【作者】 孙涛；

【导师】 董秀军；

【作者基本信息】 成都理工大学 ， 地质工程， 2021， 硕士

【摘要】 我国西南地区山高坡陡,地形地质条件复杂,新构造运动强烈,灾难性滑坡时有发生,由于这些滑坡往往地处高位,隐蔽性强,卫星光学等常规技术手段很难进行早期识别。近年来,可有效剔除植被影响获取真实地面信息的高精度机载LiDAR遥感测量技术的迅猛发展,为这类高位、高隐蔽性滑坡隐患识别提供了新的方法。但机载LiDAR直接获取的DEM数据地形可读性较差,需利用地形可视化技术提高滑坡地形特征的判释能力。目前常规可视化方法基本存在滑坡特征信息遗漏、误判等问题,对于一种可提高滑坡识别效果的增强显示方法的研究一直是学界研究热点。因此滑坡隐患机载LiDAR增强显示与识别方法研究对有助于提高滑坡识别能力。本文以大量机载LiDAR数据为基础,通过针对于滑坡灾害的机载LiDAR采样密度、综合增强显示方法、三维场景搭建、三维图谱构建研究,以多源数据融合为指导思想,建立了一套适用于滑坡识别的三维增强显示方法。研究成果既提高了滑坡的识别效果,又具有一定的快捷性与可行性,研究成果主要包括:(1)提出了适用于滑坡识别的采样密度。以现行的技术标准为基础,选取了植被郁闭度与原始点云密度为实验对象,采用了多层次分析的方法,论述了不同郁闭度与原始点云密度影响程度,总结并提出了适用于滑坡识别的采样密度。(2)分析比较了现有比较常用的几种地形可视化技术,针对滑坡的识别重在提取地形变形破坏迹象的特点,提出了采用基于图像融合的综合增强显示方法,它可以增强地形特征的识别度,最大限度地保留机载LiDAR获取的微小地形信息,并利用多源数据搭建了三维增强显示场景,为高效率、多角度与交互式识别滑坡提供了新的途径。(3)构建了西南山区滑坡典型识别标志三维图谱。选取了代表性强、易于识别的边界、后缘等滑坡基本发育特征要素作为识别标志,并对识别标志的增强显示与光学特征特征进行了归纳,针对不同识别标志构建了包含识识别标志名称、三维光学模型、三维增强显示模型、立体结构图解与文字描述五个部分的滑坡识别标志三维图谱。(4)以丹巴县城及周边区域进行了具体应用。通过机载LiDAR增强显示方法对应用区的滑坡隐患点进行了快速的识别定位,采用地面核查情况、已有入库滑坡数据与卫星In SAR形变监测成果对识别结果进行了验证分析,识别出了47处滑坡且其中有9处是新发现滑坡隐患,识别结果与实际状况达到了较好的一致性。更多还原

【Abstract】 Southwest mountainous areas have steep mountain Gradients,complex terrain and geological conditions,and strong neotectonic movements.Catastrophic landslide events have occurred at intervals,and it is difficult to identify these landslides early by conventional technical means such as satellite optics due to their high position and strong concealment.In recent years,the rapid development of high-precision airborne LiDAR remote sensing technology,which can effectively eliminate the influence of vegetation to obtain real ground information,has provided the possibility for early identification of such high-level and high-concealed landslide hazards.However,the terrain readability of DEM data directly obtained by airborne LiDAR is not very well,so terrain visualization technology should be used to improve the interpretation ability of landslide terrain characteristics.At present,the conventional visualization methods basically have the problems of missing and misjudging landslide characteristic information,and the research on an enhanced visualization method which can improve the recognition effect of landslide has been a hot topic in the academic circle.Therefore,the research on airborne LiDAR enhanced visualization and identification method for potential landslide hazards is helpful to improve the ability of landslide identification.Based on a large number of airborne LiDAR data,this paper establishes a set of three-dimensional enhanced visualization methods for landslide identification by means of airborne LiDAR sampling density,comprehensive enhanced visualization method,three-dimensional scene creation and three-dimensional map construction research for landslide disasters,taking multi-source data fusion as the guiding principle.The research results not only improve the identification effect of landslide,but also have certain shortcut and feasibility.The research results mainly include:(1)The sampling density suitable for landslide identification is proposed.Based on the current technical standards,the vegetation canopy density and original point cloud density are selected as experimental objects,the multi-level analysis method is adopted,the influence degree of different canopy density and original point cloud density are discussed,and the sampling density suitable for landslide identification is summarized and proposed.(2)Several commonly used terrain visualization techniques are analyzed and compared.Aiming at the characteristics of landslide identification,which focuses on extracting landslide deformation and destruction signs,a comprehensive enhanced visualization method based on image fusion is proposed,which can enhance the recognition degree of terrain features and retain mini-terrain information obtained by airborne LiDAR to the maximum.Three-dimensional enhanced visualization scene is created using multi-source data,which provides a new way for landslide identification with high efficiency,multi-angle and interaction.(3)Three-dimensional map of landslide identification markers in southwest mountainous area was constructed.The basic development characteristics of landslide,such as representative strong and easily identifiable boundary and trailing edge,are selected as identification marks,and the enhanced display and optical characteristics of identification marks are summarized.For different identification marks,three-dimensional maps of landslide identification marks are constructed,which include five parts: identification keys name,three-dimensional optical model,three-dimensional enhanced display model,three-dimensional structure illustration and text description.(4)The application has carried in Danba County Town and surrounding areas.Through the airborne LiDAR enhanced visualization method,the potential landslide hazards in the application area are identified and positioned quickly and accurately.The identification results are verified and analyzed by ground check,existing landslide data and satellite In SAR deformation monitoring results.The recognition has identified 47 landslides and 9 of them were newly discovered potential landslide.The identification results were in good agreement with the actual situation.更多还原