We developed Object Based Image Analysis (OBIA) workflows to map and quantify glacial ice characteristics including ice cover, brash ice, open water, and iceberg size and angularity.

The (OBIA) classification, also referred to as object-oriented (O-O) classification, was carried out using the eCognition suite of software packages (eCognition Architect, Developer, and Server) that together provide an automated and highly scalable solution to image analysis of very large data volumes.

In this study, besides reflectance (color), shape, size, spatial arrangement characteristics of ice blocks provided a basis for the classification of different ice conditions. A key step in an OBIA workflow is the segmentation of the image into coherent image objects.

Schematic of the project workflow, including steps followed in object-based image analysis.