Published Paper Details:

Images and videos captured in hazy or foggy weather conditions are seriously degraded by the scattering of atmospheric particles, which directly influences the performance of outdoor computer vision systems. The existing system mainly focuses on a portion of an image which is based on retinex method and wavelet-transform methods. The problems faced by these methods are illumination issues, over brightness and edge preserving issues. Both these techniques are complex. So it cannot be implemented in real time systems. In our paper, a fast algorithm for image and video dehazing is implemented based on linear transformation by assuming that a linear relationship exists in the minimum channel between the hazy image/video frames and the haze-free image/video frames. To accurately estimate the atmospheric light, an additional channel method is proposed based on quad-tree subdivision. This algorithm can clearly and naturally recover the image and video frames especially at the edges. It can thus achieve a good effect for single image and video dehazing. This paper is applicable in computer vision systems such as urban transportation, driver assistance systems and outdoor video surveillance.