REAL-TIME AUTOMATIC TARGET DETECTION AND TRACKING MODEL BASED ON CONVOLUTIONAL NEURAL NETWORK FOR MACHINE GUN SYSTEM

Abstract

In this paper, we construct a simulation model of an automatic target
detection and tracking machine gun application for military aircraft using 3D
printing technology. Simultaneously, we propose an algorithm for automatic
target detection and tracking using a convolutional neural network model,
specifically based on the YOLOv8 model, to enhance the speed and accuracy
of target identification. We utilize an STM32 microcontroller to control the
target tracking motion system. Experimental results demonstrate that the
target recognition system achieves a high accuracy rate of 97.86% and
effectively tracks moving targets with speeds below 0.5m/s.