This project aims to identify defects in railway tracks in Sri Lanka using a machine learning-based approach. The system leverages computer vision and deep learning techniques to automatically detect and classify defects from video footage of railway tracks. The primary goal is to improve railway safety and maintenance efficiency by providing real-time or near-real-time alerts about potential issues on the tracks.
The codebase is designed to run on the Nvidia Jetson Nano board, The Jetson Nano processes video streams from cameras mounted on railway inspection vehicles, performs inference using trained machine learning models, and manages data storage, streaming, and event logging.
Key Features:
- Automated detection of railway track defects using deep learning
- Real-time video processing and statistics collection
- Modular pipeline for inference, file saving, streaming, and event logging
- Designed for deployment on Nvidia Jetson Nano for edge AI applications
- Configurable and extensible for different camera sources and operational requirements
Target Users:
- Railway maintenance teams