CeADAR Online Tech Talk: Applications of RADARSAT Constellation Mission (RCM) for Crop Monitoring – Saeid Homayouni, 11th March (Online Recording Available)

Date: 11th March at 3 PM

Speaker: Saeid Homayouni – INRS

Title: Potential Applications of RADARSAT Constellation Mission (RCM) Compact Pol SAR EOs for Crop Monitoring and Mapping

Abstract:

Early- and mid-season crop information is essential to promote agricultural and environmental sustainability, mitigate, and decrease global and regional food security threats. Accordingly, the spatial distribution of agronomy activities, environmental conditions such as soil moisture and temperature, and up-to-date crop inventories are critical for sustainable management of this crucial economic sector. Adopting optical vegetation indices for this purpose is widespread, yet cloud cover impedes the acquisition of these data. Although early research using scatterometers and aircraft hinted at the sensitivity of Synthetic Aperture Radar (SAR) responses to crop development, the implementation of satellite SAR observations in operational crop condition monitoring is limited. Despite recent research on the potential of dual- (DP) and full-polarimetry (FP) SAR data for crop mapping, the capability of compact polarimetry (CP) SAR data has not yet been thoroughly investigated. Furthermore, a much-needed investigation is required to fully understand the differences and similarities between extracted featured from various SAR polarization data (i.e., DP, CP, and FP) for agricultural applications. This is of particular concern, given the availability of such data from the upcoming RADARSAT Constellation Mission (RCM).
In this paper, two agriculture applications of C-band Compact Polarimetric SAR imagery are presented. The first application is to model a SAR-based vegetation indicator, i.e., volume-to-surface (V/S) scattering ratio, derived from simulated CP C-band RADARSAT-2 imagery. The second application uses several multitemporal CP features for crop type mapping using the advanced machine learning approaches. To this end, both pixel- and object-based Random Forest Classifier (RFC) are applied to these CP features. These features showed advantages for a large-scale crop mapping. Because CP SAR EOs collected by RCM will have a wider swath width and an enhanced temporal resolution, these characteristics are of great significance for producing annual crop inventories on regional and national scales, compared to the FP SAR data collected by RADASAT-2.

Online Recording: https://youtu.be/s0L59c0zPHw