CN-122015767-A - Unmanned aerial vehicle refined mapping method for digital elevation of natural ice and snow surface

CN122015767ACN 122015767 ACN122015767 ACN 122015767ACN-122015767-A

Abstract

The invention relates to the technical field of ice and snow mapping, and discloses an unmanned aerial vehicle fine mapping method for digital elevation of a natural ice and snow surface. The method comprises the steps of deconstructing a mapping area into an airspace flight layer, an ice and snow coverage layer and a bedrock topography layer, obtaining unmanned aerial vehicle flight attitude parameters and illumination intensity distribution data of the airspace flight layer, collecting multispectral image sequences and laser point cloud data sets of the ice and snow coverage layer according to the flight attitude parameters and the illumination intensity distribution data, carrying out multi-resolution self-adaptive decomposition on the laser point cloud data sets to generate ice and snow layer point cloud data component sets, extracting ice and rock boundary spectral characteristics according to the multispectral image sequences, predicting a reference elevation model based on the characteristics and the bedrock topography layer historical data, calculating Gao Chengcan difference values of the ice and snow layer point cloud data component sets and the reference elevation model, identifying and removing abnormal point cloud data according to preset topography abnormal criteria, reconstructing an ice and snow surface point cloud topological structure, and generating a digital elevation surface model according to the multispectral image sequences, and realizing high-precision and fine mapping of a natural ice and snow surface.

Inventors

TIAN BO
CHENG XUYU
SUN PENG
LI SILI
JIANG WEIYE
MAO XIAOHONG

Assignees

交通运输部公路科学研究所
中国极地研究中心(中国极地研究所)
清华大学苏州汽车研究院（吴江）

Dates

Publication Date: 20260512
Application Date: 20260227

Claims (10)

1. The unmanned aerial vehicle refined mapping method for the digital elevation of the natural ice and snow surface is characterized by comprising the following steps of: Deconstructing the mapping area into three observation levels, wherein the observation levels comprise an airspace flight layer, an ice-snow covering layer and a bedrock topography layer; Acquiring unmanned aerial vehicle flight attitude parameters and illumination intensity distribution data of the airspace flight layer; acquiring a multispectral image sequence and a laser point cloud data set of the ice and snow cover layer according to the unmanned aerial vehicle flight attitude parameters and the illumination intensity distribution data; Performing multi-resolution self-adaptive decomposition on the laser point cloud data set to generate an ice and snow layer point cloud data component set; extracting the ice rock boundary spectral characteristics according to the multispectral image sequence and the ice and snow layer point cloud data component set; Predicting a reference elevation model of the bedrock topographic layer based on the characteristics of the iced rock interface spectrum and the historical data of the bedrock topographic layer; Calculating Gao Chengcan difference values between the ice and snow layer point cloud data component sets and the reference elevation model; Identifying abnormal point cloud data according to the Gao Chengcan difference values and a preset topographic anomaly criterion; Removing the abnormal point cloud data and reconstructing a point cloud topological structure on the ice and snow surface; and generating a digital elevation surface model according to the reconstructed ice and snow surface point cloud topological structure and the multispectral image sequence.
2. The unmanned aerial vehicle fine mapping method of the digital elevation of the natural ice and snow surface according to claim 1, wherein the step of obtaining unmanned aerial vehicle flight attitude parameters and illumination intensity distribution data of the airspace flight layer comprises the following steps: acquiring real-time pitch angle data, roll angle data, course angle data and positioning data of the unmanned aerial vehicle; synchronously acquiring solar altitude angle data, atmospheric transmittance data and cloud layer reflectivity data; And generating illumination intensity distribution data by adopting the real-time pitch angle data, roll angle data, course angle data, positioning data, solar altitude angle data, atmospheric transmittance data and cloud layer reflectivity data.
3. The unmanned aerial vehicle refined mapping method of the digital elevation of the natural ice and snow surface according to claim 1, wherein the step of performing multi-resolution self-adaptive decomposition on the laser point cloud data set to generate an ice and snow layer point cloud data component set comprises the following steps: acquiring point density distribution characteristics and echo intensity distribution characteristics of the laser point cloud data set; Determining an optimal decomposition scale parameter based on the point density distribution characteristics and the echo intensity distribution characteristics; Performing multi-resolution decomposition on the laser point cloud data set according to the optimal decomposition scale parameters by adopting a spatial spectrum decomposition algorithm; and generating an ice and snow layer point cloud data component set containing different spatial frequency characteristics.
4. The unmanned aerial vehicle refined mapping method of the digital elevation of the natural ice and snow surface according to claim 1, wherein the step of extracting the characteristics of the ice and rock boundary spectrum comprises the following steps: performing snow cover reflectivity correction processing on the multispectral image sequence; acquiring characteristics of an iceberg demarcation curve of the corrected multispectral image sequence in a near infrared band; extracting gradient mutation position data and reflectivity transition data of the characteristics of the rock demarcation curve; And merging the gradient mutation position data and the reflectivity transition data to generate the rock boundary spectral characteristics.
5. The unmanned aerial vehicle refined mapping method of the digital elevation of the natural ice and snow surface of claim 4, wherein the step of predicting the reference elevation model of the bedrock topography layer comprises: obtaining geological structure type data and lithology distribution data of a historical bedrock topography layer; Inputting the rock boundary spectral characteristics and geological structure type data and lithology distribution data into a terrain to generate a network model; and outputting the predicted value of the reference elevation model of the bedrock topographic layer.
6. The unmanned aerial vehicle refined mapping method of the digital elevation of the natural ice and snow surface according to claim 1, wherein the step of calculating Gao Chengcan difference values of the ice and snow layer point cloud data component set and a reference elevation model comprises the following steps: projecting the ice and snow layer point cloud data component set to a reference elevation model coordinate system; Calculating elevation difference values of positions of each point of the ice and snow layer point cloud data component set corresponding to the reference elevation model point by point; and generating an elevation residual value distribution matrix according to the elevation difference value.
7. The unmanned aerial vehicle fine mapping method of digital elevation of natural ice and snow surface of claim 6, wherein the step of identifying outlier cloud data comprises: Obtaining local variance characteristics and space autocorrelation characteristics of the elevation residual value distribution matrix; Judging whether the local variance characteristics and the spatial autocorrelation characteristics meet a preset terrain abnormality criterion or not; and if the preset topographic anomaly criterion is met, marking the corresponding point cloud data as the abnormal point cloud data.
8. The unmanned aerial vehicle refined mapping method of the digital elevation of the natural ice and snow surface according to claim 7, wherein the step of eliminating the abnormal point cloud data and reconstructing the ice and snow surface point cloud topology comprises the steps of: Establishing an ice and snow layer point cloud data space index after abnormal point cloud data are removed; filling elevation data of the position of the missing point cloud by adopting an inverse distance weighted interpolation algorithm; and performing Deloney triangulation processing on the filled ice and snow layer point cloud data to generate an ice and snow surface point cloud topological structure.
9. The unmanned aerial vehicle refined mapping method of the digital elevation of the natural ice and snow surface according to claim 1, wherein the step of generating the digital elevation surface model according to the reconstructed ice and snow surface point cloud topology and the multispectral image sequence comprises the following steps: performing Kriging space interpolation processing on the reconstructed ice and snow surface point cloud topological structure; generating rasterized elevation data by fusing texture feature data of the multispectral image sequence; Optimizing local topography relief features of the rasterized elevation data by adopting a multi-fractal filtering algorithm; And outputting the digital elevation surface model of the natural ice and snow surface.
10. The unmanned aerial vehicle fine mapping method of the digital elevation of the natural ice and snow surface according to claim 9, further comprising: when the void ratio of the point cloud topological structure on the ice and snow surface exceeds a preset threshold value, re-triggering and collecting a multispectral image sequence and a laser point cloud data set of the ice and snow cover layer; And updating the digital elevation surface model according to the newly acquired multispectral image sequence and the laser point cloud data set.

Description

Unmanned aerial vehicle refined mapping method for digital elevation of natural ice and snow surface Technical Field The invention relates to the technical field of ice and snow mapping, in particular to an unmanned aerial vehicle fine mapping method for digital elevation of a natural ice and snow surface. Background In the works such as the investigation of ice and snow resources in polar regions, mountains and the like, the dynamic monitoring of glaciers, the engineering construction of cold regions and the like, the acquisition of accurate digital elevation information of the ice and snow surfaces is very important. At present, technical means for mapping ice and snow surfaces mainly comprise satellite remote sensing mapping, ground artificial mapping, traditional unmanned aerial vehicle mapping and the like. The satellite remote sensing mapping technology can realize coverage observation of a large-scale ice and snow area, but is limited by satellite imaging resolution and revisit period, fine elevation data of the ice and snow surface in a small scale range are difficult to obtain, and in areas with more cloud cover, the data acquisition quality is easily affected, so that the high-precision monitoring requirement on local ice and snow topography change cannot be met. The ground manual mapping technology can obtain high-precision elevation data in a specific area, but the ice and snow covered area is often severe in environment, complex in terrain and difficult to reach by personnel, so that the operation efficiency is low, a large safety risk exists, and meanwhile, the rapid mapping of a large-area ice and snow area cannot be realized, so that the requirement of large-scale ice and snow resource investigation is difficult to adapt. Traditional unmanned aerial vehicle survey and drawing technique has compensatied satellite remote sensing and ground manual survey and drawing's not enough to a certain extent, can realize the ice and snow regional survey of mid-range, but still has a great deal of problem in practical application. Traditional unmanned aerial vehicle survey and drawing only focuses on single flight observation layer face generally, and the correlation between ice and snow overburden and the bedrock topography layer below can not be fully considered, so that accuracy is insufficient when handling problems such as ice and snow layer thickness change, ice and rock juncture area discernment. In addition, after data are collected, the traditional method mostly adopts a unified fixed resolution decomposition mode for processing laser point cloud data, self-adaptive adjustment cannot be carried out according to the topographic features of different areas on the ice and snow surface, the condition that point cloud data are redundant or key information is lost easily occurs, and the generation precision of a subsequent digital elevation model is affected. Meanwhile, in the denoising process of the point cloud data, the traditional method mostly adopts a simple threshold filtering mode, so that abnormal point cloud data generated by factors such as reflection of the ice and snow surface, abrupt change of the terrain and the like are difficult to effectively identify and reject, the reliability of a final digital elevation surface model is further reduced, and the actual requirements of fine and high-precision mapping on the natural ice and snow surface cannot be met. Disclosure of Invention The invention aims to provide an unmanned aerial vehicle refined mapping method for digital elevation of a natural ice and snow surface, which aims to solve the problems in the background technology. In order to achieve the above purpose, the invention provides an unmanned aerial vehicle fine mapping method for digital elevation of a natural ice and snow surface, which comprises the following steps: Deconstructing the mapping area into three observation levels, wherein the observation levels comprise an airspace flight layer, an ice-snow covering layer and a bedrock topography layer; Acquiring unmanned aerial vehicle flight attitude parameters and illumination intensity distribution data of the airspace flight layer; acquiring a multispectral image sequence and a laser point cloud data set of the ice and snow cover layer according to the unmanned aerial vehicle flight attitude parameters and the illumination intensity distribution data; Performing multi-resolution self-adaptive decomposition on the laser point cloud data set to generate an ice and snow layer point cloud data component set; extracting the ice rock boundary spectral characteristics according to the multispectral image sequence and the ice and snow layer point cloud data component set; Predicting a reference elevation model of the bedrock topographic layer based on the characteristics of the iced rock interface spectrum and the historical data of the bedrock topographic layer; Calculating Gao Chengcan difference values between the ice and snow layer