A DLG anomaly detection and correction method and system based on laser point cloud data

Abstract

The present invention discloses a DLG anomaly detection and correction method and system based on laser point cloud data, the method comprising: collecting oblique photography data, making a real-scene three-dimensional model, and generating an original DLG; acquiring airborne laser point cloud data; performing filtering and classification, and screening out ground point cloud data; constructing a digital elevation model; extracting elevation points in the original DLG, and determining the corresponding positions of the elevation points in the original DLG in the digital elevation model; judging whether there is an abnormality in the elevation difference between the elevation value of the elevation point in the original DLG and the elevation value of the point cloud interpolation at the corresponding position of the digital elevation model, and performing correction; and performing precision analysis and storage processing on the corrected DLG. The present invention can effectively detect and correct abnormal elevation points in DLG, has the advantages of low cost, high efficiency, high precision, and high degree of automation, and has high application value.

Description

Technical Field

[0001] The present invention relates to the field of geographic information technology, and in particular to a DLG anomaly detection and correction method and system based on laser point cloud data. Background Art

[0002] Digital Line Graphic (DLG) is the basic data for urban construction management and development activities such as construction planning, resource management, and investment environment analysis. In recent years, the use of real-life 3D models to digitize 2D vector maps has become a new trend. Figure 1 It has always been a research hotspot. Realistic 3D is a real, three-dimensional, and time-sequential reflection of the spatiotemporal information of human production, life, and ecological space. It is a standardized product of new basic surveying and mapping. Realistic 3D models have the characteristics of real land display, clear and rich texture, measurability, easy update, and high precision. At the same time, this method can save time and cost, and is more convenient for storage management. However, due to problems such as oblique photography angles, the real-scene 3D model has problems such as stripes and blurs on the roads in densely populated areas, making it difficult to extract the correct elevation value, especially in densely populated areas with trees, shrubs, and grass. It is almost impossible to extract the correct ground elevation points on the real-scene 3D model, so there will be a large number of abnormal elevation points in DLG.

[0003] At present, abnormal elevation points in DLG can only be detected manually, which requires operators to have rich surveying and mapping experience. The correction methods mainly include field collection of elevation points and low-altitude photogrammetry using drones. Field collection of elevation points can use real-time kinematic (RTK), leveling or total station measurement, which has high accuracy but low efficiency and consumes a lot of manpower, financial resources and material resources. Using drone low-altitude photogrammetry technology to build a real-life 3D model and extracting elevation points from the model has the advantages of low cost and high accuracy. However, for areas covered by vegetation, it is impossible to extract the correct ground elevation points on the real-life 3D model, and there is a problem of price difference in modeling effect. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art. The present invention provides a DLG anomaly detection and correction method and system based on laser point cloud data, which can effectively detect and correct abnormal elevation points in digital line map DLG, has the advantages of low cost, high efficiency, high precision and high degree of automation, and has high application value.

[0005] The present invention provides a DLG anomaly detection and correction method based on laser point cloud data, the method comprising:

[0006] Based on the oblique photography of the drone, the oblique photography data of the area to be detected is collected, the real-scene three-dimensional model of the area to be detected is made using the oblique photography data, and the original DLG of the area to be detected is generated;

[0007] Acquire airborne laser point cloud data of the area to be inspected based on airborne laser radar equipment;

[0008] Filter and classify the acquired airborne laser point cloud data to filter out the ground point cloud data;

[0009] Building a digital elevation model based on the ground point cloud data;

[0010] Extracting elevation points in the original DLG, and determining corresponding positions of the elevation points in the original DLG in the digital elevation model;

[0011] Determine whether there is an abnormality in the elevation point in the original DLG based on the relationship between the elevation point in the original DLG and the corresponding position in the digital elevation model, and correct the abnormal elevation point in the original DLG based on the determination result;

[0012] After the correction is completed, the accuracy of the corrected DLG is analyzed and the data is put into storage.

[0013] Furthermore, the method of collecting oblique photography data of the area to be detected based on oblique photography of a drone, making a real-scene three-dimensional model of the area to be detected using the oblique photography data, and generating an original DLG of the area to be detected includes:

[0014] Collecting oblique photography data of multiple perspectives in the area to be detected based on oblique photography of unmanned aerial vehicles, and preprocessing the oblique photography data;

[0015] Performing dense matching processing on the preprocessed oblique photography data to obtain dense matching point cloud data of the preprocessed oblique photography data;

[0016] Automatically reconstructing the densely matched point cloud data into three dimensions to generate a real-scene three-dimensional model of the area to be detected;

[0017] The real-scene three-dimensional model of the area to be detected is three-dimensionally mapped to generate an original DLG of the area to be detected.

[0018] Furthermore, filtering and classifying the acquired airborne laser point cloud data to screen out ground point cloud data includes:

[0019] Setting a ground reference elevation index, filtering out the airborne laser point cloud data that is higher than the ground reference elevation index or lower than the ground reference elevation index in the airborne laser point cloud data based on the ground reference elevation index, and acquiring filtered airborne laser point cloud data;

[0020] Setting a ground point cloud classification parameter index, automatically classifying the filtered airborne laser point cloud data based on the ground point cloud classification parameter index, and screening out ground point cloud initial data, vegetation point cloud initial data, and building point cloud initial data;

[0021] The ground point cloud classification parameter index is adjusted, and the ground point cloud initial data, the vegetation point cloud initial data, and the building point cloud initial data are reclassified based on the adjusted ground point cloud classification parameter index to filter out the ground point cloud data.

[0022] Furthermore, constructing a digital elevation model based on the ground point cloud data includes:

[0023] The ground point cloud data is processed by weighted average interpolation method to construct a digital elevation model.

[0024] Furthermore, the step of extracting the elevation points in the original DLG and determining the corresponding positions of the elevation points in the original DLG in the digital elevation model includes:

[0025] Dividing the original DLG into a grid network consisting of a plurality of grids;

[0026] Encode the elevation points of each grid in the grid network;

[0027] Determine the position coordinates of the corresponding elevation point according to the elevation point code;

[0028] The same position of the elevation point of the original DLG in the digital elevation model is matched according to the position coordinates.

[0029] Further, judging whether there is an abnormality in the elevation point in the original DLG based on the relationship between the elevation point in the original DLG and the corresponding position in the digital elevation model includes:

[0030] Calculating the elevation values ​​of the elevation points in the original DLG;

[0031] Calculate the elevation difference between the elevation value of the elevation point in the original DLG and the elevation value of the point cloud interpolation at the same position in the digital elevation model;

[0032] An abnormal threshold is preset, and whether there is an abnormality at the elevation point at the position in the original DLG is determined based on the comparison result between the elevation difference and the abnormal threshold.

[0033] Further, judging whether there is an abnormality at the elevation point at the position in the original DLG based on the comparison result between the elevation difference and the abnormal threshold value includes:

[0034] When the elevation difference is greater than the abnormal threshold, it is determined that the elevation point at the position in the original DLG is abnormal;

[0035] When the elevation difference is not greater than the abnormal threshold, it is determined that there is no abnormality in the elevation point at the position in the original DLG.

[0036] Further, the correcting of abnormal elevation points in the original DLG based on the judgment result includes:

[0037] For abnormal elevation points in the original DLG, elevation replacement is performed using the point cloud interpolation elevation value of the corresponding position of the abnormal elevation point in the digital elevation model.

[0038] Furthermore, after the correction is completed, the accuracy analysis of the corrected DLG is performed, and the storage processing includes:

[0039] The corrected DLG is measured, the measured value is compared with the corrected value, the accuracy analysis result is obtained, and the corrected DLG is put into storage.

[0040] The present invention also provides a DLG anomaly detection and correction system based on laser point cloud data, which is used to implement the above-mentioned DLG anomaly detection and correction method based on laser point cloud data. The system includes:

[0041] An original map generation module, the original map generation module is used to collect oblique photography data of the area to be detected based on oblique photography of the drone, use the oblique photography data to make a real-scene three-dimensional model of the area to be detected, and generate an original DLG of the area to be detected;

[0042] A laser point cloud data acquisition module, which is used to acquire airborne laser point cloud data of the area to be detected based on an airborne laser radar device;

[0043] A filtering and classification module, wherein the filtering and classification module is used to filter and classify the acquired airborne laser point cloud data to filter out the ground point cloud data;

[0044] A digital elevation model construction module, wherein the digital elevation model construction module is used to construct a digital elevation model based on the ground point cloud data;

[0045] An elevation point extraction module, the elevation point extraction module is used to extract the elevation points in the original DLG and determine the corresponding positions of the elevation points in the original DLG in the digital elevation model;

[0046] An elevation point detection and correction module, the elevation point detection and correction module is used to determine whether there are abnormalities in the elevation points in the original DLG based on the relationship between the elevation points in the original DLG and the corresponding positions in the digital elevation model, and correct the abnormal elevation points in the original DLG based on the judgment result;

[0047] The precision analysis and warehousing module is used to perform precision analysis on the corrected DLG after correction is completed, and to perform warehousing processing.

[0048] The present invention provides a DLG anomaly detection and correction method and system based on laser point cloud data. A real-scene three-dimensional model is produced based on unmanned aerial vehicle oblique photography to generate an original DLG. At the same time, airborne laser point cloud data is collected based on an airborne laser radar device, and abnormal elevation points in the original DLG are extracted by analyzing the airborne laser point cloud data. The characteristics of airborne laser radar technology, such as high spatial resolution, large dynamic detection range, laser being able to effectively penetrate vegetation obstructions, and obtaining high-precision three-dimensional information of the real surface, are fully utilized. The abnormal elevation points in the DLG of the digital line map are effectively detected and corrected. The method has the advantages of low cost, high efficiency, high precision, and high degree of automation, and has high application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0050] Figure 1 is a flow chart of a DLG anomaly detection and correction method based on laser point cloud data in Embodiment 1 of the present invention;

[0051] Figure 2 is a flow chart of generating the original DLG of the area to be detected in the first embodiment of the present invention;

[0052] Figure 3 is a flow chart of filtering and classifying airborne laser point cloud data in Embodiment 1 of the present invention;

[0053] Figure 4 is a flow chart of constructing a digital elevation model in the first embodiment of the present invention;

[0054] Figure 5 It is a flowchart of extracting the elevation points of the original DLG and the corresponding positions of the digital elevation model in the first embodiment of the present invention;

[0055] Figure 6is a flow chart of abnormal detection and correction of elevation points in the original DLG in the first embodiment of the present invention;

[0056] Figure 7 is a schematic diagram of the original DLG elevation points in the first embodiment of the present invention;

[0057] Figure 8 It is a schematic diagram of elevation points of the corrected digital line map DLG in the first embodiment of the present invention;

[0058] Fig. 9 It is the accuracy analysis comparison diagram in the first embodiment of the present invention;

[0059] Fig.10 It is an architecture diagram of a DLG anomaly detection and correction system based on laser point cloud data in Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0060] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] In the present invention, it should be understood that terms such as "include" or "have" are intended to indicate the existence of features, numbers, steps, behaviors, components, parts or a combination thereof disclosed in the specification, and are not intended to exclude the possibility that one or more other features, numbers, steps, behaviors, components, parts or a combination thereof exist or are added.

[0062] It should also be noted that, in the absence of conflict, the embodiments and features of the embodiments of the present invention can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0063] Embodiment 1

[0064] Embodiment 1 of the present invention provides a DLG anomaly detection and correction method based on laser point cloud data, the method comprising: collecting oblique photography data of an area to be detected based on oblique photography of an unmanned aerial vehicle, using the oblique photography data to make a real-life three-dimensional model of the area to be detected, and generating an original DLG of the area to be detected; acquiring airborne laser point cloud data of the area to be detected based on an airborne laser radar device; filtering and classifying the acquired airborne laser point cloud data to screen out ground point cloud data; constructing a digital elevation model based on the ground point cloud data; extracting elevation points in the original DLG, and determining corresponding positions of the elevation points in the original DLG in the digital elevation model; judging whether there are abnormalities in the elevation points in the original DLG based on the relationship between the elevation points in the original DLG and the corresponding positions in the digital elevation model, and correcting the abnormal elevation points in the original DLG based on the judgment result; after the correction is completed, performing accuracy analysis on the corrected DLG and performing storage processing.

[0065] In an optional implementation of this embodiment, Figure 1 As shown, Figure 1 A flow chart of a DLG anomaly detection and correction method based on laser point cloud data in Embodiment 1 of the present invention is shown, comprising the following steps:

[0066] S101, collecting oblique photography data of the area to be detected based on oblique photography of a drone, making a real-scene three-dimensional model of the area to be detected using the oblique photography data, and generating an original DLG of the area to be detected;

[0067] In an optional implementation of this embodiment, Figure 2 As shown, Figure 2 The flowchart of generating the original DLG of the area to be detected in the first embodiment of the present invention is shown, including the following steps:

[0068] S201, collecting oblique photography data of multiple perspectives in the area to be detected based on oblique photography of a drone, and preprocessing the oblique photography data;

[0069] In an optional implementation of this embodiment, oblique photography data of multiple perspectives in the area to be detected is collected by oblique photography using an unmanned aerial vehicle device, and the collected oblique photography data is preprocessed.

[0070] In an optional implementation of this embodiment, the preprocessing includes automatic registration processing.

[0071] Specifically, the automatic registration process is to perform automatic registration on the oblique photography data based on the SIFT algorithm or the MSER algorithm, and the specific process is to perform feature extraction, feature description, similarity comparison, overall registration and strict adjustment on the oblique photography data based on the algorithm.

[0072] It should be noted that the SIFT (Scale Invariant Feature Transform) algorithm is a point feature registration algorithm, and the MSER (Maximally Stable Extremal Regions) algorithm is a surface feature registration algorithm, which can be set according to actual registration requirements.

[0073] Since the method used here is to use drones for high-altitude oblique photography, the collected oblique photography data will have serious image deformation problems. Therefore, it is necessary to automatically align the collected oblique photography data to solve the serious image deformation problem caused by large inclination angles and multiple viewing angles in the oblique photography data.

[0074] In an optional implementation of this embodiment, the preprocessing further includes multi-view spatial triadjustment processing.

[0075] Specifically, the specific process of the multi-view spatial three-adjustment processing is to detect the gross errors of the matching points of the oblique photography data and build a free network, and then perform the regional network adjustment of the oblique photography data. The regional network adjustment of the oblique photography data, that is, the three adjustments, includes unconstrained regional network adjustment, additional constrained regional network adjustment and direct orientation regional network adjustment.

[0076] Here, since there will be gross registration error connection points after the automatic registration process, and there is a large overlap, it is necessary to perform multi-view spatial three-adjustment processing on the oblique photography data after the automatic registration process to eliminate the existing gross registration error connection points.

[0077] S202, performing dense matching processing on the pre-processed oblique photography data to obtain dense matching point cloud data of the pre-processed oblique photography data;

[0078] In an optional implementation of this embodiment, dense matching processing is performed on the pre-processed oblique photography data based on the CMVS (Clustered Multi-View Stereo) algorithm to obtain dense matching point cloud data of the pre-processed oblique photography data.

[0079] Specifically, since the method of using a drone to perform oblique photography at high altitude is adopted, the collected oblique photography data will still have the problem of serious obstruction by buildings or structures. Therefore, dense matching processing is required here. After preprocessing in step S201, the local deformation information of the feature points of the preprocessed oblique photography data is obtained based on the solution of the external orientation elements, and the camera points obstructed by the preprocessed oblique photography data are densely matched according to the deformation information. After the dense matching is completed, dense matching point cloud data is generated.

[0080] S203, automatically performing three-dimensional reconstruction processing on the dense matching point cloud data to generate a real-scene three-dimensional model of the area to be detected;

[0081] In an optional implementation of the present embodiment, the automatic three-dimensional reconstruction processing includes, first, performing point cloud fusion, meshing and denoising on the densely matched point cloud data; then performing point cloud simplification processing to blur the central point cloud data, highlight the edge point cloud data, and maintain the geometric shape of the model; finally, performing automatic texture mapping processing to calculate the projection texture mapping coordinates according to the projection effect, and generating construction data of the real-scene three-dimensional model according to the projection texture mapping coordinates, thereby generating a real-scene three-dimensional model of the area to be detected.

[0082] It should be noted that there is still a problem of repeated textures in the oblique photography data of the area to be detected based on the oblique photography of drones. The automatic three-dimensional reconstruction processing here can effectively filter out a large number of repeated textures and generate a more accurate real-scene three-dimensional model in the area to be detected.

[0083] S204: Perform three-dimensional mapping on the real-scene three-dimensional model of the area to be detected to generate an original DLG of the area to be detected.

[0084] In an optional implementation of this embodiment, three-dimensional mapping is performed on the real-scene three-dimensional model of the area to be detected to generate an original DLG of the area to be detected.

[0085] In an optional implementation of this embodiment, a manned aircraft (Pilatus PC-6) equipped with a SWDC-max6 digital aerial camera is used to perform oblique photogrammetry to complete the production of a real-life 3D model of the area to be inspected, and 3D mapping is performed based on the real-life 3D model to generate a 1:500 scale digital line map DLG of approximately 112 square kilometers.

[0086] In an optional implementation of this embodiment, iData software is used to generate the original DLG of the area to be detected.

[0087] S102, acquiring airborne laser point cloud data of the area to be detected based on the airborne laser radar device;

[0088] In an optional implementation of this embodiment, the airborne laser radar device performs a full coverage scan on the area to be detected to obtain airborne laser point cloud data of the area to be detected.

[0089] It should be noted that due to the existing oblique photography data collected by drones, the original DLG generated has low accuracy for vegetation-covered areas. Here, airborne laser radar equipment is used with high spatial resolution, large dynamic detection range, the ability to partially penetrate trees and grass, and the ability to directly obtain high-precision three-dimensional information of the real surface, so as to obtain more accurate airborne laser point cloud data in the area to be detected.

[0090] In an optional implementation of this embodiment, a Rigel VQ-1560i airborne laser radar is used to perform a full coverage scan of the area to be detected, and the point cloud density is about 30 points / square meter.

[0091] S103, filtering and classifying the acquired airborne laser point cloud data to filter out ground point cloud data;

[0092] In an optional implementation of this embodiment, Figure 3 As shown, Figure 3 The flowchart of filtering and classifying airborne laser point cloud data in the first embodiment of the present invention is shown, including the following steps:

[0093] S301, setting a ground reference elevation index, filtering out the airborne laser point cloud data that is higher than the ground reference elevation index or lower than the ground reference elevation index in the airborne laser point cloud data, and acquiring filtered airborne laser point cloud data;

[0094] In an optional implementation of the present embodiment, a ground benchmark elevation index is determined based on an RTK level meter, and based on the ground benchmark elevation index, the airborne laser point cloud data that is higher than the ground benchmark elevation index or lower than the ground benchmark elevation index, that is, the airborne laser point cloud data that is obviously higher than the ground or lower than the ground is filtered out to obtain filtered airborne laser point cloud data.

[0095] Here, the airborne laser point cloud data that is obviously above or below the ground is filtered out to improve the accuracy of the airborne laser point cloud data.

[0096] S302, setting a ground point cloud classification parameter index, automatically classifying the filtered airborne laser point cloud data based on the ground point cloud classification parameter index, and screening out ground point cloud initial data, vegetation point cloud initial data, and building point cloud initial data;

[0097] In an optional implementation of this embodiment, ground point cloud classification parameter indicators are preset, including but not limited to non-vegetation point cloud parameter indicators, vegetation point cloud parameter indicators, ground point cloud parameter indicators, and building or structure point cloud parameter indicators.

[0098] In an optional implementation of this embodiment, the filtered airborne laser point cloud data is automatically classified based on the set ground point cloud classification parameter index to screen out ground point cloud initial data, vegetation point cloud initial data and building and structure point cloud initial data.

[0099] S303, adjusting the ground point cloud classification parameter index, reclassifying the ground point cloud initial data, vegetation point cloud initial data and building point cloud initial data based on the adjusted ground point cloud classification parameter index, and filtering out the ground point cloud data.

[0100] In an optional implementation of the present embodiment, the ground point cloud classification parameter index is refined in accuracy, a ground point cloud classification parameter index with higher point cloud classification accuracy is set, the ground point cloud initial data, vegetation point cloud initial data and building point cloud initial data are reclassified, and the ground point cloud data is screened out.

[0101] Here, automatic classification errors are considered, so it is necessary to set more accurate classification parameter indicators for reclassification to obtain more accurate ground point cloud data.

[0102] S104, constructing a digital elevation model based on the ground point cloud data;

[0103] In an optional implementation of this embodiment, the ground point cloud data is processed using a weighted average interpolation method to construct a digital elevation model.

[0104] Specifically, the weighted average interpolation method is to use the value obtained by weighted averaging the elevation values ​​of the discrete elevation points in the ground point cloud data as the elevation value of the interpolation point in the corresponding digital elevation model. The weight of the discrete elevation point to the interpolation point is generally inversely proportional to the distance, that is, the closer the discrete elevation point is to the interpolation point, the greater the influence of the elevation value of the interpolation point on the elevation value of the interpolation point. Conversely, the farther the discrete elevation point is from the interpolation point, the smaller the influence of the elevation value of the interpolation point on the elevation value of the interpolation point.

[0105] In an optional implementation of this embodiment, Figure 4 As shown, Figure 4 The flowchart of constructing a digital elevation model in the first embodiment of the present invention is shown, which includes the following steps:

[0106] S401, extracting discrete elevation points from the ground point cloud data;

[0107] In an optional implementation of this embodiment, a series of discretely distributed discrete elevation points are extracted from the ground point cloud data.

[0108] S402, obtaining the projection point of the discrete elevation point on the plane;

[0109] In an optional implementation of this embodiment, projection points of a plurality of discrete elevation points in the ground point cloud data on the XOY plane are obtained.

[0110] Specifically, obtain the discrete projection points of several discrete elevation points in the ground point cloud data on the XOY plane , these discrete projection points Each with elevation .

[0111] S403, constructing a regular grid digital elevation model based on the projection points;

[0112] In an optional implementation of this embodiment, based on the discrete projection points of a plurality of discrete elevation points in the ground point cloud data on the XOY plane , a regular grid digital elevation model (DEM) is constructed based on the weighted average interpolation method.

[0113] S404: Calculate point cloud interpolation elevation values ​​of interpolated regular grid intersections in the regular grid digital elevation model.

[0114] In an optional implementation of this embodiment, for any interpolated regular grid intersection in the regular grid digital elevation model (i.e. interpolation point), through the distance between the intersection of the interpolation regular grid and each point cloud nearby, the inverse ratio of the distance is taken as the weight of each point cloud, and the interpolation elevation value of the point cloud is equal to the sum of the products of the elevation values ​​of each point cloud and the weight.

[0115] Specifically, assuming is the discrete projection point The elevation value on , is the interpolated regular grid intersection With discrete projection points The distance between ), then each interpolated regular grid intersection The elevation value It can be expressed as:

[0116] ;

[0117] In the above formula, the weight function It can be expressed as:

[0118] ;

[0119] It is known that the elevation value of each discrete projection point can be expressed as:

[0120] ;

[0121] Direction correction It is expressed as:

[0122] ;

[0123] Discrete Projection Points Intersection with interpolated regular grid The cosine of the angle It is expressed as:

[0124] ;

[0125] Among them, when When , the interpolation result is more consistent with the actual situation, because:

[0126] ;

[0127] It can be seen from the above formula that the interpolation regular grid intersection The elevation value is inversely proportional to the square of the distance from the discrete projection point to the intersection of the interpolated regular grid. According to the above calculation, the distance between each discrete projection point and the intersection of the interpolated regular grid is calculated to calculate the weight. .

[0128] It should be noted that The number of data points to be used for calculation is generally set at least 4 and at most 10. Too many points will significantly increase the amount of calculation and will not be of much help to the calculation results. Too few points will make the calculation results inaccurate.

[0129] S105, extracting elevation points in the original DLG, and determining corresponding positions of the elevation points in the original DLG in the digital elevation model;

[0130] In an optional implementation of this embodiment, Figure 5 As shown, Figure 5 The flowchart of extracting the elevation points of the original DLG and the corresponding positions of the digital elevation model in the first embodiment of the present invention is shown, and includes the following steps:

[0131] S501, dividing the original DLG into a grid network consisting of a plurality of grids;

[0132] In an optional implementation of this embodiment, the original DLG is divided into a grid network consisting of a plurality of grids with the same or different areas according to actual needs.

[0133] S502, performing elevation point coding on the elevation point of each grid in the grid network;

[0134] In an optional implementation of this embodiment, elevation point coding is performed on the elevation points in each grid.

[0135] Specifically, LiDAR encoding is used here. After full coverage scanning using airborne laser radar equipment, the elevation points in the original DLG are encoded into point cloud data, using overall X, Y, Z coordinates, or grid encoding plus split X, Y, Z coordinates for representation.

[0136] In addition, TIN (Triangular Irregular Network) coding or Raster coding can be used for elevation point coding, which can be set according to actual needs.

[0137] S503, determining the position coordinates of the corresponding elevation point according to the elevation point code;

[0138] In an optional implementation of this embodiment, the position coordinates of the corresponding elevation point in the original DLG are determined according to the elevation point code obtained in step S502.

[0139] In an optional implementation of this embodiment, it is necessary to perform an accuracy check on the position coordinates of the corresponding elevation points in the original DLG.

[0140] Specifically, RTK measurement is performed on the position coordinates of the corresponding elevation points in the original DLG, and the position coordinates with a mean error of less than 2 cm in the plane position and a mean error of less than 3 cm in the elevation position are considered to be accurate.

[0141] S504: Match the same position of the elevation point of the original DLG in the digital elevation model according to the position coordinates.

[0142] In an optional implementation of this embodiment, the position coordinates of the elevation points in the original DLG are matched to the same positions in the digital elevation model.

[0143] Specifically, the digital elevation model adopts a regular rectangular grid, which is a data set of plane coordinates X, Y and elevation coordinates Z of terrain points arranged at equal intervals in the X and Y axis directions on a Gaussian projection platform. Here, the coordinates of the data set of the regular rectangular grid are matched based on the position coordinates of the elevation points in the original DLG, thereby determining the same positions of the elevation points in the original DLG in the digital elevation model.

[0144] It should be noted that the positions of the elevation points in the original DLG correspond one-to-one to the positions in the digital elevation model.

[0145] S106, judging whether there are abnormalities in the elevation points in the original DLG based on the relationship between the elevation points in the original DLG and the corresponding positions in the digital elevation model, and correcting the abnormal elevation points in the original DLG based on the judgment result;

[0146] In an optional implementation of this embodiment, Figure 6 As shown, Figure 6 The flowchart of abnormal detection and correction of elevation points in the original DLG in the first embodiment of the present invention is shown, including the following steps:

[0147] S601, calculating the elevation value of the elevation point in the original DLG;

[0148] In an optional implementation of this embodiment, the elevation values ​​of all elevation points in the original DLG are calculated, and the point cloud interpolation elevation values ​​of the interpolated regular grid intersections corresponding to the same positions in the digital elevation model that correspond one-to-one to the elevation points in the original DLG are extracted.

[0149] S602, calculating the elevation difference between the elevation value of the elevation point in the original DLG and the elevation value of the point cloud interpolation at the same position in the digital elevation model;

[0150] In an optional implementation of this embodiment, the elevation difference between the elevation value of each elevation point of the original DLG and the point cloud interpolation elevation value of the interpolation regular grid intersection corresponding to the same position in the digital elevation model is calculated.

[0151] It should be noted that the elevation difference is the absolute value of the elevation difference between the elevation value of each elevation point of the original DLG and the point cloud interpolation elevation value of the interpolation regular grid intersection corresponding to the same position in the digital elevation model.

[0152] S603, preset an abnormal threshold, and determine whether there is an abnormality at the elevation point at the position in the original DLG based on the comparison result between the elevation difference and the abnormal threshold;

[0153] In an optional implementation of the present embodiment, the setting of the abnormality threshold refers to the basic technical regulations for national basic scale map surveying and mapping. In the 1:500 scale DLG in the present embodiment, the regulations require that the error in the elevation of the flat-bottom elevation annotation point is 0.2m. Therefore, 0.2m is used here as the elevation difference abnormality threshold between the elevation value of the elevation point of the original DLG and the point cloud interpolated elevation value of the interpolated regular grid intersection corresponding to the same position of the digital elevation model.

[0154] In an optional implementation of this embodiment, the difference between the elevation difference and the abnormal threshold is determined to determine whether there is an abnormality in the elevation point at the position in the original DLG.

[0155] S604: There is an abnormality in the elevation point of the original DLG;

[0156] In an optional implementation of this embodiment, when the elevation difference in step S603 is greater than the abnormal threshold, it is determined that an abnormality exists in the corresponding elevation point of the original DLG.

[0157] Specifically, when the elevation difference is greater than the abnormal threshold, it is determined that an abnormality exists at the elevation point at the position in the original DLG.

[0158] S605: There is no abnormality in the elevation point of the original DLG;

[0159] In an optional implementation of this embodiment, when the elevation difference in step S603 is greater than the abnormality threshold, it is determined that there is no abnormality in the corresponding elevation point of the original DLG.

[0160] Specifically, when the elevation difference is not greater than the abnormal threshold, it is determined that there is no abnormality in the elevation point at the position in the original DLG.

[0161] S606: For abnormal elevation points in the original DLG, elevation replacement is performed using the point cloud interpolation elevation value of the corresponding position of the abnormal elevation point in the digital elevation model.

[0162] In an optional implementation of this embodiment, when it is determined that there is an abnormality in the elevation point at the position in the original DLG, the elevation value of the abnormal elevation point in the original DLG is replaced by the point cloud interpolation elevation value of the interpolated regular grid intersection corresponding to the same position in the digital elevation model.

[0163] In an optional implementation of this embodiment, after all elevation points with abnormalities in the original DLG are replaced with their elevations, a revised digital line map DLG is generated.

[0164] In an optional implementation of this embodiment, Figure 7 and Figure 8 As shown, Figure 7 Schematic diagram of the original DLG elevation point in the first embodiment of the present invention. Figure 8 The schematic diagram of the modified digital line map DLG elevation points in the first embodiment of the present invention is shown in the figure, wherein the elevation points marked with a short solid line below the elevation value are within the vegetation range, the elevation points marked with a short solid dashed line below the elevation value are within the non-vegetation range, and the elevation points marked with no line below the elevation value are within the range of near buildings or structures, as shown in the figure. Figure 8 The revised digital line map DLG is compared with Figure 7Some elevation values ​​in the original digital line map DLG have been corrected.

[0165] S107: After the correction is completed, the corrected DLG is subjected to precision analysis and then put into storage.

[0166] In an optional implementation of this embodiment, the corrected digital line map DLG is measured, the measured value is compared with the corrected value, the accuracy analysis result is obtained, and the corrected digital line map DLG is stored in the database.

[0167] Specifically, field RTK measurements are performed on abnormal elevation points on the digital line map DLG before correction, and the measured values, i.e., true values, are used as the corrected values ​​of the elevation points at the same positions in the corrected digital line map DLG, and the measured values ​​are compared with the corrected values, and accuracy analysis is performed based on the comparison results.

[0168] In an optional implementation of this embodiment, the three indicators of DLG original elevation as the original value, point cloud extracted elevation as the correction value, and RTK measured elevation as the measured value are used for accuracy analysis. The accuracy analysis results are shown in Tables 1 and Fig. 9 , Fig. 9 It is the accuracy analysis comparison diagram in the first embodiment of the present invention, that is, the comparison diagram of DLG original elevation, point cloud extracted elevation, and RTK measured elevation.

[0169] Table 1

[0170]

[0171] It can be seen that for elevation points within the vegetation range and far away from buildings: the point cloud extracted elevation is generally lower than the DLG original elevation, with a maximum difference of -0.35m and a difference error of 0.25m. Ten of the 12 elevation points were detected as abnormal points; the DLG original elevation is generally higher than the RTK measured elevation, with a maximum difference of 0.39m and a difference error of 0.24m, which verifies the accuracy of the detection; the point cloud corrected elevation is compared with the RTK measured elevation, with a maximum difference of 0.05m and a difference error of 0.05m, indicating that the two are well consistent. For elevation points not within the vegetation range and far away from buildings: the DLG original elevation, point cloud extracted elevation, and RTK measured elevation are not much different, with a maximum difference of 0.05m and a maximum difference error of 0.03m, indicating that the three have comparable accuracy, no abnormal values ​​are found in the detection, and the original elevation is retained. For elevation points near buildings: the elevation extracted from point cloud is generally lower than the original elevation of DLG, with the maximum difference of -0.30m and the median error of the difference of 0.24m. Six of the eight elevation points were detected as abnormal points. Compared with the RTK measured elevation, the original elevation of DLG is generally higher, with the maximum difference of 0.29m and the median error of the difference of 0.23m, indicating that there is indeed a large difference between the two. Compared with the RTK measured elevation, the maximum difference between the elevation extracted from point cloud and the RTK measured elevation is 0.05m and the median error of the difference is 0.06m, indicating that the two are in good agreement.

[0172] In summary, the following conclusions can be drawn:

[0173] (1) This method can accurately and effectively detect and correct abnormal DLG elevation;

[0174] (2) Airborne laser point cloud data can penetrate trees, grass and other vegetation to reach the ground. Therefore, the elevation points within the green area and far away from the buildings will be generally smaller than the original elevation points after the elevation points are corrected by the elevation extracted from the point cloud data, and will be closer to the true value.

[0175] (3) Generally, no correction is required for elevation points that are not within vegetation areas and far away from buildings, such as roads and hardened surfaces.

[0176] (4) For locations close to buildings, due to factors such as tree obstruction and aerial photography blind spots, the model may have quality problems such as streaking, adhesion, and holes, making it impossible to obtain accurate elevation values. After correcting the abnormal elevation values ​​using a method based on point cloud data to extract elevation, the elevation values ​​will be closer to the ground, the values ​​will become smaller, and closer to the actual elevation.

[0177] In an optional implementation of the present embodiment, in terms of time efficiency, compared with the traditional elevation measurement method, the DLG abnormal elevation detection and correction method based on airborne laser point cloud data proposed in the present embodiment has obvious efficiency advantages. After actual time statistics, if the traditional method is adopted, two engineers need to spend 1.5 hours in the field to complete the elevation data collection of 1 square kilometer of DLG, and it takes 1 hour to complete the detection and correction of abnormal elevation in 1 square kilometer of DLG in the office, and there is no guarantee that the detection and correction are completely accurate. It takes a total of 1.5*112+1*112=280 hours to complete the DLG abnormal elevation detection and correction of 112 square kilometers of the monitored area in the present embodiment. If the method in the present embodiment is adopted, two engineers only need to spend 21 hours in the field to complete the airborne laser scanning of 112 square kilometers of the area to be detected, and it takes 40 hours in the office to complete the DLG abnormal elevation detection and correction after laser point cloud data processing, which takes a total of 61 hours, and the efficiency improvement rate is as high as 78.2%. It can be seen that the method in the present embodiment is more time efficient.

[0178] In an optional implementation of this embodiment, after the accuracy analysis is performed, the corrected digital line map DLG is stored in the database.

[0179] In summary, Embodiment 1 of the present invention provides a DLG anomaly detection and correction method based on laser point cloud data, which produces a real-scene three-dimensional model based on drone oblique photography to generate an original DLG, and collects airborne laser point cloud data based on an airborne laser radar device. The airborne laser point cloud data is analyzed to extract abnormal elevation points in the original DLG. The method makes full use of the characteristics of airborne laser radar technology, such as high spatial resolution, large dynamic detection range, laser being able to effectively penetrate vegetation obstructions, and obtaining high-precision three-dimensional information of the real surface, to effectively detect and correct abnormal elevation points in the DLG of a digital line map, which has the advantages of low cost, high efficiency, high precision, and high degree of automation, and has high application value.

[0180] Embodiment 2

[0181] Embodiment 2 of the present invention provides a DLG anomaly detection and correction system based on laser point cloud data, which is used to implement the DLG anomaly detection and correction method based on laser point cloud data in Embodiment 1. The system includes an original DLG generation module, a laser point cloud data acquisition module, a filtering and classification module, a digital elevation model construction module, an elevation point extraction module, an elevation point detection and correction module, and an accuracy analysis and warehousing module.

[0182] In an optional implementation of this embodiment, Fig.10 As shown, Fig.10 The following is a diagram showing the architecture of a DLG anomaly detection and correction system based on laser point cloud data in Embodiment 2 of the present invention, including the following modules:

[0183] The original map generation module 10 is used to collect oblique photography data of the area to be detected based on oblique photography of the drone, use the oblique photography data to make a real-scene three-dimensional model of the area to be detected, and generate an original DLG of the area to be detected;

[0184] A laser point cloud data acquisition module 20, wherein the laser point cloud data acquisition module 20 is used to acquire airborne laser point cloud data of the area to be detected based on an airborne laser radar device;

[0185] A filtering and classification module 30, wherein the filtering and classification module 30 is used to filter and classify the acquired airborne laser point cloud data to filter out the ground point cloud data;

[0186] A digital elevation model building module 40, wherein the digital elevation model building module 40 is used to build a digital elevation model based on the ground point cloud data;

[0187] An elevation point extraction module 50, the elevation point extraction module 50 is used to extract elevation points in the original DLG and determine corresponding positions of the elevation points in the original DLG in the digital elevation model;

[0188] An elevation point detection and correction module 60, the elevation point detection and correction module 60 is used to determine whether there are abnormalities in the elevation points in the original DLG based on the relationship between the elevation points in the original DLG and the corresponding positions in the digital elevation model, and to correct the abnormal elevation points in the original DLG based on the judgment result;

[0189] The precision analysis and storage module 70 is used to perform precision analysis on the corrected digital line map DLG after the correction is completed, and perform storage processing.

[0190] In summary, the second embodiment of the present invention provides a DLG anomaly detection and correction system based on laser point cloud data, which is used to implement the DLG anomaly detection and correction method based on laser point cloud data in the first embodiment, and produces a real-scene three-dimensional model based on drone oblique photography to generate an original DLG. At the same time, airborne laser point cloud data is collected based on airborne laser radar equipment, and abnormal elevation points in the original DLG are extracted by analyzing the airborne laser point cloud data. The characteristics of airborne laser radar technology, such as high spatial resolution, large dynamic detection range, laser being able to effectively penetrate vegetation obstructions, and obtaining high-precision three-dimensional information of the real surface, are fully utilized to effectively detect and correct abnormal elevation points in the digital line map DLG. The system has the advantages of low cost, high efficiency, high precision, and high degree of automation, and has high application value.

[0191] A person skilled in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the storage medium may include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0192] In addition, the embodiments of the present invention are described in detail above. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A DLG anomaly detection and correction method based on laser point cloud data, characterized in that: The method comprises: Based on the oblique photography of the drone, the oblique photography data of the area to be detected is collected, the real-scene three-dimensional model of the area to be detected is made using the oblique photography data, and the original DLG of the area to be detected is generated; Acquire airborne laser point cloud data of the area to be inspected based on airborne laser radar equipment; Filter and classify the acquired airborne laser point cloud data to filter out the ground point cloud data; Building a digital elevation model based on the ground point cloud data; Extracting elevation points in the original DLG and determining corresponding positions of the elevation points in the original DLG in the digital elevation model; The extracting of the elevation points in the original DLG and determining the corresponding positions of the elevation points in the original DLG in the digital elevation model comprises: dividing the original DLG into a grid network composed of a plurality of grids; performing elevation point coding on the elevation points of each grid in the grid network; determining the position coordinates of the corresponding elevation points according to the elevation point coding; and matching the same positions of the elevation points of the original DLG in the digital elevation model according to the position coordinates; Determine whether there is an abnormality in the elevation point in the original DLG based on the relationship between the elevation point in the original DLG and the corresponding position in the digital elevation model, and correct the abnormal elevation point in the original DLG based on the determination result; The step of judging whether an elevation point in the original DLG is abnormal based on the relationship between the elevation point in the original DLG and the corresponding position in the digital elevation model includes: calculating the elevation value of the elevation point in the original DLG; calculating the elevation difference between the elevation value of the elevation point in the original DLG and the elevation value of the point cloud interpolation at the same position in the digital elevation model; presetting an abnormal threshold, and judging whether an elevation point at the position in the original DLG is abnormal based on the comparison result of the elevation difference and the abnormal threshold; The correcting of the abnormal elevation points in the original DLG based on the judgment result includes: when there are abnormal elevation points in the original DLG, using the point cloud interpolation elevation value of the corresponding position of the abnormal elevation point in the digital elevation model to replace the elevation; After the correction is completed, the corrected DLG is analyzed for accuracy and then put into storage.

2. The DLG anomaly detection and correction method based on laser point cloud data according to claim 1, characterized in that: The method of collecting oblique photography data of the area to be detected based on oblique photography of a drone, making a real-scene three-dimensional model of the area to be detected by using the oblique photography data, and generating an original DLG of the area to be detected includes: Collecting oblique photography data of multiple perspectives in the area to be detected based on oblique photography of unmanned aerial vehicles, and preprocessing the oblique photography data; Performing dense matching processing on the preprocessed oblique photography data to obtain dense matching point cloud data of the preprocessed oblique photography data; Automatically reconstructing the densely matched point cloud data into three dimensions to generate a real-scene three-dimensional model of the area to be detected; The real-scene three-dimensional model of the area to be detected is three-dimensionally mapped to generate an original DLG of the area to be detected.

3. The DLG anomaly detection and correction method based on laser point cloud data according to claim 1, characterized in that: The filtering and classification of the acquired airborne laser point cloud data to screen out the ground point cloud data includes: Setting a ground reference elevation index, filtering out the airborne laser point cloud data that is higher than the ground reference elevation index or lower than the ground reference elevation index in the airborne laser point cloud data based on the ground reference elevation index, and acquiring filtered airborne laser point cloud data; Setting a ground point cloud classification parameter index, automatically classifying the filtered airborne laser point cloud data based on the ground point cloud classification parameter index, and screening out ground point cloud initial data, vegetation point cloud initial data, and building point cloud initial data; The ground point cloud classification parameter index is adjusted, and the ground point cloud initial data, the vegetation point cloud initial data, and the building point cloud initial data are reclassified based on the adjusted ground point cloud classification parameter index to filter out the ground point cloud data.

4. The DLG anomaly detection and correction method based on laser point cloud data according to claim 1, characterized in that: The constructing of a digital elevation model based on the ground point cloud data comprises: The ground point cloud data is processed by weighted average interpolation method to construct a digital elevation model.

5. The DLG anomaly detection and correction method based on laser point cloud data according to claim 1, characterized in that: The step of judging whether there is an abnormality at the elevation point at the position in the original DLG based on the comparison result between the elevation difference and the abnormal threshold value includes: When the elevation difference is greater than the abnormal threshold, it is determined that the elevation point at the position in the original DLG is abnormal; When the elevation difference is not greater than the abnormal threshold, it is determined that there is no abnormality in the elevation point at the position in the original DLG.

6. The DLG anomaly detection and correction method based on laser point cloud data according to claim 1, characterized in that: After the correction is completed, the accuracy analysis of the corrected DLG is performed, and the storage processing includes: The corrected DLG is measured, the measured value is compared with the corrected value, the accuracy analysis result is obtained, and the corrected DLG is put into storage.

7. A DLG anomaly detection and correction system based on laser point cloud data, characterized in that: The DLG anomaly detection and correction system based on laser point cloud data is used to implement the DLG anomaly detection and correction method based on laser point cloud data according to any one of claims 1 to 6, and the system includes: An original map generation module, the original map generation module is used to collect oblique photography data of the area to be detected based on oblique photography of the drone, use the oblique photography data to make a real-scene three-dimensional model of the area to be detected, and generate an original DLG of the area to be detected; A laser point cloud data acquisition module, which is used to acquire airborne laser point cloud data of the area to be detected based on an airborne laser radar device; A filtering and classification module, wherein the filtering and classification module is used to filter and classify the acquired airborne laser point cloud data to filter out the ground point cloud data; A digital elevation model construction module, wherein the digital elevation model construction module is used to construct a digital elevation model based on the ground point cloud data; An elevation point extraction module, the elevation point extraction module is used to extract the elevation points in the original DLG and determine the corresponding positions of the elevation points in the original DLG in the digital elevation model; The extracting of the elevation points in the original DLG and determining the corresponding positions of the elevation points in the original DLG in the digital elevation model comprises: dividing the original DLG into a grid network composed of a plurality of grids; performing elevation point coding on the elevation points of each grid in the grid network; determining the position coordinates of the corresponding elevation points according to the elevation point coding; and matching the same positions of the elevation points of the original DLG in the digital elevation model according to the position coordinates; An elevation point detection and correction module, the elevation point detection and correction module is used to determine whether there are abnormalities in the elevation points in the original DLG based on the relationship between the elevation points in the original DLG and the corresponding positions in the digital elevation model, and correct the abnormal elevation points in the original DLG based on the judgment result; The step of judging whether an elevation point in the original DLG is abnormal based on the relationship between the elevation point in the original DLG and the corresponding position in the digital elevation model includes: calculating the elevation value of the elevation point in the original DLG; calculating the elevation difference between the elevation value of the elevation point in the original DLG and the elevation value of the point cloud interpolation at the same position in the digital elevation model; presetting an abnormal threshold, and judging whether an elevation point at the position in the original DLG is abnormal based on the comparison result of the elevation difference and the abnormal threshold; The correcting of the abnormal elevation points in the original DLG based on the judgment result includes: when there are abnormal elevation points in the original DLG, using the point cloud interpolation elevation value of the corresponding position of the abnormal elevation point in the digital elevation model to replace the elevation; The precision analysis and warehousing module is used to perform precision analysis on the corrected DLG after correction is completed, and to perform warehousing processing.

Images