Railway Safety Monitoring System
A real-time video monitoring platform for railway safety, built around multi-stream RTSP ingestion, edge AI inference, and message-based data pipelines.
Overview
The system processed 30+ RTSP streams per site and ran AI models on Jetson edge devices for equipment state detection. Detection outputs were passed through a Kafka pipeline for monitoring and integration.
Slides
Key Work
- Built real-time monitoring using 30+ RTSP streams per deployment site.
- Deployed AI models on Jetson edge devices.
- Implemented YOLOv7, SVM, and OCR for equipment state detection.
- Integrated inference output with a Kafka-based pipeline.
