Online GeoAI Service to Classify Sea Ice Aerial Imagery for Polar and Geospatial Studies

Presenters: Dexuan Sha, Anusha Srirenganathan, Xin Miao, Alberto Mestas-nunez, Hongjie Xie, Chaowei Yang
The polar regions have become increasingly important since they could provide significant potential natural resources, and function as a sensitive indicator to human activities and global, environment, and climate changes, and a key driver of the Earth’s climate. For better understanding, utilizing, and protecting the polar regions, our team have developed 1) a cloud computing-based Arctic cyberinfrastructure (ArcCI) to collect, search, explore, visualize, organize, analyze and share the High Spatial Resolution images which are discrete in time and space; 2) a prototype of sea ice image online service for domain scientists to classify image and extract geophysical parameters. The developed ArcCI is be an ideal platform for integrating existing time-series images. Specifically, the functionalities of ArcCI include image data management, user management, batch image processing, results review and spatiotemporal visualization modules. We invite the communities to help test and evaluate this new platform and provide your professional reviews, comments and real demands.
This session is a hands-on session about how to use the service to process images for sea ice classification through the ArcCI science gateway. Specifically, User will 1) extract customized image collection and load into ArcCI module for management and cloud-sharing; 2) explore and review the pre-calculated geophysical parameters of sea ice in result analysis and overall spatiotemporal view; 3) train and compare object-based image classification results based on different training sets. And more advanced training and labelling request will be collected for further function expansion.