Deep Learning Calibration of the High-Frequency Airborne Microwave and Millimeter-Wave Radiometer (HAMMR) Instrument

Abstract

Calibration plays an important role in improving the accuracy of the microwave and millimeter-wave radiometric measurements. Several calibration techniques have been used in radiometers including external calibration targets, vicarious sources, and internal calibrators such as noise diodes or matched reference load. A new calibration technique based on deep learning has recently been developed to calibrate microwave and millimeter-wave radiometers. The deep-learning calibrator has been previously demonstrated on a computer noise-wave modeled Dicke-switching radiometer. This article applies the new deep-learning calibration technique for the calibration of the high-frequency airborne microwave and millimeter-wave radiometer (HAMMR) instrument. A deep-learning neural network model is built to calibrate the 2014 West Coast Flight Campaign antenna temperature measurements of the HAMMR. The deep-learning calibrator antenna temperature estimates are obtained from the radiometric measurements. The deep-learning calibration results are compared with the existing conventional calibration techniques used in HAMMR 2014 field campaign. The results have shown that the deep-learning calibrator is in agreement with the conventional calibration techniques. In this article, it is demonstrated that the deep-learning calibrator can be employed for calibrating the radiometers with high accuracy.