基于暗原色先验去雾的改进型算法研究

The Study of the Improved Algorithm for Haze Removal Based on Dark Channel Prior

分析讨论了原暗原色先验去雾技术,提出暗原色去雾算法的改进。针对在雾天情况下能见度很低,论文先根据暗原色先验得到雾天图像的每个小区域的最暗点,以这些点作为置信点,通过把这些置信点扩展膨胀来得到整幅图像的深度图,最后得到清晰的图像。采用理论推导和公式演算的方法,对改进型去雾算法的理论进行了分析,通过将原有的暗原色去雾算法与改进型去雾算法从理论上和实验结果的仿真上进行了比较,从而验证了改进型去雾算法的可行性和一定程度上的高效性。

Analysis and discussion of the original dark channel prior technology,improved dark channel defogging algorithm.According to thevisibility in the fogcondition is very low,firstly,according to the dark channel prior to obtain the mostdarkfoggy images of each small area,with these points as confidence,by putting these confidence point through the expansion of expansion to obtain the depth map of the whole image.used theory is derived and formula calculus of method,on improved type to fog algorithm of theory for has analysis,in analysis introduced of process in the,will original of dark primaries to fog algorithm and improved type to fog algorithm for has theory of compared and experiment results of simulation compared,obtained improved type to fog algorithm of superiority.