Thomas More Geel has a large parking lot with many entrances. This is good for making the campus accessible, but some control is lost here. There is a place where cars come, which was originally meant for buses and freight. There are already several solutions to this kind of problem, but it would be cool to add AI to this as well.
The solution here is to install a barrier that opens only for buses and truck traffic. Freight traffic would now still have to access manually through the reception, but buses should be able to pass through.
Through a camera pointed at the street in front of the barel, we can use object detection to start seeing which vehicle is there. The images are livestreamed to a server that will perform the detection. When a bus is detected a signal will be sent to a Raspberry pi pico which provides a pulse signal to open the barel.
There was not enough time to train a custom model that recognizes buses. So we decided to use the YOLO model to recognize the vehicles. We use only the output of the detected vehicles and send a signal to the barrier only when a certain bus detection threshold is reached.
