Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BARC) » Beltsville Agricultural Research Center » Hydrology and Remote Sensing Laboratory » Research » Publications at this Location » Publication #380088

Research Project: Integrating Remote Sensing, Measurements and Modeling for Multi-Scale Assessment of Water Availability, Use, and Quality in Agroecosystems

Location: Hydrology and Remote Sensing Laboratory

Title: Performance evaluation of UAVSAR and simulated NISAR data for crop/non-crop classification over Stoneville, MS

item KRAATZ, S. - University Of Massachusetts, Amherst
item ROSE, S. - University Of Massachusetts, Amherst
item Cosh, Michael
item TORBICK, N. - Applied Geosolutions, Llc
item HUANG, X. - Applied Geosolutions, Llc
item SIQUEIRA, P. - University Of Massachusetts, Amherst

Submitted to: Earth and Space Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 12/1/2020
Publication Date: 12/8/2020
Citation: Kraatz, S., Rose, S., Cosh, M.H., Torbick, N., Huang, X., Siqueira, P. 2020. Performance evaluation of UAVSAR and simulated NISAR data for crop/non-crop classification over Stoneville, MS . Earth and Space Science. 8, e2020EA00136.

Interpretive Summary: Synthetic Aperture Radar is a valuable remote sensing method for monitoring agricultural status. There is currently a need to develop algorithms for assessing if land is cropland or non-cropland from radar. Using a recent airborne campaign in Mississippi, various algorithms were assessed for accuracy against an in situ validation dataset. Strong accuracies were achieved for resolutions of 30 and 100 meters, while the accuracy of the 10 meter product was not within standard specifications. Radar, while having a high resolution, also contains noise within its signal which must be overcome with resolution degradation.

Technical Abstract: Synthetic Aperture Radar (SAR) data are well-suited for change detection over agricultural fields, owing to high spatiotemporal resolution and sensitivity to soil and vegetation. The goal of this work is to evaluate the science algorithm for the NASA ISRO SAR (NISAR) Cropland Area product using UAVSAR and simulated NISAR data (129A). The NISAR algorithm uses the coefficient of variation (CV) to perform crop/non-crop classification at 100 m. We evaluate classifications using three accuracy metrics (overall accuracy, J-statistic, Cohen’s Kappa) and spatial resolutions (10, 30 and 100 m) for crop/non-crop delineating CV thresholds (CVthr) ranging from 0 to 1 in 0.01 increments. All but the 10 m 129A product exceeded the mission accuracy requirement of 80%. The UAVSAR 10 m data performed best, achieving maximum overall accuracy, J-statistic, and Kappa values of 85%, 0.62 and 0.60. The same metrics for the 129A product respectively are: 77%, 0.40, 0.36 at 10 m; 81%, 0.55, 0.49 at 30 m; 80%, 0.58, 0.50 at 100 m. We found that using a literature recommended CVthr value of 0.5 was suboptimal (65%) and that optimal CVthr values monotonically decreased with decreasing spatial resolution.