The Research On Machine Vision Of Pneumatic Robot

Machine vision is regarded as the most important branch of the research of intelligent robots , virtual instrument is a fresh software flat in the field of the application of computer . Based on the technology of virtual instrument and the study of the status of robot vision in recent years , this thesis design the pneumatic robot vision orientation system . The thesis firstly analyzes the camera calibration and its adjustment for the robot vision . Further we particularly discuss the noise filtering and feature extraction .The thesis analyses the physical model of camera and the geometrical model of grille imaging in camera plane. According to the feature of 2-D plane image, we presented a method for the camera adjusting that can reduce the camera distort and improve the precision of orientation based on BP Neural Networks . For the image process , the study is illustrated with a hexangular nut experiments relating to the method development and analysis . For the feature extraction, based on the basic shape feature inspection in the work piece inspection, the thesis presents a method to get the center coordinate of the image with the area.On the aspect of the realization of robot vision servo control system, this thesis based the advantage of the visual programming , combined with the software IMAQ Vision , developed a completed vision servo control system . And the route of robot line interpolation also programmed.Practice have proved that combined the virtual instrument with the machine vision to design the robot vision system shortened the designing period , and also improved the reliability and the ratio to capability and cost. Moreover , combined the virtual instrument with the machine vision , in term of the logic of the machine vision system to analyse the whole system , explore the way to develop the machine vision . Therefore , it has profundity significance in both project and theoretics .