Microwave Radiometer RFI Detection Using Deep Learning

Priscilla N. Mohammed; Jeffrey R. Piepmeier

Home/Archive/NASA NTRS

NASA NTRS · 20210018059

Microwave Radiometer RFI Detection Using Deep Learning

Priscilla N. Mohammed·Jeffrey R. Piepmeier

View original source ↗Export metadata ↓

Abstract

Radio frequency interference (RFI) is a risk for microwave radiometers due to their requirement of very high sensitivity. The Soil Moisture Active Passive (SMAP) mission has an aggressive approach to RFI detection and filtering using dedicated spaceflight hardware and ground processing software. As more sensors push to observe at larger bandwidths in unprotected or shared spectrum, RFI detection continues to be essential. This article presents a deep learning approach to RFI detection using SMAP spectrogram data as input images. The study utilizes the benefits of transfer learning to evaluate the viability of this method for RFI detection in microwave radiometers. The well-known pretrained convolutional neural networks, AlexNet, GoogleNet, and ResNet-101 were investigated. ResNet-101 provided the highest accuracy with respect to validation data (99%), while AlexNet exhibited the highest agreement with SMAP detection (92%).

Explore related subjects

Microwave radiometry radio frequency interference (RFI)deep learning transfer learning

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Priscilla N. Mohammed, Jeffrey R. Piepmeier. 2021-06-23. Microwave Radiometer RFI Detection Using Deep Learning. https://ntrs.nasa.gov/citations/20210018059

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related reports

NASA NTRS2021

Microwave Radiometry at Frequencies from 500 to 1400 MHz: An Emerging Technology for Earth Observations

Microwave radiometry has provided valuable spaceborne observations of Earth's geophysical properties for decades. The recent SMOS, Aquarius, and SMAP satellites have demonstrated the value of measurements at 1400 MHz for observing surface soil moisture, sea surface salinity, sea ice thickness, soil freeze/thaw state, and other geophysical variables. However, the information obtained is limited by penetration through the subsurface at 1400 MHz and by a reduced sensitivity to surface salinity in cold or wind-roughened waters. Recent airborne experiments have shown the potential of brightness temperature measurements from 500–1400 MHz to address these limitations by enabling sensing of soil moisture and sea ice thickness to greater depths, sensing of temperature deep within ice sheets, improved sensing of sea salinity in cold waters, and enhanced sensitivity to soil moisture under vegetation canopies. However, the absence of significant spectrum reserved for passive microwave measurements in the 500–1400 MHz band requires both an opportunistic sensing strategy and systems for reducing the impact of radio-frequency interference. Here, we summarize the potential advantages and applications of 500–1400 MHz microwave radiometry for Earth observation and review recent experiments and demonstrations of these concepts. We also describe the remaining questions and challenges to be addressed in advancing to future spaceborne operation of this technology along with recommendations for future research activities.

Microwave radiometry↗

NASA NTRS2021

Evaluation of Stereology for Snow Microstructure Measurement and Microwave Emission Modeling: A Case Study

Reliable microstructure measurement of snow is a requirement for microwave radiative transfer model validation. Snow specific surface area (SSA) can be measured using stereological methods, in which snow samples are cast in the field and photographed in the laboratory. Processing stereology photographs manually by counting intersections of test cycloids with air–ice boundaries reduces the problems in binary segmentation. This paper is a case study to evaluate the repeatability of the manually stereology interpretation by two independent research groups. We further assessed how uncertainty in snow SSA influences simulated brightness temperature(TB) driven by the Microwave Emission Model of Layered Snowpacks (MEMLS), and how stereology compares to Near Infrared (NIR) camera and hand lens. Data was obtained from two alpine snow profiles from Steamboat Springs, Colorado. Results showed that stereological SSA values measured by two groups are highly consistent, and the ground radiometer measured TBat 19 and 37 GHz was successfully predicted (RMSE<3.8 K);simulations using NIR SSA and hand-lens geometric grain size (Dg)measurements have larger errors. This conclusion was not sensitive to uncertainty in the free parameters of TB modeling.

Microwave radiometry↗

NASA NTRS2019

Impact of Microwave Sounder Calibration on Precipitation for the Global Precipitation Measurement Mission

Cross-track microwave sounders make up a significant percentage of the radiometers included in the Global Precipitation Measurement (GPM) constellation. Therefore, it is important to properly assess the calibration of each sounder instrument and to understand the impact of the calibration on the derived precipitation rates. This ensures an accurate precipitation product is produced for the entire constellation. This paper will use data from past and current microwave sounders to show how offsets in the calibration can impact the precipitation using the GPM Level 2 GPROF algorithm. Potential improvements to the instrument calibration will be assessed by analyzing how they would positively impact the precipitation trends and agreement among the constellation sensors.

Microwave radiometry↗