Image Processing for Ice Floe Analyses in Broken-ice Model Testing

Abstract

The ice floe shape and size distribution are important ice parameters in ice-structure analyses. Before performing an analysis at full scale, the dynamic positioning (DP) experiments in model ice at the Hamburg Ship Model Basin (HSVA) allow for the testing of relevant image processing algorithms. A complete overview image of the ice floe distribution in the ice tank was generated from the experiments. An image processing method based on a gradient vector flow (GVF) snake and a distance transform is proposed to identify individual ice floes. Ice floe characteristics such as position, area, and size distribution are obtained. A model of the managed ice field's configuration, including identification of overlapping floes, is also proposed for further studies in ice-force numerical simulations. Finally, the proposed algorithm is applied to an ice surveillance video to further illustrate its applicability to ice management.