| 基于区域分割合并的建筑物半自动提取方法 |
| |
| 引用本文： | 周俊,晏非,孙曼.基于区域分割合并的建筑物半自动提取方法[J].海洋测绘,2005,25(1):58-60. |
| |
| 作者姓名： | 周俊 晏非 孙曼 |
| |
| 作者单位： | 西安测绘信息技术总站,陕西,西安,710054 |
| |
| 摘 要： | 提出了一种遥感图像半自动提取建筑物的方法。首先通过对建筑物边缘进行折线逼近和感知编组，可以得到其顶部大体结构。然后按照“二叉树”的数据结构，利用“超线”对建筑物影像进行区域分割。最后通过对分割后的区域进行相似性的统计并按照“二叉树”记录的分割顺序进行生长合并，从而得到了屋顶的边缘，实现了对平顶直角建筑物的提取。该方法能够充分利用区域的性质进行合并，稳定精确地提取建筑物。
|
| 关 键 词： | 像片判读 建筑物提取 区域分割 超线 二叉树 |
| 文章编号： | 1671-3044(2005)01-0058-03 |
| 修稿时间： | 2004年9月2日 |
Semi-automatic Extraction of Building Based on Region Segmenting and Merging |
| |
| ZHOU Jun,YAN Fei,SUN Man.Semi-automatic Extraction of Building Based on Region Segmenting and Merging[J].Hydrographic Surveying and Charting,2005,25(1):58-60. |
| |
| Authors: | ZHOU Jun YAN Fei SUN Man |
| |
| Abstract: | The paper presents a method for semi-automatic extractin of building . We arrange the line segments of (building's) edges to reconstruct the structure of the building roof by means of line closing and perceptual organization method first. And then, a binary space partitioning tree is adopted as an efficient data structure to aid recursive partitioning of a region with hyperline. When the image of the building is segmented into statistically homogeneous sub-regions, we can merge the regions by using pre-defined homogeneous predicates. Through the above steps, the roof of the building can be extracted accurately and steadily. |
| |
| Keywords: | photo interpretation building extraction region segmenting hyperline binary tree |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
