基于视觉引导的工业机器人快速分拣系统研究

摘要

Combined with MELFAF RV-13F industrial robot,vision guiding technique is applied to object sorting.After filtering the source image and ORB feature detection,the best matching feature points can be used to calculate the pixel coordinates and rotation angle to implement fast detection of the object.The results show that the error between calculating and actual coordinates in X,Y is within 1mm and in Z is within 3 mm.It ensures that the industrial robot can grasp object precisely to complete object sorting fast.%将三菱RV-13F工业机器人与视觉引导技术相结合应用于目标分拣.对原始图像进行滤波处理后再利用ORB算法实现对目标物体的快速检测,并利用最佳匹配特征点计算出目标物体的像素坐标和旋转角度.实验结果表明:计算坐标与实际坐标在X、Y坐标上的误差在1 mm之内,可以保证工业机器人准确地运动到目标物体的上方,Z坐标上的误差在3 mm之内,可以保证能将目标物体抓取到,快速地完成分拣任务.