Machine Vision Identification of Vehicles and Impact on Corridor-Level Energy Modeling

The composition of on-road vehicle fleets is changing with purchases of next-generation vehicles. While these vehicles are increasingly replacing internal combustion engine vehicles over time, this likely will not happen before 2050. Therefore, these vehicles will co-exist on transportation corridors over the next 30 years. This has important implications for any understanding of vehicle air quality and energy use impacts, as planning and impact assessment models are functions of on-road vehicle fleet composition. In other words, to assess whether vehicle fleet penetration and use models are accurate, it is necessary for researchers to be able to accurately and reliably quantify on-road fleet compositions characterized by a mix of vehicle types. This project addresses that need by proposing a machine vision vehicle identification model. Current vehicle identification is often conducted by capturing vehicle license plates and comparing them to vehicle registration data. However, machine vision techniques are now sufficiently advanced to identify vehicles based on video monitoring alone. The proposed vehicle identification model using object recognition and detection methods on video-monitored transportation corridors will be used to develop on-road fleet compositions.