Hyper-spectral Image Denoising Using Sparse Representation

Abstract

Sparse representation, statistical and probabilistic approach have been used in image processing applications. Here, simple technique is used to denoise Hyper-spectral image by using sparse representation. Main focus is to transform the given image into another form which is combination of dictionary and sparse vector. So, basic statistical methods are used for the updation of dictionary and also for sparse coding. Then, probabilistic approach is used to determine new size of dictionary and this new dictionary is used to achieve denoised image and finally, peak signal to noise ratio (PSNR) is used to measure performance of denoising methods.