Research On Methods Of Road Extraction From Remote Sensing Images

With the progress of science and technology,domestic series satellites have gradually broken the dominant position of foreign commercial satellites.Remote sensing image technology has also been supported by high resolution remote sensing satellites,and began to develop to the civil industry.At present,the spatial resolution of civil satellites is continuously improving,which lays the foundation for the effectiveness of road information extraction.The research of road extraction algorithm can show its advantages in making road network diagram and updating database,especially in GIS data acquisition.Based on the road features of high resolution remote sensing images,this paper studies the extraction techniques of roads in high resolution remote sensing images,such as improved watershed,support vector machine and iterative segmentation.In this paper,the algorithm of gray threshold,edge detection operator segmentation and traditional watershed segmentation is studied.An improved watershed algorithm is proposed to solve the problem of over segmentation of the traditional watershed.The algorithm is processed before segmentation,and the maximum component is considered from the multi-scale morphological gradient.Gradient,obtain the relative minimum value from the low frequency region,form the two value graph and calibrate the original gradient map,then use the watershed segmentation method to divide the remote sensing image,and extract the road information according to the road shape feature.In view of the segmentation problem of the small road network,a support vector machine which adopts the iterative segmentation and a morphological optimization is proposed.From the image classification,the algorithm uses support vector machine to classify the high resolution remote sensing images,and aims at the problem that the road information is not complete after classification.Then the image is segmented by iterative method and the morphological optimization is carried out.Finally,the extraction of road network is completed.This paper studies the road extraction of high resolution remote sensing images,and the algorithm is verified by experiments.The experimental results show that the extraction effect of the road of remote sensing image is better combined with the support vector machine and iterative method,and the road network information can be extracted completely.