基于空间矩的视觉图像基元特征提取方法研究  

Extraction of primitive feature of vision images based on spatial moment

作　　者：崔彦平[1,2] 葛杏卫[1] 

机构地区：[1]河北科技大学机械电子工程学院,河北石家庄050018 [2]天津大学精密测试技术及仪器国家重点实验室,天津300072

出　　处：《光电子．激光》2010年第10期1565-1569,共5页Journal of Optoelectronics·Laser

基　　金：国家自然科学基金资助项目(51075119;11002046);教育部博士点专项科研基金资助项目(20030056017);河北省自然科学基金资助项目(E2010000868);河北省科学技术研究与发展计划资助项目(10212111D)

摘　　要：建立了三级灰度图像边缘模型的空间矩算子。首先利用LOG(Laplacian of Gaussian)算子定位速度快的特点,确定图像像素级边缘;然后在包含边缘点的邻域内利用空间矩进行边缘的亚像素定位,由Hough变换提取直线和椭圆边缘像素点;最后采用基于最小二乘原理的直线拟合边缘提取方法,得到亚像素级被检测直线和椭圆。对空间矩算子亚像素定位算法与像面直线和椭圆亚像素提取算法的有效性和精度进行了实验研究,实验结果表明,在测量速度相当的情况下,本文提出的算法具有较高的精度和稳定性。A spatial moment operator of image edge model for the three-gray-level image is proposed.Firstly the Laplacian of Gaussian（LOG） operator is used to locate the edge at the level of pixel quickly,and then the spatial moment is used for subpixel location of the edge in the domain containing edge points.The beeline and ellipse edge pixel points are extracted by the Hough transform.The least square fitting edge extraction is adopted to obtain the subpixel beeline and ellipse.The validity and precision of spatial moment subpixel edge location and the extraction of beeline and ellipse on image plane are studied.The experimental results show that the algorithm proposed in this paper has better precision and stability under a certain measurement speed.

关 键 词：边缘检测 直线提取 椭圆提取 空间矩 HOUGH变换 最小二乘算法 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]