基于双目视觉的分拣机器人立体匹配方法研究
Based on Binocular Vision of Sorting Robot Research on Stereo Matching Method
黄维 1王超越2
作者信息
- 1. 武汉职业技术学院,湖北 武汉 430074
- 2. 武汉理工大学,湖北 武汉 430070
- 折叠
摘要
针对目前汽车零部件分拣机器人的分拣速度慢、精度低等问题,提出了一种将SIFT算法和Harris角点算法相结合的汽车零件图像特征点提取方法.通过阈值自适应方法对筛选后的高对比度特征点进行匹配,通过改进RANSAC算法对误匹配点进行消除,完成分拣机器人立体匹配.通过实验进行对比分析,验证所提方法的有效性.实验结果表明,该方法在保证准确的基础上,简化了计算且有效消除了误匹配,实际测量与计算之间的最大误差为1.21mm,符合分拣机器人的精度要求.这项研究为分拣机器人的发展提供了一定的参考.
Abstract
Aiming at the problems of slow speed and low precision of the current auto parts sorting robot,a method for extract-ing feature points of auto parts images that combines SIFT algorithm and Harris corner point algorithm is proposed.By match-ing high contrast feature points after screening through threshold adaptive method,by improving the RANSAC algorithm to eliminate mismatched points,complete the three-dimensional matching of sorting robots.By conducting comparative analysis through experiments,verify the effectiveness of the proposed method.Experimental results show that,this method simplifies the calculation and effectively eliminates mismatches on the basis of ensuring accuracy,the maximum error between actual mea-surement and calculation is 1.21mm,which meets the accuracy requirements of the sorting robot.This research provides a cer-tain reference for the development of sorting robots.
关键词
分拣机器人/汽车零部件/SIFT算法/Harris角点算法/RANSACKey words
Sorting Robot/Auto Parts/SIFT Algorithm/Harris Corner Algorithm/RANSAC引用本文复制引用
基金项目
中国高教学会职业技术教育分会项目(ZJGA202007)
出版年
2024
NETL
NSTL
维普
万方数据