Radar quantitative rainfall estimation method based on deep learning

Abstract

The invention discloses a radar quantitative rainfall estimation method based on deep learning. The method comprises the following steps: acquiring weather radar data and ground station rainfall data; performing data screening and coordinate correction on the reflectivity data of the elevation angle of the weather radar data; performing data quality control on the rainfall data of the ground station to obtain a relatively accurate rainfall intensity label; carrying out space-time matching on the weather radar data and the ground station rainfall data after pre-data processing, cutting the weather radar data into three-dimensional radar data structures, and respectively taking the three-dimensional radar data structures as input samples and labels of a radar quantitative rainfall estimation model; and generating a complete quantitative rainfall estimation result, and superposing the result on the topographic file to generate accurate regional rainfall information. The rainfall intensity can be accurately estimated based on the radar data, and high-precision regional rainfall intensity estimation is realized.

Description

technical field [0001] The invention belongs to the technical field of remote sensing information processing, and in particular relates to a radar quantitative precipitation estimation method based on deep learning. Background technique [0002] More accurate and real-time Quantitative Precipitation Estimation (QPE) provides a reliable data source for hydrological, meteorological and disaster forecasting. However, the spatial distribution of rainfall processes is very complex, and 70–80% of the uncertainty in land hydrological processes is attributed to the temporal and spatial variability of precipitation, which poses a serious challenge to QPE. The spatial resolution of radar data is about 1 km, which is the best among all remote sensing data, and is often used in the study of QPE. The QPE methods based on radar data are mainly divided into two types: one uses the Z-R relationship, and the other is based on the combined radar-raingauge (R-G) method. The Z-R relationship ...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the problems existing in the prior art, the present invention provides a radar quantitative precipitation estimation method based on deep learning. The radar quantitative precipitation estimation method estimates the precipitation intensity by constructing a deep learning regression model to obtain more accurate regional precipitation intensity information , to solve the problem of large error in the existing radar precipitation estimation method

Images