Detecting peat extraction related activity with multi-temporal Sentinel-1 InSAR coherence time series
| Autor: | Jaan Praks, Tauri Tampuu, Tanel Tamm, Rivo Uiboupin, Kaupo Voormansik, Ain Kull |
|---|---|
| Přispěvatelé: | University of Tartu, Jaan Praks Group, Tallinn University of Technology, KappaZeta, Department of Electronics and Nanoengineering, Aalto-yliopisto, Aalto University |
| Rok vydání: | 2021 |
| Předmět: |
Synthetic aperture radar
Global and Planetary Change Coherence time Peat Peatland Monitoring Backscatter Coherence (statistics) Management Monitoring Policy and Law SAR interferometry Bog Interferometric synthetic aperture radar Change detection Environmental science Computers in Earth Sciences Coherence Decorrelation Earth-Surface Processes Remote sensing |
| Zdroj: | International Journal of Applied Earth Observation and Geoinformation. 98:102309 |
| ISSN: | 1569-8432 |
| Popis: | Monitoring of when, where and in which quantity peat is harvested is currently based on manual declarations. Synthetic Aperture Radar (SAR) is a powerful tool for change detection and monitoring. The aim of this study was to evaluate whether Sentinel-1 6-day interferometric SAR (InSAR) temporal coherence could allow peat extraction monitoring from satellite. We demonstrate that temporal median coherence enables to detect harvest related surface altering works and therefore also spatially explicitly determine active and inactive extraction areas. A polygon-based multi-orbit time series approach is sufficient for the task. Hereby, vertical–vertical polarisation (VV) is more sensitive to the changes compared to vertical-horizontal (VH). During the main harvest season the peat extraction area has median VV coherence lower than 0.2 while the abandoned area and open bog which serve as reference for undisturbed extraction area have close to 0.6. Also, the potential for coherence based milled peat extraction intensity estimation is demonstrated and an indication is given how partially extracted areas could be distinguished from fully harvested and not harvest areas, by the use of coherence standard deviation. Regarding the influence of rainfall, only heavy rain on one of the acquisitions of the image pair whereas the other is from dry conditions seems to cause decorrelation comparable to surface altering works. Moreover, deploying images from multiple consecutive orbits or introducing backscatter intensity σ 0 or reference polygons of undisturbed area helps to reduce risk for rain induced false positives. Developing an operational algorithm for peat extraction identification could be undertaken in future studies. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect