ALOS-PALSAR performances on a multiple sensor DInSAR scenario for deformation monitoring applications
Roman Arbiol(1), Pablo Blanco(1) and Vicenç Palà(1)
(1) Intitut Cartografic de Catalunya, Parc de Montjuic s/n, 08038, Spain
The Institut Cartogràfic de Catalunya (ICC) has developed its own DInSAR processing chain during the past years. Initial interferometry topographic applications have evolved to differential interferometry with many images for terrain deformation monitoring purposes. These developments have allowed the successful detection of different deformation phenomena employing ERS and ENVISAT stacks of interferograms.
The appearance of new sensors, as ALOS-PALSAR, in combination with the already existent, encourages the information merging for exploiting their different capabilities in order to maximize the amount of derived useful knowledge. Especially in non-urban environments, the employment of the L-band becomes important for getting more coherent areas, which are generally rejected when working with shorter wavelengths. Furthermore, ALOS-PALSAR offers the possibility of using different polarimetric channels.
Consequently, the use of ALOS-PALSAR interferograms stacks open new observation challenges both individually and also combined to other sensors information. According to the ICC’s acquired experience working with C-band (ERS, ENVISAT) and recently with X-band (TerraSAR-X) sensors, the ALOS-PALSAR performance is investigated within a multiple sensor framework for deformation monitoring applications. Several aspects will be addressed:
- Band influence on coherent area detection. Depending on the terrain characteristics, it is expected to get a different return from each of the employed bands. This opens new possibilities in order to increase the coverage of the observed area.
- Exploitation of the different ALOS-PALSAR polarimetric channels. Coherence maps in each channel may differ from one polarization to the other. Consequently, the one with higher coherence could be chosen for retrieving useful information. The impact of this selection can be quantified in the final deformation map.
- ALOS-PALSAR combination with different sensor information. The first approximation will consist in comparing ALOS-PALSAR results to ERS-ENVISAT results, both obtained as independent data-sets. From this starting point, compatibility conditions will be investigated. In order to exploit temporal redundancy, it is convenient to process the whole data set (coming from different sensors) simultaneously on the multiple images DInSAR algorithm. According to the first part conclusions, combination strategies will be investigated and addressed.