基于长距离TLS强度数据的潮滩含水量反演研究

Water content inversion for tidal flats is of great significance for the studies of coastal creatures, material transport, etc. However, the existing methods have certain limitations. In the present project, long-range terrestrial laser scanners (TLS) are used to simultaneously acquire the three-dimensional geometrical data and intensity data of the tidal flats. First, the influencing mechanisms and correction methods of various factors on long-range TLS intensity data are studied. Standard diffuse reflection targets and natural homogeneous targets are utilized to quantitatively estimate the relationships between the incidence angle, distance, and intensity data and to establish an overall model for the correction of long-range TLS intensity data. Second, based on the differences of the geometry and laser reflection characteristics between the point clouds of the tidal flat and vegetation, the geometry information and corrected intensity data are combined to filter the vegetation and extract the tidal flat. Then, according to the absorption characteristics of water to the emitted laser of long-range TLS, the relationship between the water content of the tidal flat samples and the corresponding corrected intensity data is analyzed to establish a quantitative function model between the corrected intensity data and the water content. Finally, the corrected intensity data of long-range TLS are used to achieve a rapid, comprehensive, and accurate method for multi-temporal and large-scale water content inversion for tidal flats. The project can offer theoretical bases for exploring the intrinsic relationships between the water content change trend and the dynamic evolution of tidal flats, biological growth habits, etc. In addition, the project can provide technical supports for the studies of resources, hydrology, ecology, environment, and disasters in tidal flats.

潮滩含水量反演对研究海岸生物及物质传输等具有重要意义，现有方法存在一定的局限性。本项目利用长距离地面三维激光扫描仪（TLS）同步获取潮滩三维几何数据和回波强度数据，研究入射角、距离等因素对强度数据的影响机理及改正方法。利用标准漫反射目标和自然同质目标定量求取入射角、距离和强度数据之间的关系，构建长距离TLS强度数据整体改正模型。根据潮滩和植被的几何和激光反射特性差异，结合点云几何数据和改正后强度数据进行植被滤波，提取潮滩点云。根据水对长距离TLS发射激光的吸收特性，分析潮滩样本含水量及其改正后强度数据之间的关系，建立改正后强度数据与潮滩含水量之间的定量函数模型。利用长距离TLS改正后强度数据，实现多时相、大范围潮滩含水量的快速、全面和精确反演，为探析含水量的变化趋势和潮滩地形动态演变、生物生长习性等的内在关联提供理论依据，为潮滩区域的资源、水文、生态、环境、灾害等研究提供技术支撑。

海岸带淤泥质潮滩含水量反演对研究海岸生物及物质传输等具有重要意义，现有野外实地调查及光学遥感解译方法在效率、精度等方面存在较大的局限性。本项目基于长距离地面激光雷达（TLS），解决了淤泥质潮滩含水量反演的技术难题，取得的主要研究成果如下。（1）构建了长距离TLS强度数据入射角和距离效应整体改正多元多次多项式模型，结合室内定标法和室外自然同质目标定标法，解决了长距离TLS强度数据校正参数估计的难题。构建了基于校正后强度数据的指数模型对潮滩含水量进行了精确反演。（2）提出了一种结合强度数据和几何数据的TLS点云非监督分类方法。首先，利用多项式模型对原始强度数据进行校正，根据校正后强度数据对点云进行初步分离；然后，利用点云密度差异，再次进行点云分类；最后，提出了一种连通性聚类方法，根据类大小对点云进行区分。（3）提出了一种新的TLS单站点云非监督分类方法。首先，利用法向量变化率对点云进行初步分离；然后，提出了一种新的方法对密度数据的距离效应进行校正，再次进行点云分类；最后提出了一种自适应连通性聚类方法，根据类的点、线、面特征及内部点云数量对点云进行区分。研究成果实现了大范围潮滩含水量的快速、全面和精确反演，为潮滩区域的资源、水文、生态、环境等研究提供了技术支撑，也为TLS点云数据处理和智能解译提供了新的方法和思路。相关方法在不同类型植被茎叶分类、海岸带垃圾识别、无人机激光雷达海岸带盐沼植被空间结构特征提取等方面也取得了很好的应用，在遥感领域ISPRS JP&RS、IEEE TGRS、JAG等知名期刊上发表论文7篇（其中SCI检索顶刊5篇）。