Building a Robot to Identify Electric Vehicles from Infrared Images

Rémi Lamoureux-Lévesque, Derek Jacoby, Marc-Antoine Drouin, Jonathan Fournier, Xavier Maldague, Yvonne Coady

Abstract: Identification of electric vehicles (EV) is important from the perspective of monitoring the environmental, social, and governance (ESG) goals around sustainable transport choices. We present steps for developing a system capable of improving image quality to help the automated process of counting EV and ICE vehicles. We have built a pan-tilt camera motion system to keep the target vehicle centered in the camera and then, collected imagery by developing a data collection robot based on the Nvidia Jetson nano. The pan-tilt system was developed in two phases. In the first place, we implemented a proportional algorithm and then went on a proportional-integral, which increase the result for placing the target in the region of interest. We will be presenting the steps made to improve the collecting system.