一种高分辨率遥感影像建筑物提取方法

A method for building extraction from high resolution color remote sensing image

采用了一种基于独特像元的快速自动提取建筑物的方法,如线与线之间的交叉点。首先,在预处理阶段采用边缘保留双边滤波器降低噪声,提高建筑物边缘的对比度;其次采用EDLines线段探测器实时准确地提取建筑物线段;最后,采用图表搜索感知分组的方法分层次组织预判线段为候选矩形建筑物。实验证明,该方法可以实现遥感影像建筑物的准确和快速提取。

This paper presents a method of quickly and automatically building extraction from high-resolution remote sensing.The proposed method conducts building extraction based on distinctive image primitives such as lines and line intersections.First,in the preprocessing stage it uses a edge-preserving bilateral filter to reduce noise and improve building edges contrast.Secondly,a real-time line segment detector called EDLines is introduced to accurately extract building segments;Finally,the paper presents a graph search-based perceptual grouping method to hierarchically group previously detected line segments into candidate rectangular buildings.The result indicates that this method can achieve accurate and rapid extraction building from remote sensing images of the building.