复杂场景下军事目标的轻量级检测方法

人工智能技术在现代战争中具有举足轻重的地位 ,复杂环境下的军事目标精准识别有 利于使我方抢占先机从而克敌制胜。秉承海空一体化联合作战的重要理念,为使目标检测方 法能够嵌入到各种军事单位中的微电脑中,提出了一种轻量且精准的军事目标检测方法。通 过分析与结合士兵、汽车、与坦克及其履带的特征,设计出一个轻量级的网络单元。使用网 络单元组成一种计算复杂度低且精准的骨干网络,用于提取目标特征信息。设计一个MSCA(multi-scale context aggregation)模块,对骨干网络中高维与低维的特征分别提取,解 决了目标遮挡的问题。实验结果表明,本文方法在军事目标测试集中的准确识别率为97.8% ,与最新的YOLOv4检测方法相比,检测准确度提高了1.1%,运行速度 提高了5倍,能够满足 嵌入式设备实时运行的要求。通过实验可得,本文方法可以实时且精准的检测多种场景下的 军事目标。

Lightweight detection method for military targets in complex scenes

Artificial intelligence technology has a pivotal position in modern war fare.The precise identification of military targets in a complex environment wi ll help us seize the opportunity to defeat the enemy.Adhering to the important concept of sea air integrated joint operation,a lightweight and accurate milita ry target detection method is proposed in order to embed the target detection me thod into the microcomputers of various military units.By analyzing and combini ng the characteristics of soldiers,cars,tanks and their tracks,a lightweight network unit is designed.The network unit is used to form a backbone network wi th low computational complexity and accuracy,which is used to extract the targe t feature information,which improves the detection speed of the method.A MSCA (multi-scale context aggregation) module is designed to improve the high-dimen si onal performance of the backbone network.Separate extraction from low-dimensio n al features solves the problem of target occlusion.Experimental results show th at the accurate recognition rate of this method in the military target test set is 97.8%.Compared with the YOLOv4detection method,the detection accuracy is i mproved by 1.1%,and the running speed is increased by 5times,which can meet t he requirements of real-time operation of embedded devices.Through experiments , the proposed method can detect military targets in a variety of scenarios in re al time and accurately.