Although the origins of radar can be traced back to the military, since its inception, civilian applications have flourished, especially those related to remote sensing. Applications such as object (e.g., ship or vehicle) detection directly translate from their military counterparts of airborne and ground-based automatic target recognition (ATR). However, most applications involving the remote sensing of the environment fundamentally reverse the way radar backscattering is perceived. In detection and recognition, scattering from any surface other than the object is regarded as “clutter”- undesirable reflections that ought to be removed or suppressed in the data so that the true nature of the object of interest can be ascertained. However, in environmental remote sensing, it is the surface or volume scattering that we seek to understand and exploit. In many cases, remote sensing aims at extracting geophysical properties, which can be related back to the way materials interact with electromagnetic waves. Examples include soil moisture or water concentration, terrain elevation, biomass, mass movement rates, hydrometeor type, plant health, drought tolerance, crop yield, ice layers, and snow thickness. Because deep learning (DL) was originally developed in consideration of real-valued data and optical images, the potential performance, architectures, and optimization of deep neural networks (DNNs) operating on radar remote sensing data must be reassessed. The relationship between geophysical properties, electromagnetic scattering, and the RF data representations used to reveal these relationships creates vast, complex, multidimensional, and time-varying datasets over which DL can be leveraged. Thus, the rich and unique qualities of remote sensing data present new challenges for DL, which has driven much research in this area.

Chapter Contents:

• 11.1 Open questions in DL for radar remote sensing
• 11.1.1 What is the best way to exploit multimodal sensing data?
• 11.1.2 How to satisfy the large data needs for training a DL system?
• 11.1.3 How can prior knowledge on sensor quality be incorporated into joint domain models?
• 11.1.4 What is the best way to leverage data, models, and prior knowledge?
• 11.1.5 How can DL aide in solving multi-temporal processing challenges?
• 11.1.6 How can the big data challenge presented by remote sensing data be addressed?
• 11.1.7 How can DL be used to leverage nontraditional data sources?
• 11.1.8 How can DL be used in automotive autonomy?
• 11.2 Selected applications
• 11.2.1 Land use and land cover (LULC) classification
• 11.2.1.1 Land cover classification
• 11.2.1.2 Land use classification
• 11.2.2 Change detection
• 11.2.3 Ship detection
• 11.2.4 Geophysical parameter estimation
• 11.2.4.1 Drought monitoring
• 11.2.4.2 Precipitation nowcasting
• 11.2.4.3 Snow depth estimation
• 11.2.4.4 Ice layer tracking
• 11.2.5 Radar aeroecology
• 11.2.5.1 Birds in ground radar data
• 11.2.5.2 Birds and insects in meteorological radar data
• 11.3 Additional resources
• 11.4 Concluding remarks
• References

Inspec keywords: radar imaging; image classification; remote sensing by radar; neural nets; geophysical image processing

Other keywords: geophysical properties; radar remote sensing data; environmental remote sensing; civilian applications; military counterparts; deep learning; way radar backscattering; airborne ground-based automatic target recognition

Subjects: Geophysical techniques and equipment; Data and information; acquisition, processing, storage and dissemination in geophysics; Radar equipment, systems and applications; Image recognition; Instrumentation and techniques for geophysical, hydrospheric and lower atmosphere research; Computer vision and image processing techniques