Research and Implementation of Human-Computer Interaction System Based on Human Body Attitude Recognition Algorithm

Abstract

Aiming at a series of problems in the application of visual human-computer interaction, such as complicated recognition process and low interaction efficiency, a gesture recognition algorithm based on human body structure is proposed. Using this algorithm and the design idea of hardware and software cooperation, a robot control system based on visual perception is implemented on FPGA platform. The attitude sensor is used to obtain the human attitude information and solve it. The attitude is determined by the information fusion method based on D-S evidence theory. combine the complex actions, simplify the operation, obtain the attitude of the remote robot through the “double virtual model” and correct the motion trajectory. Aiming at the defect of single sensor in information detection dimension and the sensitivity to noise data, a multi-sensor fusion method for human posture estimation is proposed. In this paper, an algorithm of human posture recognition (i.e. human joint position recognition) based on model constraints is proposed to solve the problem that the robot system can accurately extract the spatial coordinates of human joints before the transfer and multiplication operation.