Traffic images enhancement based on vanishing point detection and atmospheric scattering model

Due to the complex illumination conditions under foggy weather, it is essential to perform an image enhancement stage to reverse the contrast-degraded. We propose a new traffic image enhancement method based on the vanishing point detection. The base principle of the method is atmosphere scattering model. Vanishing points are used to estimate the depth of field. The key feature of vanishing points algorithm is that extract the corner points from thinning the edges, which is based on the Curvelet transform. Five favorable image sharpness evaluation functions are combined to evaluate the effect of the method. Experimental results on an open foggy traffic images set, which includes 359 color images. Using this method image contrast increased by 29.79% compare with histogram equalization and 52.06% compare with Multi-scale Retinex enhancement methods.