Abstract
The security system in night is very important, and the system that watches a person's movements is useful in the security. However, animals may cause the false detection on the system in night. This paper describes a classification system of human, animal-like, and the other object from distance distribution video image measured by an infrared laser camera. This camera radiates infrared waves (850nm) and receives reflected ones. It calculates the distance distribution based on the time of flight principle. To classify the objects, firstly, we do background subtraction and noise rejection in each frame of the distance distribution. Secondly, we extract features from each object such as the height, roughness, aspect ratio. Thirdly, we make fuzzy if-then rules from knowledge of human and quadrupedal being so as to classify the object to human, animal-like and the other object. Here, we made the fuzzy membership function with respect to each feature. Finally, we classify the object to one with the highest fuzzy degree among human, animal-like and the other object. In the experiment, we set up the camera in room and tested some cases. The method successfully classified them in real time processing.
