Corner Reflectors — TEXAS STYLE
by Chris Wyatt
In the spring of 2004, a different sort of equipment was put to work in the south Texas brush country. Alongside stock tanks, cattle guards and centerpivot irrigators, five ASF corner reflectors were deployed in an array nearly 150 km in diameter.
And though you might hesitate to label them as state-of-the-art agricultural technology, these corner reflectors and the Radarsat-1 data being acquired over the region will be valuable components for water resources research and management in this region dominated by multi-thousand acre private farms.
The corner reflector installation is part of an ASF effort to support the work of NASA-funded researcher Ni-Bin Chang at the Texas A&M University – Kingsville (TAMUK) Department of Environmental Engineering.
Chang’s research at TAMUK includes using Radarsat-1 Standard Beam SAR data to study soil moisture within the watershed above the Choke Canyon Reservoir southwest of San Antonio. Chang’s Data Acquisition Request (DAR) specifies repeatpass data acquired during 16 consecutive cycles beginning in April 2004.
Coincident with the SAR data collection, TAMUK student, staff and faculty researchers will be in the field to obtain a grid of soil moisture measurements at selected calibration sites.
Prior to the initiation of Chang’s Radarsat-1 campaign, technical staff from ASF and TAMUK discussed the need to coregister the SAR and soil moisture datasets, agreeing that geolocation accuracy was critical to the success of this project.
Ground control points (GCPs) must be identifiable in the imagery, as well as being accessible on foot, in order to obtain Global Positioning System (GPS) data at that site. Because 99 percent of Texas land is privately owned, access to existing features with strong radar signatures observed in previously acquired data could not be taken for granted.
ASF had recently removed several Generation-1 corner reflectors from the field in Alaska and agreed to loan five of these devices to TAMUK for the duration of Chang’s DAR. A corner reflector, when precisely pointed at the satellite in azimuth and elevation, will appear as a very distinct bright object in the resulting imagery.
Using a Geographic Information System (GIS) with cultural, physiographic, and SAR image layers, TAMUK and ASF staff developed an approximate deployment pattern to provide sufficient spatial distribution for coregistering the image data to the GPS points.
The final step in site identification was to find cooperative landowners near the proposed GCPs who were willing to host an inverted 8-foot aluminum trihedron on their property for a year and a half.
Mark Beaman, a research associate at TAMUK, accomplished this via truck, foot and cell phone. With a laptop computer at his side, Beaman, a Texas native, sat down with farm owners and ranch managers, often at the kitchen table, and gave PowerPoint presentations to introduce the project. More than once, the hosts impressed Beaman with their own GIS familiarity, and shared data and shapefiles from their precision agriculture GIS applications.
Corner reflector panels were shipped from Alaska to Texas in March 2004. ASF Advanced Product Developer Chris Wyatt and Calibration Technician Charley Slater traveled to Texas in mid-April, where they joined Beaman and the equipment in time to install the reflectors prior to the first Radarsat-1 acquisition on April 20.
Slater generated reflector-pointing parameters in the field for each reflector using real-time GPS coordinates and a laptop computer running software developed at ASF.
The deployment went smoothly, and then began the waiting as the data was acquired, transcribed and shipped from the Canadian ground station, where it was downlinked, to ASF for processing.
Preliminary analysis of the SAR data collected in April and June indicates that the corner reflector ground control effort will achieve the desired improvement in geolocation accuracy. (Chang’s May data was affected by the Radarsat-1 orbit anomaly.)
All five reflectors are visually identifiable in the slant range and ground range image data. The imagery, once calibrated and projected using ASF software, has been observed to deviate by as much as 500 m from the differentially corrected GPS data and other reliable GIS layers.
By carefully adjusting the image data to the known latitude and longitude of the GCPs, the mean planimetric error can be reduced to less than 1 pixel (12.5 m) without significant impact to the absolute radiometry. This will allow the in situ soil moisture measurements to be confidently correlated with the coincident SAR data.
The work by Dr. Chang and his students will attempt to model soil moisture at scales ranging from a few acres at calibration sites to a watershed-wide model spanning nearly 15,000 km2.
Click here to download a copy of the newsletter featuring this article