Multi-Kinect Human-Computer Interaction Technology And Software Based On Unity3D

With the development of Internet technology, all kinds of man-machine interaction are emerging. Meanwhile, people are increasingly demanding for good user experiences in applications. As the human-computer nature interaction, the motion-sensing technology has become the research hotspot. At present, the applications based on Microsoft Kinect are almost confined to single Kinect equipment. It limits the interaction space. Therefore, aiming to extend the space of somatosensory interaction, we study how to use multiple Kinects in human-computer interaction in this paper.We developed a software package that is able extract the spatial location and depth of the aimed objects in the view of multiple Kinects on the basis of OpenNI2 and OpenCV. We proposed a depth-based target recognition algorithm. Furthermore, we developd three recognition modes, which are head, foot, polygon mode, according to the characteristics of the parts of body. Then, we used the C++/C# mixed programming technique to link up our motion-sensing library and the Unity3D engine. Then, we use the scene construction and 3D rendering properties of Unity3D platform, acquisition of real-time depth data and spatial information by transmission Kinect and make the data utilized, thus wen developed a demo interactive application by using 2 Kinects.As a significant extension, this technology shows great potential in many applications. It is worth to be further researched.