基于数学形态学的交通图像滤波处理方法研究  

Algorithm for filtering traffic video images based on mathematical morphology

作　　者：崔宝侠[1] 李爱军[1] 苏桂华[1] 

机构地区：[1]沈阳工业大学系统工程研究所,沈阳110023

出　　处：《沈阳工业大学学报》2007年第2期192-196,217,共6页Journal of Shenyang University of Technology

摘　　要：针对智能交通视频图像中常见的几种噪声源,阐述了数学形态学在交通图像滤波中的应用.结合实际拍摄的动态交通图像,提出了基于数学形态学和帧差法的交通图像噪声滤除方法.根据提出的算法方案,编写了Visual C++程序,进行了实验验证,并在此基础上比较了不同形态学算法对同一组交通图像的处理结果.从实验结果得出,该算法有效分离并较完整地保留了交通图像中的运动车辆,取得了良好的降噪处理效果.The application of mathematical morphology in filtering traffic images was discussed with respect to a number of noise sources common in intelligent traffic video images. A noise-filtering algorithm of real traffic video images based on mathematical morphology and frame difference was proposed. Visual C ＋ ＋ programs for the algorithm were developed and experiments were conducted to test the proposed algorithm. The processing results of different morphologic algorithms for the same set of traffic images were compared. The comparison results show that the proposed algorithm effectively separates and preserves the moving vehicles in the traffic images and achieves good noise filtering results.

关 键 词：智能交通系统 交通流量 视频图像处理 数学形态学 帧差法 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]