Vehicle Class Recognition of Street-Parking Vehicles from Side-View Range Images

This paper describes a novel method for recognizing the classes of street-parking vehicles. We have already developed the following two systems: one is vehicle recognition system based on local-feature configuration, and the other is detecting street-paring vehicles from side-view range images. In this paper, we combine these two systems to develop a new system with which we can not only count the number of street-parking vehicles but also recognize their class of vehicle type such as sedan, wagon, mini-van or so. We have confirmed that our classification algorithm is still robust on range images, performing outdoor experiments. Our system can recognize four vehicle classes of sedan, wagon, mini-van and hatchback from outdoor range images with accuracy of about 80%.