摘要
从图像中准确提取光条中心对于主动视觉系统十分重要。分析微小物体内表面视觉检测系统中圆形结构光光条图像特征,将干扰光条分为漫反射和镜面反射两类。利用Hessian矩阵确定图像中光条法线方向,在光条法线方向上求解光条中心亚像素位置,用8方向搜索法剔除干扰光条。实验证明,光条提取和剔除干扰光条方法是有效的。
Accurately extracting light stripes center from an image is very important for an active vision system. The image characteristics of circular structured light stripe in a micro-object inner surface vision inspection system are analyzed and the false stripes can be divided into diffuse reflection and regular reflection. The normal directions of light stripes in an image are determined by using Hessian matrix. The sub-pixel positions of light stripes can be found in normal directions. The false light stripes can be eliminated with 8-direction searching method. The experiments demonstrate that light stripes extraction and false light stripes rejection are effective.
出处
《光电工程》
EI
CAS
CSCD
北大核心
2004年第4期46-49,共4页
Opto-Electronic Engineering
基金
国防基础科研计划资助项目
