LiDAR后向散射回波强度的辐射校正及应用研究

LiDAR systems can record the backscattered echo, i.e., the laser intensity, from the scanned target. Laser intensity demonstrates the reflectivity ability of the target with respect to the emitted laser. Utilizing the intensity data can compensate for the deficiency of single geometrical data and facilitate the extraction of target geometric shape, surface property, internal structure, change of movement and characteristic parameter. However, the intensity value is affected by multiple variables, such as the instrumental effects, atmospheric effects, target scattering characteristics, and scanning geometry, which results in the phenomenon that different parts of a homogeneous surface are represented as different values or different surfaces are represented as the same values in the measured intensity data. Therefore, the original intensity data should be corrected, i.e., radiometric calibration. This proposed project aims to utilize the intensity data to extract target properties, such as reflectance, grain size, roughness, and moisture, by analyzing, modeling, and quantitatively correcting the influencing factors, which can provide a data and theory base for the visualization, registration, and classification of point cloud, target feature extraction, and the fusion of point cloud and image data. This project belongs to the theoretical field of LiDAR. The research results are expected to have a wide range of applications in tunnel deformation and leakage monitoring, target identification, melting glaciers detection, and disaster assessment.

LiDAR通过光电接收系统还能记录目标对发射激光的“后向散射回波强度”，也称为“激光强度”。激光强度表征目标对激光的反射光谱特性，利用强度数据可弥补单一几何数据的不足，对目标几何形状、表面特性、内部结构、运动和特征参量等进行提取和反演。但是，激光强度受扫描仪系统特性、大气衰减、目标特性、扫描环境等多种因素的影响，导致“同物异谱”、“异物同谱”现象，因此，需要对影响强度数据的各种因素进行改正，即“辐射校正”。本课题旨在利用LiDAR的激光强度信息，对其影响因素进行分析、建模及定量改正，从改正后强度数据中提取目标特征，如反射率、纹理、粗糙度、含水量等，为点云可视化、配准与分类，目标特征精细提取，点云与影像数据的融合等提供理论与数据基础。本课题属于LiDAR领域的基础理论研究，预期的研究成果可拓展LiDAR在隧道变形与渗水检测、地物识别、冰川融化检测、灾害评估等多个领域的应用。

除获取目标的三维空间几何信息外，LiDAR系统还能同步记录一个表征目标对发射激光反射能力的值，即激光强度值。理论上，不同目标表面由于对特定激光波长的反射率不同，获取的强度值存在差异，因此激光强度是反映目标表面特性及提取目标特征的重要物理量。利用强度数据可直接、精确、快速地对扫描目标结构、材质、含水量等进行提取和反演，弥补传统方法利用单一几何数据的缺陷。由于激光强度受到扫描仪特性、大气传输特性、目标表面特性、扫描几何构造等众多因素的影响，因此并不能从原始激光强度数据中直接获取目标表面特性。在利用强度数据进行点云分类和目标特征提取等应用之前，需要对其进行改正，得到仅与目标表面特性有关的改正后强度数据。针对强度数据在LiDAR领域中应用的巨大潜力，本课题围绕地基LiDAR强度数据的改正与应用展开研究：分析了入射角与距离效应的影响，提出了多项式、基于参考目标及基于Oren-Nayar和参考目标的三种改正模型；研究了仪器与环境效应所致的镜面高光现象，提出了经验模型和Phong模型两种方法对其进行消除；拓展了改正后强度数据在目标提取与点云分类方面的应用，将改正后的激光强度值成功应用于破损区域检测、快速边缘提取、室外点云分类、潮滩含水量反演等多个领域中。