Hand Gesture Recognition for Intelligent Support in Agricultural Field

Abstract

The development of a communication system between workers and machines is effective to achieve an intelligent working system in collaboration with machines in agricultural field. In this paper, we developed a hand gesture recognition method as one of communication methods, which could extract and track moving objects by inter-frame difference method, and could detect and recognize the arc of the hand gesture. The hand gestures of six tested subjects in natural scene were recognized experimentally. When the size of the tested subject to the image height was about 60% and the speed of hand gesture was about once in a second, the method could recognize the hand gesture if only test subjects could gesture two times. The development of a robust recognition method against both the case that the size of a worker to the image height was too small such as 30%, 20%, and the case that the speed of hand gesture such as once in two seconds was too slow was essential to apply to real situation.