摘要
提出了一种在视觉检测中用于多圆提取的改进随机Hough变换算法,该算法采用一种新的取点方法来进行计算,避免了随机Hough变换由于随机采样造成的大量无效累积。并在计算圆的几何参数时充分利用了圆的几何特征,因此检测速度快,占用内存空间小,精度高,具有良好的抗噪和抗干扰能力。最后文中还分别给出了人工合成图和实物图的实验结果。
An improved randomized Hough transform （RHT） algorithm is proposed for detecting multi-circles in vision inspection. This algorithm uses a new way to get points for calculating the parameters of the circle which avoids the mass useless accumulation caused by RHT, and it uses the geometric features of the circle a lot when calculating the parameters. The algorithm is fast, high precision and takes only a little storage space, performs well against noise as well. At the end of the paper, it also gives the experimental results of synthetic image and real image.
出处
《微计算机信息》
北大核心
2007年第28期288-290,共3页
Control & Automation
基金
广东省自然科学基金资助项目(032482)
关键词
计算机视觉
视觉检测
随机HOUGH变换
圆检测
Computer Vision
Vision Inspection
Randomized Hough Transform （RHT）
Circle Detection
