Application And Development Of Vision-Based Gesture Interaction In The Automotive Flat Operation

With the rapid development of modern automotive electronics industry, the traditional mode of operation has been unable to meet the needs of the people interact with the vehicle-mounted device, vision-based gesture recognition technology is the key technology in the new generation of somatosensory interaction, through integration with the car multimedia applications can greatly improve HCI(Human-Computer Interaction), and enhance the driving experience and safety.Firstly, from the perspective of human-computer interaction in the car, analysis of the vision-based gesture interaction technology applications in the automotive flat in the background and significance, then introduces the gesture recognition research status at home and abroad; By comparing several common somatosensory device, select Leap Motion as a development device; Then elaborated Leap Motion device structures, recognition principle and technical characteristics. At the same time, this thesis introduces the relevant basic theoretical knowledge of the Android system. In this thesis, the development of based on C/S structure is developed from two aspects, after the analysis of the overall demand for the definition of gestures and functions, the server-side through processing to frame data collection from the Leap Motion device, realization of gesture recognition based on vision, and through the Socket communication based on WIFI technology to send commands; In client-side, the application software architecture is designed based on the Android platform, and completed the development of on-board flat music applications, maps applications, and related functions, can receive instructions and perform the operations required by the driver. Map of the development of the use of AMap SDK, hybrid positioning technology can be used to deal with complex situations.Finally, simulation experiments on company cars, Non contact hand gesture operation test for vehicle plate application, the results show, vision-based gesture interaction is more natural and flexible, and has good stability, real-time performance and higher accuracy.