Identifying Sea Ice Ridging in SAR Imagery using various Machine Learning Models

Abstract

Sea ice ridging presents a great challenge to ships navigating the
arctic. In this paper, we examine the capabilities of various machine
learning methods in predicting regions of high ridge density from
SAR imagery of Hudson Strait. Our results showed that although
ridging in Hudson Strait may be difficult to distinguish even with the
human eye, machine learning can give some insight into potentially
dangerous regions of Hudson Strait.