基于纹理提取与注入的交通图像去雾算法研究

Research on the defog algorithm of traffic image based on texture extraction and injection

雾天户外视觉系统所拍摄的交通图像质量下降、特征模糊直接影响到后续智能交通系统的监控精度,研究交通图像的去雾问题具有重要的实际意义.不同于一般的可视场景图像,交通图像的去雾在关注其去雾后视觉效果的同时,更应重视对交通图像纹理、边缘等特征信息的保持.提出一种基于纹理提取与注入的交通图像去雾算法,利用引导滤波从含雾交通图像的非下采样Contourlet变换的高频子带中提取交通图像的纹理、边缘信息,然后将其注入利用暗通道模型的去雾图像中,在有效去雾的同时一定程度上增强了去雾后交通图像的纹理细节信息.实验结果验证了算法的有效性.

The declining quality and the fuzzy features of the traffic image taken by the outdoor vision system in foggy days directly affect the monitoring accuracy of the subsequent intelligent transportation system.It is of great practical significance to study the defogging problem of the traffic image.Different from the general visual scene image,the defogging of traffic image should not only pay attention to the visual effect after defogging,but also pay more attention to the maintenance of the traffic image texture,edge and other feature information.In this paper,a defogging algorithm of traffic image based on texture extraction and injection is proposed.The texture and edge information of traffic image is extracted from the high frequency sub-band of the non sampling Contourlet transform using guided filtering,and it is injected into the defogged image based the Dark Channel Model.The proposed algorithm can further enhance the texture details of the defogged traffic image while removing fog effectively.The experimental results verify the effectiveness of the algorithm.