高分遥感影像道路提取方法研究

高分辨率遥感影像道路提取是当前研究的一个热点,传统的提取方法多基于光谱特征,不能充分利用高分影像的形状、纹理等特征。文章提出了一种利用光谱特征和形状特征提取高分遥感影像道路的方法。首先利用道路的光谱特征,采用阈值分割法提取可能的道路段并去除小面积区域,再根据道路的线性特征去除非线性区域,之后进行形态学闭合运算填补由道路光谱差异造成的孔洞,并进行修剪,实现道路的提取。

Scores road extraction method of remote sensing images

Road Extraction from High Resolution Remote Sensing is currently a hot research topic, the traditional multi-spectral feature extraction methads, Not take advantage of the image score shape, texture and other characteristics. This paper presents a shape using the spectral features and feature extraction of high The method of remote sensing image path. Firstly, the spectral characteristics of the road, using threshold segmentation method to extract and remove possible road segments Small areas, Linear characteristics according to the road and then remove the non-linear region, morpholog- ical closing operation after fill light by the road Caused by differences in spectral holes and trimming to achieve road ex- traction