Research On Intelligent System Of Gestures Control With Visual Perception

With the deepening of human-computer interaction technology,the gesture recognition technology of visual perception has become a research hotspot in the field of human-computer interaction.The gesture recognition technology based on visual perception is used to the control of switch cabinet in power plant and other aspects of industrial production,it can free workers from dangerous working environment and complex labor,and improve work efficiency,ensue the safety of suff.Therefore,the development of visual perception gesture intelligent control system not only has important theoretical significance,but also has broad application prospects and promotion value.Based on the study of video image acquisition,gesture image preprocessing,gesture segmentation,HOG feature extraction,SVM classifier training and binocular ranging technology,this paper designs and implements the intelligent control system of visual perception.The system uses a dual camera which has parallel optic axis to capture the operator’s gestures,extract frame,recognize gestures and analyze the gesture intention and control the target action to achieve the remote operation with non-contact gesture control.The research contents of this thesis are mainly divided into the following aspects: Firstly,Using the SVM algorithm to train the testing classifier and recognition classifier,in order to segment the gesture area effectively and recognize the gesture information correctly.The method has good robustness under the complex background and illumination change.Secondly,Through the known size checkerboard to calibrate the camera,and the stereoscopic correction of the camera is realized by adopting the pole line constraint.Using SGBM algorithm to the stereo matching for the image collected by camera,according to the feature point disparity between left and right images,it can conclude the three-dimensional coordinates of image,analysis the three-dimensional coordinates information and measure the depth of the close shot object information.Thirdly,An open source computer vision library is opencv3.0 is used to construct a intelligent system of gestures control with visual perception,use gestures 1-6 to select the switch cabinet of number 1-6,and use the palm or fist to control the disconnection or closure of the switch cabinet.And set the safe operation distance within 1-2 meters,if the gesture instruction meet the range,the corresponding switch cabinet will be responsed and present the state of disconnected or closed.Finally,completed the construction of experimental platform,after the test passed,we applied the intelligent system of gestures control with visual perception to the actual control of switch cabinet in power plant.