摘要
针对机械手对无mark点成品的视觉定位算法运算量过大的问题,提出了一种图像定位NMAK算法。该算法避开了计算量较大的经典图像匹配算法,实现机械手在目标区域快速定位抓取或放置物品,尤其适合对单一标记区域或多标记区域质心重合或无mark点的圆角目标物进行视觉定位。
In view of the problem that the manipulator has a large amount of calculation in the visual positioning algorithm of the finished product without mark,an image positioning NMAK algorithm was proposed. In the algorithm,the classic image matching algorithm was avoided,and the manipulator rapidly positioning fetching or placing objects in the target area was realized. It is especially suitable for the target object that has single tag area or region centroid coincidence or rounded corners with no mark points.
引文
[1]冈萨雷斯.数字图像处理(MATLAB版)[M].北京:电子工业出版社.2007.
[2]江燕,王敬东.图像处理在平面零件分拣控制中的应用[J].电气传动,2015,45(2):76-80.
[3]康牧.图像处理中几个关键算法的研究[D].西安:西安电子科技大学,2009.
[4]桂振文,吴侹,彭欣.一种融合多传感器信息的移动图像识别方法[J].自动化学报,2015,41(8):1394-1404.
[5]张志涌.精通MATLAB 6.5版[M].北京:北京航天航空大学出版社,2009.
[6]晏祖根,李明,徐克非,等.高速机器人分拣系统机器视觉技术的研究[J].包装与食品机械,2014,32(1):28-31.
[7] Patrícia R Oliveira,Roseli A F Romero. Improvements on ICA Mixture Models for Image Pre-processing and Segmentation[J]. Neurocomputing,2007,71(10/11/12):2180-2193.
[8]李志军,傅蓉浣.应用于机器视觉的图像自动比对算法研究[J].机电工程,2009,26(7):53-56.