Abstract:

It is of great practical significance to extract road information from high-resolution remote sensing images.In view of the fact that the existing image classification methods can not directly obtain the high-precision road network information and the degree of automation is low,this paper presents a road extraction method based on the OSM (OpenStreetMap) vector road network guidance to achieve fast and accurate automatic road extraction of high resolution remote sensing images.Firstly,remote sensing images are preprocessed by the gray morphological corrosion,expansion and opening and closing operations.Through the prior information provided by the OSM network,the fuzzy C-means algorithm is improved and the input remote sensing images are roughly divided into three categories.Then the rough classification results are taken as the classification features,and the road samples are automatically selected by the OSM vector network.Using support vector machines for fine classification,and the optimal classification parameters are selected by means of the particle swarm optimization.Finally,we post-process the classification results to get a complete road network.Experiments using two groups of Google Earth images show that the proposed algorithm outperforms the comparison algorithm in the accuracy of road network extraction.

Key words: road extraction, vector guidance, fuzzy C-means, support vector machines, particle swarm optimization