基于投影变换的窗户提取方法  

Window Extraction Based on Projection Transformation

作　　者：汪晓楚 梁栋栋 吴旭[3] WANG Xiao-chu;LIANG Dong-dong;WU Xu(School of Geography and Tourism, Anhui Normal University, Wuhu 241003,China;Geographic Big Data Research Center, Anhui Normal University, Wuhu 241003,China;School of Computer and Information, Anhui Normal University, Wuhu 241003, China)

机构地区：[1]安徽师范大学地理与旅游学院,安徽芜湖241003 [2]安徽师范大学地理大数据研究中心,安徽芜湖241003 [3]安徽师范大学计算机与信息学院,安徽芜湖241003

出　　处：《电脑知识与技术》2019年第2期257-260,共4页Computer Knowledge and Technology

基　　金：安徽省智慧城市与地理国情监测重点实验室2016年度开放性基金重点课题

摘　　要：现代建筑物融合了越来越多的曲面元素,提高了窗户等细节特征的提取难度。本文提出了一种基于投影变换的曲面窗户的提取方法,本方法首先对点云曲面进行旋转变换和距离平移以选取简单的投影平面,将三维点云投影为平面点云,再利用alpha shapes算法提取点云边界点后,将边界点投影逆变换为曲面点云,最后基于特征提取方法进行分类进而提取出窗户边界点。本文以安徽师范大学花津校区图书馆圆形报告厅为例进行算法验证,证实了该算法的有效性。提出了一种基于投影变换的窗户提取算法,实现了曲面点云的窗户提取。Modern buildings integrate more and more cambered elements, which make the extraction of windows and other features more difficult. Therefore, this paper proposes a method of cambered window extraction based on projection transformation. First, the point cloud surface is transformed by rotation and distance translation to select a simpler projection plane, and the 3D point cloud is projected as a planar point cloud. Then, after that the alpha shapes algorithm extracts the boundary points of the point cloud, the boundary points are inversely projected into the cambered point cloud. Finally, it classifies and extracts the window boundary points based on the method of feature extraction. In this paper, we use the circular auditorium of Huajin Campus Library of Anhui Normal University as an example to test the algorithm, which proves the validity of the algorithm. A window extraction al. gorithm based on projection transformation is proposed, which realizes the window extraction of point cloud surface.

关 键 词：点云曲面 投影变换 平面点云 ALPHA shapes算法 窗户提取 

分 类 号：P232[天文地球—摄影测量与遥感]