Font Size: a A A

Based On Stereo Vision Fingertips Positioning And Human-computer Interaction Research

Posted on:2014-11-08Degree:MasterType:Thesis
Country:ChinaCandidate:Y Y XiFull Text:PDF
GTID:2268330422961829Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
The human-computer interaction system based on computer vision is an importantresearch direction of human-computer interaction field, and has a extensive applicationprospect in natural human-computer interaction, hand gesture recognition, virtual reality,smart home system, body sense game and so on. Most of traditional computer vision touchsystem has problems of need to wearing special equipment or touch panel requires specialmaterials production and other issues. In this case, this paper proposes an interactive systembased on infrared cameras and projectors, and implements interactive touch operation on anordinary desktop. The main contents are as follows:1. Develop a human-computer interaction system based on infrared camera and projector.System uses a projector as the output device, and the content of the computer monitorwill be displayed in the general desktop. Two infrared cameras for the video captureinput device, through visual detection algorithms and principles of stereoscopic visionpositioning fingertips. When the fingertips touch the desktop, start the touch event andachieve the appropriate touch operation.2. In-depth study and analysis of the fingertip detection algorithm in infrared imagingsystem. According to the infrared camera imaging principle and characteristics, ofproposed based on the convex hull and the convex defect detection algorithm formulti-finger, which able to adapt changing projection environment, with bettertimeliness and robustness.3. In order to get accurate finger positioning, the paper researched the distortioncorrection and three-dimensional calibration of double-camera system, designed thecalibration procedure of double-camera system. It solved the problem ofthree-dimensional measurement of double-camera under multi-angle and the imagedistortion because of manufacture of camera.4. According to the characteristics of the fingertips, the paper proposed a fast andefficient block-matching algorithm. According to the principle of binocular stereovision and the characteristics of the standard three-dimensional test bed, the papercompleted fingertips matching in a completely coplanar and rows aligned image, it reduced algorithm complexity and improved the system real-time. According to thematching result, the paper can compute three-dimensional coordinates of fingertip.The prototype system which developed has made a large number of field trials, the resultsshowed that: the interactive system which the paper designed and developed can realize avariety of interactive entertainment features on the desktop, such as table games, virtualkeyboard input and web browsing operations. Meanwhile, the system has a friendly interface,real-time and high robustness, portability and low cost.
Keywords/Search Tags:Computer Vision, Human-Computer Interaction, Infrared Imaging, FingertipPositioning, Binocular Three-Dimensional Measurement
PDF Full Text Request
Related items