Performance Evaluation of Deep Learning Image Classification Modules in the MUN-ABSAI Ice Risk Management Architecture
The retreat of Arctic sea ice has opened new maritime routes, offering faster shipping opportunities; however, these routes present significant navigational challenges due to the harsh ice conditions. To address these challenges, this paper proposes a deep learning-based Arctic ice risk management a...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-01-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/25/2/326 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832587547067285504 |
|---|---|
| author | Ravindu G. Thalagala Oscar De Silva Dan Oldford David Molyneux |
| author_facet | Ravindu G. Thalagala Oscar De Silva Dan Oldford David Molyneux |
| author_sort | Ravindu G. Thalagala |
| collection | DOAJ |
| description | The retreat of Arctic sea ice has opened new maritime routes, offering faster shipping opportunities; however, these routes present significant navigational challenges due to the harsh ice conditions. To address these challenges, this paper proposes a deep learning-based Arctic ice risk management architecture with multiple modules, including ice classification, risk assessment, ice floe tracking, and ice load calculations. A comprehensive dataset of 15,000 ice images was created using public sources and contributions from the Canadian Coast Guard, and it was used to support the development and evaluation of the system. The performance of the YOLOv8n-cls model was assessed for the ice classification modules due to its fast inference speed, making it suitable for resource-constrained onboard systems. The training and evaluation were conducted across multiple platforms, including Roboflow, Google Colab, and Compute Canada, allowing for a detailed comparison of their capabilities in image preprocessing, model training, and real-time inference generation. The results demonstrate that Image Classification Module I achieved a validation accuracy of 99.4%, while Module II attained 98.6%. Inference times were found to be less than 1 s in Colab and under 3 s on a stand-alone system, confirming the architecture’s efficiency in real-time ice condition monitoring. |
| format | Article |
| id | doaj-art-14866b5062d64cf99e80cb37e948ccfc |
| institution | Kabale University |
| issn | 1424-8220 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj-art-14866b5062d64cf99e80cb37e948ccfc2025-01-24T13:48:31ZengMDPI AGSensors1424-82202025-01-0125232610.3390/s25020326Performance Evaluation of Deep Learning Image Classification Modules in the MUN-ABSAI Ice Risk Management ArchitectureRavindu G. Thalagala0Oscar De Silva1Dan Oldford2David Molyneux3Faculty of Engineering and Applied Science, Memorial University of Newfoundland (MUN), St. John’s, NL A1B 3X5, CanadaFaculty of Engineering and Applied Science, Memorial University of Newfoundland (MUN), St. John’s, NL A1B 3X5, CanadaAmerican Bureau of Shipping (ABS), St. John’s, NL A1B 3X5, CanadaFaculty of Engineering and Applied Science, Memorial University of Newfoundland (MUN), St. John’s, NL A1B 3X5, CanadaThe retreat of Arctic sea ice has opened new maritime routes, offering faster shipping opportunities; however, these routes present significant navigational challenges due to the harsh ice conditions. To address these challenges, this paper proposes a deep learning-based Arctic ice risk management architecture with multiple modules, including ice classification, risk assessment, ice floe tracking, and ice load calculations. A comprehensive dataset of 15,000 ice images was created using public sources and contributions from the Canadian Coast Guard, and it was used to support the development and evaluation of the system. The performance of the YOLOv8n-cls model was assessed for the ice classification modules due to its fast inference speed, making it suitable for resource-constrained onboard systems. The training and evaluation were conducted across multiple platforms, including Roboflow, Google Colab, and Compute Canada, allowing for a detailed comparison of their capabilities in image preprocessing, model training, and real-time inference generation. The results demonstrate that Image Classification Module I achieved a validation accuracy of 99.4%, while Module II attained 98.6%. Inference times were found to be less than 1 s in Colab and under 3 s on a stand-alone system, confirming the architecture’s efficiency in real-time ice condition monitoring.https://www.mdpi.com/1424-8220/25/2/326sea ice risk mitigationice classificationsea ice imagesdeep learningYOLOv8 |
| spellingShingle | Ravindu G. Thalagala Oscar De Silva Dan Oldford David Molyneux Performance Evaluation of Deep Learning Image Classification Modules in the MUN-ABSAI Ice Risk Management Architecture Sensors sea ice risk mitigation ice classification sea ice images deep learning YOLOv8 |
| title | Performance Evaluation of Deep Learning Image Classification Modules in the MUN-ABSAI Ice Risk Management Architecture |
| title_full | Performance Evaluation of Deep Learning Image Classification Modules in the MUN-ABSAI Ice Risk Management Architecture |
| title_fullStr | Performance Evaluation of Deep Learning Image Classification Modules in the MUN-ABSAI Ice Risk Management Architecture |
| title_full_unstemmed | Performance Evaluation of Deep Learning Image Classification Modules in the MUN-ABSAI Ice Risk Management Architecture |
| title_short | Performance Evaluation of Deep Learning Image Classification Modules in the MUN-ABSAI Ice Risk Management Architecture |
| title_sort | performance evaluation of deep learning image classification modules in the mun absai ice risk management architecture |
| topic | sea ice risk mitigation ice classification sea ice images deep learning YOLOv8 |
| url | https://www.mdpi.com/1424-8220/25/2/326 |
| work_keys_str_mv | AT ravindugthalagala performanceevaluationofdeeplearningimageclassificationmodulesinthemunabsaiiceriskmanagementarchitecture AT oscardesilva performanceevaluationofdeeplearningimageclassificationmodulesinthemunabsaiiceriskmanagementarchitecture AT danoldford performanceevaluationofdeeplearningimageclassificationmodulesinthemunabsaiiceriskmanagementarchitecture AT davidmolyneux performanceevaluationofdeeplearningimageclassificationmodulesinthemunabsaiiceriskmanagementarchitecture |