Method of Foggy Image Removing Algorithm Based on Retinex Theory
Lin Xiaoju · Video Engineering · 2013
A new curvelet transform image enhancement algorithm based on the Retinex is proposed to improve the center surround Retinex algorithm whose choices is limited about the scale. The proposed algorithm can remove fog effectively while enhance details in the applications of foggy images enhancement. Firstly,the algorithm decomposes the original image into the incidence image and the reflectance image using the Gaussian function. Secondly,the curvelet transform is used to decompose the reflectance into high frequency coefficient and low frequency coefficient. The high frequency coefficient is disposed by auto-commutation threshold. The low frequency coefficient is disposed. At last the curvelet coefficients are reconstructed to obtain the enhancement image. The experimental results show that the proposed method can improve the details,the PSNR,the information entropy and the visual effects. It can decrease the image distortion.