Raster land cover product derived from TerraSAR-X imagery for the Komakuk Beach study site on the Beaufort Coast, Canada
The dataset contains a high-resolution (5m) land cover classification product for the Komakuk Beach study site located on the Beaufort Coast, Canada. Land cover types were classified using a Random Forest classifier with predictors derived from the Kennaugh elements K0 and K1 of seven dual-polarimet...
Main Authors: | , , , , , , , , |
---|---|
Format: | Dataset |
Language: | English |
Published: |
PANGAEA
2021
|
Subjects: | |
Online Access: | https://doi.pangaea.de/10.1594/PANGAEA.937687 https://doi.org/10.1594/PANGAEA.937687 |
Summary: | The dataset contains a high-resolution (5m) land cover classification product for the Komakuk Beach study site located on the Beaufort Coast, Canada. Land cover types were classified using a Random Forest classifier with predictors derived from the Kennaugh elements K0 and K1 of seven dual-polarimetric HH/HV TerraSAR-X images acquired between July and December 2019. Land cover in-situ data were collected in August 2019 and upscaled to image objects (314 to 4766 m2) using segmentation of a WorldView-3 image and ArcticDEM elevation data. The SAR pixel values within these objects were used as reference data for the land cover classification. This dataset was produced to examine the potential of the Kennaugh Element Framework applied on dual-pol SAR data for Arctic tundra land cover classification. |
---|