Application of Coarse to Fine Level Set Segmentation in Satellite Images

Segmentation of vegetated areas is inevitable in precision agriculture and has high importance in urban area with social and environmental aspects. This paper addresses the segmentation of fused high resolution multispectral satellite images into distinct regions such as vegetation, buildings and barren land. Even though the IHS based fusion of satellite imagery improves the visual interpretation, it results in color distortion which can be nullified using vegetation indexes(VI).The vegetated area can be depicted using high resolution normalized vegetation index and to detect the distribution of vegetated area, soil classes, buildings and barren land. For that the coarse-to-fine level set method is used. Undecimated wavelet transform is adopted to separate focused areas from the background. Homogeneity metric is used to measure the variation inside and outside the contours. The weight distribution ratio is proposed to adaptively tune the relative weight of the features. Based on the homogeneity metric and the weight distribution ratio, a novel energy functional is developed to solve the contour extraction problem and a coarse-to-fine scheme is applied to progressively extract contours in finer scale which also reduces the computational burden.