Time-series Meteorological Scene Recognition Method in Terminal Area Based on Spectral Clustering

Abstract

The invention belongs to the technical field of civil aviation traffic control, and in particular relates to a time-series meteorological scene recognition method in a terminal area based on spectral clustering, which includes: obtaining weather avoidance area data within the calculation range of the target terminal area and airport visibility data; defining and calculating Rasterization with weights; forming time series RSI Data set and time-series visibility data set; construct weather scene similarity matrix; and train spectral clustering model for identifying weather scene to obtain the final time-series weather scene recognition result, realizing that the terminal area weather scene recognition can not only be operated as a historical flight The basic support of analysis can also provide decision-making basis for the implementation of air traffic flow management strategies under the influence of convective weather and other meteorological conditions. The identification of meteorological scenes in the terminal area can help improve the safety of flight operations and reduce the abnormality of flight operations.

Description

technical field [0001] The invention belongs to the technical field of civil aviation traffic control, and in particular relates to a time-series meteorological scene recognition method in a terminal area based on spectral clustering. Background technique [0002] Meteorology is an environmental characteristic of aircraft operation, which will have a significant impact on airspace and ground flight operations. Severe convective weather affects the capacity and flow of the airspace, and flights will be diverted, delayed, and canceled due to a series of traffic control strategies issued by the air traffic control department, such as ground waiting and crossing restrictions. According to the impact of weather on the operation of the terminal area, the terminal area can be identified as different operation scenarios to represent the different types and degrees of interference and influence brought by the weather on the operation of the terminal area. The meteorological scene re...

AI Technical Summary

Problems solved by technology

However, these techniques do not consider the time-series changes of meteorological impacts, and can only give instantaneous meteorological impact scenarios

However, air traffic flow management strategies, controller work shifts, and historical flight operation analysis are all carried out in units of hours. The existing meteorological scene recognition methods in the terminal area cannot fully meet the needs of the civil aviation traffic control field, and the actual reference value is relatively low. weak

Moreover, the previous methods lack reasonable dimension reduction or information processing schemes, and it is difficult to fully utilize ultra-high-dimensional radar weather information.

Images