Single Image Haze Removal using Contextual Regularization

Abstract

Haze is a condition where our visibility gets affected due to particles, smoke, dust and moisture which are suspended in air. Haze reduces the visibility. A contextual regularization de-hazing algorithm is proposed which uses a single frame image for enhancing the foggy image using multilevel transmission mapping. Quantitative parameters such as MSE and PSNR are considered to assess the superiority of the proposed method. The proposed method is fast and free from noise when compared with other existing methods.