【摘要】 基于机器视觉的MOTOMAN机器人轨迹控制方法具有较强的灵活性,可以很好地适用于MOTOMAN机器人涂胶、切割等作业。使用张正友标定法对CCD相机的内参与外参矩阵进行求解,得到图像坐标系与机器人坐标系之间的转换关系;分析背景差分法与二值化法各自的优缺点,提出使用背景差分法与二值化结合检测法对CCD相机所捕捉的目标轨迹图像进行特征提取;对提取的特征图像进行细化与去噪处理;针对提取目标处理后得到的点序,对Motocom32库函数进行封装;机器人调用Motocom32库函数完成对MOTOMAN工业机器人的实时控制。通过试验验证了该方法的准确性。更多还原

【Abstract】 Trajectory control method for MOTOMAN robots based on machine vision has strong flexibility,it can be well applied to glue coating, cutting and other operations of MOTOMAN robot.By using Zhang’s calibration method,the inernal and external parameter matrixes of the CCD camera were calculated, and the transformation relation between image coordinate system and robot coordinate system was analyzed;the advantages and disadvantages of background difference method and binarization were analyzed,the background difference method and binarization detection method were used to extract the features of target trajectory images captured by CCD camera; the feature image was refined and denoised; for the ordered points obtained after processing the extracted target, the Motocom32 library function was encapsulated; the real-time control of MOTOMAN industrial robot was realized by the robot calling the Motocom32 library function. The accuracy of the method was verified by experiments.更多还原