无人平台人机交互视线追踪技术OA北大核心CSTPCD

A head-mounted eye tracker based human-machine interaction system for unmanned platform is designed.The focus is on analyzing a two-dimensional line of sight estimation algorithm based on spatial projection principle.By using multi-camera registration,the three-dimensional line of sight vector output from the eye tracker is mapped into a two-dimensional interface of the known coordinate.Through bidirectional recognition,the line of sight and the screen interactive interface are recognized separately,the coordinate conversion of the point of view of human eyes in the software interface of the human-machine interaction is achieved.The interaction experiment of target tracking of unmanned platform is designed,the applicable situations in the simple click interaction task between eye gaze tracking and touch interaction are compared.The results show that compared with touch interaction,eye gaze tracking can reduce the time for simple click interaction events by 1 to 1.5 seconds.