Space inversion method and system for space-borne remote sensing water vapor based on neural network

Abstract

The invention provides a neural network-based space-borne remote sensing water vapor space inversion method and system, which belongs to the technical field of water vapor inversion prediction, and the scheme includes: obtaining space-based remote sensing observation data, land cover type data, and ground-based GNSS water vapor Observational data; based on the linear interpolation algorithm, the three types of data obtained are matched globally in time and space to obtain matching data on a global scale; based on the location information of the area to be retrieved by water vapor and the matching data on a global scale, determine the current Matching data corresponding to the location information; input the location information and the corresponding matching data into the pre-trained deep learning model, and output the water vapor inversion result of the area to be water vapor inversion.

Description

technical field [0001] The invention belongs to the technical field of water vapor space inversion prediction, and in particular relates to a space-borne remote sensing water vapor space inversion method and system based on a neural network. Background technique [0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art. [0003] Although a large number of neural network algorithms for PWV (Precipitable Water Vapor) inversion have been proposed and have achieved good results, most of these algorithms are obtained through such as NCEP (National Center for Environmental Prediction) and ECMEF (European Central Scale Weather Forecast Center) reanalysis data established. [0004] Ground-based GNSS (Global Navigation Satellite System) observations can provide higher accuracy and time resolution PWV values ​​at fixed stations than reanalyzed data. However, the inventors found that the ...

AI Technical Summary

Problems solved by technology

However, the inventors found that the scattered distribution of GNSS stations reduces the spatial versatility of the algorithm. The existing joint spaceborne remote sensing data and ground-based GNSS data, and the neural network algorithm based on ground-based GNSS stations are usually only in fixed GNSS stations. However, for areas without GNSS stations, the neural network algorithm cannot be used to perform accurate PWV inversion due to the lack of historical data of GNSS stations; at the same time, due to the complex land cover in space Types, especially in microwave inversion, the complexity of land cover types leads to the variability of microwave emissivity parameters, which in turn leads to the establishment of algorithms based on fixed stations are still point-domain algorithms, not extended to the space domain

Images