郑州大学学报（理学版）2018，Vol.50Issue(1)：77-80,81,5.DOI:10.13705/j.issn.1671-6841.2017058
改进的RANSAC圆检测算法
Improved RANSAC Algorithm for Circle Detection
摘要
Abstract
To solve the problem of long sampling time,poor accuracy and stability of test results yielded by traditional RANSAC(random sample consensus)circular detection algorithm,an improved RANSAC algorithm was proposed to detect circle.Firstly,the image borders were filtered,only leaving longer bor-ders for sampling.Then,when the candidate circle was firstly picked,invalid points were further elimi-nated based on this candidate circle,which greatly reduced the required sampling times.Moreover, the least squares method were used to optimize parameters and choose the best circle among the candidates. Thus,the stability and accuracy of the final results were improved.According to the experimental re-sults,it was shown that the algorithm could fully meet the needs in actual visual inspection with fast speed and good accuracy.关键词
RANSAC算法/圆检测/候选圆/最小二乘法Key words
RANSAC algorithm/circle detection/the candidate circles/the least square method分类
信息技术与安全科学引用本文复制引用
邓仕超,高阳,韩海媚..改进的RANSAC圆检测算法[J].郑州大学学报（理学版）,2018,50(1):77-80,81,5.基金项目
广西制造系统与先进制造技术重点实验室项目:新型无CaF2平场复消色差显微物镜研发(14-045-15-007Z) （14-045-15-007Z）
桂林电子科技大学教育教学改革示范中心专项:布氏硬度压痕直径自动化测量仪器的研制. （）
