基于视频识别数据融合的场面监视增强方法

Method of Surface Surveillance Enhancement Based on Fusion of Video Recognition Data

针对场面监视雷达、多点等传统机场场面监视传感器探测机场场面航空器、车辆等目标,出现目标虚假、分裂等现象,造成空管系统目标航迹不可靠影响机场管制运行安全问题。提出了一种基于视频识别数据融合的机场场面目标监视增强方法,方法采用YOLO v5网络模型的深度学习智能视频识别技术识别场面航空器、车辆、行人目标,并对存疑目标再利用云台相机捕捉特写,对初次比对分析的结果进行二次对比分析确认,然后与A-SMGCS系统中传统监视数据融合对比处理,自动对场面目标做真伪、增减处理,剔除虚假目标、弥补丢失目标,增加场面目标监视的可信度、可靠性。为塔台管制员提供真实、可靠、全面的场面航空器、车辆、人员运行态势“一幅图”,提高航班管制运行安全。

For traditional airportsurface surveillance sensors such as SMR and MLAT detection of aircraft,vehicles and other targets on the aerodrome surface,the phenomena of false targets and split targets prone to occur,resulting in the unreliable target track of air traffic control systems,which affects the safety of airport control operations.This research proposed a method of aerodrome surface target surveillance enhancement based on the fusion of video recognition data.This method adopted the deep learning intelligent video recognition technology of the YOLO v5 network model to identify the surface aircraft,vehicles,and pedestrian targets,and the pan-tilt camera was used to capture close-up of the suspicious targets to confirm the results of the first comparison and analysis with a second comparison and analysis.Then the final results were compared with the traditional surveillance data of the A-SMGCS system,and the authenticity,increase and decrease of surface targets were automatically processed,so as to eliminate false targets and make up for lost targets,and improve the credibility and reliability of surface target surveillance.This method provides tower controllers with a real,reliable,and comprehensive "picture" of surface aircraft,vehicles,and personnel operating situation to improve the safety of flight control operations.