YOLO-Rail: An Improved YOLO Model for Obstacle Detection on Railway Tracks

YOLO-RAIL is an enhanced object detection system designed for real-time obstacle detection on railway tracks using an improved YOLO (You Only Look Once) model. The system integrates a Raspberry Pi as the central controller, connected with a web camera, ultrasonic sensor, LCD display, buzzer, and speaker. The web camera continuously captures live video from the railway track, and the YOLO-based model processes these images to detect obstacles such as humans, animals, or vehicles. In parallel, the ultrasonic sensor measures the distance of nearby objects to ensure accurate detection even in low-visibility conditions. When an obstacle is identified, the system triggers a buzzer and also provides a voice alert through the speaker to warn about the abnormal condition. The detected distance and status are displayed on the LCD for monitoring purposes. Additionally, if the detected object is a vehicle, the system captures its number plate for identification purposes. This solution aims to improve railway safety by providing a low-cost, real-time, and intelligent monitoring system that can help prevent accidents and enhance track surveillance efficiency.