抵近侦察中的智能目标发现与识别前置应用

Pre-application of intelligent target discovery and recognition in close-in reconnaissance

在远程或近程火力打击中,遥感卫星、远程无人机等可对静止的大型地物及目标进行有效侦察。但是,对于存在伪装的目标或可移动目标,这些侦察设备要么很难有效发现,要么发现了也难以提供及时有效的信息,因此,依托人力或无人平台的抵近侦察必不可少。抵近侦察中,现有“前端获取”和“后端处理”的应用模式在效率上存在时效性问题,而快速高效地发现、匹配及识别伪装或移动目标,是先敌构建“观察、判断、决策、行动”(OODA)回路并形成作战优势的必要条件。为此,本文提出一种新的由智能技术赋能的目标发现、匹配与细粒度识别的前置应用模式,并就该应用模式下的关键技术及进展进行了综述分析,展望探讨了该应用模式存在的可能问题及应对策略。所提模式是新技术、新设备推动下面向陆战抵近侦察应用的前瞻性构想。

In a long-range or a short-range fire attack,remote sensing satellites and long-range UAVs can carry out effective reconnaissance on large stationary ground objects and targets.However,the use of this type of reconnais‐sance equipment is inefficient for effectively finding and providing timely and effective information for camouflaged or mobile targets.Close-in reconnaissance relying on a human or unmanned platform is therefore essential.In close-in reconnaissance,the existing“front-end acquisition”and“back-end processing”application modes have serious prob‐lems regarding time efficiency.To build an“Observation,Orientation,Decision and Action”(OODA)loop and gain an operational advantage,it is necessary to discover,match and identify camouflaged or mobile targets faster than the enemy.In this paper,we propose a new pre-application mode for target discovery,matching and fine-grained recogni‐tion enabled by intelligent technology.The key technologies and their literature are summarized and analyzed,as well as possible problems with the mode and countermeasures are proposed and discussed.The proposed mode is an ad‐vanced concept for land combat close-in reconnaissance applications driven by new technologies and new equipments.