No abstract

No abstract

No abstract

Aiming at the problems of poor consistency and low efficiency in the manual plugging test of 5G RF(Radio Frequency) device, an intelligent plugging system based on robotic manipulation and machine vision is proposed. The mathematical model of the RF device is firstly established. The pixel size calibration and hand-in-eye calibration algorithms are then designed. A positioning algorithm is proposed, consisted of monocular vision and binocular vision. The monocular vision is employed to calculate the position information of the 5G RF device, and consequently, the binocular vision facilitates more accurate positioning, which reduces the positioning error between the actual model and the theoretical model and improves the positioning accuracy. In the binocular ranging, this algorithm improves the traditional Census transform by adopting a double cross window and utilizing the median value of the pixels in the window as the reference value, effectively reducing the impact of noise and meanwhile preserving detailed information. The experimental results show that the average absolute error of the coordinate axis X direction is 0.33mm, the average absolute error of the coordinate axis Y direction is 0.46mm, the average absolute error of the coordinate axis Z direction is 0.27mm, and the plugging success rate is 100%.

No abstract