舰船识别在UUV电子侦察作战中应用

Application of ship and warship identification in UUV electronic reconnaissance operation

水下无人航行器(UUV)在电子侦察作战中对舰船目标的快速识别有助于掌握战场态势、获取战场主动权,智能化UUV将在海战中扮演更加重要的角色。为了提高UUV的目标识别侦察能力,充分利用其无人、无声、无形等作战优势,发挥非对称作战能力。将具有实时检测、目标识别准确率高并对UUV计算性能要求不高的YOLOv3算法搭载于UUV中,并将YOLOv3算法网络结构进行微调,针对舰船目标特点进一步优化网络模型提高识别能力,并应用于UUV对舰船目标识别侦察,将会大大提高UUV电子侦察作战能力。将模型经过训练优化并在公开舰船数据集上测试,得到有效的舰船目标识别模型,针对多种复杂环境下的目标识别准确率为92.76%。实验数据表明,该模型应用于UUV电子侦察作战检测与识别舰船目标具有显著优势。

The rapid identification of the ships and warships(the targets)by unmanned underwater vehicles(UUV)in electronic reconnaissance operations is helpful to grasp the battlefield situation and gain the initiative on the battlefield.Intelligent UUV will play a more important role in naval warfare.In order to improve the target recognition and reconnaissance capability of UUV,its unmanned,silent,invisible and other combat advantages should be fully taken to exert its asymmetric combat capabilities.The YOLOv3 algorithm with real-time detection ability,high target recognition accuracy and low requirements for UUV computing performance is installed in UUV.And then,the network structure of the YOLOv3 algorithm is fine-tuned.According to the characteristics of the ships and warships(the targets),the network model is further optimized to improve the recognition ability and applied to UUV to identify and reconnoiter the ships and warships(the targets),which can greatly improve the electronic reconnaissance combat capability of UUV.The model is trained,optimized and tested on the public ship data set,and an effective ship and warships(the targets)recognition model is obtained.Its target recognition accuracy rate in various complex environments is 92.76%.The experimental data show that the model has significant advantages when applied in UUV electronic reconnaissance combat detection and identification of the ships and warships(the targets).