Abstract
Sea ice is important to study as it reflects climate change and the state of the environment. ArcCI is a utility which allows users to upload images of sea ice from online databases which can be cropped into smaller resolutions, which are segmented using a variety of algorithms via borders (e.g. the border between ice and water). The user can then categorize each segment as a different property (e.g. ice or water). Most standard methods have difficulty determining the difference between melt ponds and submerged ice, but through Polygon Neighbor Analysis, the regions surrounding a body of water with ice underneath can be used to determine the feature. This method can only be implemented with HSR (high resolution images), which are much larger in size and therefore more challenging to optimally process. The user interface is being updated alongside the development of a desktop variant of the app. Additionally, an artificial intelligence model is being trained by interns and the ArcCI development team to automate the process of segment labeling with the hope that the full process could be automated. With the assistance of Abishek Kanthan, an export script was added, giving users a method to extract the local data into a zip file to be shared for training.
How to Cite:
, P., , A., , T., , G. & , C.,
(2023) “ArcCI Development ”,
Journal of Student-Scientists' Research 5.
doi: https://doi.org/10.13021/jssr2023.4002
Files
Downloads are not available for this article.