A 1D synthetic aperture microwave radiometer SST inversion method based on deep learning

Abstract

The invention discloses a one-dimensional synthetic aperture microwave radiometer SST inversion method based on deep learning, and the method comprises the following steps: firstly, building a data set for training and verifying a deep learning model, and providing data support for sea surface temperature inversion; Secondly, constructing a deep learning model coupled by an auto-encoder and a fullconnection layer to train the auto-encoder by a training data set, and enabling the auto-encoder to achieve the effect of reducing data errors; Then, connecting a full connection layer behind a decoding layer of the auto-encoder, constructing a complete deep learning model, taking the output of the auto-encoder as the input of the full connection layer, and finely adjusting the whole full connection layer through a minimization loss function so as to invert the sea surface temperature; Finally, verifying the whole deep learning model through the verification data set. The problem that inversion of the sea surface temperature is difficult under the condition of multiple incidence angles is solved, and the inversion efficiency is improved.

Description

technical field [0001] The invention relates to the technical field of remote sensing, in particular to a deep learning-based one-dimensional synthetic aperture microwave radiometer SST inversion method. Background technique: [0002] Sea surface temperature (SST) plays an important role in global climate change and long-term weather processes, among which passive microwave remote sensing can be observed continuously all day and all day long. One of the representative instruments of passive microwave remote sensing is the real-aperture microwave radiometer, which can provide products of various marine environmental elements including sea surface temperature. However, because the spatial resolution of the real-aperture microwave radiometer is limited by the size of the antenna, its spatial resolution is relatively low. Aiming at this shortcoming, a one-dimensional synthetic aperture microwave radiometer is designed, which is different from the mechanical scanning imaging met...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a one-dimensional comprehensive aperture microwave radiometer SST inversion method based on deep learning, which breaks the traditional thinking of real aperture microwave radiometers, and uses the powerful nonlinear fitting of deep learning to solve the problem of one-dimensional comprehensive Aperture microwave radiometer is difficult to retrieve sea surface temperature with multiple incident angles, and it can retrieve sea surface temperature efficiently and with high precision

Images