Classification assessment of real vs. simulated Compact and Quad-Pol Modes of ALOS-2
Compact polarimetry (CP) offers a tradeoff with fully polarimetric modes in terms of swath width, power budget, and polarimetric information content. In this letter, a classification comparison is made among real CP, simulated CP (SCP), and quad polarimetric (QP) data acquired from the L-band SAR system onboard the ALOS-2 satellite. The Wishart supervised classification scheme is used to compare data modes over two regions of a mixed test site in India. The quantitative classification assessment indicates that the QP data have higher classification accuracy than any other polarimetric combinations for both regions. The comparative classification accuracy of real versus SCP data is different for the two regions. The overall accuracy of the real CP data is slightly higher ~1% than SCP for region 1, which is dominated by urban and rice classes, whereas it is lower by ~9% for the agricultural crop dominated region 2.
Fig: Wishart supervised classified image of agriculture-dominated area.V. Kumar, Y. S. Rao, A. Bhattacharya and S. R. Cloude, “Classification Assessment of Real Versus Simulated Compact and Quad-Pol Modes of ALOS-2,” in IEEE Geoscience and Remote Sensing Letters, vol. 16, no. 9, pp. 1497-1501, Sept. 2019.
An investigation of inversion methodologies to retrieve the LAI from C-band SAR data
Studies on the sensitivity of microwave scattering to vegetation canopies have led the researchers to conclude that crop biophysical parameters can be modeled from Synthetic Aperture Radar (SAR) backscatter. In this study, we assess different methods of modeling the Leaf Area Index (LAI), an important biophysical indicator of crop productivity, from dual-polarized SAR. Particularly, we evaluate the performance of the Water Cloud Model (WCM) to estimate the LAI of corn using VV and VH backscatter derived from RADARSAT-2 and Sentinel-1 satellites over two test sites (Canada and Poland). We tested the performance of four different approaches to invert the WCM. These are: (a) iterative optimization (IO), (b) Look-up table (LUT) search, (c) Support Vector Regression (SVR) and (d) Random Forest Regression (RFR).
Fig: Validation plots of LAI for IO, LUT search, SVR and RFR based inversion approaches using both RADARSAT-2 and Sentinel-1 data.
The accuracy of each inversion was measured by comparing the estimates from the WCM to the LAI of corn measured in-situ. Our results indicated that the inversion of the WCM using the SVR method delivered the best performance, yielding a correlation (r-value) between estimated and measured LAI of 0.92 and a root mean square error (RMSE) of 0.677 m2 m−2. The other approaches produced higher errors, with the LUT search resulting in the greatest error (RMSE of 0.977 m2 m−2). This study will be of interest to the agricultural sector as this community works towards developing robust methods for tracking crop productivity from SAR technologies across multiple sites and using data from multiple satellite platforms.Dipankar Mandal, Mehdi Hosseini, Heather McNairn, Vineet Kumar, Avik Bhattacharya, Y.S. Rao, Scott Mitchell, Laura Dingle Robertson, Andrew Davidson, Katarzyna Dabrowska-Zielinska, “An investigation of inversion methodologies to retrieve the leaf area index of corn from C-band SAR data”, International Journal of Applied Earth Observation and Geoinformation, Volume 82, 2019, 101893, DOI: 10.1016/j.jag.2019.06.003.