A Video Based Hierarchical Vehicle Classification System

The orange lines at the top show the canonical structures for different vehicles. Bottom pictures show the rear views of corresponding passenger vehicles (a) Sedan, (b) Pickup truck, (c) Minivan, and (d) SUV

Background:

Transportation and vehicle classification systems are becoming smarter and more automated. For example, electronic toll collection systems have been introduced and drivers are not required to stop, eliminating road delays. New technologies have also been added to these systems that enable service providers to acquire data on what type of vehicles are utilizing their amenities as well as vehicle identification for safety & control purposes.

 

Brief Description:

UCR Researchers have developed a method and system for vehicle classification using video imaging. This novel invention entails a vehicle ground clearance measurement system along with a video camera that captures a travelling vehicle and categorizes it into a vehicle class. The cameras on current methods and systems rely on side views of the vehicle, which can easily be obstructed by other vehicles.

• Classifies vehicles into one of six classes: sedans, pickups, minivans, SUVs, buses and trucks
• Camera imaging of the rear end of vehicles for a more accurate classification

• Vehicle identification readers for toll roads – provides information on the type of vehicles that are using the tolls as well as an added security measure for violators
• Substitute parking attendants – automatic classification of vehicle for parking entrance & exit payment systems
• Other safety and control system products