亚象元边缘定位的实现  

Implementation on a Subpixel Edge Locating Method

作　　者：孔丹[1] 李介谷[1] 

机构地区：[1]上海交通大学图象处理与模式识别研究所

出　　处：《上海交通大学学报》1997年第12期52-56,共5页Journal of Shanghai Jiaotong University

基　　金：中科院自动化所模式识别国家重点实验室资助

摘　　要：亚象元方法能够突破离散象素采样的限制，使边缘定位精度高于图象分辨率．基于连续的理想边缘模型，利用邻域空间矩拟合来计算象素中可能存在边缘的参数，从而获得边缘的亚象元定位．这种方法精度高，抗噪声，而且具有多尺度性和目标外形拟合简便等特点．The subpixel methods are not limited in discrete pixel sampling and can lead to a higher locating precision than image resolution.Based on a continuous ideal edge model a subpixel edge detection and locating method is proposed.The parameters of potential edges,which exist in every pixel neighborhood,can be estimated by space moment fitting.The method is robust and highly precise and has advantages of multi scale measurement and edge fitting simplicity of object contour.

关 键 词：边缘检测 边缘定位 图象处理 图象处理 亚象元 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]