| With the rapid development of computer technology and popularization. The methodof communication with human and computer has become a new hot research area.People pursue the most nature way of interaction. Currently in the market, the mostpopular new interactive technology is multi-touch technology. It has replaced thetraditional mouse and keyboard operation mode, with finger drag click can be a seriesof complex orders for the computer. Although the interface is easy to understand, butstill a distance with the most natural way of interaction.Since the depth camera device Kinect was public, Kinect allows not only canreceiving two-dimensional image information, also make the image pixels with depthinformation. The data receiving from the camera also is3d information not2dinformation.In this paper, we will use Kinect to interact with virtual reality scene objects inunity3d. The method is mainly divided into three parts: using depth cameras for imageacquisition, the body region detection and background segmentation, the extractionskeleton of human body, the positioning and skeleton redrawing of bone joints, arevised get stable skeleton information. And recorded in quaternion movementinformation, using real-time tracking, filtering to get stable motion frame data. Finallyby improving the spherical skin for making3d automatic skin role model, andcalculate the weight, the last in unity with the bone data matching, realize the rolecontrol the movement of the3d characters.The experimental results show that the method can overcome the external factors suchas environment light influence on camera to collect data, and has good recognitioneffect. With stable skeleton data we can use it to control the movement of virtualreality character, achieve natural interaction.This paper adopts the method of depthcameras identification technology popularization, and the application of Kinectrecognition technology in the virtual reality interaction has a certain significance. |
PDF Full Text Request