货车超限差分算法图像分割  

A Difference Algorithm on Image Segmentation of Gauge Monitoring

作　　者：王颋[1] 肖世德[1] 朱为鹏[1] 

机构地区：[1]西南交通大学机械工程学院智能机电技术研究所,成都610031

出　　处：《计算机工程与应用》2004年第11期212-215,共4页Computer Engineering and Applications

基　　金：铁道部预研项目(02-03)

摘　　要：该文讨论了在铁路货物超限监测中,利用数字图像处理技术提取货车轮廓,从而用于货车超限判断。其中的核心问题是运动货车图像的提取。该文提出了基于差分算法下自动图像分割的方法,在一定天气与背景条件下,测试结果表明,该算法适应性较好,检测精度和速度基本能满足要求。The paper presents an algorithm for freight train ima ge segmentation based on the image segmentation theory.The profile of the freight train and cargoes on the image is obtained with the arithmetic and whether car goes exceeding the gauge or not is concluded by the location of the freight tra in and cargoes on the image.Auto image segmentation is presented based on diffe rence algorithm.It is proved that the algorithm has good adaptability,acceptab le precision and satisfied processing speed in normal climate.

关 键 词：图像分割 货车超限监测 差分 运动检测 阈值分割 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]