基于大气散射模型和曲波变换的交通图像增强研究

摘要

According to the characteristics of color foggy images, an image enhancement algorithm based on vanishing points detection is proposed. First by means of the atmosphere scattering model, a curvelet transform is adopted to thinning the edges to obtain the vanishing lines. Then the road vanishing points are estimated in lines comers. Final the images are enhanced by calculated depth of fields. Five favorable image sharpness evaluation functions are combined to evaluate the effects of the method. Experimental results show the better performance compare with histogram equalization and multiscale Retinex methods.%通过研究雾天彩色降质图像的特征,提出以大气散射模型为基础,采用基于曲波变换的消失点检测算法进行边缘稀疏化以提取角点,并根据计算的景点深度进行图像增强运算.选取5种图象锐化函数对实验结果进行评估,对比直方图均衡化图像增强法和多尺度Retinex图像增强法,显示出基于大气散射模型和曲波变换的方法在图像增强方面效果更佳.