基于方向滤波的监控图像边缘检测  

Edge detection of monitoring image based on directional filter

作　　者：刘付民[1] 张治斌[1] 

机构地区：[1]河南理工大学计算机科学与技术学院,河南焦作454000

出　　处：《传感器与微系统》2011年第4期127-130,共4页Transducer and Microsystem Technologies

基　　金：河南省科技计划资助项目(092300410216);河南省软科学研究计划资助项目(102400450034)

摘　　要：监控图像的边缘检测是进行高级图像处理和模式识别的基础工作。将方向滤波理论引入监控图像的边缘检测中。通过Roberts算子的四方向滤波扩展,介绍了方向滤波用于边缘检测的一般过程;针对监控图像中斜坡边缘丰富的特点,引入Petrou-Kittler滤波器作为基础滤波器;用方向滤波理论改造Petrou-Kittler滤波器,形成了三方向Petrou-Kittler边缘检测器。实验结果表明:监控图像经过三方向Petrou-Kittler边缘检测器处理,检出的边缘信息完整、细滑、分割区域正确。Edge detection of monitoring image is basic work for advanced image processing and pattern recognition.Theory of directional filter is introduced into edge detection of monitoring image.Process of edge detection based directional filter is elaborated by taking Roberts operator as an example.Petrou-Kittler filter is introduced as basic filter,aiming to ramp edge in monitoring image.Petrou-Kittler filter is reconstructed by directional filter theory,and tri-directional Petrou-Kittler filter came into beings.Experimental results show that edge information is complete,thin and correct by processing of tri-directional Petrou-Kittler filter.

关 键 词：监控图像 边缘检测 斜坡边缘 方向滤波 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]