In this paper, an algorithm for detecting position for model trains is presented. It uses a single affordable video camera, a software which utilizes computer vision algorithms to track objects and a simple server to share that data with model embedded devices. This approach has proven that it provides accurate enough (several centimeter precision) positional data to control the models during testing. The system is easy to setup and allows to migrate the software code to use GPS without significant changes.
Autonomous vehicles | Speed estimation | Computer vision| Safety system | Satellite positioning simulation| Autonomous vehicle modelling