基于对比度增强和形态学的遥感影像道路提取

Road Extraction of Remote Sensing Image Based on Contrast Enhancement and Morphology

提出了一种基于对比度增强和形态学的遥感影像道路边界与特征点提取的方法。先对遥感影像进行对比度变换增强,通过对比分析直方图均衡化和对比度分段线性增强两种方法获取的增强影像,选取区分度大的分段线性增强方法进行影像增强,然后运用数学形态法进行影像分割,实现道路和其他图像信息的有效分离。利用Krisch算子进行边缘检测提取道路的边缘信息,并基于边缘特征利用改进的Harris算子提取特征点,将提取的特征点进行拟合并用函数模型描述图像道路信息,用于后期制图中道路信息的矢量化。

A method of road boundary and feature point extraction for remote sensing image was proposed based on contrast enhancement and morphology. Firstly, it carries on contrast enhancement of remote sensing image. Piecewise linear enhancement method with larger discriminative power was selected to do image enhancement after comparative analysis difference between histogram equalization and contrast piecewise linear enhancement. In order to realize the separation of road and other information effectively, image segmentation was done using mathematical morphological method. Then, using edge detection method by Krisch operator to extract road edge information, and feature points extracted by improved Harris operator based on edge feature. Last, fitted the extracted feature points and the function model was used to describe the road information of the image and applied to vector road information in the later period of drawing.