ALOS-Indonesia POLinSAR Experiment (AIPEX): First Result
Mahmud Raimadoya(1), Ludmila Zakharova(2), Bambang Trisasongko(3) and Nurwadjedi Fahmi(4)
(1) Bogor Agricultural University, P.O. Box 2049, Bogor 16020, INDONESIA
(2) Russian Academy of Sciences (RAS), Vvedensky sq., 1, Fryazino 141190, Russian Federation
(3) Bogor Agricultural University, P.O. Box 2049, Bogor 16020, Indonesia
(4) National Mapping Co-ordination Agency, Jalan Raya Jakarta Bogor Km46, Cibinong, Indonesia
The analysis of polarimetric SAR data in the developing countries depend largely on : (1) access to the polarimetric SAR imageries, and (2) availability of the analysis tools. ALOS/PALSAR provides an excellent opportunity for Indonesia to work for the first time with Quad-Pol spaceborne SAR data. Such an opportunity was never available, since there was no access to SIR-C Quad-Pol data of Indonesia (1994).
Since ALOS/PALSAR Quad-Pol mode (PLR) is limited to twice records per two years, this study is directed for maximum exploration of the available PLR. The following approach is implemented : (1) calibration study for the available selected Level 1.1 PLR imageries, (2) first training exercise of POLinSAR application on whatever available of the limited early 2007 PLR pairs (Cycles: 10-11), (3) identification of 2007 PLR pairs coverage/gap in relation to plantation forest and natural forest target plots, to be considered as POLinSAR “super-sites” for 2009 PLR pairs observation request (Cycles: 26-27), and (4) application of POLinSAR analysis upon the available 2007 PLR pairs on a certain super-sites if available. This approach was a flexible adaptation to compromise between the available target plots in accordance to the given ALOS Systematic Observation Plan.
This paper updates the first result on the implementation of calibration, POLinSAR first training exercise and PLR super-sites identification. Faraday rotation effect and quality check (Quegan algorithm) of polarimetric data were applied for calibration purpose on two level 1.1 PLR scenes. The result showed that the Faraday angles derived from models and measured PALSAR scenes were small in general and close enough to each other for both scenes. Furthermore, range plots of amplitude values for cross-talks terms on receive and transmit as well as imbalance on transmit were plotted and compared against value provided by JAXA. Calibration figures extracted from PALSAR scenes were close to JAXA value and considered to be correct. JAXA distorting matrices, therefore, were applicable for Indonesia and no need to uncalibrate the PALSAR scenes and re-calibrate them again. There was no obvious trend of sigma-nought in range for cross-polarized component. Subtle decrease of sigma-nought existed in co-polarized component but within 1 dB absolute calibration error which is acceptable for vegetation application. Based on this positive result, first training exercise of POLinSAR was implemented for an early arbitrary PLR pair of natural forest. The selection of POLinSAR super-sites were completed and to be announced for JAXA attention.