Detection And Tracking System Based On Multi-touch Based On Computer Vision Research

Now most of the computer input device is mouse and keyboard. Compared withthe control of computer directly with your fingers, it is less convenient, so there is atouch-screen technology. Most bank ATMs have touch screen; and the touch screen alsohas been in hospitals and libraries hall. Many kinds of mobile phones, Mp3and digitalcamera also support the touch-screen function. But these are all single-pointtouch-screen. In order to make the operation more convenient, the multi-touchtechnology comes, which is an important innovation of the human-computer interaction(HCI) technology.In this paper, it will introduce the multi-touch technology which is based oncomputer vision to identity simultaneously more than one finger on the touch-screenand react accordingly. The multi-touch technology which based on computer visiontechnology has lower cost and is easy to extend. Therefore it is more suitable forlarge-scale demand. In this paper, the research work is as follows:Study and compare several different principles based on optical multi-touchtechnology. I improve and achieve a complete multi-touch system. And then I completea multi-touch control application program with client/server mode to control somemulti-touch programs.Research the camera model and homograph matrix. Using the nine-pointcalibration technique accomplishes the manual calibration of the touch plane. But themanual calibration is less convenient than automatic calibration. So the corner detectiontechnology is researched. A cross template matching algorithm is advanced for boardinner corner point detection to achieve the automatic calibration of the projection plane.Study the extraction and tracking of the touch-point. With the computer visionalgorithms, achieve the extraction of the touch-point; and with the shape of thetouch-point, make sure it is finger of object; and then using bipartite graph in graphtheory, track and match the touch-points.Research the Various touch gestures and the detection and tracking of touch-point. Analyze the trajectory of the touch-point moving. And with it, realize the gesturerecognition and the controlling the computer to react accordingly on the gesture.