陆战场装甲目标细粒度识别技术

Research on fine-grained recognition of armored targets in land battlefield

智能化作为新的改变战争游戏规则的颠覆性技术,成为世界各军事强国的发展重点。当前,智能化可以实现对陆战场军事装甲目标的自动识别,但是现有识别任务主要是粗粒度的目标类别识别,缺乏细粒度的目标识别。本文针对陆战场军事装甲目标细粒度识别任务,提出了一种适用于无人作战平台的轻量化细粒度目标识别方法,并初步验证了方法的可行性和有效性,基于此,提出了无人作战平台的新的作战运用。该方法可辅助操作人员进行陆战场装甲目标识别的决策判断,或提供无人打击武器自主打击的判断能力。

Intelligent warfare,as a new disruptive technology to change the rules of war games,has become the focal point of the development of the world’s powerful military countries.At present,intelligent technology can realize the automatic recognition of military armored targets in land battlefield.However,the existing recognition tasks are mainly coarse-grained category recognition and lack of fine-grained sub-category recognition.Therefore,this paper explores the task of fine-grained recognition of military armored targets in land battlefield,a light-weighted fine-grained recognition method suitable for unmanned combat platform is proposed,and preliminarily verify the feasibility and effectiveness of the method.Besides,a new operational usage is proposed for unmanned combat platform.The proposed method can assist operators in decision-making and judgment of armored target recognition in land battlefield,or provide the judgment ability of unattended attacking weapon.