An innovative hybrid technique for road extraction from noisy satellite images

Abstract Remote sensing takes an important role in the detection and monitoring of natural resources. It is also used in abundant fields, including geography, topographical survey, and geoscience disciplines such as land monitoring, forest management, crop identification, soil mapping, identification of ocean resources, road extraction, etc. This road extraction has a vital application in which it plays a key role in the development of GIS. Automatic updating of GIS information has become very important in our daily life. Semi-automated road extraction had a problem of computational complexity, high rate; consume a large amount of manpower and time. A fully Automated process is better than this but due to the quality of RSI, pre-processing techniques, segmentation, etc., the extraction differs. This paper proposes a new hybrid technique for the extraction of roads with better accuracy. Gaussian filter and CLAHE are used to improve the quality of the image, and to reinforce the features with the neighboring pixels. This approach of extraction from RSI images is based on Histogram equalizer, fuzzy – c mean algorithm. This is achieved by mat-lab and applies to high-resolution satellite images.