复杂环境下的多工件定位研究  

Research of multiple workpieces positioning under complex environment

作　　者：刘念[1] 邹湘军[1] 叶敏[1] 罗陆锋[1,2] 王成琳[1] 

机构地区：[1]华南农业大学工程学院,广东广州510642 [2]天津职业技术师范大学机械工程学院,天津300222

出　　处：《电子技术应用》2015年第12期120-123,128,共5页Application of Electronic Technique

基　　金：国家自然科学基金(31571568);佛山市经济科技项目(佛经信〔2015〕5号);广州市科技一般项目(201510010140)

摘　　要：为了获取工业流水线上工件精确的位姿信息,提出一种在复杂环境下同时定位多个工件的方法。首先在复杂环境下采用基于边缘梯度的匹配算法及图像金字塔搜索策略快速识别多个目标,得到目标在图像坐标系中的精确位置信息。然后通过张正友棋盘标定法对单目摄像机进行标定,获得摄像机的内参数及畸变参数。最后,通过优化最小二乘法建立了一种图像坐标系到世界坐标系间的转换关系,实现目标由图像坐标系到世界坐标系中的精确位姿定位。实验表明该方法能在杂乱背景、部分遮挡及非线性光照等环境下,准确地定位工件,鲁棒性较高;识别目标的平均时间在0.5 s以内,工件在x-o-y平面的定位误差小于1 mm、姿态角度定位误差小于1°,可以满足工业应用中复杂环境下的实时性和精度要求。In order to obtain accurate position information of the target on the industrial assembly line,a method of multiple work- pieces positioning under complex environment concurrently was proposed. Firstly, combined with image pyramids search strategy, used a matching algorithm based on edge gradient to identify the target under complex environment and obtain the accurate located information of the objects on the image-coordinate system quickly. Then, using the ZHANG Z Y chess board method for calibration of monocular camera,got the camera＇s internal parameters and the lens distortion coefficient. Finally, establishing a simple, accurate coordinate transformation method, to transform the position of the target in the image coordinate system to the world coordinate sys- tem.Experimental results show that this method can accurately position the target in a complex background, partial occlusion and nonlinear illumination, so it was highly robust.Beside,it can identify industrial targets within an average time of 0.5 s, the position- ing error of the target on x-o-y plane is less than 1 mm and the attitude angle positioning error is less than 1°,which obviously meet the requirements of real-time and accuracy under complicated environment in industrial applications.

关 键 词：单目视觉 边缘梯度 摄像机标定 坐标转换 

分 类 号：TP394.1[自动化与计算机技术—计算机应用技术] TH691.9[自动化与计算机技术—计算机科学与技术]