2026 Volume 5 Issue 1 Pages 333-340
Heavy snowfall can significantly impact railway operations, requiring train patrol to monitor snow accumulation on tracks. However, conventional methods impose a significant burden on personnel, highlighting the need for operational efficiency. This study developed a snow accumulation assessment system that combines forward-facing image capture using smartphones with an AI-based image classification model. The system automatically classifies snow conditions on railway tracks into six categories and achieves high accuracy under various environmental conditions. Field trials confirmed the effectiveness of real-time visualization and information sharing via a web application, demonstrating its potential to reduce manual workload. Future enhancements, such as detecting side snow and snow cornices, are expected to further improve operational efficiency.
