Single Image Dehazing based upon Modified Image Enhancement Algorithm

Abstract

The images acquired in the bad weather such as smoke, fog and haze are often disturbed to a large extent. Therefore, dehazing has become a challenging task in many fields. By combining the quadtree atmospheric light estimation model with the guided filtering algorithm, this paper proposes an image enhancement algorithm, which can significantly improve the image contrast and preserve the details. This method can significantly improve color image and gray image without prior information. The multi-scale decomposition module provides basic image and detail image for the subsequent smoothing module and enhancement module. Through the atmospheric light intensity estimation algorithm and transmittance estimation algorithm described above, the accuracy of image restoration can be significantly improved. The details of the image can be saved by processing the residual image by guided filtering algorithm. Finally, the final image can be obtained by color correction algorithm.