Low Lightness Image Enhancement Using HSV Color Based on DCP with Color Restoration and Lightning Stretch

The poor visibility of low-light images makes them unsuitable for computer vision algorithms and human observation. Many image-enhancing methods have been suggested as techniques for this problem. But it does not restore colors and improve lighting properly. Therefore, in this study, we propose an algorithm that restores colors and improves lightness for low-light images using LOL data, which contains 485 low-light images. The proposed algorithm includes three main steps. The first is to use the brightness low light area technique, and the second is to restore colors by applying two filters, the third is to improve the lighting component based on the HSV color model, where the color components are isolated from the lighting, and then maps are drawn using lightness stretch and adaptive histogram equalization. We have used reference quality standards Peak Signal to Noise Ratio (PSNR) and non-reference quality metrics Entropy (EN)‚ Average Gradient (AG) and Natural Image Quality Evaluator(NIQE). Through analyzing the results, the proposed method succeeded in improving low-light images and color restoration, and the quality standards reached EN (7.004), AG (11.286), NIQE(3.502), and PSNR (14.013) values.