continuity of the data within the area is quantified. In a low-density verification area grid cell, such as a plain or open grassland, the point cloud density may be only 50 data points per square meter, indicating high data continuity but low density.

[0117] Using point cloud density and spatial continuity as indicators, the regional error weighting factor for each grid cell is calculated. The regional error weighting factor is defined as a quantitative indicator reflecting the data integrity and reliability level within the grid cell. The calculation method is as follows: the point cloud density and data continuity indicators within the grid cell are normalized to obtain standardized data within the range of 0 to 1, and a comprehensive indicator is calculated using a linear weighting method. Taking a high-density verification grid cell in an urban building area as an example, if the normalized data density indicator is 0.9 and the normalized data continuity indicator is 0.85, then the regional error weighting factor can be calculated as density indicator multiplied by 0.6 plus continuity indicator multiplied by 0.4, equaling 0.88. In a grassland grid cell in a low-density verification area, the density indicator is 0.4 and the continuity indicator is 0.95, therefore the error weighting factor is 0.4 × 0.6 + 0.95 × 0.4 = 0.62. The above method is applied to each grid cell in both the high-density and low-density verification areas to obtain a set of regional error weighting factors.

[0118] Based on the regional error weighting factor, the regional error weighting factor corresponding to each grid cell is assigned to the data point cloud within the corresponding grid cell.

[0119] For the data point cloud within each grid cell in the full-scene reconstruction data frame, a regional error weighting factor is assigned to all data points within the grid cell, ensuring that each data point has error weighting factor information. For example, in a high-density verification grid cell within an urban building area, the calculated regional error weighting factor is 0.88. Therefore, all data points within the grid cell (e.g., 2000 point cloud data points) are recorded with a regional error weighting factor of 0.88. In a grassland grid cell within a low-density verification area, the regional error weighting factor is 0.62. Therefore, each data point cloud within this area is recorded with an error weighting factor of 0.62.

[0120] The differential error correction coding method is used to perform differential coding on the data point cloud after assigning regional error weight factors to generate UAV mapping results data;

[0121] Differential error correction coding is an effective coding method that simultaneously achieves data compression and error correction, suitable for UAV mapping data frame storage management. First, differential coding is applied to the data point cloud after assigning regional error weight factors. Specifically, the data point clouds within each grid cell are sorted according to a preset spatial order (e.g., sorted in ascending order of spatial coordinates). The spatial coordinate difference and error weight factor difference between adjacent data points are calculated one by one, reducing the amount of stored data. For example, if the coordinates of the first data point in a grid cell are (X=10.000, Y=20.000, Z=5.000) with an error weight factor of 0.88, and the coordinates of the second data point are (X=10.005, Y=20.003, Z=5.002) with the same error weight factor of 0.88, then when recording the second data point, only the difference information (ΔX=+0.005, ΔY=+0.003, ΔZ=+0.002, Δweight factor=0) needs to be stored, significantly reducing data redundancy.

[0122] Secondly, differential error correction coding methods include error correction mechanisms, such as CRC checksums or Hamming codes. Specifically, an error correction checksum is calculated for each group of consecutive differentially coded data (e.g., every 256 data points per group) to ensure that errors during data storage and transmission can be detected and corrected in a timely manner. For example, a 32-bit CRC checksum is calculated for every 256 differentially coded data points and stored at the end of the data file. When UAV mapping data is extracted and used, a checksum value is calculated using the checksum to confirm data integrity or achieve automatic data error correction.

[0123] Finally, the encoded point cloud data is stored in a standard file format to form a UAV mapping result data file, such as a LAS or GeoTIFF file, recording differential coding coordinate values, error weighting factors, and error correction coding information, thus forming a UAV mapping data result file.

[0124] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0125] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0126] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0127] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0128] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0129] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0130] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0131] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0132] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0133] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A data management system based on unmanned aerial vehicle (UAV) mapping, characterized in that, include: Data acquisition module: Acquires the set of raw mapping data frames obtained by the UAV platform and outputs the processed mapping data set; Region division module: Based on the spatial characteristics of the surveying and mapping dataset, the surveying and mapping area is divided into high-density verification area and low-density verification area by means of terrain elevation variation coefficient and obstacle projection coverage; Feature matching module: Uses a local feature matching algorithm to calculate the matching degree of SIFT feature points in the overlapping area of ​​adjacent frames for the high-density verification region; Redundancy check module: When the matching degree is lower than the first threshold, the sliding window CRC check is started and the first correction data is output; Data compensation module: Performs data compensation on low-density verification areas, performs spatial interpolation compensation on areas occluded by obstacles, and reconstructs the point cloud topology structure by cubic spline interpolation on areas with abrupt terrain changes, and outputs second correction data; Data alignment module: Based on the first and second correction data, it generates a full-scene reconstruction data frame by aligning the boundary data of the high-density verification area and the low-density verification area; The differential error correction module performs integrity management on the full-scene reconstruction data frames, dynamically allocates regional weight factors, performs differential error correction coding, and outputs UAV mapping results data.

2. The data management system based on UAV mapping according to claim 1, characterized in that, The system collects the raw mapping data frames acquired by the UAV platform and outputs the processed mapping data set, specifically: Simultaneously collect 3D point cloud data and 2D image data within the target area; The collected 3D point cloud data and 2D image data are time-stamped and tagged according to the recorded flight trajectory information; Based on the timestamp synchronization marker, the collected 3D point cloud data and 2D image data are respectively subjected to spatial coordinate registration and data fusion processing to generate a standard format mapping data frame set containing complete spatial coordinate system information and image feature information; The standard format surveying data frame set is uniformly encoded according to the spatial data format specification, and the uniformly encoded surveying data set is output.

3. A data management system based on UAV mapping according to claim 2, characterized in that, Based on the spatial characteristics of the surveying dataset, the surveying area is divided into high-density verification zones and low-density verification zones using the terrain elevation variation coefficient and obstacle projection coverage. Specifically: The mapping data frame set is divided into spatial grids, and multiple grid units are generated according to a preset grid size. Calculate the terrain elevation variation coefficient for the point cloud data within each grid cell according to the height value distribution; The optical image data within each grid cell is mapped to the point cloud data through a projection model to calculate the obstacle projection coverage. The terrain elevation variation coefficient and obstacle projection coverage are compared using thresholds. Grid cells with both variation coefficient and coverage higher than the predetermined threshold are marked as high-density verification areas, while grid cells with both variation coefficient and coverage lower than the predetermined threshold are marked as low-density verification areas.

4. A data management system based on UAV mapping according to claim 3, characterized in that, For the high-density verification region, a local feature matching algorithm is used to calculate the matching degree of SIFT feature points in the overlapping areas of adjacent frames, specifically as follows: Two-dimensional image data of the overlapping area between adjacent frames is extracted from the mapping data frames within the grid cells marked as high-density verification areas; Detect and extract spatial scale-invariant feature points from two-dimensional image data in overlapping regions; The feature vectors of the extracted spatial scale-invariant feature transformation feature points are calculated, and the feature vectors between adjacent frames are matched. The matching degree is calculated based on the ratio between the number of successfully matched feature vectors and the total number of extracted feature vectors.

5. A data management system based on UAV mapping according to claim 4, characterized in that, When the matching degree is lower than the first threshold, the sliding window CRC check is initiated, and the first correction data is output, specifically: Set the first threshold as the minimum allowable value for feature vector matching degree; For mapping data frames within grid cells of high-density verification zones where the feature vector matching degree is lower than the first threshold, a sliding window cyclic redundancy check is initiated. The sliding window method is used to sequentially extract continuous data segments from the mapping data frame to be verified, and a cyclic redundancy check operation is performed on each continuous data segment. Output the first correction data of the high-density check area grid cells generated after cyclic redundancy check.

6. A data management system based on UAV mapping according to claim 5, characterized in that, Data compensation is performed on the low-density verification area. This is achieved by spatial interpolation compensation for obstacle-occluded areas and cubic spline interpolation reconstruction of the point cloud topology for areas with abrupt terrain changes. The second correction data is then output, specifically as follows: Identify obstacle-occluded areas and terrain abrupt change areas within the low-density verification area; When performing data compensation for point cloud data within the area obstructed by obstacles, a spatial interpolation algorithm based on neighboring point clouds is used for interpolation compensation. When performing data compensation on point cloud data in areas of abrupt terrain change, a cubic spline curve interpolation algorithm is used to reconstruct the point cloud topology. Output the second correction data generated by the low-density check area grid cells.

7. A data management system based on UAV mapping according to claim 6, characterized in that, Based on the first and second correction data, a full-scene reconstruction data frame is generated by aligning the boundary data of the high-density verification region and the low-density verification region, specifically as follows: Based on the spatial coordinates of the grid cells, the boundary regions of adjacent grid cells between the high-density and low-density verification regions are determined. Based on the overlapping point cloud data characteristics of adjacent grid cells within the boundary region, a spatial coordinate registration algorithm is used to align the boundary data of the first correction data of the grid cells in the high-density verification area and the second correction data of the grid cells in the low-density verification area. Based on the aligned boundary data, all correction data of the high-density verification area and the low-density verification area are merged to generate a full-scene reconstruction data frame covering the entire surveying area.

8. A data management system based on UAV mapping according to claim 7, characterized in that, Integrity management is performed on the full-scene reconstruction data frames, regional weight factors are dynamically allocated, differential error correction coding is performed, and UAV mapping results data are output, specifically: Data integrity assessment is performed on the reconstructed data frames of the entire scene, and the regional error weighting factor corresponding to each grid cell in the high-density and low-density verification areas is calculated. Based on the regional error weighting factor, the regional error weighting factor corresponding to each grid cell is assigned to the data point cloud within the corresponding grid cell. The differential error correction coding method is used to perform differential coding on the data point cloud after assigning regional error weight factors to generate UAV mapping results data.

Images