A wavelet-based approach to improve foggy image clarity

Abstract

Under foggy viewing conditions, image contrast is often significantly degraded by atmospheric aerosols, which makes it difficult to quickly detect and track moving objects in intelligent transportation systems (ITS). A foggy image visibility enhancing algorithm based on an imaging model and wavelet transform technique is proposed in this paper. An optical imaging model in foggy weather conditions is established to determine the image degradation factors and compensation strategies. Based on this, the original image is firstly transferred into YUV color space of a luminance and two chrominance components. Then the luminance component is decomposed through wavelet transform into low- and high-frequency subbands. In the low-frequency subband, the medium scattered light component is estimated using Gaussian blur and removed from the image. Nonlinear transform for enhancement of foggy images is applied to high-frequency subbands. In the end, a new image is recovered by combining the chrominance components and the corrected luminance component altogether. Experimental results demonstrate that this algorithm can handle the problem of image blurring caused by atmospheric scattering effectively, and has a better real-time performance compared with a standard model-based procedure.