着陆阶段航空器航迹检测与风险识别方法

进近着陆是整个飞行过程中最易发生危险的阶段.目前着陆阶段的风险识别 只能依靠塔台管制员的目视观察.在飞行员技术不熟练、起降频繁的通用航空飞行训练 中，目视观察很难准确掌握情况，也因此蕴含了巨大的安全风险.利用视频监视航空器着 陆行为，同时提出了一种适合复杂背景下小目标分割算法，从视频中提取航空器着陆航 迹；并以此为基础进一步识别降落过程中的运行风险.通过对多组不同角度的视频监视数 据验证，该算法可在复杂背景下实现航空器着陆航迹的跟踪、落地时刻的判断及危险滑 行航迹的准确识别，对于保障飞行安全具有重要的现实意义.

Aircraft Flight Path Detection and Risk Identification in Landing Phase

The landing is the most dangerous stage of the flight. At present, the risk recognition in landing stage only rely on visual observation of tower controller. In general aviation training when the pilot is unskilled and contains frequently taking off and landing, it is very difficult for visual observation to completely handle the situation, therefore it contains a huge safety risk. In this paper, the data of video surveillance is used to track the landing of aircraft, a small target segmentation algorithm is proposed which is suitable for the complex background at the same time; and based on this, the operational risk of the landing process is further identified. It is verified by the video surveillance data from different point of view, the algorithm can achieve the tracking of aircraft landing path under complex background, judgment of landing moments and accurate recognition of dangerous sliding path, which has important practical significance to ensure flight safety.