Research On Kinect Based Virtual Touch System

With the development of technology,artificial intelligence and human-computer interaction might take a pivotal role in the coming future.Now,touch screen has gradually taken place of mouse and keyboard becoming a major method of human-computer interaction.However,monitor equipped with touch screen existed several defects including high price,maintenance cost and breakdown rate.Meanwhile,it is impossible to realize touch control function on projection screen.To fulfill the demands of normal user and to realize touch control on projection screen,this paper proposed a new Kinect based virtual touch interaction system.The gesture was collected by Kinect and analyzed by skin color segmentation to collect the image of hand.Then,relative algorithm was designed to recognize fingertip and judge the movement by analyzing the coordinate between every image frame to call the relative command in computer.Lastly,the system used the CA model of Kalman filter to filter the gesture movement to reduce error.This system separated the monitor and touch system to realize virtual touching function.Two major parts of this system are filter optimization of the fingertip track and the algorithm of fingertip recognition.According to the motion feature of gesture,we used CA model to track target with a high accuracy.Meanwhile,our algorithm can recognize multiple move status.Then,the system used the CA model of Kalman filter to filter the gesture movement to reduce error.This system can meet the usage requirement both in accuracy and response speed during the test.Meanwhile,the accuracy is higher than some touch screen device in the market.