Study on human detection system using deep neural network and alternative learning for autonomous flying drones

Abstract

An alternative learning and its application to construct an overviewing human detection system (OHDES-V2) of flying drone for emergency rescue and investigation is presented in this paper. In this system, a deep neural network and alternative learning are used key techniques for object recognition from a free viewpoint. Simple appearance-based characteristics is determined from captured images, and the system uses a deep neural network to automatically classify human body, automobiles and so forth. The proposed system shows that several objects can be recognized from a bird’s-eye view. Experimental results show that the system can effectively recognize four types of objects and walking persons with accuraces of 98.5% and 97.12%, respectively.