No abstract

No abstract

Ship traffic in ice exposed areas is increasing, and ice navigation is largely a manual task. Despite the progress in machine learning and computer vision algorithms, little focus has been given to computer-aided scene understanding in icy waters to help modernize navigational support systems. This work lays the foundation for the automated identification of ice for surface vessels using modern deep learning (DL) algorithms. The focus is on locating and classifying multiple ice objects within images from a surface vessel travelling through icy waters. The following categories of surface ice features are considered: level-ice, deformed ice, broken-ice, icebergs, floebergs, floebits, icefloes, pancake-ice, and brash-ice. In the first phase, we used DL algorithms to classify the ice objects in an image. For this task, seven state-of-the-art residual network (ResNet) models have been tested and include ResNet18, ResNet34, ResNet50, SE_ResNet50, Xception-Cadene, Inception-v4, and Inception-ResNet-v2. During the second phase, we used DL algorithms to locate and classify ice objects. For these tasks, we used the UNet architecture combined with conditional random fields (CRFs) and analysed the effects of using fully connected CRF and convolutional CRF. We have trained and validated the models using the close-range optical ice imagery, and then the promising models were used to classify and locate the different ice features in images from the bridge of the US Coast Guard icebreaker Healy and the nuclear-powered icebreaker 50 Let Pobedy. This paper provides the main findings and lessons that were learned from the execution of this study.

Ice charts play an important role in the planning of marine operations, including navigational guidance among other use cases (e.g., climate monitoring and model validation). With a growing number of vessels operating in dynamic sea ice cover and considering the November 2021 events when several ships were stuck and delayed in the Arctic waters, it becomes ever more important to have accurate and timely ice information as well as to account for the underlying uncertainties in the sea ice products. To this end, the present study evaluates the variability in estimated total sea ice concentration in ice charts of the Russian Arctic and Antarctic Research Institute (AARI) and the Norwegian Meteorological Institute (MET Norway). The weekly ice charts from AARI were compared with several daily charts from MET Norway for the corresponding week to discover any discrepancies in the reported sea ice concentration. Preliminary results of this study indicate seasonal as well as spatial trends in the absolute difference in total ice concentration between the two sea ice products. A higher difference in concentrations was observed in the western and the central regions of the Kara Sea which see a lot of ship traffic. Sensitivity of the results to the comparison approach is conducted and the found discrepancies between the two ice products are placed in the context of operational route planning.

This study investigates whether the vessels remain within their operational limitations in ice using the risk index calculated based on the Polar Operational Limitations Assessment Risk Indexing System (POLARIS) — an acceptable methodology for the assessment of operational limitations in ice infested waters, referenced in the Polar Code of the International Maritime Organization (IMO). The speeds and positions of the vessels in the Kara Sea region were analyzed from January through April for 2017–2019 using the navigational data provided by the Northern Sea Route Administration. For each vessel, except for the icebreakers, the risk index based on POLARIS was calculated using the open-access ice information that was provided by the Arctic and Antarctic Research Institute in Russia. The variation of risk index was analyzed with respect to various parameters such as the ice-class of the vessel, the reported operating speed of the vessel, and the built year of the vessel. Furthermore, we explored the limitations of the risk assessment system as well as the limitations of the available ice information and its implications on the risk assessment system. This paper reports preliminary results from the analysis.