System
Home » Traffic Survey »
System Description
This system uses an original concept from Greg Tinkers, subsequently enhanced by Tim Hodges. Both authors give excellent descriptions of how computer vision can be used to track moving objects in their project write-ups.
Written in Python3, running on a Raspberry Pi 4B, the program uses Eclipse Mosquitto to handle network data transfers with MQTT. The drifting web server stores the received MQTT data in a MySQL database. Charts embedded in the site’s WordPress pages are served by a locally hosted version of Grafana. Grafana is reverse proxied behind the Apache2 web server.
The traffic survey is only indicative and is also conservative. The monitor stops measuring speed when vehicles cross over or pedestrians enter the camera field of view. Speed calculations continue as vehicles cross the camera’s field of view. The program rejects any calculations which, when compared, have a high standard deviation.
The program stores images of vehicles exceeding a set speed. The image has a speed, date and time stamp. Currently, the threshold speed is high to limit the number of images stored.