东川小江泥石流迹地的多尺度遥感探测试验分析研究

Remote sensing is an important technique to support for geologic hazards’ investigation and monitoring. Debris flow is a common kind of the geologic hazards which is of impact and compulsion to the man-land system safety in mountain area. In macroscopic view，Xiaojiang is selected as the study area for this research project work since its highlighted debris flow geologic hazards. Based on remote sensing and GIS technique integrated utilization, at first the debris flow status in Xiaojiang basin is detected and surveyed by application of satellite remote sensing image technique. Then analysis research will be completed in basin scale centered to the debris flow’s spatial distribution and the different characteristics, as well as drive effect that come from the other environment factors..Under the medium research scale, the typical debris flow valley is selected as an experiment study area for the research on Unmanned Aerial Vehicle remote sensing technique application. In case the UAV-image data is got, based on the debris flow’s valley 3D scenario analysis, the research work can carry out to explore the debris flow’s surface structure feature deeply and reveal its disaster form topographic feature mechanism..Finally, in the microscopic view scale, center point at the accompanied or symbiosis disasters in the typical debris flow valley, multi-temp on-the-spot surveying and investigation by use of 3D laser scanning technique are carried out. Based on the 3D point cloud dada model analysis, try to depict accurately the accompanied or symbiosis disasters’ surface structure characteristics and reveal the disaster change dynamic condition feature..The purpose of this research project is with the view to form the integrated application on “space-sky-earth” multi-scale debris flow theme technique hierarchy framework for Xiaojiang area. This research work is of science experiment significance and has technology application service perspective effect. The outcome will provide science monitoring technique hierarchy framework which being benefit for disaster prevention and reduction management in vast mountain region.

遥感是区域地质灾害调查监测的重要技术手段，泥石流是胁迫广大山区人地系统安全的地质灾害类型。项目以泥石流地质灾害突出的东川小江为研究区，综合利用多尺度遥感技术应用与GIS技术支持，组织面向小江流域的泥石流迹地卫星遥感综合调查，进行区域整体尺度下的泥石流迹地空间分异特性与相关环境驱动作用分析；选择典型泥石流沟谷开展无人机低空遥感摄影探测试验，利用三维场景构建方法深化面向典型泥石流沟谷结构形态特征及灾害形成条件机制的分析研究；针对典型泥石流沟谷中伴生或共生地质灾体，组织进行地面三维激光扫描探测试验，基于3D点云数据处理，探索精准分析地质灾害体特征及其动态过程。研究目的在于搭建形成基于“天地空一体”的面向小江泥石流专题地质灾害探测分析的多尺度遥感技术框架体系，对小江泥石流灾害获得系统认知。项目工作具有科学试验研究意义与科技服务应用前景，可为广大山地区域减灾防灾管理提供科学监测技术方法体系支撑。

东川小江流域发育有区域性的泥石流山地灾害，综合天-空-地多尺度遥感探测技术研究，形成面向山地泥石流灾害的多平台探测技术框架与实证研究，为揭示滇中区域特定山区地质灾害机制机理研究提供技术支撑与方法研究。围绕天-空-地遥感进行泥石流迹地探测研究内容：1）利用开源卫星遥感进行了研究区宏观尺度LUCC提取技术方法试验，完成了地表景观空间分布特性研究；利用GF-2影像完成面向对象分类和卷积神经网络的泥石流迹地信息提取技术对比研究，提出利用多维影像的卷积网络训练特征，实证卷积神经网络方法精准提取泥石流迹地优势；开展了流域整体地表环境脆弱性评价及其空间分异特征制图分析，深化了人地关系空间关联分析及其协调发展探索研究。2）进行了三个时段低空无人机遥感摄影测量数据采集试验与预处理分析，组织不同飞行路径的多方案设计比选优化，提出针对峡谷型泥石流沟谷的立面航线飞行摄影成像技术方法，完成了低空无人机航线规划评价分析方法及其实证对比研究；进行了基于无参考图像质量评价的无人机影像降冗方法研究，利用无人机影像3D建模技术，构建高精度三维场景模型，深化对泥石流迹地中典型泥石流伴生滑坡灾害单体的三维特征精准探测及过程机制分析研究。3）完成不同采集模式下的带状沟谷点云空间配准处理技术要点和精度研究，提出一种基于地形特征精化的TIN 渐进加密点云去噪处理方法；提取两个时段点云数据地表特征，形成了基于点、线、面状特征变化的3D探测分析技术流程，实现多方位变化探测技术研究；面向复杂山地泥石流沟谷迹地GIS数字模拟分析，提出了综合多种算法的点云滤波优化策略，基于特征点和SIFT关键点提取的点云数据精简处理技术研究，以及对不同地形地貌复杂度进行点云曲面构建技术方法。 项目研究取得了相应的理论方法探索与技术应用试验研究结果，公开发表期刊论文42篇，会议论文6篇，专利3项，其中SCI 论文12篇，EI 论文9 篇；培养相关研究方向的博士4人，硕士21人，较好完成了项目研究预期目标。