Research Of The Landslide Recognition Based On LiDAR Technology In The Complex Geological Environment Area

The landslide geological hazards have group-occurring, multiple and sudden features. According to the sorting of damage degree, the landslide has become the second largest geological disaster next only to earthquake, and it is a serious threat to people’s life and property security. Therefore, raising the research level of the landslide recognition has important theoretical and practical significance for the forecast and early warning of geological hazards and disaster prevention and reduction.The Three Gorges area, which is strong dissected topography and heavily forested terrain and has a large number of landslide geological disasters distributing along both sides of the Yangtze River, is typical complex geological environment area. This geological background conditions decide that traditional field investigations have been unable to meet large areas of landslide investigation and monitoring in the Three Gorges Reservoir area. Due to the influence of high vegetation coverage and mountain shadows, the optical remote sensing landslide recognition technology and the interference radar measurement technology are difficult to eliminate the influence of vegetation and mountain shadow and unable to obtain real terrain information of the bare surface. It seriously hampers the landslide recognition application ability of the optical and interference radar technology in the Three Gorges area.Laser radar has certain vegetation penetration ability, can obtain multiple echoes information which can accurately measure the microtopography parameters of the complex geological background areas, and can get more terrain information and the bare surface information that have eliminated the influence of vegetation and mountain shadow. The information has very important significance for landslide research.In the above background, the paper selected the Zigui segment of the Three Gorges area as the study area, and carried out the landslide recognition capability research using airborne LiDAR data as the primary data source and combining with other multi-source geoscience data. It can provide new technical support for the identification, forecast and early warning of the landslide geological hazards in the complex geological background area.1. The airborne LiDAR data processing and analysis of the study areaThe paper formed a set of scientific and efficient airborne LiDAR data processing method in the complex geological background area. (1) Date resolving and calibration of the systematic error:resolve the original data, calibrate the systematic error, and then obtain the laser point cloud data with three-dimensional coordinates of the surface.(2) Check and correction of the systematic error of the flight strip:The point cloud data obtained by the above resolving can still produce remnant systematic error of the flight strip or regional error. It needs to check and correct before classification. Only by this we can filter the point cloud and use the distinguished ground point to generate DEM products.(3) Classification of the laser point cloud data:According to research objectives of the paper, the classification target of the laser point cloud mainly is to distinguish the ground points and the non-ground points. Finally we can use the ground points to generate DEM products.(4) Analysis of the produce results:Classify the produce results and do amount of the fine manual editing to get the final product of the mainstream and tributaries bank slope. The evaluation results of the plane precision and height accuracy showed that the LiDAR data in the accuracy verification area is fully able to meet the nominal accuracy requirements.2. Research of landslide occurrence conditions and distribution law based on LiDAR data in the study areaUsing LiDAR DEM of the study area and generating slope, aspect image, and combining with other data sources such as formation lithology, analyze the relationship of landslide occurrence and space distribution and these factors. The main conclusions are as follows:(1) In the study area landslides distributed mainly in the elevation range from119m to380m. It would not develop landslides when the height is greater than920m.(2) In the study area landslides distributed mainly in the slope range from15°to35°.(3) In the study area landslides occurred mainly in three aspect direction:east, southeast and northeast direction, accounting for71%of the total landslides number and80%of all landslides area.(4) In the study area about84.8%of landslides distributed in the range of less than1200m from the river. The farther away from the water system, the landslide density was lower.3. Recognition technology research of the typical genetic type landslidesIn this paper we selected landslide (landslide group) of different genetic types such as rock landslide and soil landslide with gravel pieces as study objects, and carried out the landslide recognition application potential research of LiDAR technology in the qualitative and quantitative levels. The main conclusions are as follows:(1) When the sun elevation angle is certain, the hillshade graphics with different azimuth angles can intuitively express and display the microtopography morphology of the landslide and non landslide regions. It can provide data support for landslide boundaries defining and main landslide elements recognition.(2) Compared with non-sliding region, the trailing edge of landslide, landslide body and the lateral margin of landslide all have significant difference characteristics in the slope and surface roughness images. It can help to accurately delineate these landslide elements. The curves of semi-variance and fractal dimension of different landslide elements have different characteristics. In the same spatial scales, the semi-variance of the trailing edge of landslide is largest, and landslide body follows. The landslide tongue is minimum. The fractal dimension characteristics are on the contrary, that is, the fractal dimension of the trailing edge of landslide is the smallest, and landslide body follows. The landslide tongue is maximum.Overall, using LiDAR data and the method proposed in this paper, can break through the bottleneck of low recognition accuracy of landslide elements caused by the rough DEM data precision. It can provide data support for the landslide recognition and analysis. It can not only recognize and depict landslide elements, but also can delineate the hidden trouble areas of geological disasters, then make landslide remote sensing research from "viewing literacy" phase to the "qualitative and quantitative combination" stages, and can improve the "digital landslide" research quality and level.