基于激光雷达点云数据的树冠结构三维空间异质性分析

发布时间：2019-03-03 12:36

【摘要】：随着科学技术的发展,激光产品已成为人们观察客观世界的”眼睛”和”神经”。目前激光雷达遥感技术在数字化信息获取方面有着重要的地位,激光雷达离散点云数据具有三维坐标信息,对于地物研究能表达其在空间中的细节信息,因而在许多领域有着广泛的应用。树冠结构是林木进行光合作用等生态学过程不可或缺的重要组成部分,研究树冠的空间结构有着重要的意义。不同树种或同一树种之间树冠的空间结构分布因外界因素和本身特性的影响会存在一定的差异性,这种差异性就是空间异质性。分析空间差异性的方法很多,因树木一般具有分形特征,最初使用的是分形维数法,但由于分形理论的始祖Mandelbrot发现分形维数相同的对象间,空间结构表现形式会存在明显的差异性,仅用分形维数这一指标不足以描述空间分布的差异性,因此需要其它指标(如空隙度指数)来区分不同物体。由于林冠的重要性,国内外学者进行了大量的研究,关于林冠方面的研究成果越来越多,但至今仍没有一个利用激光雷达点云数据优势研究树冠三维空间结构的相关报道,大多数研究都是基于光学影像或研究区的某片林分提出的。在上述背景下,文章介绍了空隙度指数的定义和三维滑动盒算法。同时,以激光雷达点云数据为数据源,提出了一种基于体素和三维凸包的空隙度指数的改进方法。本文主要对单个树冠空隙度指数的算法及其在林业研究中的应用进行了初步的研究和改进。研究所开展的工作有： (1)对国内外近些年来树冠结构的研究现状进行了全面的综述,并分析了不同数据源和方法的主要应用领域和存在的主要问题； (2)详细介绍了激光雷达点云数据的基本形式和优缺点,以及作为本文研究树冠结构数据源的优势； (3)详细阐述了空隙度指数的基本定义和原始的计算方法,并在其基础上做了改进。同时,用改进的方法计算点云树冠的三维空隙度指数,继而根据实验果分析了该方法的有效性； (4)利用三维形状方法分别检测实验中各点云树冠的三维形状,判断树冠间的形状表面差异性,从而与本文中提出的空隙度指数分析法进行比较； (5)阐述了空隙度指数在树冠空间异质性分析研究中的作用,并对其应用范围和前景作了展望。文中利用实测的树冠点云数据检验了改进算法应用在树冠空间结构上的有效性。实验结果表明,激光雷达离散点云数据作为研究树冠三维空间结构分布的数据源是有效的,为研究不同树种树冠的空间分布异同与生态变化过程在森林经理方面的研究提供数据支持,同时提供了一种区分不同树种的新方法。
[Abstract]:With the development of science and technology, laser products have become "eyes" and "nerves" to observe the objective world. At present, lidar remote sensing technology plays an important role in digital information acquisition. Lidar discrete point cloud data has three-dimensional coordinate information, and it can express the detailed information in space for the research of ground objects. Therefore, it has a wide range of applications in many fields. Canopy structure is an indispensable part of forest photosynthesis and other ecological processes. It is of great significance to study the spatial structure of tree crown. The spatial distribution of canopy between different tree species or the same tree species will be different due to the influence of external factors and its own characteristics, and this difference is spatial heterogeneity. There are many methods to analyze spatial differences, because trees generally have fractal characteristics, and the fractal dimension method was used at first, but because Mandelbrot, the ancestor of fractal theory, found that the fractal dimension is the same among objects. There are obvious differences in the representation of spatial structure. The fractal dimension is not enough to describe the difference of spatial distribution, so other indexes (such as void index) are needed to distinguish different objects. Because of the importance of canopy, scholars at home and abroad have done a lot of research, and more research results have been made on canopy, but there is still no relevant report to study 3D spatial structure of tree crown by using the advantage of point cloud data of lidar. Most of the studies are based on optical images or a stand in the study area. Under the above background, the definition of void index and the algorithm of three-dimensional sliding box are introduced in this paper. At the same time, an improved method of voidage index based on voxel and 3-D convex hull is proposed based on lidar point cloud data. This paper mainly studies and improves the algorithm of single tree crown void index and its application in forestry research. The research work is as follows: (1) the research status of crown structure at home and abroad in recent years is comprehensively reviewed, and the main application fields and main problems of different data sources and methods are analyzed; (2) the basic forms, advantages and disadvantages of Lidar point cloud data are introduced in detail, and the advantages of Lidar point cloud data as the data source of tree crown structure are introduced in detail. (3) the basic definition of void index and the original calculation method are described in detail, and the improvement is made on the basis of it. At the same time, the improved method is used to calculate the three-dimensional void index of the crown of point cloud, and then the validity of the method is analyzed according to the experimental results. (4) using the three-dimensional shape method to detect the three-dimensional shape of the crown of each point cloud in the experiment, and to judge the difference of the shape surface between the crowns, so as to compare it with the void index analysis method proposed in this paper. (5) the role of void index in spatial heterogeneity analysis of tree crown is expounded, and its application scope and prospect are prospected. In this paper, the validity of the improved algorithm in tree crown spatial structure is verified by using the measured tree crown point cloud data. The experimental results show that the Lidar discrete point cloud data can be used as a data source to study the three-dimensional spatial distribution of tree crown. This paper provides data support for the study of spatial distribution and ecological change process of different tree species in forest management, and provides a new method to distinguish different tree species.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN958.98

【参考文献】