An Integration of 2-D Mulyi-Looked Differential Interferograms with Persistent Scatterer Interferometry
Kourkouli, Penelope¹; Wegmüller, Urs¹; Wiesmann, Andreas¹; Tansey, Kevin^2
1Gamma Remote Sensing, Gümligen, SWITZERLAND; ²University of Leicester, Department of Geography, Leicester, UNITED KINGDOM

1. INTRODUCTION

Ground deformation can lead to significant problems. Subsidence can cause damage to structures like roads, bridges, wells and dams and can lead to loss of life. The impacts of land subsidence cannot be underestimated because the environmental and economic effects of land subsidence phenomena can vary from negligible to severe depending on the land use nature of the affected area and the subsidence magnitude and spatial extent [1].

Traditionally, techniques such as GPS, precision leveling surveys and extensometer wells were, in the main, used for land subsidence monitoring. In the last two decades Differential SAR Interferometry (DInSAR) was developed providing a new way for low-cost, time-effective, precise ground-motion monitoring, covering both small and extended areas [2][3][4].

Furthermore, in the last years, Persistent Scatterer Interferometry (PSI), which is a pixel based method, has been developed. DInSAR and PSI both have their advantages and disadvantages. One important limitation of PSI is that it does not perform well over non-urban areas. For that reason, spatial coverage is a limitation in some cases and in others it may not be possible to get any interferometric information. To address such limitations, we develop a methodology combining elements from DInSAR and PSI techniques, aiming at achieving a better spatial coverage and subsequently mitigating limitations. Specifically, the approach presented in this paper improved the spatial coverage when compared to the PSI only result.

The Advanced Land Observing Satellite (ALOS) lunched on the 24th January of 2006 [5], carries three instruments, one of which is the Phase Array type L-band Synthetic Aperture Radar (PALSAR). L-band (e~23 cm) is of particular interest as the interferometric signal remains more coherent even over vegetated areas. For the present study, a dataset of 20 PALSAR scenes was used. The data acquired in Fine Beam Single (FBS) and Fine Beam Dual (FBD) mode with HH polarization, in descending orbit and covering the time period between August of 2007 and February of 2011.

A test area, in a semi-arid region in Jordan with rocky parts and steppe areas was selected. The criterion for this selection came up from the research question, namely how well can we improve the spatial coverage of the PSI result in non-urban areas? Thus, the main objective of the current study was to integrate the DInSAR and PSI approaches aiming at a high quality (low phase standard deviation values) interferometric result to achieve better spatial coverage.

Firstly, the DInSAR only approach was applied using one scene as reference. Hence, a data stack of 20 2-D interferograms was generated. Strong multi-looking was performed on the interferograms in order to increase the spatial coverage and subsequently keep an adequate number of possible distributed scatterers into the later analysis. Thereafter, the multi-looked differential phases were extracted into the same vector data format as used for the point phases of a normal PSI processing, in order to make the integration with the PSI approach possible. Moreover, the PSI approach was applied individually using the same reference scene as the DInSAR. The final step was the performance of the combined methodology, which was the combined processing and interpretation of the 2-D multi-looked differential phases and the point based phase in a PSI approach. Finally, in order to determine how well the result was improved using the combined methodology, we compared its result with the PSI only technique.

From the generated first results derived from the PSI and the integrated approach, we noticed that both methodologies worked well. It is observed that compared with the PSI result, the combined result shows a better spatial coverage in many parts of the region (Figure 1). Moreover, it is also noted that the deformation pattern appears similar for both methodologies. Nevertheless, the spatial coverage could be improved in some areas, for instance in the mountainous regions. Future work will focus on multi-reference stack techniques.

Penelope Kourkouli gratefully acknowledges the Marie-Curie fellowship funded by the Marie Curie People Actions under the European Seventh Framework Programme (GIONET project). Data courtesy: PALSAR © Jaxa.

[1] Abidin, H. Z. , Andreas H., Djaja, R., Darmawan D., and Gamal, M., "Land subsidence characteristics of Jakarta between 1997 and 2005, as estimated using GPS surveys," GPS Solutions, vol. 12, no. 1, pp. 23-32, Mar. 2007.

[2] Goldstein, R. M. and Zebker H. A., "Interferometric radar measurement of ocean surface currents", Nature, vol. 328, pp. 707-709, 1987.

[3] Gabriel, A. K., Goldstein, R. M., and Zebker, H. A., "Mapping small elevation changes over large areas: Differential radar interferometry," Journal of Geophysical Research, vol. 94, no. B7, p. 9183, 1989.

[4] Canova, F., Tolomei, C., Salvi, S., Toscani, G., and Seno, S., "Land subsidence along the Ionian coast of SE Sicily (Italy), detection and analysis via Small Baseline Subset (SBAS) multitemporal differential SAR interferometry," Earth Surface Processes and Landforms, vol. 37, no. 3, pp. 273-286, Mar. 2012.

[5] Rosenqvist, A., Shimada, M., and Watanabe, M., "ALOS PALSAR : Technical outline and mission concepts," vol. 1, no. 7, pp. 1-7, 2004.

Figure 1. Subset area over Jordan showing the PSI result (on the left) and the integrated result (on the right).