Swarm Intelligence for Feature Identification in Natural Terrain Environment

One of the most important problems in remote sensing studies is the classification of features in the satellite images, as it provides land use/ land cover information of the area under study. The Land Cover information like vegetation, water bodies, rocky area, sandy area etc. and its change over a period of time greatly affect local, regional and global environmental changes. Land Use information like buildings, roads and railway tracks are important features of urban infrastructure which crucially affect the life of people in cities. Recent developments in biologically inspired optimization techniques have motivated the researchers to explore the application of these techniques to the problem of satellite image feature classification. In this paper particle swarm optimization (PSO) along with the morphological operators is used for the identification of urban features like road and railway network, as well as land cover types, contained in the image. The concept of this paper is to explore and utilize the neighborhood information of the swarm computing algorithm to accurately identify features in the natural terrain environment. The test areas used are located in urban environment of Chandigarh and Saharanpur. Google Earth images of these areas were acquired and processed. The approach adopted has yielded satisfactory results.