A New Method of Visibility Enhancement for single Foggy or Hazy Image

摘要

Outdoor images captured in bad weather conditions often have poor qualities in terms of visibility and contrast. In this paper, a new method to enhance visibility of fog or haze-degraded images was proposed. By testing the real scene images taken in foggy weather with prevailing dehazing algorithms, the design concept of this method was as following: the local detail factors of each sub-block on the input image was acquired firstly, then local detail factors and power-law transformation were applied to modify the global histogram based on which the adaptive brightness preserving histogram equalization was conducted. With the proposed algorithm, output image gained excellent visibility and acceptable visual effect. Furthermore, the problem of uneven light transmission in a defective optical system could be compensated to suppress facula effect in output image.