一种基于动态图像处理的邮件检测方法  

Method of post detection based on dynamic image processing

作　　者：王建勇[1] 周晓光[1] 

机构地区：[1]北京邮电大学自动化学院,北京100876

出　　处：《计算机工程与设计》2005年第4期1059-1061,共3页Computer Engineering and Design

摘　　要：利用动态图像处理技术,设计了一种实时处理的邮件检测算法。采用了一种快速的中值滤波算法进行降噪处理,提出了一种简洁的边界提取算法,可以快速地提取出单像素边界。然后以自适应多直线匹配的方法实现两帧图像的配准,并利用差分图像信息实现邮件图像的提取,最后计算出邮件的相关参数。仿真实验表明,该算法能有效检测出运动状态下邮件的相关参数,实时性良好。A real-time algorithm of post detection is proposed using dynamic image processing technology. First, a fast median filter is adopted to denoise. A laconic algorithm for edge extraction is presented. The edge with the width of single pixel can be extracted quickly using this method. Then two frames are matched with the method of adaptive multi-line matching. With the information of dif- ference image, the post is segmented from the image. The parameter of post is figured out in the end. The experiment on computer shows this algorithm can detect post efficiently and quickly.

关 键 词：动态图像 邮件分检机 差分图像 图像处理 邮件检测方法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]