基于二次曲面拟合的子像素边缘检测算法  

Based on Quadric Surface Fitting of Sub Pixel Edge Detection Algorithm

作　　者：巨志勇[1] 王超男[2] 

机构地区：[1]上海理工大学上海现代光学重点实验室,上海200093 [2]上海理工大学光电信息与计算机学院,上海200093

出　　处：《电子科技》2016年第6期8-10,14,共4页Electronic Science and Technology

基　　金：国家自然科学基金资助项目(81101116)

摘　　要：针对传统边缘检测算法边缘信息保留不完整与边缘点定位不精确的问题,提出了一种基于二次曲面拟合的子像素边缘检测算法。利用拟合法求得子像素图,其抗干扰性强,检测位置较准确,并在此基础上利用改进的Canny算子精确地判断图像边缘位置。实验结果表明,该算法不仅提高了图像的边缘信息保留能力,且图像边缘光滑,边缘点定位精确,细节信息完全。In view of the traditional edge detection algorithm and the peripheral edge information incomplete point positioning inaccurate problem,put forward a kind of based on quadric surface fitting of sub pixel edge detection algorithm. Using the sub pixel figure obtained by fitting,its anti-jamming is strong,position accurately and on this basis,using the improved Canny operator accurate judgement of the image edge position. The experimental results show that the algorithm not only improves the image edge information ability,and the image edge smooth,edge point positioning precision,detail information completely.

关 键 词：子像素 CANNY算子 边缘检测 二次曲面拟合 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]