Towards Pan-Arctic Sea Ice Type Retrieval using Sentinel-1 TOPSAR modes
Sea ice monitoring has been subject to intense attention over the last few decades. Besides the scientific interest in sea ice, the operational aspect of ice charting is becoming more and more important due to increasingly ice free Arctic, resulting in growing navigational possibilities. Widely used...
| Main Author: | |
|---|---|
| Format: | Conference Object |
| Language: | English |
| Published: |
IEEE
2021
|
| Subjects: | |
| Online Access: | https://elib.dlr.de/134597/ https://elib.dlr.de/134597/1/2021%20EUSAR%20Singha.pdf https://ieeexplore.ieee.org/document/9472667 |
| Summary: | Sea ice monitoring has been subject to intense attention over the last few decades. Besides the scientific interest in sea ice, the operational aspect of ice charting is becoming more and more important due to increasingly ice free Arctic, resulting in growing navigational possibilities. Widely used daily pan-Arctic sea ice concentration maps are mainly derived from space-borne microwave radiometer data, with a typical spatial resolution of dozens of kilometers which are rather inadequate for navigational purposes. Since last few years, Sentinel-1a/b providing unprecedented spatial and temporal coverage of entire Arctic in C-band with its Extended Interferometic Wide Swath (EW) mode. Despite proven sea ice classification achievements on ’ScanSAR’ type Synthetic Aperture Radar (SAR) images, a fully automated, operational classifier for has not been established due to large variation in sea ice manifestations and incidence angle induced impacts. Here we propose a methodology for Pan-Arctic sea ice type retrieval using Sentinel-1 (EW, HH-HV) dataset which accounts for the noises and incidence angle related impacts. Proposed supervised classification algorithm consists of two steps: The first step comprises of preprocessing, mosaicing and texture based feature extraction, the results of which are used to train a Support Vector Machine based classifier in the second step and used for subsequent sea ice type retrieval at pan-Arctic scale. Test results from the dataset acquired over the Northeast Greenland and Fram Strait showed that the classifier is capable of retrieving three broad ice types (Open Water, First Year Ice, Young Ice) with an overall accuracy of 99%. |
|---|