• Ayishath Nafiya B
• Ambili M
• Gayathri T
• Dr. Parvathraj K M M
• Prof. Daya Naik

Monitoring sea ice is vital for understanding climate change, ensuring safe marine navigation, and supporting polar research. This project presents an AI-based system that detects and classifies sea ice concentrations using Sentinel-2 satellite imagery focused on the Hudson Bay region. By employing a U-Net convolutional neural network, the model performs detailed image segmentation to categorize sea ice into meaningful classes. The workflow includes automated data acquisition, preprocessing to manage cloud cover and class imbalance, and the use of augmented data to strengthen model learning. The model achieves reliable performance with 83% accuracy and a mean Intersection over Union (IoU) of 0.44. A web-based interface has been developed to make this system accessible for real-time chart generation. The approach demonstrates a scalable and adaptable framework that enhances existing methods of sea ice analysis and supports environmental monitoring and operational decision-making in polar regions.

Sea Ice Segmentation, Sentinel-2 Imagery, U-Net Architecture, Satellite Image Processing, Climate Monitoring, Deep Learning, Convolutional Neural Network, Arctic Navigation, Image Classification, Environmental Mapping.

"SEA ICE DETECTION USING SATELLITE IMAGES", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 7, page no.741-748, July-2025, Available :http://www.jetir.org/papers/JETIRGX06138.pdf

"SEA ICE DETECTION USING SATELLITE IMAGES", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 7, page no. pp741-748, July-2025, Available at : http://www.jetir.org/papers/JETIRGX06138.pdf