基于地面激光雷达的单株立木几何结构特征参数测量

发布时间：2018-05-31 08:57

本文选题：地面激光雷达 + 单株立木　；参考：《福州大学》2014年硕士论文

【摘要】：森林资源的精准调查是国家制定林业政策、森林资源管理规划的重要基础及参考依据。长期以来,传统的森林资源调查方法耗时费力、工作量大、工作效率低,越来越无法适应精准林业资源管理应用需求。地面激光雷达技术自出现以来,以其高精度、高效率等优越性能被国内外学者广泛地运用于森林资源调查,尤其是单株立木几何特征参数测量方面。基于地面激光雷达可快速、高精度地获取单株立木点云数据及几何特征参数,本文从地面激光雷达获取的单株立木点云数据出发,基于简单曲面和自由曲面两种方法实现了单株立木几何结构建模,并以立木几何建模成果为基础提取单株立木几何特征参数。本文的主要研究内容如下：(1)单株立木点云数据获取及预处理。一颗高度约18m的木棉树为研究目标,针对单株立木几何建模需求,设计了单株立木点云数据获取方案,并对收集的原始点云数据及配准后的点云数据开展了质量评价,针对点云数据量大的特点,有针对性地实施了点云数据抽稀。(2)基于简单曲面及自由曲面方法的单株立木表面几何建模。简单曲面方案首先以Kd-tree数据结构作为点云数据管理、组织方法,应用最短路径方法对单株立木点云数据分段,并计算各分段几何中心,将几何中心所在平面点云数据旋转变换并投影至参照水平面,以最小二乘法拟合二次曲线,如圆、椭圆,建立分段不同的几何模型,如圆柱、圆台等,将各分段连接成完整的单株立木几何模型。自由曲面方案采用Crust算法作为曲面构建方法,该方案以Voronoi和Delaunay三角网为基础,首先对点云集构建Voronoi图,之后将构建的Voronoi图集和原始点云集合并并形成Delaunay三角网,最后保留三个顶点都是采样点的三角形。以木棉树为例,简单曲面方案简明高效,但在不同级别枝条连接时误差较大,与真实木棉树曲面模型有一定误差,与简单曲面方案相比,自由曲面方案能够较为真实的还原木棉树的几何表面形态特征,但基于表面三角网如何提取单株立木几何特征参数仍然具有挑战。(3)单株立木点云数据的几何结构特征参数测量。通过构建曲面模型来估算单株立木的几何结构参数。估算的树高、胸径与野外实测的误差分别为2.1%、1.1%,通过胸径代入经验公式计算的体积与曲面模型估算的误差为-4.58%、-3.55%、2.13%。
[Abstract]:The accurate investigation of forest resources is an important basis and reference for the government to formulate forestry policy and forest resource management planning. For a long time, the traditional forest resources survey methods are time-consuming, heavy workload, low efficiency, and can not meet the needs of precision forestry resource management applications. Since its emergence, ground lidar technology has been widely used in forest resource survey by domestic and foreign scholars for its high precision and high efficiency, especially in the measurement of geometric characteristic parameters of single standing tree. Based on the fact that the ground lidar can obtain the point cloud data and geometric characteristic parameters of a single stand tree quickly and accurately, this paper starts from the point cloud data of a single standing tree obtained by the ground laser radar. Based on simple surface and free-form surface, the geometric structure modeling of single standing tree is realized, and the geometric characteristic parameters of single standing tree are extracted based on the result of geometric modeling. The main contents of this paper are as follows: 1) acquisition and preprocessing of point cloud data of single stand tree. A cotton tree with a height of about 18 m was used as the research object. According to the requirement of geometric modeling of single standing tree, a single tree stand point cloud data acquisition scheme was designed, and the quality evaluation of the collected original point cloud data and the registered point cloud data was carried out. In view of the large amount of point cloud data, point cloud data thinning. 2) based on simple surface and free-form surface method, the geometric modeling of single standing wood surface is implemented. The simple surface scheme firstly uses Kd-tree data structure as point cloud data management, organization method, applies the shortest path method to segment the single stand tree point cloud data, and calculates the geometric center of each segment. The data of point cloud in the plane of geometric center is rotated and projected to the reference plane, and the quadratic curve, such as circle and ellipse, is fitted by least square method, and different geometric models are established, such as cylinder, circle, etc. Each segment is connected into a complete geometric model of single standing tree. The free-form surface scheme uses the Crust algorithm as the surface construction method. The scheme is based on Voronoi and Delaunay triangulation. Firstly, the Voronoi graph is constructed for the point cluster, then the constructed Voronoi graph and the original point cloud are merged to form the Delaunay triangulation. Finally, three vertices are reserved for triangles that are sample points. Taking the kapok tree as an example, the simple surface scheme is simple and efficient, but the error is large when the branches of different grades are connected, and there is a certain error compared with the real kapok tree surface model, compared with the simple surface scheme. The free-form surface scheme can restore the geometric surface features of kapok tree, but how to extract geometric characteristic parameters of single tree based on surface triangulation is still a challenge. 3) the geometric structure characteristic parameters of point cloud data of single tree stand are measured. A curved surface model was constructed to estimate the geometric structure parameters of a single standing tree. The estimated errors of tree height, DBH and field measurements are 2.1 and 1.1.The errors of volume and curved model estimated by the empirical formula of DBH are -4.58and -3.55g / 2.13, respectively.
【学位授予单位】：福州大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN958.98

【相似文献】