DETECT AND ISOLATE THE LINEAR COMPOSITIONS IN SATELLITE IMAGES

Abstract

Remote sensing data is a source for deriving a spatial graph of the Earth's surface, which is mainly composed of linear faults, drainage systems, road networks, and other natural and artificial landscapes. Their appearance clearly and accurately often requires algorithms and image processing methods to detect edges. On this basis, the subject of the research included identifying and automatically isolating the spatial characteristics of the earth's surface, such as linearity, drainage systems, and the boundaries of the earth's coverings. Robert and Sobel filters were used for the purpose of edge enhancement. . In order to isolate the edges clearly from the rest of the image contents, the adaptive obfuscation method was used in order to obtain a binary image that can be used in post-processing operations. For the purpose of isolating the edges better and closer to their true shape, a new algorithm was corrected to join and slim the edges, and this algorithm included the processes of extension and stripping. The final algorithm has been applied to different scenes from Nineveh Governorate captured by satellite sensors and aerial photography, and the results have proven reasonable success in detecting linear structures.