基于暗原色理论和自适应双边滤波煤矿尘雾图像增强算法.docVIP

基于暗原色理论和自适应双边滤波煤矿尘雾图像增强算法 　　摘要：针对煤矿井下存在大量煤尘、水雾导致获取的视频图像伴有大量的噪声、分辨率低、模糊的问题，提出了一种基于暗原色理论和自适应双边滤波的煤矿尘雾图像增强算法。基于暗原色先验理论，采用自适应双边滤波代替softmatting过程来求取精细透射率图，并根据煤矿井下特殊环境，从新的角度求取全球大气光值、粗略透射率图，并根据图像退化模型实现图像的去噪。实验结果表明，对于分辨率为1024×576的图像处理时间为1.9s，与He算法（HE K， SUN J， TANG X. Single image haze removal using dark channel prior. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2011，33（12）：1-13.）相比，运行效率提高了5倍。与直方图均衡法等算法相比，所提算法有效增强了图像细节、边缘，整体上更加适合人类视觉和视频监控的要求。 　　关键词：暗原色先验理论；自适应双边滤波；图像去雾；介质传播函数；大气物理散射模型 　　中图分类号： TP391.41 文献标志码：A 　　Abstract：Concerning the problem that videos images captured from coal mines filled with coal dust and mist are often with quality problems such as lots of noise， low resolution and blur. To solve this problem， an enhancement algorithm for fog and dust images in coal mine based on dark channel prior theory and bilateral adaptive filter was proposed. On the basis of dark channel prior， the softmatting process was replaced with the adaptive bilateral filtering to obtain fine transmittance map. Then according to the special circumstances of coal mines， the global atmosphere light and the rough transmittance map were got from new perspective and image denoising was realized on the basis of the image degradation model. The experiment results show that the image processing time for a resolution of 1024×576 is 1.9s. Compared with He algorithm （HE K， SUN J， TANG X. Single image haze removal using dark channel prior. IEEE Transactions on Pattern Analysis and Machine Intelligence， 2011，33（12）：1-13.）， the efficiency increased 5 times.Compared with other algorithms such as histogram equalization method， the proposed algorithm is effective to enhance the image detail. In this way， images can be more suitable for human vision as a whole. 　　Key words： dark channel prior theory； bilateral adaptive filtering； fog removing of image； medium transmission function； atmospheric scattering model 　　0 引言 　　随着煤炭技术的发展，煤炭产业在我国工业产业中所占的比例越来越大， 同时煤矿安全生产问题也引起了社会的广泛关注，国家煤