Very Low Illumination Image Enhancement via Lightness Mapping

Abstract

Image taken in low light suffers from poor quality, low contrast, and color distortion, so this image with too low light is difficult to improve due to lack of color information and lack of lightness. Therefore, enhancing images is an important topic because it is included in many important applications such as surveillance, tracking, and medical images. In this paper, we introduce a new method for enhancing very low illumination images. This method consists of two steps, first step includes color restoration by using adding a median filter and min filter, the color restoration applies on each color component R, G, B then converted to HSV color model to extract lightness component (v) then apply the second step on component (v) which include illumination enhancement using sigmoid function and Contrast Limited Adaptive Histogram Equalization (CLAHE). Then recombine with chromatic components (H, S). The HSV model is converted to the RGB model to get the final enhanced image. The proposed method is applied on eight images taken from the DICM database, and the proposed method is compared with other methods such as Parallel Nonlinear Adaptive Enhancement (PNAE), New Nonlinear Adaptive Enhancement (NNAE), Naturalness (NAT) preserved Fusion-based-enhancing (FU), new contrast enhancement (CN), and image enhancement by fuzzy (EF) using image quality metrics as entropy and naturalness image quality assessment evaluator (NIQE). Experimental results show that the proposed method has the highest value than other methods, where the proposed method achieves average quality value (6.65) and (3.80) for eight images of NIQE and entropy metrics.