OPERA RTC Product, Algorithm, and Validation Approach
Abstract
The Observational Products for End-Users from Remote Sensing Analysis (OPERA) project at the Jet Propulsion Laboratory (JPL) will provide a near-global land-surface Radiometric Terrain Corrected (RTC) product derived from Copernicus Sentinel-1 (S1) synthetic aperture radar (SAR) data.
The baseline algorithm to generate the OPERA RTC product relies on a new area-based projection (AP) algorithm developed for the NASA-ISRO Synthetic Aperture Radar (NISAR) mission and implemented within the open-source ISCE3 framework. The workflow consists of two main steps: 1. radiometric terrain correction (RTC-AP) and 2. geocoding with adaptive multilooking (GEO-AP). Our analysis suggests that the RTC-AP algorithm performs at least comparably, but usually better, compared to state-of-the-art algorithms in terms of accuracy with a significantly shorter run time (6.5 times faster for multi-looked S1 data and 26.3 times faster for single-look S1 data). The geocoding GEO-AP averages all radar samples available for each element of the output grid according to the local topography and radar geometry. Careful verification of results indicates significant improvement of GEO-AP in preserving fine features in the backscatter image compared to previous algorithms. The OPERA RTC algorithm and product are evaluated through a 2-step algorithm verification and product validation (V&V) process. We verify the algorithm by comparing: the normalization factor applied to the RTC product with those obtained from other algorithms; RTC products over ascending and descending satellite tracks; and the flatness of RTC backscatter of different types of target and polarizations with respect to the local topography. The RTC product validation includes assessing absolute and relative geolocation errors. Preliminary algorithm verification shows that gamma-naught depends strongly on the land type and only exhibits near constant behavior in certain forest land types. For that reason, we also validate the OPERA RTC product in forest areas by evaluating the linear regression of the RTC backscatter against the local incidence angle and comparing the backscatter difference over foreslope and backslope areas. The OPERA RTC product will be publicly distributed through the ASF.DAAC without charge, with a release date scheduled for September 2023.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2022
- Bibcode:
- 2022AGUFM.G41A..08S