CRITICAL ANALYSIS OF IMAGE RESOLUTION ENHANCEMENT TECHNIQUES ON SAR IMAGES

Abstract

In the current thesis a critical analysis is performed on the image resolution enhancement techniques for improving the resolution of the low resolution MODIS and PALSAR satellite images. The wavelet based framework for image processing which has received a lot of attention in the recent past and has proved to be a better alternative to the conventional interpolation techniques due to the processing in transform domain, has been explored for the satellite images. The fusion of multi sensor images is a very arduous and challenging task. MODIS and PALSAR images belong to the optical and microwave bands of the spectrum respectively. The different spectral bands highlight different land cover classes. Thus integrating the data available in both the images can enhance the features and as such improve the classification accuracy. The wavelet framework for resolution enhancement has been coupled with the fusion frame work to obtain images with enhanced features and improved accuracy. The edges in the image strongly influence the perceptual quality of the images. In addition to that they can simplif' several important task such as change detection, image compression, - coastal and glacial region monitoring. This thesis presents a wavelet based edge enhancement and edge detection technique where the process of detection occurs in two phases. The first phase is concerned with identifying the edge regions using wavelets. The second stage performs the task of the image segmentation using the image developed during the first stage to obtain faster convergence and greater accuracy. The thesis also throws light on the future prospects in the current domain and the direction for the further research.