面向装配机器人的零件视觉识别与定位方法

摘要

Aiming at the problem of low accuracy of recognition and localization caused by com-plex backgrounds and parts shape diversity in the assembly process,a method of part recogni-tion and localization combining SURF algorithm,FREAK algorithm,Hamming distance and affine transformation is proposed.Firstly,the feature points of the assembly parts are described by SURF and FREAK algorithms,followed by the Hamming distance as the similarity meas-ure of the matching points.Finally,the RANSAC algorithm is used to eliminate the pairs of in-valid matching points.Through the template images and the image of the assembly parts,four parameters affine transformation model is set;combined with affine transformation,the center of mass assembly coordinates is obtained;and then the calibration parameters of the visual sys-tem are used to get the center of mass world coordinates.The experimental results show that this method can realize the rapid and accurate recognition and localization of the assembly parts.%针对装配过程中背景复杂、零件形状多样性等因素导致的识别与定位精度较低的问题,结合加速稳健特征(SURF,speeded up robust feature)算法、快速视网膜特征点(FREAK,fast reti-na keypoint)算法、汉明距离及仿射变换等,给出了的一种零件识别与定位方法.首先通过 SURF与FREAK算法对装配零件的特征点进行描述,其次以汉明距离作为匹配点的相似性度量,最后采用随机采样一致性(RANSAC,random sample consensus)算法进行二次匹配去除无效匹配点对.通过模板图像与装配零件图像建立四参数的仿射变换模型,结合仿射变换求解装配零件的质心坐标,再利用视觉系统的标定参数得到零件质心的世界坐标.实验结果表明,该方法实现了对装配零件快速精准的识别与定位.