| Title | Wetland monitoring using the curvelet-based change detection method on polarimetric SAR imagery |
| |
| Author | Schmitt, A; Brisco, B |
| Source | Water vol. 5, no. 3, 2013 p. 1036-1051, https://doi.org/10.3390/w5031036  Open Access |
| Year | 2013 |
| Alt Series | Natural Resources Canada, Contribution Series 20181098 |
| Publisher | MDPI AG |
| Document | serial |
| Lang. | English |
| Media | paper; on-line; digital |
| File format | pdf |
| Subjects | geophysics; remote sensing |
| Program | Climate Change
Geoscience |
| Released | 2013 07 11 |
| Abstract | One fundamental task in wetland monitoring is the regular mapping of (temporarily) flooded areas especially beneath vegetation. Due to the independence of weather and illumination conditions, Synthetic
Aperture Radar (SAR) sensors could provide a suitable data base. Using polarimetric modes enables the identification of flooded vegetation by means of the typical double-bounce scattering. In this paper three decomposition techniques-Cloude-Pottier,
Freeman-Durden, and Normalized Kennaugh elements-are compared to each other in terms of identifying the flooding extent as well as its temporal change. The image comparison along the time series is performed with the help of the Curvelet-based Change
Detection Method. The results indicate that the decomposition algorithm has a strong impact on the robustness and reliability of the change detection. The Normalized Kennaugh elements turn out to be the optimal representation for Curvelet-based
change detection processing. Furthermore, the co-polarized channels (same transmit and receive polarization in horizontal (HH) and vertical (VV) direction respectively) appear to be sufficient for wetland monitoring so that dual-co-polarized imaging
modes could be an alternative to conventional quad-polarized acquisitions. |
| GEOSCAN ID | 311452 |
|
|