The Research Of Intelligent Wheelchair Humancomputer Interaction System Based On Head Pose Recognition

In recent years,with the improvement of living standards and the continuous development of medical and health technology,the quality of life of the disabled and the elderly has received more attention.The emergence of intelligent mobile robots has facilitated the travel of them,and has become the current research hotspot in the medical field.Among them,for the physically handicapped people,the head-pose control has become a new choice,it has favored by many consumers and researchers and has high market value and research value.Based on the research of head pose recognition,this paper applies the head recognition to the mobile robot interaction system,and realizes the motion control of mobile service robot by head movement.The main research results are as follows:(1)Head recognition is the basis of the whole human-computer interaction system.Based on the existing head pose researches,the machine vision algorithm is used to identify the head depth image.In this paper,we use the Kinect depth sensor to collect the depth image of the head.Because of the defects such as noise and holes in the original depth image,the original depth image is preprocessed by the median filter algorithm and morphological filtering for following recognition.(2)According to the current situation of domestic and foreign research,the random forest and ICP algorithm are used to calculate the head posture.After the two algorithms are implemented respectively,the limitations of the two algorithms are analyzed in precision and real-time.In this paper,the idea of random forest combined with ICP algorithm is proposed,the random forest algorithm is used as the rough registration of ICP algorithm to reduce the number of iterations of ICP algorithm in fine registration stage.The two algorithms complement each other's defects,making the head pose recognition with real-time and accuracy.(3)The mapping between the head pose and the moving state of the mobile robot is the bridge of the human-computer interaction system.This paper puts forward the control mode of the headoriented vector simulation of the joystick,which conforms to the natural interaction mode and has good control effect.According to the above analysis and development of intelligent mobile service robot interaction system based on head attitude recognition,Kinect is the image acquisition device,PC is the computing control core,and the mobile service robot is the application platform,and the whole human-computer interaction system is realized.