OBSTACLE DETECTION AT RAILWAY CROSSROADS USING IMAGE PROCESSING
Ahmad Sugiana, Bandiyah Sri Aprilia, Muhammad Naufan Rifki
Telkom University
Abstract
There is a potential of collision hazard on the railroad crossing then the crossing guard will give the signal and call the control station to stop the train that passes by, but there is so many railroad crossing with the passive level crossing which have not guards. In this study focus on the collision detection system comprised single board computer, LORA communication module, and Image processing machine with HOG & SVM method embedded within it as the cheap and effective solution to tackle such a problem. From Prototype testing with a 20% confidence threshold, the outcome has 89% true rate for the classification and 53% successful rate of the communication on urban city test condition.
Keywords: railroad crossing, HOG and SVM, Hazard Detection
Topic: Electrical Engineering