基于观测与视频数据的机场能见度预测研究

Research on Airport Visibility Prediction Based on Observation and Video Data

能见度预测将直接影响交通运输的安全和效率。能见度检测一般需用昂贵的激光能见度仪,但该仪器对团雾检测精度不高,探测的范围很小,维护成本高,从而在无设备区域易出现安全隐患。因此,如何高效的检测能见度是高速管理部门和航空公司十分关注的问题。文章以大数据技术为基础,以预测各种场景中的能见度为最终目的,分别利用测量数据、视频、图像、时间序列四种数据进行分析和预测,最终得到普适性强、精确度高的最佳方案。实验结果表明模型均达到了较好的表现,为未来机场能见度的人工智能预测系统提供了有益的借鉴。

Visibility prediction will directly affect the safety and efficiency of transportation,and detection often requires an expensive laser visibility meter,which has low detection accuracy for cloud fog,a small detection range,and high maintenance costs with potential safety hazard in equipment-free areas.Therefore,how to efficiently detect visibility is a matter of great concern to high-speed management departments and airlines.Based on big data technology,with the ultimate goal of predicting the visibility in various scenes,using four kinds of data,namely measurement data,video,image,and time series,this paper analyzes,predicts,and finally obtains a universal and high-accuracy solution.The experimental results show that the model has achieved good performance,which lays the foundation for the future artificial intelligence prediction system of airport visibility.