基于小型无人机LiDAR与光学遥感信息融合的水库滑坡变形监测及时空规律研究

Colluvial landslides under the effect of reservoir water level fluctuation and rainfall often behave the characteristics of intermittent movement and nonuniform deformation. The synthetic analysis of temporal-spatial deformation law is the key guarantee of accurate landslide warning. Traditional GPS and total station can only acquire a small amount of point data, while the spatial and temporal resolution of InSAR technology is relatively low. Therefore, based on the micro-UAV platform equipped with LiDAR and optical sensor, the deformation monitoring method of landslide surface is to be built, which is characterized with monocular visual assistant, linear constraints of landslide cracks. Then, the monitoring method is to be valuated and improved by comparison with landslide geometry model and the field monitoring data. Finally, based on the surface displacement field, GIS and spatial data mining technology will be used to analyze the deformation laws (magnitude, rate, and vector angle) of different parts of the landslide surface during different activity stages under the effect of reservoir water level fluctuation and rainfall. This study can realize low cost, high resolution and high frequency investigation of reservoir landslide, which is an important supplement to traditional means and InSAR. It can promote the intersection fusion of remote sensing technology and landslide disaster prevention and mitigation, and has important scientific significance and application value for such landslide warning and prevention.

水库堆积层滑坡在库水位波动与降雨作用下多呈现间歇性活动和非均匀变形特征。综合分析滑坡时空变形规律是准确预警预报的重要保证。传统GPS和全站仪只能实现少量“点”数据采集；InSAR技术“面”监测的时空分辨率相对较低。为此，本项目基于小型无人机遥感平台，融合机载LiDAR和光学影像的数据优势，构建一套单目视觉辅助下、以滑坡地表裂缝线性约束为主、基于无人机LiDAR的滑坡地表变形监测方法；联合滑坡几何模型和现场监测数据，验证和完善形变测量理论和方法；基于无人机“面”监测结果，采用GIS和空间数据挖掘方法，分析不同活动阶段滑坡地表不同部位变形特征（变形量、变形速率、变形矢量角）受库水位波动和降雨的时空影响规律。本项目可实现低成本、高分辨率、高频次的水库滑坡监测，是传统手段和InSAR的重要补充，可促进遥感技术与滑坡灾害防灾减灾的交叉融合，对此类滑坡预警预报及防治工程设计具有重要科学意义和应用价值。

水库滑坡备受国内外学者关注，许多水利工程的重大事故都与滑坡灾害有关。综合分析滑坡时空变形规律是准确预警预报的重要保证。传统GPS和全站仪只能实现少量“点”数据采集；InSAR技术“面”监测的时空分辨率相对较低。本项目以库岸堆积层滑坡为研究对象，以现场无人机调查、室内模型试验及理论分析为主要研究手段，开展了水库滑坡变形面监测及时空规律研究。本项目主要研究内容和成果包括：基于小型无人机遥感平台开展了三峡库区典型滑坡灾害调查，建立了基于立体视觉算法的滑坡地形三维重建方法；基于典型库区滑坡灾害调查结果，开发了适用于水库滑坡变形特征模拟的室内模型试验系统，并基于点监测和面监测数据分析了不同库水位高度和降水速率下滑坡地表变形规律；基于三维激光扫描的面监测数据，采用GIS和空间数据挖掘方法，进一步探究了不同活动阶段滑坡地表不同部位变形特征受库水位波动的时空影响规律。本项目研究成果可作为传统地面监测手段和InSAR技术的重要补充，可为水库滑坡预警预报及防治工程合理布设提供参考依据。