多回波地面激光扫描点云的叶簇遮挡目标特性建模

The scanning and detecting of foliage covered targets using the small footprint multi-echo terrestrial laser is a cutting edge technique. The key problems are the complicated scattering of targets and the deficiency of local features. Laser radar is able to penetrate the foliage and small footprint multi-echo terrestrial laser has the characteristic of high spatial resolution and high frequency. Therefore, it can record the information of sample points with high density. These characteristics of laser radar provide new solutions for classification and sensing of points cloud targets. To start the research of foliage covered targets' reflection characteristics of laser scanning in 3D space specifically and provide scientific basis for target perception, it is necessary to determine whether there are any target that cannot be identified by visible light. This project intends to introduce small footprint multiple-echo information. And through the study of its echo characteristics, the research will focus on three aspects: the full waveform TLS point cloud data acquisition method, echo waveform analysis based on small footprint target, modeling target characteristics of multiple-echo point cloud data. The research aims to establish a mathematical model and calculation methods of covered target perception in the mass point clouds, and seek breakthroughs of 3D object perception. The research results of this project will provide basic theory and core technology for target detection and recognition.

利用小光斑多次回波扫描成像激光雷达来感知被可穿透性叶簇遮挡的目标是三维点云处理的前沿技术。三维点云数据中被遮挡的目标散射波形背景复杂、单一的局部特征不能够完整描述目标状态是其中重要的科学问题。激光雷达具有穿透树丛的特点，小光斑全波形激光扫描回波具有高空间分辨率和高频率的特点, 能够记录高密度的采样点信息，为解决点云目标分类及感知问题提供了新的思路。为了有针对性地开展对三维空间中被叶簇遮挡目标的激光雷达特性分析工作，为目标感知提供科学依据，确定是否有可见光无法识别的目标是很必要的。本项目拟引入小光斑多回波信息，研究其点云的目标特性，重点针对小光斑目标的全回波波形特性分析、多回波模型的遮挡目标特性建模、基于叶丛稀疏程度穿透的多回波目标可探测性分析三个方面展开研究，建立大规模点云遮挡目标特性的数学模型和计算方法，寻求对三维目标感知的突破，项目研究成果将为目标检测及识别理论提供基础理论和核心技术。

利用小光斑多次回波扫描成像激光雷达来感知被可穿透性叶簇遮挡的目标是三维点云处理的前沿技术。三维点云数据中被遮挡的目标散射波形背景复杂、单一的局部特征不能够完整描述目标状态是其中重要的科学问题。激光雷达具有穿透树丛的特点，小光斑全波形激光扫描回波具有高空间分辨率和高频率的特点, 能够记录高密度的采样点信息，为解决点云目标分类及感知问题提供了新的思路。为了有针对性地开展对三维空间中被叶簇遮挡目标的激光雷达特性分析工作，为目标感知提供科学依据，确定是否有可见光无法识别的目标是很必要的。本项目引入小光斑多回波信息，研究其点云的目标特性，重点针对小光斑目标的全回波波形特性分析、多回波模型的遮挡目标特性建模、基于叶丛稀疏程度穿透的多回波目标可探测性分析三个方面展开研究，建立大规模点云遮挡目标特性的数学模型和计算方法，对三维目标感知进行突破，项目研究成果为目标检测及识别理论提供基础理论和核心技术。本项目所采用的数据为RIEGL VZ1000针对森林及城市树木的数据。因RIEGL的激光雷达扫描仪具有多回波的特点，仪器相对昂贵，但是其多回波的特征为解决森林参数提取以及被遮挡目标检测提供了解决方案。