The Application Of Action Recognition And Gesture Recognition Based On Kinect Sensor In HCl

In recent years,the problems that brought by the aging of the population have gradually emerged,a large number of elderly and disabled people,making Helping the Disabled has become a very important and difficult problem in society.The people who need careful care not only brings great mental and economic pressure to the family,but also brings a heavy burden to the country.Meanwhile,the rapid development of the robot industry has put forward new requirements for HCI(human-computer interaction).The first stage in the history of HCI——Interface phase already can't meet the needs of people's lives.People are looking forward to the second stage of HCI——the natural language interface.Solving the problems caused by aging population by using the dividend generated by the robot industry can not only improve people's quality of life,but also provide a feasible plan for caring for the elderly.I taked this as an opportunity and designed and developed an algorithm for human action recognition and gesture recognition.And I designed the robot verification system to verify the algorithm.The specific work is as follows:(1)Design and optimization of action recognition algorithm.An action recognition system is designed based on Kinect obtained human skeleton data.And according to the characteristics of Kinect and the characteristics of human actions,the algorithm of action recognition is optimized.(2)Design of gesture recognition algorithm.In this paper,Kinect is used to get depth of field data,and a gesture recognition algorithm based on recognizing the number of finger points in single gesture is designed and developed.(3)Design an helping old robot verification system.The above two kinds of research results are applied to the helping old robot.This robot system takes Arduino ATmega2560 microcontroller as the core control chip,and uses the WIFI module of the Open Wrt system as the communication module,the mechanical arm of two configurations is designed.(4)The verification results of the robot verification system are statistically analyzed.The success rate of various action recognition and gesture recognition rate of robot verification system is over 90%.This shows that the two interaction modes designed in this paper have good practicability and research value.At the end of the article,the system is summarized and prospected,and the shortcomings of the system are pointed out,and the improvement scheme is put forward.