Iceberg Detection in Satellite Images Using IBM Watson Studio

Abstract

Ship navigation and offshore installations are seriously harmed by icebergs. As a consequence of this, there is a significant interest in timely and extensive localization. Satellite Synthetic Aperture Radar (SAR) images are one of the most commonly used data sources for operational ice conditions and iceberg occurrences due to their independence from daylight and cloud cover. The most common image spatial resolution for iceberg monitoring is between a few and 100 meters. Processed SAR data are characterized by speckle noise, which gives the images a grainy appearance and makes it extremely difficult to identify icebergs. The methods of satellite monitoring of dangerous ice formations, such as icebergs in the Arctic Sea, pose a threat to the safety of Arctic Shelf navigation and economic activity. The creation of a model that automatically determines whether a remotely sensed object is an iceberg is the primary objective of this project. An iceberg is frequently mistakenly categorized as a ship. Because lives and billions of dollars in energy infrastructure are at risk, the algorithm needed to be extremely accurate