Development of Tracking and Control System Based on Computer Vision for RoboMaster Competition Robot
Xinyang Tang, Chuntao Leng, Yiheng Guan, Hao Li, Shukun Wu · 2020
In this paper, we present a tracking and control system used to auto-aiming and shooting some given targets. By combining the traditional computer vision method and the neutral network method, our detection for some specific objects achieves real-time (100fps+) and it is robust enough to work on most normal situations. To achieve better performance in shooting moving targets, we fuse the IMU data to do the motion prediction. This makes the system outperforms the human players in hit rate. This system is applied for the RoboMaster competition as an aim-support system and shows good results.