Multi-Class Military Target Detection Algorithm Based on Improved YOLOv8

In this paper, a multi-category military target detection algorithm based on improved YOLOv8 is proposed, combining PP-LCNet lightweight convolutional neural network and ParNet attention module in order to improve the detection accuracy and efficiency. For the military dataset containing multi-class targets such as ‘artilleriya’ and ‘tank’, the model is optimised and trained in this paper. The experimental results show that the improved algorithm outperforms the traditional YOLOv8 in terms of accuracy and real-time performance, and is suitable for target detection tasks in real military applications.