基于机器视觉和力反馈的自动装配技术研究

Research on Automatic Assembly Technology Based on Machine Vision And Force Feedback

针对工业生产中自动装配技术装配精度不高的问题,提出了一种基于机器视觉和六维力传感器的自动装配控制方法。使用两个单目摄像头对目标进行两次定位,通过改进的NCC匹配算法进行一次定位,通过基于边缘特征的模板匹配进行二次定位,利用六维力传感器获取装配过程中的力与力矩变化情况,并基于反馈的力与力矩提出了直线运动与螺旋线运动两种装配轨迹规划策略。在搭建的机器人平台上对两种策略进行了对比实验,实验结果表明,在轴孔间隙较大时直线运动装配效率较高,但当轴孔间隙小于0. 1 mm时,直线运动的装配效率和成功率均大幅下降,而螺旋线运动的装配时间主要与装配孔直径有关,对于不同轴孔间隙装配表现稳定,并能以较高成功率实现精度0. 05 mm的轴孔装配。

In order to solve the problem of low assembly precision of automatic assembly technology in industrial production, an automatic assembly control method based on machine vision and six-dimensional force sensor is proposed. Two monocular cameras were used to locate the target twice. The NCC matching algorithm was used for positioning, and the template matching based on edge features was employed for secondary positioning.The force and torque changes during assembly were obtained by the six-dimensional force sensor, and two assembly trajectory planning strategies for linear motion and spiral motion were proposed based on the feedback of force and torque. The two strategies were compared on the built robot platform. The experimental results show that the assembly efficiency of linear motion is higher when the shaft hole clearance is relatively large, but when the shaft hole clearance is less than 0.1 mm, the assembly efficiency and success rate of the linear motion are greatly reduced. However, the assembly time of the spiral motion is mainly related to the diameter of the assembly hole, the assembly performance is stable for different shaft hole clearances, and the spiral motion can achieve shaft hole assembly with an accuracy of 0. 05 mm at a high success rate.