Development of multi-band image fusion technique

Abstract

Satellite images are quite useful for information regarding weather forecasting, map making,security purpose and environmental monitoring but this need costly equipment to build,monitor and maintains of satellite. Therefore, it is compulsory utilize data regarding satellitein a best way. A best example of fusion is pan-sharpening. Pansharpening is process of combing high spatial and low spectral resolution image with low spatial and high spectral image. This technique is quiet useful for remote sensing. Basic aim of this thesis is to solve problems of pan-sharpening. The thesis is comprised of five parts. In the first part by using radiometric and geometric index function for extracting edge map of panchromatic image to get improved version of high resolution panchromatic image. In the second step, further enhance the high resolution panchromatic image by using function of local synthesis.Then done final fusion by considering technique of Bayesian fusion. By the end of second step details of image quiet visible in a better way. In the fourth step of algorithm apply Intensity Hue Saturation transformation for enhancing color of image. In the last step consider various qualitative metric i.e MSE, RMSE, CC, PSNR,SNR and MAE to evaluate final performance of technique.