Rain, snow and hail classification monitoring method based on semi-supervised domain adaptation

Abstract

The invention discloses a rain, snow and hail classification monitoring method based on semi-supervised domain adaptation. Preprocessing data of each type of precipitation particle weather are obtained according to radar wave reflectivity under various types of precipitation particle weather, a first data set carrying a label and a second data set not carrying a label are constructed, a first covariance matrix and a second covariance matrix are calculated, a first feature subspace is determined, a second feature subspace is determined, a kernel function is determined according to the first feature subspace and the second feature subspace, according to the kernel function, an initial classifier is trained by taking the first data set as a training sample set, unsupervised learning is carried out on the initial classifier, so that the initial classifier adapts to a target field, obtains and determines an adjacency graph and optimizes a target function to determine a final classifier, thefinal classifier is adopted to classify rain, snow and hail, accurate classification can be carried out on rain, snow and hail, and a corresponding classification monitoring scheme has higher accuracy.

Description

technical field [0001] The invention relates to the technical field of surface meteorological detection, in particular to a rain, snow and hail classification monitoring method based on semi-supervised domain adaptation. Background technique [0002] For regions or countries with concentrated rainy seasons and frequent rainstorms, the abnormal temporal and spatial distribution of precipitation is an important factor that causes natural disasters such as floods, landslides, and debris flows. The research on precipitation has already exceeded the scope of scientific research. When measuring and studying precipitation, it is first necessary to distinguish the types of precipitation particles—rain, snow, hail, etc. At present, the identification of rain, snow, and hail is mainly based on the method of weather radar volume scan data and dual polarization Doppler radar polarization parameters. Among them, the method of body scanning with weather radar is relatively simple, but th...

AI Technical Summary

Problems solved by technology

The existing microwave identification methods for rain, snow, and hail require a large amount of labeled microwave attenuation data using traditional machine learning methods. However, in actual scenarios, the amount of microwave attenuation data with precipitation type labels is often not sufficient, so Classified monitoring solutions for rain, snow, and hail often have low accuracy

Images