Sea ice boundary detection in SAR satellite images using conflicting strength

Abstract

In this paper, a boundary detection algorithm based on conflicting strength between edge and boundary in SAR (synthetic aperture radar) sea ice images is presented. In SAR sea ice images, different ice types can have the same intensity signature but the floe size and shape can be different. To measure the flow size and shape, proper boundary detection is crucial. Due to the inherent speckle noise with SAR satellite images, boundary detection in SAR images can be challenging. The implemented technique eliminates the unwanted edges from the overestimated boundary to obtain the desired boundaries by measuring the strength of each edge and the boundaries to which it belongs. Edges are removed if edge strength does not conflict with the boundary strength. Test results using operational SAR sea ice imagery are presented and results are encouraging.