摘要
在精密测量、自动化装配和机器人等诸多领域,往往需要对圆形器件或图标进行识别和定位,而目前传统的检测方法是Hough变换,计算复杂,对资源需求大,且不利于实时控制.本文利用圆形几何对称的性质,采用基于颜色分类方法,提出一种非Hough变换的圆的检测方法,从而达到对彩色图像中圆形目标进行快速识别的目的.设计了算法的流程,编制了相应的圆识别程序,通过对足球机器人定位的验证,表明该算法具有运算速度快及对畸变的圆形目标适应性好等优点,为图像处理中圆目标的快速识别与定位提供了一种借鉴.
Precision measurement,automated assembly and robotics fields often require an identification and location to a circular object or icon.The current detection method is based on Hough transform,which is complex in computing and requires much resource.A non-Hough method is proposed,which can recognize circular targets in color image by making use of the symmetrical property of circles and adopting the segmentation method based on color.The experiments of applying this method to the location of soccer robot show that the method can calculate quickly and adjust to deformed circle.The algorithm of quickly identifying circular target provides a reference in image processing.
出处
《河南理工大学学报（自然科学版）》
CAS
2010年第4期498-501,共4页
Journal of Henan Polytechnic University（Natural Science）
基金
国家自然科学基金资助项目(10802027)
河南省教育厅自然科学项目(科技攻关)(2009A460005)
河南理工大学博士基金资助项目(648163)
关键词
圆检测
非Hough变换
图像识别
机器视觉
circle detection
non-Hough transform
image recognition
machine vision
