Optimally sectioned and successively reconstructed histogram sub-equalization based gamma correction for satellite image enhancement

Abstract

This paper presents an overall quality enhancement approach especially for dark or poorly illuminated images with a core objective to re-allocate the processed pixels using recursive histogram sub-division. An information preserved and image content based behavioral reconstruction inspired adaptive stopping criterion based on pixel-wise relative L2−norm basis (which itself is intuitively related to optimal PSNR value) is proposed in this paper, so that highly adaptive gamma value-set can be derived out of it for sufficient enhancement. Due to this adaptive behavior of the intensity distribution the gamma value-set when derived from it, is obviously highly adaptive and here individual gamma values are evaluated explicitly raised over reconstructed intensity values, unlike conventional gamma correction methods. This adaptiveness makes the entire methodology highly capable for covering a wide variety of images, due to which robustness of the algorithm also increases. The proposed methodology has been verified on various dark images. The simulation results authenticate the overall enhancement (contrast as well as entropy enhancement along with sharpness enhancement) achieved by the proposed has been found superior to other dark image enhancement techniques.