中文摘要：

针对在烟、雾、霾等恶劣的户外条件下,由于大气粒子的散射作用导致捕获的景物图像严重退化的问题.本文提出一种新的单幅图像去雾算法,该算法从大气散射模型出发,首先对降质图像进行白平衡,从而简化了大气散射模型;然后利用非线性空间滤波方法估计粗糙的大气幕亮度,用联合双边滤波方法进一步细化大气幕亮度;最后采用色度调和得到恢复图像.该算法的时间复杂度是图像像素的线性函数,运算速度较快.试验结果表明,该算法有效地恢复了场景的对比度和颜色,从而明显地提高了图像的视见度.

英文摘要：

In the smoky, foggy, hazy and other harsh outdoor conditions, scene image captured is severe- ly degraded by the scattering of atmospheric particles. Based on atmospheric scattering model, the de- graded image needs a white balance operation to simplify the atmospheric scattering model; then, a new nonlinear spatial filtering algorithm is put forward to estimate the rough luminance of the sky, and joint bilateral filter is introduced to further refine the luminance of the sky. Finally, image is restored by using tone mapping. The complexity of the proposed algorithm is only a linear function of the input image pixels numbers, and this allows a very fast implementation. Experiments show that the proposed algorithm can effectively restore the fog-degraded images. With this algorithm, the images is restored with more natural, realistic, and higher-definition.