Hyper spectral images restoration approach using LRMR and LDA

The most important factor in quality measurement of any technique or system is the efficiency. The hyper spectral images are the complex ones those are taken from the low resolution cameras. These images are often degraded in the acquisition process due to the integrated so many different noises in it. These can include various types of noises as impulse noise, Gaussian noise, deadlines and stripes etc. This paper defines new hybrid image restoration method based on Linear Discriminative Analysis (LDA) and Low Rank Matrix Recovery (LRMR). The proposed approach works in two phase, first it converting an image into number of patches and treating each patch as a separate image, performance can be measured out and that patches induced in LRMR framework In second phase, for further improvement in image quality LDA approach is used.