Study on Human-machine Interactive System based on Binocular Vision and Depth Camera

Abstract

The human-machine interactive system to be introduced in this paper is based on a binocular-vision imaging system[1-3] and a Leapmotion depth camera[4]. The system is developed and studied to use non-contactive gesture-recognition technology to directly interact with the virtual stereoscopic object synthesized in the real space. And a method is also presented to accurately integrate the gesture and the virtual stereoscopic image synthesized in the real space. In this way, in terms of different-gesture recognition, the operator can conduct different interactions with the virtual 3D object in the real space and realize vivid multiple-gesture operation. The operational and interactive effects of the system are demonstrated to be close to real-object operation through trials on 5 testers and the results will be presented in the last part of the paper.