An improved image clearness algorithm based on dark channel prior

摘要

An improved image clearness method based on the theory of dark channel prior is proposed. Firstly, the single channel image coming from the minimum value of each pixel among the three color channels is computed; secondly, the atmospheric light is estimated by using the wavelet threshold method; thirdly, the transmission maps are calculated by using the atmospheric light, and then the image is recovered by using the atmospheric light. The experiment results show that this method can restore the fogging image effectively and reduce the time complexity which brought by Soft Matting algorithm greatly.