Road Extraction in Remote Sensing Images Using a New Algorithm

Road extraction research has always been an active research on automatic identification of remote sensing images. In this paper, improving the traditional algorithm, the author presents a new method which uses multi-weighted terms to judge the road edges, makes full use of the physical characteristics of the road, and makes the context of road edge pixels as a judgment, in order to recognize the edge segments of the road. Bridge-link mode accord to the characteristics of the road: two parallel lines and certain width. To connect the two marginalized segment of the corresponding pixels in order to achieve road extraction .Experimental results demonstrate the algorithm can eliminate noise, effectively improve the accuracy and velocity of the road extraction.