A Fuzzy Moving Object Estimation Using Infrared TOF Camera

This paper describes an object classification method using infrared TOF camera. The method detects moving objects from a distance distribution data. The objects are extracted by k-means clustering method based on fuzzy inference. To classify them, fuzzy if-then rules and fuzzy membership functions are derived from the knowledge of human characteristics. The system classifies them to toddler, child, adult and the other object such as baggage and animal. In the experiment, we employed seven volunteers, two dogs and a box. The system successfully classified them.