Reserch On Control System Of Hand Gestrue Recognition For 6-dof Manipulator Based On Binocular Vision

Currently,in the field of industrial production and manufacture,the robot arm is widely applied and played a more and more important role as it could work quickly accurately and efficiently instead of artificial operation,in high-risk working conditions or complex and variable space environments.Gesture information obtained by vision,as the motion control signal for robot arm,is natural intuitive flexible and concise,so that the threshold of operating and controlling the robot arm is declined and telecontrol operation could realize.Taking gesture gained by vision as the control signal for robot arm refers to the technology of computer vision and image processing.The paper analyzed the working requirement of controlling manipulator by gesture,and proposed an overall plan of controlling robot arm which used binocular vision to gain the gesture information considering specific conditions and actual situation,and translated image of gesture using related technology of image processing to obtain the meaning of gesture.The methods of image preprocessing and gesture segmentation were analyzed.The segmentation method of skin color threshold was selected to perform hand gesture segmentation.The morphological processing method is analyzed.Some methods of contour extraction for segmented gesture images was compared,and the detection method with Canny operator was used to extract hand gesture contour.The methods of hand gesture feature extraction were analyzed,and the fingertip detection algorithm with approximation defect circle contour based on D-P algorithm was improved,and the recognition of ten kinds of gesture was realized.This paper analyzed the dynamic tracking method,discussed the importance of using binocular vision,introduced the principle of binocular vision,and proposed a method of "pushing-pulling" dynamic gesture recognition.The whole scheme design was analyzed in detail,and the robot model was established based on the D-H method,the positive and inverse solutions of the robot kinematics were analyzed,the motion simulation were performed through the MATLAB toolbox.The gesture recognition application and the human-computer interaction interface was developed based on the Visual Studio 2012 platform using the OpenCV function library,the robot motion simulation track based on dynamic gesture trajectory was given and the control system was studied,and the software tool environment configuration was completed,the communication of the camera,computer and experiment equipment was established,and the "pushing-pulling" dynamic gesture algorithm was experimentally tested to verify the feasibility of the control system.