摘要
研究零件尺寸亚像素测量问题。目前存在的亚像素检测算法精度低、实时性差,不能实现零件图像边缘的精准定位。为提高检测速度、检测精度,提出一种基于Zernike正交矩的亚像素级边缘定位检测的改进算法。采用机器视觉技术获取零件的图像数据,首先利用数学形态法中的四邻域腐蚀法进行边缘点的像素级粗定位,然后利用Zernike正交矩算法对边缘点进行亚像素级重新定位,分析误差并进行误差补偿,以实现高精度的图像亚像素边缘检测。实验结果表明,改进算法能够快速有效完成亚像素级边缘检测。
Study the problem of part size sub - pixel measurement. The precise positioning of the edge of the image of the part can not be achieved using the existing low accuracy and poor real - time algorithms. Start with the key factors that affect the mechanical parts from visual inspection applications - detection rate and accuracy, an improved algorithm was presented based on Zernike orthogonal moments sub - pixel edge location. Machine vision techniques have been introduced to capture dig ital image of parts. Firstly, the algorithm located pixel - level edge points for coarse positioning using four - neighborhood corrosion of the mathematical morphology method, then re - located the sub - pixel level edge points by means of Zernike orthogonal moments algorithm. Finally, the errors were analyzed and dealt with, and the sub - pixel level edge detection of the image was attained. The experimental results show that the algorithm can quickly and efficiently complete the sub - pixel edge detection.
出处
《计算机仿真》
CSCD
北大核心
2014年第2期288-292,共5页
Computer Simulation
基金
国家自然基金(61172185)
天津市高等学校科技发展基金项目(20100705)
关键词
机器视觉
尺寸测量
亚像素
边缘检测
Machine Vision
Size measurement
Sub - pixel
Edge detection