一种基于视觉测量的亚像素边缘检测算法  

A Subpixel Edge Detection Algorithm Based on Visual Measurements

作　　者：李杰[1] 耿学贤[1] 

机构地区：[1]武汉大学电子信息学院,武汉430072

出　　处：《科学技术与工程》2015年第19期32-36,共5页Science Technology and Engineering

基　　金：国家"973计划"(2012CB725301);国家测绘地理信息公益性行业科研专项(201412015)资助

摘　　要：图像边缘检测精度决定了目标实际尺寸的测量精度。利用亚像素边缘检测技术来解决工业生产中芯片管脚的测量和检测问题。针对传统空间矩算子存在计算量大和边缘定位精度不足等缺点,首先利用改进的Prewitt算子快速提取连续性较好的像素级边缘,然后推导了空间矩算子原理误差并利用改进的空间矩算子进行亚像素级边缘定位。实验结果表明该方法速度快,定位精度高,并且具有较强的抗噪能力。Edge detection accuracy determines the accuracy of the actual size of the target. The using of subpixel edge detection techniques aims to solve the problems of measurement and detection in industrial chips production. In terms of these shortcomings of traditional spatial moment operator, such as intensive computation and deficiency of edge positioning accuracy. Firstly, an improved Prewitt operator was used to extract edge with better continuity. Then principle errors of spatial moment operator weve derived, and improved spatial moment operator was used for subpixel edge location. Experimental results show that the method is fast and has a strong anti-noise capability as well as high positioning accuracy.

关 键 词：边缘检测 一阶微分算子 亚像素 空间矩 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]