Improved Hyperspectral Image Denoising Employing Sparse Representation

Hyper spectral imaging is the latest imaging technologies which observe the Earth's surface using some hyper spectral sensors. The aim of this imaging is to obtain the spectrum for each detailed pixels carrying the image of the scene, for classifying objects, identifying materials. Noisy elements can decrease the classification accuracy by affecting the boundaries of particular objects so it will affect on the important part of interest. So extraction of valuable information becomes difficult and can say that noise can create artifacts and loss of spatial resolution. To overcome this problem HSI (hyper spectral imaging) denoising (removing noise) is an important preprocess step before these images/data are analyzed. It has been proved that there is much global and local redundancy and correlation (RAC) in both spatial/spectral dimensions on HSI. So this proposed system is focused on jointly using the global and local RAC in spatial/spectral domains and to eliminate spectral distortion, low-rank constraint is used which deal with the global RAC in the spectral domain. The system aims to reduce the time complexity in achieving improved denoising of the hyper spectral image than other existing methods. The system focuses on improving values of peak signal-to-noise ratio (PSNR), structural similarity index measurement (SSIM) and feature similarity index measurement (FSIM) indices.