Research Of Positioning Technology Of Assembly Hole Group Based On Machine Vision

Based on its non-contact, high speed, high precision and automation, positioning technology based on binocular stereo vision has become an effective means of 3d measurement. It has been widely applied in the modern industrial production and detection fields. In this paper, according to the principle of binocular stereo vision, the target artifact recognition and positioning technology is studied. It mainly relates to the four contents of the binocular vision system calibration, target recognition, feature point extraction and stereo matching,3d reconstruction.(1) Using the improved linear calibration method based on the perspective model to calibrate the binocular vision system in the Halcon software, precision of 0.2 pixels. Then an image pair will be rectified to epipolar images and the epipolar constraint will be checked, error is 0.6 pixels.(2) Studied the recognition method based on template matching, to achieve the recognition of the workpiece image of rotation, scaling and shade. Improved the correlation function of similarity calculation, reduces the computational complexity.(3) Choose the Harris operator to extract the workpiece image feature points. Using the based on improved polar constraints feature point matching combined with morphological Euclidean distance method to carry on the stereo matching.(4) Matching point pair for 3d reconstruction based on linear calibration method, in order to position the assembly hole group.The centers of the circle are 3d reconstructed. In order to verify the accuracy of binocular stereo vision system, this article has carried on the center distance and the height of the gauge block measurement experiment.