City surveying and mapping data processing method based on point cloud data

Abstract

The invention relates to the technical field of surveying and mapping data processing, in particular to a point cloud data-based urban surveying and mapping data processing method, which comprises the following steps of: determining an expected evaluation value of a project for each category; obtaining original point cloud data; determining a plurality of intercepted sub-regions; matching the category of each surveying and mapping feature based on the similarity, and determining a clustering characterization value for each category; determining whether city surveying and mapping are qualified or not based on the classification deviation value, and correcting the similarity tolerance parameter when it is determined that city surveying and mapping are abnormal; grouping the classified point clouds, and constructing a three-dimensional model; and carrying out denoising and smoothing processing on the preliminary model, and complementing a missing part. And the processing efficiency of the urban surveying and mapping data is improved.

Description

Technical Field

[0001] This invention relates to the field of surveying and mapping data processing technology, and in particular to a method for processing urban surveying and mapping data based on point cloud data. Background Technology

[0002] With the rapid development of laser scanning and oblique photogrammetry technologies, acquiring large-scale, high-precision 3D point cloud data of cities has become a reality. This point cloud data forms the foundation for urban 3D modeling, planning management, and smart city construction.

[0003] Traditional point cloud data processing workflows typically include steps such as preprocessing, feature extraction, land cover classification, and 3D reconstruction.

[0004] Chinese Patent Publication No. CN117237557A discloses a method for processing urban surveying and mapping data based on point cloud data, including the following steps: setting a first acquisition area and a second acquisition area; obtaining first point cloud data and second point cloud data from the first acquisition area and the second acquisition area; preprocessing the first point cloud data and the second point cloud data to obtain first usable data and second usable data; obtaining the overlapping area of ​​the first acquisition area and the second acquisition area; obtaining a set of usable data from the overlapping area; fusing the usable data set to obtain overlapping area data; obtaining urban surveying and mapping data based on the first usable data, the second usable data, and the overlapping area data; and modeling and visualizing all areas based on the urban surveying and mapping data. It is evident that the above technical solution has the following problems: it does not consider real-time evaluation of surveying and mapping data quality, and cannot adaptively adjust processing parameters based on evaluation results, thus affecting the processing efficiency of urban surveying and mapping data. Summary of the Invention

[0005] To address this, the present invention provides a method for processing urban surveying and mapping data based on point cloud data, which overcomes the problem in the prior art that does not consider real-time evaluation of surveying and mapping data quality and cannot adaptively adjust processing parameters based on evaluation results, thus affecting the processing efficiency of urban surveying and mapping data.

[0006] To achieve the above objectives, the present invention provides a method for urban surveying and mapping data processing based on point cloud data, comprising: Determine the expected assessment values ​​for each project category, including buildings, vegetation, and topography; The test area is scanned using a lidar to obtain raw point cloud data; The original point cloud data is subjected to noise reduction, filtering, sieving, and cloud registration to obtain the registered complete scene point cloud data; Feature extraction is performed on the registered point cloud data to determine several truncated sub-regions, each of which includes a single mapping feature; Based on similarity, the categories of each mapping feature are matched, and cluster representation values ​​for each category are determined; Determining the adequacy of a city's surveying based on classification bias values ​​includes: Once the city's surveying and mapping is deemed satisfactory, a model will be created. Alternatively, identify mapping anomalies for the city and adjust the similarity tolerance parameters accordingly; The classified point cloud is grouped and a 3D model is constructed. The initial model is denoised and smoothed, and missing parts are filled in.

[0007] Furthermore, determining the expected rating value for a single category of projects includes: Obtain theoretical planning data for the test area to obtain a set of three-dimensional coordinates for each theoretical feature of a single category; For a single feature point in a three-dimensional coordinate set, calculate its Euclidean distance to each feature point in the three-dimensional coordinate set, and determine the minimum value among all Euclidean distances as the nearest neighbor distance for the single feature point. The average nearest neighbor distance is obtained by calculating the average nearest neighbor distance for each feature point in a single category. The variance of the nearest neighbor distance for each feature point in a single category is calculated to obtain the nearest neighbor distance difference. The distribution coefficient is obtained by calculating the ratio of the nearest neighbor distance difference to the average nearest neighbor distance; The product of the distribution coefficient and the feature point number weighting coefficient is used to obtain the uniformity index. The difference between 1 and the uniformity index is calculated to obtain the expected evaluation value of the project.

[0008] Further, cluster representation values ​​for individual categories are determined, including: For a single category that has been classified, extract the coordinate set of all its mapping feature points and count the total number of cloud data points for that single category; The average mapping proximity distance is obtained by calculating the average nearest neighbor distance of each mapping feature in a single category. The variance of the nearest neighbor distance for each mapping feature in a single category is calculated to obtain the nearest neighbor distance difference. The mapping distribution coefficient for a single category is obtained by calculating the ratio of the nearest mapping neighbor distance difference to the average mapping neighbor distance; The sum of the total number of cloud data points and the preset benchmark number of points is used to obtain the point evaluation coefficient. The ratio of the total number of cloud data points to the point evaluation coefficient is used to obtain the total point correction factor. The test density is obtained by calculating the ratio of the total number of cloud data points of each category to the area of ​​the test region. The uniformity index is obtained by calculating the ratio of the difference between the test density and the theoretical planned density to the theoretical planned density. Calculate the difference between 1 and the uniformity index to obtain the density consistency factor; The mapping reference value is obtained by multiplying the total number of points correction factor, density consistency factor and mapping distribution coefficient. Calculate the difference between 1 and the mapping reference value to obtain the cluster characterization value.

[0009] Furthermore, the classification deviation value is used to determine whether the city's mapping is qualified, including: For a single category, calculate the absolute value of the difference between the expected evaluation value of the project and the corresponding cluster characterization value, and solve for the ratio of the absolute value to the cluster characterization value to obtain the classification bias value for a single category; When a classification deviation value is greater than the preset classification deviation value, mapping anomalies for the city are identified, and the total number of points in the point cloud data for each category will be used to determine the similarity tolerance parameter of the classification standard and adjusted to the corresponding value.

[0010] Furthermore, the process of determining similarity includes: For the mapping features within a single extracted sub-region, determine the three-dimensional coordinates of each point cloud data within the sub-region, in order to determine the number of regional points and the distribution dispersion of each point cloud data within the sub-region; The distribution dispersion is the standard deviation of the distance between each point cloud data point in a sub-region and the centroid of a mapping feature within a single extracted sub-region; The similarity is obtained by weighted summation of the number of points in the region and the dispersion of their distribution.

[0011] Furthermore, the total number of points in the point cloud data for each category will be used to determine the similarity tolerance parameter for the classification criteria, which will be adjusted to the corresponding value. The decrease in the similarity tolerance parameter is positively correlated with the total number of points in each category of point cloud data.

[0012] Furthermore, after adjusting the similarity, the parameters for determining the truncated sub-region are determined based on the optimized quantization value, including: The optimized quantification value is obtained by calculating the ratio of the difference between the historical average classification deviation value and the newly determined average classification deviation value to the historical average classification deviation value. When the optimized quantization value is less than or equal to the preset optimized quantization value, the initial voxel size is corrected based on the optimized quantization value, and after the initial voxel size is corrected, the data correction is determined to be qualified based on the newly determined classification deviation value. When the optimized quantization value is greater than the preset optimized quantization value, the data correction is deemed acceptable based on the newly determined classification deviation value.

[0013] Furthermore, the initial voxel size is corrected based on the optimized quantization value, wherein, The reduction in initial voxel size is negatively correlated with the optimized quantization value.

[0014] Furthermore, the adequacy of data correction is determined based on the redefined classification bias value, including: If a newly determined classification deviation value is greater than the preset classification deviation value, the data correction is identified as abnormal, and the point cloud density in the scanning parameters is adjusted to the corresponding value based on the newly determined classification deviation value.

[0015] Furthermore, based on the redefined classification bias value, the point cloud density in the scanning parameters is adjusted to the corresponding value, wherein, The increase in point cloud density is positively correlated with the average value of the redefined classification deviation values. Compared with the prior art, the beneficial effect of this invention is that it determines the expected evaluation value of the project by acquiring theoretical planning data of the test area. This reflects the distribution of feature points in the theoretical planning and serves as a reference standard for subsequent actual mapping results. The three-dimensional coordinate set of theoretical features represents the positional information of each feature point in the theoretical planning of the test area. The minimum Euclidean distance from a single feature point to all other feature points reflects the density of feature points around that feature point. The average nearest neighbor distance of each feature point reflects the overall distribution density of the feature points. The variance of the nearest neighbor distance of each feature point reflects the dispersion of the feature point distribution. The ratio of the nearest neighbor distance difference to the average nearest neighbor distance comprehensively reflects the uniformity of the feature point distribution. The product of the distribution coefficient and the feature point quantity weighting coefficient considers the influence of the number of feature points on the distribution uniformity. The difference between 1 and the uniformity index serves as a quantitative indicator of the uniformity of feature point distribution in the theoretical planning. The quantity weighting coefficient is calculated based on the total number of feature points and is used to adjust the influence of the distribution coefficient on the evenness index. The more feature points there are, the closer the weighting coefficient is to 1. It provides a quantitative theoretical reference standard for evaluating the accuracy of actual surveying results and improves the processing efficiency of urban surveying data.

[0016] Further, cluster representation values ​​are determined by extracting the coordinate sets of each classified mapping feature from the point cloud data. These values ​​reflect the distribution and density of feature points in the actual mapping. The total number of points in each category of cloud data represents the total number of points in the point cloud data, reflecting the amount of information in the actual mapping. The area of ​​the test region is the actual area of ​​the test region, used to calculate the density. The average nearest neighbor distance for each mapping feature reflects the density of feature point distribution in the actual mapping. The variance of the nearest neighbor distance for each mapping feature reflects the dispersion of feature point distribution in the actual mapping. The ratio of the nearest neighbor distance difference to the average nearest neighbor distance comprehensively reflects the uniformity of feature point distribution in the actual mapping. The sum of the total number of cloud data points and the preset baseline number of points is used to adjust the influence of the total number of points on the total number of point correction factors. The ratio of the total number of cloud data points to the point evaluation coefficient considers the influence of the total number of cloud data points on the cluster representation value. The ratio of the total number of cloud data points to the area of ​​the test region reflects the density of the actual mapping. The ratio of the difference between the tested density and the theoretical planned density to the theoretical planned density reflects the difference between the actual surveyed density and the theoretical planned density. The difference between 1 and the uniformity index reflects the consistency between the actual surveyed density and the theoretical planned density. The product of the total point number correction factor, the density consistency factor, and the surveyed distribution coefficient comprehensively considers factors such as the uniformity of feature point distribution, the total number of points, and density consistency. The difference between 1 and the surveyed reference value serves as a quantitative indicator of the distribution of feature points in the actual survey. Quantifying the distribution and density of feature points in the actual survey and comparing them with the expected evaluation value of the project helps determine whether the actual survey results conform to the theoretical plan. This allows for the timely detection of deviations between the survey results and the theoretical plan, improving the accuracy of the survey. It also improves the processing efficiency of urban surveying data.

[0017] Furthermore, the classification deviation value is used to determine whether the surveying is qualified. This value reflects the degree of deviation between the actual surveying results and the theoretical plan. Based on the degree of deviation between the actual surveying results and the theoretical plan, the qualification of the surveying is judged to determine whether modeling or adjustments are necessary. This allows for the timely detection of anomalies in the surveying process, ensuring the accuracy of the results and preventing unqualified surveying data from entering the subsequent modeling stage. This improves the processing efficiency of urban surveying data.

[0018] Furthermore, when the classification deviation value exceeds the preset value, it indicates a significant discrepancy between the actual surveying results and the theoretical plan. In this case, the similarity tolerance parameter used to determine the classification standard is adjusted. By adjusting the similarity tolerance parameter, the classification standard of the surveying features is changed, making the classification results more consistent with reality. The similarity tolerance parameter determines the allowable error range for whether a surveying feature belongs to a certain preset category. The larger the parameter, the more lenient the classification standard; the smaller the parameter, the more stringent the classification standard. When surveying results are abnormal, adjusting the similarity tolerance parameter corrects the classification deviation, improves the accuracy of classification, and makes the surveying results more consistent with reality. Adaptively adjusting the classification standard improves classification accuracy, makes the surveying results more consistent with reality, and ensures the reliability of subsequent modeling. This improves the processing efficiency of urban surveying data.

[0019] Furthermore, based on the optimized quantification value, it is determined whether to correct the parameters for the selected sub-region. The optimized quantification value measures the degree of optimization of classification bias. When the optimized quantification value is small, it indicates that the classification bias has not been effectively improved after adjusting the similarity tolerance parameter. In this case, the division of the selected sub-region is further optimized to improve classification accuracy. When the optimized quantification value is large, it indicates that the current adjustment and improvement effect is satisfactory. Further judgment is made on whether the data correction is satisfactory to determine whether other measures should be taken. Through the evaluation of the optimized quantification value, subsequent processing steps are determined in a targeted manner to further optimize the data processing process. If the optimization effect is abnormal, the parameters for the selected sub-region are fine-tuned to improve classification accuracy; if the effect is satisfactory, a judgment is directly made on whether to take other adjustment measures, which improves the processing efficiency of urban surveying data.

[0020] Furthermore, the initial voxel size is adjusted based on the optimized quantization value. A smaller optimized quantization value indicates a lower optimization effect on classification bias. At this point, the sub-region division is further refined, making the reduction in initial voxel size negatively correlated with the optimized quantization value. By adjusting the initial voxel size, the point cloud data is divided more precisely, improving classification accuracy. The initial voxel size is the initial size when dividing the voxel grid, used to control the granularity of sub-region division. Dynamically adjusting the initial voxel size based on the optimized quantization value further refines the sub-region division, improving classification accuracy and providing more accurate feature information for subsequent modeling. This improves the completeness and accuracy of data processing. Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating the steps of the urban surveying data processing method based on point cloud data according to an embodiment of the present invention. Figure 2 This is a logic diagram illustrating the determination of whether a city's surveying is qualified based on classification deviation values, according to an embodiment of the present invention. Figure 3This is a logic diagram illustrating whether to correct the parameters for determining the truncated sub-region based on optimized quantization values ​​in an embodiment of the present invention. Figure 4 This is a logic diagram for determining whether data correction is qualified based on the redefined classification deviation value in an embodiment of the present invention. Detailed Implementation

[0022] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0023] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0024] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0025] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0026] Please see Figure 1 The diagram illustrates the steps of an urban surveying data processing method based on point cloud data according to an embodiment of the present invention. The urban surveying data processing method based on point cloud data of the present invention includes: S1, determine the expected assessment values ​​for each category of the project, including buildings, vegetation, and terrain; S2, use LiDAR to scan the test area to obtain raw point cloud data; S3 performs noise reduction, filtering, sieving, and cloud registration processing on the original point cloud data to obtain the registered complete scene point cloud data; S4. Feature extraction is performed from the registered point cloud data to determine several truncated sub-regions, each of which includes a single mapping feature. S5, based on similarity, match the categories of each mapping feature and determine the clustering representation value for each category; Determining the adequacy of a city's surveying based on classification bias values ​​includes: Once the city's surveying and mapping is deemed satisfactory, a model will be created. Alternatively, identify mapping anomalies for the city and adjust the similarity tolerance parameters accordingly; S6: Group the classified point clouds and construct a 3D model; S7 performs denoising and smoothing on the initial model and fills in any missing parts.

[0027] Specifically, urban mapping requires acquiring spatial information about the urban environment. Point cloud data is a collection of numerous 3D coordinate points obtained through laser scanning, used to describe the surface shape and location of various objects in the city. Scanning records information about the city's topography, buildings, and vegetation. Raw point cloud data contains noisy points and redundant data. Noise reduction is used to remove noisy points, filtering and sifting remove useless data, and registration aligns point cloud data from different locations to the same coordinate system. Edge contour features are extracted from the registered point cloud data, and then the point cloud data is classified into different categories based on the similarity of these features. Similarity measures the degree of similarity between different mapping features and is used to determine the category. The number of points in a region is the number of point cloud data points within a single cropped sub-region, reflecting the density of objects within that region. Distribution dispersion is the standard deviation of the distance from each point cloud data point within a sub-region to the centroid of the mapping features within that region, reflecting the dispersion of the point cloud data within that region. Classifying the point cloud data allows for better identification and differentiation of different objects and terrains in the city, providing more targeted data for subsequent modeling and analysis.

[0028] Specifically, determining the expected rating for a single category of projects includes: Obtain theoretical planning data for the test area to obtain a set of three-dimensional coordinates for each theoretical feature of a single category; For a single feature point in a three-dimensional coordinate set, calculate its Euclidean distance to each feature point in the three-dimensional coordinate set, and determine the minimum value among all Euclidean distances as the nearest neighbor distance for the single feature point. The average nearest neighbor distance is obtained by calculating the average nearest neighbor distance for each feature point in a single category. The variance of the nearest neighbor distance for each feature point in a single category is calculated to obtain the nearest neighbor distance difference. The distribution coefficient is obtained by calculating the ratio of the nearest neighbor distance difference to the average nearest neighbor distance; The product of the distribution coefficient and the feature point number weighting coefficient is used to obtain the uniformity index. The difference between 1 and the uniformity index is calculated to obtain the expected evaluation value of the project.

[0029] Specifically, the quantity weight coefficient Q = [1 - 1 / (1 + ln(n))], where n is the total number of theoretical features within a single category.

[0030] Specifically, the expected evaluation value of the project is determined by acquiring theoretical planning data of the test area. This reflects the distribution of feature points in the theoretical planning and serves as a reference standard for subsequent actual mapping results. The three-dimensional coordinate set of the theoretical features represents the location information of each feature point in the theoretical planning of the test area. The minimum Euclidean distance from a single feature point to all other feature points reflects the density of feature points around that feature point. The average nearest neighbor distance of each feature point reflects the overall density of the feature point distribution. The variance of the nearest neighbor distance of each feature point reflects the dispersion of the feature point distribution. The ratio of the nearest neighbor distance difference to the average nearest neighbor distance comprehensively reflects the uniformity of the feature point distribution. The product of the distribution coefficient and the feature point quantity weighting coefficient considers the influence of the number of feature points on the uniformity of distribution. The difference between 1 and the uniformity index serves as a quantitative indicator of the uniformity of feature point distribution in the theoretical planning. The quantity weighting coefficient is calculated based on the total number of feature points and is used to adjust the influence of the distribution coefficient on the uniformity index; the more feature points there are, the closer the weighting coefficient is to 1. This provides a quantitative theoretical reference standard for evaluating the accuracy of actual mapping results. It improves the efficiency of processing urban surveying and mapping data.

[0031] Specifically, determining the cluster representation value for a single category includes: For a single category that has been classified, extract the coordinate set of all its mapping feature points and count the total number of cloud data points for that single category; The average mapping proximity distance is obtained by calculating the average nearest neighbor distance of each mapping feature in a single category. The variance of the nearest neighbor distance for each mapping feature in a single category is calculated to obtain the nearest neighbor distance difference. The mapping distribution coefficient for a single category is obtained by calculating the ratio of the nearest mapping neighbor distance difference to the average mapping neighbor distance; The sum of the total number of cloud data points and the preset benchmark number of points is used to obtain the point evaluation coefficient. The ratio of the total number of cloud data points to the point evaluation coefficient is used to obtain the total point correction factor. The test density is obtained by calculating the ratio of the total number of cloud data points of each category to the area of ​​the test region. The uniformity index is obtained by calculating the ratio of the difference between the test density and the theoretical planned density to the theoretical planned density. Calculate the difference between 1 and the uniformity index to obtain the density consistency factor; The mapping reference value is obtained by multiplying the total number of points correction factor, density consistency factor and mapping distribution coefficient. Calculate the difference between 1 and the mapping reference value to obtain the cluster characterization value.

[0032] Specifically, cluster representation values ​​are determined by extracting the coordinate sets of each classified mapping feature from the point cloud data. These values ​​reflect the distribution and density of feature points in the actual mapping. The total number of points in each category of cloud data represents the total number of points in the point cloud data, reflecting the amount of information in the actual mapping. The area of ​​the test region is the actual area of ​​the test region, used to calculate the density. The average nearest neighbor distance for each mapping feature reflects the density of feature point distribution in the actual mapping. The variance of the nearest neighbor distance for each mapping feature reflects the dispersion of feature point distribution in the actual mapping. The ratio of the nearest neighbor distance difference to the average nearest neighbor distance comprehensively reflects the uniformity of feature point distribution in the actual mapping. The sum of the total number of cloud data points and the preset baseline number of points is used to adjust the influence of the total number of points on the total number of point correction factors. The ratio of the total number of cloud data points to the point evaluation coefficient considers the influence of the total number of cloud data points on the cluster representation value. The ratio of the total number of cloud data points to the area of ​​the test region reflects the density of the actual mapping. The ratio of the difference between the tested density and the theoretical planned density to the theoretical planned density reflects the difference between the actual surveyed density and the theoretical planned density. The difference between 1 and the uniformity index reflects the consistency between the actual surveyed density and the theoretical planned density. The product of the total point number correction factor, the density consistency factor, and the surveyed distribution coefficient comprehensively considers factors such as the uniformity of feature point distribution, the total number of points, and density consistency. The difference between 1 and the surveyed reference value serves as a quantitative indicator of the distribution of feature points in the actual survey. Quantifying the distribution and density of feature points in the actual survey and comparing them with the expected evaluation value of the project helps determine whether the actual survey results conform to the theoretical plan. This allows for the timely detection of deviations between the survey results and the theoretical plan, improving the accuracy of the survey. It also improves the processing efficiency of urban surveying data.

[0033] Please see Figure 2 The diagram shown illustrates the logic for determining the qualification of a city's surveying based on classification deviation values, according to an embodiment of the present invention. The process of determining the qualification of a city's surveying based on classification deviation values ​​includes: For a single category, calculate the absolute value of the difference between the expected evaluation value of the project and the corresponding cluster characterization value, and solve for the ratio of the absolute value to the cluster characterization value to obtain the classification bias value for a single category; When the classification deviation values ​​for all categories are less than or equal to the preset classification deviation values, the city's surveying and mapping is deemed qualified, and modeling is then carried out. When a classification deviation value is greater than the preset classification deviation value, mapping anomalies for the city are identified, and the total number of points in the point cloud data for each category will be used to determine the similarity tolerance parameter of the classification standard and adjusted to the corresponding value.

[0034] Specifically, the preset classification deviation value is selected within the range [0.04, 0.06]. Those skilled in the art can select and determine it themselves, and can adjust it according to the actual surveying accuracy requirements. In this embodiment, the preset classification deviation value is preferably 0.5.

[0035] Specifically, the classification deviation value is used to determine whether a survey is qualified. This value reflects the degree of deviation between the actual survey results and the theoretical plan. Based on this deviation, the qualification of the survey is determined, in order to decide whether to proceed with modeling or make adjustments. This allows for the timely detection of anomalies in the surveying process, ensuring the accuracy of the results and preventing unqualified survey data from entering the subsequent modeling stage. This improves the efficiency of processing urban surveying data.

[0036] Specifically, the process of determining similarity includes: For the mapping features within a single extracted sub-region, determine the three-dimensional coordinates of each point cloud data within the sub-region, in order to determine the number of regional points and the distribution dispersion of each point cloud data within the sub-region; The distribution dispersion is the standard deviation of the distance between each point cloud data point in a sub-region and the centroid of a mapping feature within a single extracted sub-region; The similarity is obtained by weighted summation of the number of points in the region and the dispersion of their distribution.

[0037] Specifically, the similarity-based classification process includes: For each preset category (building, vegetation, terrain), there are corresponding preset similarity and similarity tolerance parameters stored. Mapping features whose similarity is within the range of preset similarity ± similarity tolerance parameter are determined as the corresponding preset category.

[0038] Specifically, the total number of points in the point cloud data for each category will be used to determine the similarity tolerance parameter for the classification criteria, which will be adjusted to the corresponding value. The decrease in the similarity tolerance parameter is positively correlated with the total number of points in each category of point cloud data.

[0039] In this embodiment, optionally, The total number of points T based on the point cloud data for each category will be adjusted to the corresponding value for the similarity tolerance parameter S used to determine the classification criteria. If T≤1×10 5 Then S is adjusted to S × 0.95; If 1×10 5 <T≤1×10 6 If so, then S is adjusted to S × 0.9; If T>1×10 6 Then S is adjusted to S × 0.85.

[0040] Specifically, when the classification deviation value exceeds the preset value, it indicates a significant discrepancy between the actual surveying results and the theoretical plan. In this case, the similarity tolerance parameter used to determine the classification standard is adjusted. By adjusting the similarity tolerance parameter, the classification standard of the surveying features is changed, making the classification results more consistent with reality. The similarity tolerance parameter determines the allowable error range for whether a surveying feature belongs to a certain preset category. The larger the parameter, the more lenient the classification standard; the smaller the parameter, the stricter the classification standard. When surveying results are abnormal, adjusting the similarity tolerance parameter corrects the classification deviation, improves the accuracy of classification, and makes the surveying results more consistent with reality. Adaptively adjusting the classification standard improves classification accuracy, makes the surveying results more consistent with reality, and ensures the reliability of subsequent modeling. This improves the processing efficiency of urban surveying data.

[0041] Please see Figure 4 As shown, this is a logic diagram illustrating the process of determining whether data correction is acceptable based on a redefined classification deviation value, according to an embodiment of the present invention. After adjusting the similarity, the present invention determines whether to correct the parameters for selecting the sub-region based on the optimized quantization value, including: The optimized quantification value is obtained by calculating the ratio of the difference between the historical average classification deviation value and the newly determined average classification deviation value to the historical average classification deviation value. When the optimized quantization value is less than or equal to the preset optimized quantization value, the initial voxel size is corrected based on the optimized quantization value, and after the initial voxel size is corrected, the data correction is determined to be qualified based on the newly determined classification deviation value. When the optimized quantization value is greater than the preset optimized quantization value, the data correction is deemed acceptable based on the newly determined classification deviation value.

[0042] The preset optimization quantization value is selected within the range [0.02, 0.04]. Those skilled in the art can select and determine it according to the actual use scenario. In this embodiment, the preset optimization quantization value is preferably 0.03.

[0043] Specifically, the redefined average classification deviation is the average of the redefined classification deviation values ​​for each category; Specifically, the historical average classification deviation is the average of the classification deviations of each category recorded before adjusting the similarity tolerance parameter.

[0044] Specifically, the optimization quantification value determines whether to correct the parameters of the selected sub-region. The optimization quantification value measures the degree of improvement in classification bias. A small optimization quantification value indicates that the classification bias has not been effectively improved after adjusting the similarity tolerance parameter. In this case, further optimization of the sub-region division is needed to improve classification accuracy. Conversely, a large optimization quantification value indicates that the current adjustment and improvement effect is satisfactory. Further assessment of the data correction is needed to determine whether other measures should be taken. Through the evaluation of the optimization quantification value, subsequent processing steps are determined in a targeted manner, further optimizing the data processing process. If the optimization effect is abnormal, the parameters of the selected sub-region are fine-tuned to improve classification accuracy; if the effect is satisfactory, a direct determination is made on whether to take other adjustment measures, thus improving the processing efficiency of urban surveying data.

[0045] Specifically, the process of extracting a sub-region includes: S311 uses adaptive voxel mesh generation, dynamically adjusting voxel sizes based on point cloud density to initially separate distinct feature regions and extract sub-regions; initial voxel size... =2 meters; S312, automatically select the most representative point from each voxel as a seed, based on the consistency of the point normal direction and the continuity of curvature, and stop when the included angle of the normal exceeds the threshold θ_max = 30°, the curvature change exceeds the threshold κ_max = 0.1, or the maximum growth distance d_max = 20 meters is reached, so as to ensure that each sub-region contains only a single feature. S313, Matching preset categories based on the similarity of a single feature; After completing the matching of the preset categories, the α-shape algorithm is used to accurately extract the boundaries in order to determine the mapping features within a single truncated sub-region; If a single feature cannot match the preset category, then S311-S313 is repeated to expand the adaptive voxel grid until a single feature within a single truncated sub-region completes the matching of the preset category.

[0046] Specifically, the voxel size is dynamically adjusted based on the point cloud density. Regions with high point cloud density use smaller voxels, and regions with low point cloud density use larger voxels, to initially separate distinct feature regions, facilitating subsequent feature extraction and classification. The initial voxel size is the default size used when initially dividing the voxels, serving as the basis for subsequent adjustments. Since different regions have significantly different point cloud densities, using a fixed voxel size would result in some regions being divided too finely or too coarsely, failing to accurately separate feature regions. Adaptive voxel grid division is used to reasonably divide the voxels according to the actual density, enabling more accurate separation of different feature regions and improving the efficiency and accuracy of subsequent feature extraction and classification.

[0047] Specifically, seed points are determined and region growing is performed. The most representative point is selected from each voxel as the seed. Region growing is based on the consistency of the point's normal direction and the continuity of curvature. Growth stops when the normal angle, curvature change, or growth distance exceeds a threshold, ensuring that each sub-region contains only a single feature. The normal angle threshold is used to determine the degree of difference in normal direction between points; exceeding this threshold indicates a significant difference in direction, belonging to different features. The curvature change threshold measures the change in the curvature of the point cloud surface; exceeding this threshold indicates a significant change in curvature within the region, belonging to different features. The maximum growth distance limits the range of region growth, preventing overgrowth that results in multiple features. Region growing groups points with similar features together to form sub-regions, and these thresholds control the growth process, ensuring that each sub-region contains only a single feature, facilitating subsequent category matching. Accurately dividing the point cloud data into sub-regions of single features provides clear feature units for subsequent category matching. Based on similarity matching and boundary extraction, this method matches the features of each sub-region with a preset category to determine its category. Then, the α-shape algorithm is used to accurately extract the boundaries, identifying the mapping features within each individual sub-region. Matching to preset categories allows for the classification of different features, while boundary extraction accurately defines the range of each feature, providing accurate feature information for subsequent modeling and analysis. This achieves the classification of point cloud data and accurate extraction of feature boundaries, providing accurate feature information for subsequent modeling and analysis. When a single feature cannot match a preset category, it indicates that the current sub-region division is unreasonable, requiring the adaptive voxel mesh to be expanded, and the sub-region division and category matching to be redone until a match with the preset category can be achieved. This ensures that each feature can accurately match a preset category, avoiding unclassified features and improving the completeness and accuracy of classification. It also improves the completeness and accuracy of data processing.

[0048] Specifically, the initial voxel size is corrected based on the optimized quantization value, where... The reduction in initial voxel size is negatively correlated with the optimized quantization value.

[0049] In this embodiment, optionally, The optimized quantized value is compared with the first optimized value and the second optimized value; When the optimized quantization value is less than or equal to the first optimized value, the initial voxel size V0 is adjusted to V0×0.76; When the optimized quantization value is less than or equal to the second optimized value and greater than the first optimized value, the initial voxel size V0 is adjusted to V0×0.81; When the optimized quantization value is greater than the second optimized value, the initial voxel size V0 is adjusted to V0×0.94; The first optimized value is 0.006, and the second optimized value is 0.013.

[0050] Specifically, the initial voxel size is adjusted based on the optimized quantization value. A smaller optimized quantization value indicates a lower optimization effect on classification bias. In this case, the sub-region division is further refined, making the reduction in initial voxel size negatively correlated with the optimized quantization value. By adjusting the initial voxel size, the point cloud data is divided more precisely, improving classification accuracy. The initial voxel size is the initial size when dividing the voxel grid, used to control the granularity of sub-region division. The initial voxel size is dynamically adjusted based on the optimized quantization value to further refine the sub-region division, improve classification accuracy, and provide more accurate feature information for subsequent modeling. This improves the completeness and accuracy of data processing.

[0051] Please see Figure 4 The diagram shown illustrates the logic for determining whether data correction is acceptable based on a redefined classification deviation value, according to an embodiment of the present invention. The process of determining whether data correction is acceptable based on a redefined classification deviation value includes: When all the newly determined classification deviation values ​​are less than or equal to the preset classification deviation values, the data correction is deemed qualified, and modeling is then performed. If a newly determined classification deviation value is greater than the preset classification deviation value, the data correction is identified as abnormal, and the point cloud density in the scanning parameters is adjusted to the corresponding value based on the newly determined classification deviation value.

[0052] Specifically, based on the newly determined classification bias value, the point cloud density in the scanning parameters is adjusted to the corresponding value, where, The increase in point cloud density is positively correlated with the average value of the redefined classification bias values.

[0053] In this embodiment, optionally, The average of the newly determined classification deviation values ​​is determined as the changed average difference value; When the average difference value is less than or equal to the first preset difference value, the moving speed of the mobile scanning platform used for scanning is adjusted to 0.92 times the initial moving speed; When the average difference value is less than or equal to the second preset change value and greater than the first preset change value, the moving speed of the mobile scanning platform used for scanning will be adjusted to 0.88 times the initial moving speed. When the average difference value exceeds the second preset value, the moving speed of the mobile scanning platform used for scanning is adjusted to 0.81 times the initial moving speed.

[0054] The first preset change value is 0.08, and the second preset change value is 0.12.

[0055] For mobile scanning platforms such as vehicle-mounted systems, reducing the driving speed allows the laser scanner to emit more pulses per unit distance, thereby increasing cloud density.

[0056] Specifically, the recalculated classification bias value is the result of a series of adjustments and reflects the accuracy of the current data classification. Based on this recalculated bias value, the adequacy of data correction is determined. If the correction is still unsatisfactory, the scanned root data is adjusted for further refinement. This improves the accuracy and reliability of modeling. Simultaneously, it allows for the timely identification and adjustment of problems during the data correction process, thus enhancing the completeness and accuracy of data processing.

[0057] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0058] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for processing urban surveying and mapping data based on point cloud data, characterized in that, include: Determine the expected assessment values ​​for each project category, including buildings, vegetation, and topography; The test area is scanned using a lidar to obtain raw point cloud data; The original point cloud data is subjected to noise reduction, filtering, sieving, and cloud registration to obtain the registered complete scene point cloud data; Feature extraction is performed on the registered point cloud data to determine several truncated sub-regions, each of which includes a single mapping feature; Based on similarity, the categories of each mapping feature are matched, and cluster representation values ​​for each category are determined; Determining the adequacy of a city's surveying based on classification bias values ​​includes: Once the city's surveying and mapping is deemed satisfactory, a model will be created. Alternatively, identify mapping anomalies for the city and adjust the similarity tolerance parameters accordingly; The classified point cloud is grouped and a 3D model is constructed. The initial model is denoised and smoothed, and missing parts are filled in.

2. The urban surveying and mapping data processing method based on point cloud data according to claim 1, characterized in that, Determining the expected rating for a single category of projects includes: Obtain theoretical planning data for the test area to obtain a set of three-dimensional coordinates for each theoretical feature of a single category; For a single feature point in a three-dimensional coordinate set, calculate its Euclidean distance to each feature point in the three-dimensional coordinate set, and determine the minimum value among all Euclidean distances as the nearest neighbor distance for the single feature point. The average nearest neighbor distance is obtained by calculating the average nearest neighbor distance for each feature point in a single category. The variance of the nearest neighbor distance for each feature point in a single category is calculated to obtain the nearest neighbor distance difference. The distribution coefficient is obtained by calculating the ratio of the nearest neighbor distance difference to the average nearest neighbor distance; The product of the distribution coefficient and the feature point number weighting coefficient is used to obtain the uniformity index. The difference between 1 and the uniformity index is calculated to obtain the expected evaluation value of the project.

3. The urban surveying data processing method based on point cloud data according to claim 2, characterized in that, Determining cluster representation values ​​for individual categories includes: For a single category that has been classified, extract the coordinate set of all its mapping feature points and count the total number of cloud data points for that single category; The average mapping proximity distance is obtained by calculating the average nearest neighbor distance of each mapping feature in a single category. The variance of the nearest neighbor distance for each mapping feature in a single category is calculated to obtain the nearest neighbor distance difference. The mapping distribution coefficient for a single category is obtained by calculating the ratio of the nearest mapping neighbor distance difference to the average mapping neighbor distance; The sum of the total number of cloud data points and the preset benchmark number of points is used to obtain the point evaluation coefficient. The ratio of the total number of cloud data points to the point evaluation coefficient is used to obtain the total point correction factor. The test density is obtained by calculating the ratio of the total number of cloud data points of each category to the area of ​​the test region. The uniformity index is obtained by calculating the ratio of the difference between the test density and the theoretical planned density to the theoretical planned density. Calculate the difference between 1 and the uniformity index to obtain the density consistency factor; The mapping reference value is obtained by multiplying the total number of points correction factor, density consistency factor and mapping distribution coefficient. Calculate the difference between 1 and the mapping reference value to obtain the cluster characterization value.

4. The urban surveying data processing method based on point cloud data according to claim 3, characterized in that, Determining the adequacy of a city's surveying based on classification bias values ​​includes: For a single category, calculate the absolute value of the difference between the expected evaluation value of the project and the corresponding cluster characterization value, and solve for the ratio of the absolute value to the cluster characterization value to obtain the classification bias value for a single category; When a classification deviation value is greater than the preset classification deviation value, mapping anomalies for the city are identified, and the total number of points in the point cloud data for each category will be used to determine the similarity tolerance parameter of the classification standard and adjusted to the corresponding value.

5. The urban surveying and mapping data processing method based on point cloud data according to claim 4, characterized in that, The process of determining similarity includes: For the mapping features within a single extracted sub-region, determine the three-dimensional coordinates of each point cloud data within the sub-region, in order to determine the number of regional points and the distribution dispersion of each point cloud data within the sub-region; The distribution dispersion is the standard deviation of the distance between each point cloud data point in a sub-region and the centroid of a mapping feature within a single extracted sub-region; The similarity is obtained by weighted summation of the number of points in the region and the dispersion of their distribution.

6. The urban surveying and mapping data processing method based on point cloud data according to claim 5, characterized in that, The total number of points in the point cloud data for each category will be used to determine the similarity tolerance parameter for the classification criteria, which will be adjusted to the corresponding value. The decrease in the similarity tolerance parameter is positively correlated with the total number of points in each category of point cloud data.

7. The urban surveying and mapping data processing method based on point cloud data according to claim 6, characterized in that, After adjusting the similarity, the parameters for determining the truncated sub-region are adjusted based on the optimized quantization value, including: The optimized quantification value is obtained by calculating the ratio of the difference between the historical average classification deviation value and the newly determined average classification deviation value to the historical average classification deviation value. When the optimized quantization value is less than or equal to the preset optimized quantization value, the initial voxel size is corrected based on the optimized quantization value, and after the initial voxel size is corrected, the data correction is determined to be qualified based on the newly determined classification deviation value. When the optimized quantization value is greater than the preset optimized quantization value, the data correction is deemed acceptable based on the newly determined classification deviation value.

8. The urban surveying and mapping data processing method based on point cloud data according to claim 7, characterized in that, The initial voxel size is corrected based on the optimized quantization value, where... The reduction in initial voxel size is negatively correlated with the optimized quantization value.

9. The urban surveying and mapping data processing method based on point cloud data according to claim 8, characterized in that, Determining whether data correction is adequate based on the redefined classification bias value includes: If a newly determined classification deviation value is greater than the preset classification deviation value, the data correction is identified as abnormal, and the point cloud density in the scanning parameters is adjusted to the corresponding value based on the newly determined classification deviation value.

10. The urban mapping data processing method based on point cloud data according to claim 9, characterized in that, Based on the newly determined classification bias value, the point cloud density in the scanning parameters is adjusted to the corresponding value, where, The increase in point cloud density is positively correlated with the average value of the redefined classification bias values.

Images