论文部分内容阅读
现实中常见的建筑物是比较规则的,外形大多为矩形或矩形综合体,现有的建筑物提取方法也主要是以这类规则建筑物为提取对象。本文针对球顶这类相对独立的特殊建筑物屋顶结构,在现有各种霍夫变换改进算法的理论基础之上,提出了一种基于扫描线的快速霍夫变换。该方法首先将滤波后的点云数据分割成各扫描线段,以局部极大值点为中心沿扫描线向两边搜寻曲率近似相等的点构成特征弧段,最后基于二维霍夫变换提取出目标弧段。试验表明,该方法对于机载LiDAR数据中球顶房屋的提取是有效的。
Common buildings in reality are relatively regular, mostly rectangular or rectangular shape of the complex, the existing building extraction method is mainly based on such rules for the extraction of buildings. In this paper, a kind of relatively independent roof structure with special independent structures such as dome is proposed. Based on the existing theories of various improved Hough transform algorithms, a fast Hough transform based on scan lines is proposed. Firstly, the filtered point cloud data is divided into scanning line segments. The local maximum point is taken as the center to search the approximate arc-like points along the scan line to form the characteristic arcs. Finally, the target is extracted based on the two-dimensional Huff transform Arc segment. Experiments show that this method is effective for the extraction of domes in airborne LiDAR data.