Open-source algorithm for detecting sea ice surface features in high-resolution optical imagery

摘要

Snow, ice, and melt ponds cover the surface of the Arctic Ocean infractions that change throughout the seasons. These surfaces control albedoand exert tremendous influence over the energy balance in the Arctic.Increasingly available meter- to decimeter-scale resolution optical imagery capturesthe evolution of the ice and ocean surface state visually, but methods forquantifying coverage of key surface types from raw imagery are not yet wellestablished. Here we present an open-source system designed to provide astandardized, automated, and reproducible technique for processing opticalimagery of sea ice. The method classifies surface coverage into three maincategories: snow and bare ice, melt ponds and submerged ice, and open water.The method is demonstrated on imagery from four sensor platforms and onimagery spanning from spring thaw to fall freeze-up. Tests show theclassification accuracy of this method typically exceeds 96 %. Tofacilitate scientific use, we evaluate the minimum observation area requiredfor reporting a representative sample of surface coverage. We provide anopen-source distribution of this algorithm and associated training datasetsand suggest the community consider this a step towards standardizing opticalsea ice imagery processing. We hope to encourage future collaborativeefforts to improve the code base and to analyze large datasets of opticalsea ice imagery.