A Remote Sensing Image Segmentation Method Based On Spectral and Texture Information

Segmentation is an important problem in remote sensing image processing. In this paper, we propose a new method for segmenting a remote sensing image that provides spectral and texture information. Laplacian of gaussian (LoG) filters are used for the removal of noise. The enhanced image uses K-Mean clustering algorithm. Local spectral histogram representation, which comprises of histograms of filter responses in a local window, provides an effective feature to capture both spectral and texture information. The SVD is calculated for error estimation depending on the size of the image. The experimental results discussed in this paper provides MATLAB implementations of gray scale image, LoG filter, K-Mean algorithm, histogram equalization, and SVD with graph plot.