A high-precision geospatial information acquisition surveying and mapping method and system

Abstract

The application relates to the technical field of geographic information measurement, in particular to a high-precision geographic spatial information acquisition and mapping method and system, which specifically comprises the following steps: constructing a fusion point cloud dataset by combining point cloud data with RGB images and thermal infrared images at each moment, and analyzing terrain complexity; analyzing the occlusion and omission characteristics of the point cloud data through the dataset to construct an effective score of a blind area; constructing a reliability score of the data at each moment based on the difference between the vertical displacement of echo intensity at each moment and a preset reference true value, combining the effective score of the blind area and the terrain complexity to construct an occlusion compensation confidence of each moment; triggering a path planning algorithm based on the occlusion compensation confidence to guide a UAV to round an obstacle to compensate for a low-confidence area; the method can significantly improve the integrity of the data in the occlusion area and the micro-feature capture rate, reduce the blind area of high-precision mapping, improve the scene data integrity, and ensure the engineering usability of complex scene data.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of geographic information measurement, in particular to a high-precision geographic spatial information acquisition and mapping method and system. BACKGROUND

[0002] The acquisition and mapping of high-precision geographic spatial information refers to a process of obtaining accurate positions, geometric shapes, attributes and spatio-temporal change information of objects on the earth surface by using advanced technology. With the development of science and technology, the precision and efficiency of high-precision geographic spatial information acquisition have been greatly improved.

[0003] However, in the acquisition and mapping of high-precision geographic spatial information in complex scenes, the unmanned aerial vehicle aerial photography is difficult to obtain the real terrain at the bottom of the building and under the tree crown due to the limitation of the visual angle, and the ground laser scanning cannot penetrate the vegetation or capture the back target due to the visual obstruction; at the same time, the physical bottleneck of the sensor leads to a large number of omissions of micro-features such as road markings, small well covers and pipeline nodes, which directly affects the engineering usability of the data. In the hidden area and detail acquisition, the multi-band laser radar with stronger penetration is used for ultra-high resolution oblique photography. Although the penetration technology can partially overcome the shielding of water bodies or sparse vegetation to obtain the riverbed or shallow surface shape, the physical penetration ability is limited by the medium properties and wavelength, and it is difficult to obtain the real structure in the deep and dense shielding area, and cannot completely replace the field verification. Thus, the existing technology cannot obtain relatively complete geographic spatial information. SUMMARY

[0004] In order to solve the above technical problems, the purpose of the present application is to provide a high-precision geographic spatial information acquisition and mapping method and system, and the technical solution is as follows:

[0005] In a first aspect, the present application provides a high-precision geographic spatial information acquisition and mapping method, which comprises the following steps:

[0006] For urban streets, the RGB image and the thermal infrared image collected at each time are used to give the RGB color value and the thermal radiation intensity value to the point cloud data collected by the laser sensor at each time, and a fusion point cloud data set at each time is constructed;

[0007] The point cloud data at each time is grid divided, the terrain parameters of each grid are constructed based on the three-dimensional distribution characteristics of the point cloud data in each grid, and the terrain complexity at each time is determined; the reliability score of the point cloud data at each time is constructed based on the difference between the vertical displacement amount of the laser sensor echo intensity and the preset reference true value;

[0008] The point cloud data set is clustered to obtain each feature cluster, the horizontal plane projection area of the point cloud data in each feature cluster is obtained, and the effective point number per unit area of each feature cluster is constructed in combination with the number of point cloud data in the feature cluster; based on each feature cluster at each time, the label value of each feature cluster is obtained in combination with the random forest algorithm; and the effective score of the blind area at each time is calculated based on the effective point number per unit area and the label value.

[0009] The prediction accuracy of the effective score of the blind area at each time is calculated based on the effective score of the blind area at all times before each time and the terrain complexity; and the compensation value at each time is obtained based on the prediction accuracy and the reliability score at all times in combination with a compensation algorithm.

[0010] The shielding compensation confidence at each time is constructed based on the compensation value and the prediction accuracy, and the motion path of the data acquisition equipment is adjusted based on the shielding compensation confidence.

[0011] In one embodiment, the process of obtaining the terrain complexity at each time is as follows:

[0012] The horizontal plane on which the point cloud data at each time is projected is divided into a grid of a preset size, the standard deviation of all point cloud elevation values in each grid is taken as a first evaluation value, the coefficient of variation of the point cloud projection point density of each grid and all adjacent grids thereof is calculated as a second evaluation value, and the weighted sum of the first evaluation value and the second evaluation value of each grid is taken as the terrain parameter of each grid. The mean value of the terrain parameters of all grids at each time is taken as the terrain complexity at each time.

[0013] In one embodiment, the process of obtaining the reliability score of the point cloud data at each time is as follows:

[0014] The relative error between the vertical displacement of the current time laser sensor echo intensity and the preset reference true value is calculated, and the difference between the natural number 1 and the relative error is taken as the reliability score of the point cloud data at the current time.

[0015] In one embodiment, the process of obtaining the effective point number per unit area of each feature cluster is as follows:

[0016] The horizontal plane projection result of all point cloud data points in each feature cluster is taken as the input of the convex hull algorithm to obtain the corresponding horizontal plane projection area, and the ratio of the number of point cloud data points in each feature cluster to the horizontal plane projection area is taken as the effective point number per unit area of each feature cluster.

[0017] In one embodiment, the process of obtaining the label value of each feature cluster is as follows:

[0018] A random forest model is trained through all feature clusters at historical times, wherein the label and label value of each feature cluster include: 0 for effective feature and 1 for shielding omission.

[0019] Taking all feature clusters of the current moment as input of the trained random forest, the output is the label value of each feature cluster of the current moment.

[0020] In one embodiment, the process of obtaining the effective score of the blind area of each moment is as follows:

[0021] Calculate the product of the effective point number per unit area of each feature cluster and the label value; take the product of all feature clusters of the current moment as the effective score of the blind area of the current moment.

[0022] In one embodiment, the process of obtaining the prediction accuracy of the effective score of the blind area of each moment is as follows:

[0023] Take the sequence composed of the terrain complexity of all moments before the current moment as the terrain complexity sequence of the current moment; take the sequence composed of the effective score of the blind area of all moments before the current moment as the blind area effective sequence of the current moment;

[0024] Take the blind area effective sequence and the terrain complexity sequence of the current moment as input of the XGBoost algorithm, and the output is the prediction accuracy of the effective score of the blind area of the current moment.

[0025] In one embodiment, the process of obtaining the compensation value of each moment is as follows:

[0026] Take the sequence composed of the prediction accuracy of all moments before the current moment as the prediction accuracy sequence of the current moment; take the sequence composed of the reliability score of all moments before the current moment as the reliability score sequence of the current moment;

[0027] Take the prediction accuracy sequence and the reliability score sequence of the current moment as input of the weighted linear compensation algorithm, and the output is the compensation value of the current moment.

[0028] In one embodiment, the process of constructing the shielding compensation confidence of each moment and adjusting the motion path of the data acquisition equipment based on the shielding compensation confidence is as follows:

[0029] Take the normalized value of the ratio of the prediction accuracy and the compensation value of the current moment as the shielding compensation confidence of the current moment;

[0030] If the shielding compensation confidence of the current moment is lower than the preset trigger threshold, trigger the path planning algorithm immediately to generate the optimal compensation path of the unmanned aerial vehicle, and perform regional rescan along the path.

[0031] In a second aspect, the embodiments of the present application also provide a high-precision geospatial information acquisition and mapping system, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, and the processor implements the steps of the method according to any one of the preceding aspects when executing the computer program.

[0032] The embodiments of the present application have at least the following beneficial effects:

[0033] The present application assigns RGB color values and thermal radiation intensity values to the point cloud data collected by the laser sensor at each moment by collecting RGB images and thermal infrared images at each moment, constructs a fusion point cloud data set at each moment, analyzes the terrain complexity at each moment through the three-dimensional distribution characteristics of the point cloud data, clusters the fusion point cloud data set to obtain each feature cluster, analyzes the occluded missing features of the point cloud data in combination with the random forest, constructs the effective score of the blind area at each moment, quantifies the effective point number per unit area of the occluded area and the recognizability, accurately locates the occluded hotspot area with missing data, and avoids the blindness of the traditional method in missing micro features; for the problem of accuracy fluctuation of the occluded area in complex terrain and performance difference of the sensor leading to data credibility decline, the reliability score of the point cloud data at each moment is constructed based on the difference between the vertical displacement of the sensor echo intensity at each moment and the preset reference true value, the prediction accuracy of the blind area effective score at each moment is calculated in combination with the blind area effective score and the terrain complexity, the occluded compensation confidence at each moment is constructed, and the penetration ability and data restoration credibility of multi-source cooperation on the occluded area are comprehensively reflected, so as to eliminate the interference of the performance fluctuation of the sensor itself on the data quality of the occluded area; based on the occluded compensation confidence at the collection moment, the path planning algorithm is triggered to guide the UAV to round the obstacle to compensate for the low confidence area, significantly improve the integrity of the data of the occluded area and the micro feature capture rate, reduce the blind area of high-precision mapping, improve the scene data integrity, and ensure the engineering usability of complex scene data. BRIEF DESCRIPTION OF DRAWINGS

[0034] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the drawings needed in the following embodiment or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0035] Figure 1 A step flowchart of a high-precision geospatial information acquisition and mapping method provided by an embodiment of the present application is shown in the following figure:

[0036] Figure 2 A schematic diagram of the terrain complexity acquisition process is shown in the following figure. DETAILED DESCRIPTION

[0037] In order to further clarify the technical means and effects taken by the present application to achieve the predetermined object of the application, the specific implementation, structure, features and effects of the high-precision geographic spatial information acquisition and mapping method and system according to the present application are described in detail as follows in combination with the drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0038] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0039] The specific scheme of the high-precision geographic spatial information acquisition and mapping method and system provided by the present application is described in detail below in combination with the drawings.

[0040] Please refer to Figure 1 which shows the step flowchart of the high-precision geographic spatial information acquisition and mapping method provided by one embodiment of the present application, which includes the following steps:

[0041] Step S1, for urban streets, through the RGB image and thermal infrared image collected at each time, the point cloud data collected by the laser sensor at each time is given RGB color value and thermal radiation intensity value, and the fusion point cloud data set at each time is constructed.

[0042] A multispectral LiDAR (Light Detection And Ranging) sensor and a five-lens oblique photography system are deployed on the aerial work platform to collect point cloud data of urban streets, specifically: centimeter-level precision point cloud data is directly obtained through the multispectral LiDAR sensor, RGB images and thermal infrared images are synchronously obtained through the five-lens oblique photography system, and the image frame rate is synchronized with the LiDAR scanning; the point cloud obtained by the LiDAR sensor is used as a spatial reference, the images obtained by the oblique photography system are projected to the point cloud coordinate system through the POS (Position and Orientation System) data and ICP registration algorithm, each LiDAR point is given RGB color value and thermal radiation intensity value, and a fusion point cloud data set with multispectral attributes is generated. In the embodiment of the present application, the aerial work platform is a UAV. The fusion point cloud data set at each acquisition time is obtained.

[0043] For the data acquisition frequency of the LiDAR sensor and the five-lens oblique photography system, preferably, in the embodiment of the present application, the data acquisition frequency is set to 10 Hz. As other embodiments of the present application, the implementer can set the data acquisition frequency by himself according to the actual situation.

[0044] In step S2, the point cloud data at each time is grid divided, the terrain parameters of each grid are constructed based on the three-dimensional distribution characteristics of the point cloud data in each grid, and the terrain complexity at each time is determined; the reliability score of the point cloud data at each time is constructed based on the difference between the vertical displacement amount of the echo intensity of the laser sensor and the preset reference true value.

[0045] (1) The point cloud data at each acquisition time is projected on a horizontal plane, and the projected horizontal plane is divided into 5m*5m grid units. Each grid unit needs to contain at least 50 point cloud projection points, otherwise it is marked as an invalid grid. For each grid, the standard deviation of the elevation values of all point clouds in the grid is calculated as the first evaluation value to quantify the terrain fluctuation intensity; the coefficient of variation of the point cloud projection point density of the grid and all its adjacent grids is calculated as the second evaluation value; the first evaluation value and the second evaluation value of each grid are weighted and summed, wherein the weights of the first evaluation value and the second evaluation value are automatically determined by using the entropy weight method, and the weighted sum result is taken as the terrain parameter of each grid. The average of the terrain parameters of all grids at each acquisition time is taken as the terrain complexity at each acquisition time, realizing the comprehensive quantification of the terrain ruggedness and morphological abruptness. Wherein, the acquisition process of elevation value, point cloud projection point density and coefficient of variation, as well as point cloud plane projection and entropy weight method, are all known technologies, and the specific process will not be repeated.

[0046] (2) At the underground key nodes such as manhole covers, pipeline interfaces and the like, a buried displacement sensor is installed, and combined with LiDAR echo intensity consistency analysis, sensor reliability data is generated. Specifically, the vertical displacement amount of the ground surface is continuously monitored by the buried displacement sensor, which is taken as the reference true value; at the same time, the vertical displacement amount of the LiDAR sensor echo intensity in continuous scanning is obtained, which is a known technology and the specific process will not be repeated; finally, the relative error between the vertical displacement amount of the echo intensity at the current time and the reference true value of the buried sensor at the current time is calculated, and the difference between the natural number 1 and the relative error is taken as the reliability score R of the overall data reliability at the current time.

[0047] (3) For all acquisition times, the terrain complexity data and reliability score data obtained at all historical acquisition times before each acquisition time are extracted, and the same kind of data is sorted in ascending order of time to obtain the terrain complexity sequence and the reliability score sequence of the sensor at each acquisition time.

[0048] Step S3, clustering the fused point cloud data set to obtain each feature cluster, obtaining the horizontal plane projection area of the point cloud data in each feature cluster, and combining the number of point clouds in the feature cluster to construct the effective point number per unit area of each feature cluster; based on each feature cluster at each time, combining a random forest algorithm to obtain a label value of each feature cluster; and based on the effective point number per unit area and the label value, calculating the effective score of the blind area at each time.

[0049] In high-precision geospatial collection of complex scenes, due to sensor view occlusion and physical penetration limit, a large number of key micro-features such as road markings, manhole covers, and pipeline nodes are missed, which seriously reduces the engineering usability of the data. The physical penetration limit in the present application refers to the upper limit of the ability of the sensor to penetrate obstacles to obtain occluded target information due to its physical properties.

[0050] (1) For each collection time, the fused point cloud data set obtained at the collection time is taken as the input of the DBSCAN clustering algorithm. In order to avoid the problem that the point cloud clustering cluster of the micro-feature cannot be identified due to too large a value, in the embodiments of the present application, the neighborhood radius is set to 0.3 and the minimum point number is set to 5. As other embodiments of the present application, the implementer can set the values of the neighborhood radius and the minimum point number according to the actual situation. The algorithm output is each clustering cluster, which is taken as each feature cluster, so as to identify sub-meter spatially aggregated independent objects, such as manhole covers which are dense point clusters with a diameter of 0.6 m, and wall corners which are linear high-density point groups, so as to achieve the effect of deconstructing the micro-topographic structure of the occluded area. The DBSCAN clustering algorithm is a known technology, and the specific process will not be described again.

[0051] It should be noted that for clustering of the fused point cloud data set, only one clustering method is provided in the embodiments of the present application, and there are many existing clustering methods, and the implementer can also use other clustering algorithms to cluster the fused point cloud data set, which is not specifically limited in the present application.

[0052] (2) For each feature cluster, the projection area of the point cloud in the horizontal plane is obtained. Specifically, first, the projection of the point cloud data points in the horizontal plane is obtained, and the projection process is a known technology, and the specific process will not be described again. Then, the projection results of all point cloud data points in the feature cluster are taken as the input of the convex hull algorithm, and the horizontal plane projection area of these point cloud data points is obtained. Further, the ratio of the number of point cloud data points in the feature cluster to the horizontal plane projection area of all point cloud data points in the feature cluster is calculated as the effective point number per unit area of the feature cluster, so as to reflect the laser penetration capability.

[0053] (3) For each feature cluster of each acquisition time in the historical acquisition task, set the label value of each feature cluster of each time through artificial marking, and the label type is valid feature and shielding omission, and the corresponding label values are set to 0 and 1 respectively; in all acquisition times in the historical acquisition task, randomly select 70% of the feature clusters of the acquisition times as a training set, and the remaining 30% of the feature clusters of the acquisition times as a test set, perform random forest training, and obtain a trained random forest model. The random forest model training process is a known technology, and the specific process will not be repeated.

[0054] The valid feature in the present application refers to that the feature cluster is not shielded, and the sensor successfully captures the target; the shielding omission refers to that the data is missing due to shielding, and the key micro feature is not captured.

[0055] Then, all feature clusters of the current time are taken as the input of the trained random forest, wherein the number of decision trees is set to 100 in the embodiment of the present application, the split criterion is the Gini index, the label value of each feature cluster of the current time is output, the valid feature is 0, and the shielding omission is 1, so as to quantify the recognizability of each feature cluster, for example, the manhole cover is marked as 0 due to high metal reflectivity, and the soil under the crown is marked as 1 due to point cloud missing, and the feature validity criterion of the shielding area is established.

[0056] (4) Based on the above analysis, the effective score of the blind area of each acquisition time is calculated, and the expression is:

[0057]

[0058] In the formula, A is the effective score of the blind area of the current time; I is the number of feature clusters of the current time; D i is the effective point number per unit area of the i-th feature cluster of the current time; l i is the label value of the i-th feature cluster of the current time.

[0059] In the DBSCAN clustering, the feature cluster is an independent feature object that is spatially aggregated at a sub-meter level, in high-precision geographic spatial acquisition, the value of D i is a measure of the sensor penetration ability and data acquisition efficiency, the larger the value is, the more the laser successfully penetrates the sparse vegetation or avoids shielding, and high-density point clouds are obtained in the blind area under the crown, at the bottom of the building, etc.; l i is a binary label of the i-th feature cluster, and the value is 0 or 1, indicating the recognizability state of the feature cluster, representing the identification of shielding effect and feature reliability, when the value is 1, it corresponds to the shielding hot spot in the real world, and when the value is 0, it indicates that the sensor information acquisition is complete, l i as a binary switch controls whether D i is included in the calculation, that is, only when the value is 1, D iOnly the problem area of the occlusion missing cluster is taken into account, and the effective feature area is not involved in the calculation.

[0060] The blind area effective score A represents the comprehensive score of the data integrity of the occlusion area, quantifies the penetration ability and data density of the laser sensor in the occlusion area, and thus evaluates the influence of the occlusion on the acquisition integrity of the key micro features: the higher the A value, the more dense data that can be acquired under the occlusion, the better the scene is measurable, and the smaller the data blind area is; otherwise, it reveals serious data loss.

[0061] In step S4, the prediction accuracy of the blind area effective score of each time is calculated through the blind area effective scores of all previous times and the terrain complexity; based on the prediction accuracy and the reliability score of all times, the compensation value of each time is obtained by combining the compensation algorithm.

[0062] Due to the uncertainty of the ground feature integrity index A of the occlusion area (such as dense forest and building dead angle) in complex terrain, the data reliability fluctuates, directly leading to the out-of-control problem of the mapping accuracy of the occlusion area.

[0063] Therefore, the blind area effective scores calculated at all acquisition times before the current time are sorted in ascending order of time to obtain the blind area effective sequence of the current time; the blind area effective sequence and the terrain complexity sequence of the current time are taken as the input of the XGBoost (eXtreme Gradient Boosting) algorithm, wherein the number of decision trees in the algorithm is set to 200 to improve the fitting ability of complex relationships; the maximum depth of the tree is set to 6 to prevent overfitting; and the learning rate is set to 0.05 to balance the convergence speed and accuracy. The output is the prediction accuracy of the blind area effective score of the occlusion area at the current time, which ranges from 0 to 1, thereby quantifying the reliability of the real data restoration of the occluded area and reflecting the penetration ability of the sensor cooperation to the occlusion.

[0064] Subsequently, the prediction accuracies of the occlusion area obtained at all acquisition times before the current time are sorted in ascending order of time to obtain the prediction accuracy sequence of the current time; the prediction accuracy sequence of the current time and the reliability score sequence of the sensor are taken as the input of the weighted linear compensation algorithm, and the dynamic compensation coefficient is set, wherein if the prediction accuracy P of the current time is less than 0.7, the dynamic compensation coefficient is set to 0.3; if the prediction accuracy P of the current time is greater than or equal to 0.7, the dynamic compensation coefficient is set to 0.1, to preferentially protect the low-accuracy area. The output is the compensation value G of the current time, which represents the final reliability score of the fusion device performance. The higher the prediction accuracy P is, the less compensation is needed, and the dynamic compensation coefficient should be lower. As other embodiments of the present application, the implementer can set the dynamic compensation coefficient according to the actual situation.

[0065] Step S5, constructing a shielding compensation confidence of each time point based on the compensation value and the prediction accuracy, and adjusting the motion path of the data acquisition equipment based on the shielding compensation confidence.

[0066] Based on the above analysis, for all data acquisition time points, the shielding compensation confidence of each acquisition time point is calculated, and the expression is:

[0067]

[0068] In the formula, B is the shielding compensation confidence of the current time mapping information; P is the prediction accuracy of the current time; G is the compensation value output by the weighted linear compensation algorithm of the current time; and Norm() is a normalization function.

[0069] P is the prediction accuracy of the shielding area, which is used to quantify the reliability of the sensor system in restoring the real ground features in the shielding environment, and directly reflects the penetration ability and data restoration reliability of the sensor system to the shielding area. The larger the value is, the more effectively the sensor system can penetrate sparse vegetation or building shadows to obtain high-fidelity point cloud and image data, and key micro features are captured completely; G is the compensation value output by the weighted linear compensation algorithm, which represents the compensation score of the reliability of the sensor system. The smaller the value is, the better the performance of the sensor itself is, and the lower the compensation demand is.

[0070] The shielding compensation confidence B quantifies the final reliability and reliability of the data obtained by multi-source sensor cooperation in the limited sensor view angle or physical shielding area, evaluates the ability to penetrate shielding and capture key micro features during geological space acquisition, and the higher the B value is, the smaller the data blind area is, the stronger the scene is, and the more reliable the engineering decision basis is; otherwise, it reveals serious data loss.

[0071] Due to the inherent line-of-sight shielding (such as dense vegetation, building bottoms, steep terrain) in complex geographical environments and the physical penetration limit of sensors, the traditional single-platform acquisition method causes serious data loss (such as ground micro-deformation, underground pipeline nodes, real topography under tree crowns) and a large number of micro-features (such as fine lines, manhole covers) in shielding areas, which directly affects the integrity of the mapping data and the engineering usability, and makes the subsequent analysis or construction have a significant error risk.

[0072] Therefore, a dynamic optimization strategy needs to be designed according to the shielding compensation confidence index B, and the specific dynamic optimization strategy design is as follows:

[0073] The trigger threshold is set, preferably, in the embodiments of the present application, the trigger threshold is set to 0.4. As other embodiments of the present application, the implementer can set the trigger threshold according to the actual situation. When the real-time calculated shielding compensation confidence is lower than the trigger threshold, it indicates that the confidence of the current area data acquisition is insufficient, there is a high risk of shielding omission or distortion, at this time, the DLite dynamic path planning algorithm is triggered immediately, the algorithm generates an optimal supplementary acquisition path online by comprehensively considering the current device position, the B value distribution of all acquisition time points and the known underground key nodes, guides the unmanned aerial vehicle to bypass the shielding object, and carries out targeted rescan along the optimal supplementary acquisition path, so as to maximize the acquisition of direct observation data of the shielded target. Therefore, the reliability of the shielding area data and the capture rate of the micro features are significantly improved.

[0074] The acquisition process of the terrain complexity is shown in the schematic diagram as Figure 2

[0075] Based on the same inventive concept as the above method, the embodiments of the present application also provide a high-precision geospatial information acquisition and mapping system, which comprises a memory, a processor and a computer program stored in the memory and running on the processor, and the processor implements the steps of any one of the methods in the above-mentioned high-precision geospatial information acquisition and mapping method when executing the computer program.

[0076] ​In summary, the embodiment of the present application provides a high-precision geographic spatial information acquisition and mapping method. RGB images and thermal infrared images collected at each time are used to give RGB color values and thermal radiation intensity values to point cloud data collected by a laser sensor at each time, and a fusion point cloud dataset at each time is constructed. In view of the problem of micro-feature omission of a shielding area due to sensor viewing angle and penetration limitation, the terrain complexity at each time is analyzed by three-dimensional distribution characteristics of the point cloud data, each feature cluster is obtained by clustering the fusion point cloud dataset, the shielding omission features of the point cloud data are analyzed in combination with a random forest, an effective score of a blind area at each time is constructed, the effective point number per unit area of the shielding area and the recognizability are quantified, the shielding hotspot area with missing data is accurately positioned, and the blindness of the traditional method for micro-feature omission is avoided. In view of the problem of accuracy fluctuation of the shielding area in a complex terrain and performance difference of a sensor, which leads to a decrease in data reliability, a reliability score of the point cloud data at each time is constructed based on the difference between the vertical displacement of the sensor echo intensity at each time and a preset reference true value, the prediction accuracy of the effective score of the blind area at each time is calculated in combination with the effective score of the blind area and the terrain complexity, a shielding compensation confidence at each time is constructed, the penetration capability of multi-source cooperation on the shielding area and the data restoration reliability are comprehensively reflected, and the interference of the performance fluctuation of the sensor itself on the data quality of the shielding area is eliminated. Based on the shielding compensation confidence at the collection time, a path planning algorithm is triggered to guide the unmanned aerial vehicle to compensate for the low-confidence area around the obstacle, the integrity of the data of the shielding area and the micro-feature capture rate are significantly improved, the blind area of high-precision mapping is reduced, the scene data integrity is improved, and the engineering usability of complex scene data is ensured.

[0077] It should be noted that the above-mentioned embodiments of the present application are in the order of description only, and do not represent the advantages and disadvantages of the embodiments. Moreover, the above-mentioned embodiments of the present application are described. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous.

[0078] Each embodiment in the present application is described in a progressive manner, and the same or similar parts of each embodiment can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0079] The above only describes the preferred embodiments of the present application, and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for collecting and mapping high-precision geospatial information, characterized in that, The method comprises the following steps: For urban streets, RGB images and thermal infrared images collected at each time are used to assign RGB color values and thermal radiation intensity values to point cloud data collected by laser sensors at each time, thereby constructing a fusion point cloud dataset at each time; Grid division is performed on the point cloud data at each time, and terrain parameters of each grid are constructed based on the three-dimensional distribution characteristics of the point cloud data in each grid to determine the terrain complexity at each time; the reliability score of the point cloud data at each time is constructed based on the difference between the vertical displacement of the laser sensor echo intensity at each time and the preset reference true value; Each feature cluster is obtained by clustering the fusion point cloud dataset, the horizontal plane projection area of the point cloud data in each feature cluster is obtained, and the unit area effective point number of each feature cluster is constructed in combination with the number of point cloud data in the feature cluster; based on each feature cluster at each time, the label value of each feature cluster is obtained in combination with the random forest algorithm; the effective score of the blind area at each time is calculated based on the unit area effective point number and the label value. The prediction accuracy of the effective score of the blind area at each time is calculated through the effective score of the blind area at all times before each time and the terrain complexity; the compensation value at each time is obtained based on the prediction accuracy and the reliability score of all times in combination with the compensation algorithm; The shielding compensation confidence at each time is constructed based on the compensation value and the prediction accuracy, and the motion path of the data acquisition equipment is adjusted based on the shielding compensation confidence.

2. The method of claim 1, wherein, The process of obtaining the terrain complexity at each time is as follows: The horizontal plane projected by the point cloud data at each time is divided into a grid of a preset size, the standard deviation of all point cloud elevation values in each grid is taken as a first evaluation value; the coefficient of variation of the point cloud projection point density of each grid and all its adjacent grids is calculated as a second evaluation value; the weighted sum of the first evaluation value and the second evaluation value of each grid is taken as the terrain parameter of each grid, and the mean value of the terrain parameters of all grids at each time is taken as the terrain complexity at each time.

3. The method of claim 1, wherein, The process of obtaining the reliability score of the point cloud data at each time is as follows: The relative error between the vertical displacement of the laser sensor echo intensity at the current time and the preset reference true value is calculated, and the difference between the natural number 1 and the relative error is taken as the reliability score of the point cloud data at the current time.

4. The method of claim 1, wherein, The process of obtaining the unit area effective point number of each feature cluster is as follows: The horizontal plane projection results of all point cloud data points in each feature cluster are taken as the input of the convex hull algorithm to obtain the corresponding horizontal plane projection area; the ratio of the number of point cloud data points in each feature cluster to the horizontal plane projection area is taken as the unit area effective point number of each feature cluster.

5. The method of claim 1, wherein, The process of obtaining the label value of each feature cluster is as follows: A random forest model is trained through all feature clusters at historical times, wherein the label and label value of each feature cluster include: 0 for effective features and 1 for shielding omissions; All feature clusters at the current time are taken as the input of the trained random forest, and the output is the label value of each feature cluster at the current time.

6. The method of claim 1, wherein, The process of obtaining the effective score of the blind area at each time is as follows: The product of the unit area effective point number and the label value of each feature cluster is calculated; the product of all feature clusters at the current time is taken as the effective score of the blind area at the current time.

7. The method of claim 1, wherein the high-precision geospatial information is collected by a plurality of sensors mounted on a plurality of unmanned aerial vehicles (UAVs) and a plurality of ground-based sensors. The acquisition process of the prediction accuracy of the effective score of the blind area at each moment is: a sequence of terrain complexity of all moments before the current moment is taken as the terrain complexity sequence of the current moment; and a sequence of blind area effective scores of all moments before the current moment is taken as the blind area effective sequence of the current moment; the blind area effective sequence and the terrain complexity sequence of the current moment are taken as the input of the XGBoost algorithm, and the output is the prediction accuracy of the effective score of the blind area of the current moment.

8. The method of claim 1, wherein, The acquisition process of the compensation value at each moment is: a sequence of the prediction accuracy of all moments before the current moment is taken as the prediction accuracy sequence of the current moment; and a sequence of reliability scores of all moments before the current moment is taken as the reliability score sequence of the current moment; the prediction accuracy sequence and the reliability score sequence of the current moment are taken as the input of the weighted linear compensation algorithm, and the output is the compensation value of the current moment.

9. The method of claim 1, wherein, The construction of the shielding compensation confidence at each moment is based on the shielding compensation confidence for data acquisition equipment motion path adjustment, and specifically, a normalized value of the ratio of the prediction accuracy and the compensation value of the current moment is taken as the shielding compensation confidence of the current moment; if the shielding compensation confidence of the current moment is lower than a preset trigger threshold, the path planning algorithm is triggered immediately to generate an optimal compensation path of the unmanned aerial vehicle, and the region is re-scanned along the path.

10. A high-precision geospatial information acquisition surveying system comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, The processor executes the computer program to realize the steps of the method of any one of claims 1-9.

Images