视觉检测中圆形标志的定位方法研究

在使用圆形人工标志的计算机视觉检测中,标志的中心定位准确度直接影响到检测的整体准确度。在利用数学形态学算法提取标志边缘的基础上,采用椭圆拟合法、灰度加权质心法、灰度平方加权质心法对标志中心进行定位,并通过实验比较了三者的定位准确度和稳定性。最后得出,在视觉检测中可采用灰度加权质心法来定位标志的中心。

Study on Center Location of Artificial Targets in Vision Detection

During the computer vision detection process of round artificial targets, the accuracy of artificial targets＇ centers strongly influences the accuracy of the whole project. The ellipse fitting method, gray centroid method and gray square centroid method were adopted to locate artificial targets＂ centers, which were detected by mathematical morphology arithmetic. For the purpose of testing the precision and stabilization of these methods, one experiment was conducted. The experimental result show that gray centroid method is best for locating artificial targets＇ centers in computer vision detection.