摘要：植被叶片与叶片、叶片与茎秆之间经常有遮挡，不能得到完整的测量数据。因此需 要根据已有测量数据，基于同类作物叶片的相似性，对被遮挡部分叶片数据进行修补， 以获得完整的作物三维数字模型描述。利用被测对象（植物叶片）之间的相似性，对部 分被测对象中缺失的点云数据进行填补，得到完整的点云数据，消除被测对象数据缺失 问题。本文针对基于几何特征的植物叶片的点云修补问题进行了深入系统的研究，围绕 点云叶片分割、叶片轮廓提取、轮廓点排序、轮廓匹配、孔洞修补等算法开展研究。针 对叶片点云分割问题，本文提出了基于距离的层次聚类的分割算法，设定距离阈值为 2 对叶片点集进行划分，实现叶片点云的分割。针对点云叶片轮廓的提取问题，本文提出 了一种提取叶片点云轮廓方法，对叶片点云按照 x 轴坐标大小进行排序、分组，提取每 个组里面 x 最大的点记录下来，对 y 轴进行同样的操作。从每组提取出的这些点就是叶 片点云轮廓点。针对点云修补问题，研究了点云轮廓的排序、轮廓匹配、点云叶片旋转、 点云数据平滑处理等算法。通过对填补的点云数据的 z 轴进行优化，使得叶片点云数据 更加平滑，实现叶片点云孔洞的修补。在提出以上算法的基础上，同时采用 MATLAB 编程 实现，验证算法的正确性以及实用型，以实现基于几何特征的点云修补技术。
关键词 点云，点云分割，轮廓提取，孔洞修补
毕 业设计 说明书 外文摘 要
Title Research on Point Cloud Repair Technology Based on Geometric Characteristics
Abstract：Vegetation leaves and leaves, leaves and stems are often blocked between, can not get a complete measurement data. Therefore, it is necessary to repair the occluded part of the leaf data based on the similarity of the similar crop leaves according to the existing measurement data to obtain the complete crop three-dimensional digital model description. Using the similarity between the measured object (plant leaf), the point cloud data missing from the measured object is filled to obtain the complete point cloud data, and the missing data of the measured object is eliminated. In this paper, the problem of point cloud repair of plant leaves based on geometric characteristics is studied deeply. The research on the algorithm of point cloud blade segmentation, leaf contour extraction, contour point sorting, contour matching and hole repair is carried out. Aiming at the problem of leaf point cloud segmentation, this paper proposes a segmentation algorithm based on distance hierarchical clustering, and sets the distance threshold of 2 to pide the leaf point set to realize the segmentation of leaf point cloud. In this paper, we propose a method to extract the leaf point cloud contour method. The point cloud is sorted and sorted according to the size of the x-axis coordinate, and the maximum point of each group is extracted. Perform the same operation. The points extracted from each group are the leaf point cloud point. In order to solve the problem of point cloud repair, the algorithm of point cloud contour, contour matching, point cloud blade rotation and point cloud data smoothing are studied. By optimizing the z-axis of the filled point cloud data, the leaf point cloud data is smoothed and the leaf point cloud hole is patched. On the basis of the above algorithm, we use MATLAB programming to realize the correctness and practicality of the algorithm to realize the point cloud repair technology based on geometric features.
Keywords point cloud, point cloud segmentation, contour extraction, hole repair 基于几何特征的点云修补技术研究:http://www.chuibin.com/wuli/lunwen_206254.html