Design and Implementation of Automatic Aiming System Based on YOLO

Automatic aiming system plays an important role in the military domain. Aiming at the present situation of automatic aiming system, a new control method for automatic aiming system is proposed. Firstly, Tiny YOLO (You Only Look Once) target detection algorithm is used to train and model for the target object, meanwhile the trained model is used to recognize and locate the target object. For the tracking failure problem caused by the short-term disappearance of the target because of high mobility, the Kalman filter algorithm is used to filter the target position and predict target motion. Also, the classical PID (Proportion Integration Differentiation) algorithm is used to control the automatic aiming system and realize the real-time tracking and aiming of the target object. Finally, the automatic aiming battery system was built for experiments and to monitor the running state of the system, and achieve accurate shooting within 3 meters of the battery. The results show that the system can accurately aim at the central position of the target object in the complex background environment, achieving expectations and having practical values.