Lake Ice Classification from Sentinel-1, 2015-2016, Teshekpuk, Alaska (US)

Weekly lake ice classifications distinguish grounded ice and floating ice in lakes based on Sentinel-1A and Sentinel-1B synthetic aperture radar time-series with a resolution of 23 m in mid to late April, 2015 & 2016. Three versions of the product are provided using three different algorithms: 1) image thresholding (THRESH), 2) unsupervised image segmentation using K-means classification (KMEANS), and 3) unsupervised image segmentation using the Iterative Region Growing with Semantics (IRGSEM) algorithm. All version contain two classes: grounded ice (CLASS 1) and floating ice (CLASS 2).

More information about image processing and classification can be found in the product guide.