Research On Robot Control Technology Based On Human Posture Recognition

With the continuous improvement of human-computer interaction technology, theapplications of this technique is becoming more and more widely. Therefore, theapplication background of the human-computer interaction is becoming more and morecomplex. The traditional human-computer interaction based on vision in strong light ordark conditions will be affected a lot, the uncertainty will happen. In this paper, to avoidthe above problems, using the depth sensor for human body posture recognition schemeis presented. Meanwhile, using human body posture to control the robot movement cannot only broaden the working environment of the robot, but also make robot motionmore intelligent. In this paper, recognize the human body posture and control themovement of robot by using depth sensor are studied.Firstly, the research background and significance of this paper are illustrated, theoverseas and domestic research status of human body posture recognition technology issummarized, the robot technology related based on somatosensory equipment isintroduced.Secondly, the hardware and software system of Kinect are introduced, and theapplication of Kinect technology is illustrated. The depth information acquisitionprinciple of Kinect and depth of information transformation are analyzed. By usingmedian filter and background difference method, the human body area is extracted fromthe complex background.Then, the related theory of skeleton extraction is expounded. To extract theskeleton, the Hilditch thinning algorithm is adopted. The feature representation andfeature extraction are introduced in detail.Finally, a sample database is established, human gesture recognition experiment iscarried out by using template matching. The identification results are obtained, and thestability and real-time of the experiment are analyzed. Also, the scheme and the motioncontrol of robot are combined, and the intelligent motion control of the robot is realized.