Fusion Based Single Image De-hazing
Shilong Liu, N. M. Kwok, Yeping Peng, Ruowei Li, Hongkun Wu · 2019
Digital images captured in hazy outdoor environments, due to particle disturbance and air-light intervention, often suffer from loss of contrast and color distortion. These images are incapable of providing valuable information to viewers. To recover the haze-free counterpart, an image-fusion based algorithm is here proposed. The transmission factor, representing the haze severity, is found closely related to the input image minimum color channel. Furthermore, to ensure the recovered haze-free image does not lose the depth sense within the allowed dynamic range, a threshold is included as the lower bound for the transmission range. Intermediate images generated from varying transmission thresholds are fed into a fusion process, where the weighting matrix is determined by the range filtered image intensity. The resultant image is further processed to maintain the original hue and realize contrast enhancement. Results are analyzed both qualitatively and quantitatively compared to other state-of-the-art image de-hazing algorithms to validate the performance and efficiency of the proposed method.