Three-dimensional Measurement Of The Workpiece Size Based On Binocular Vision

In the current industrial production, there are often some problems on human vision measure, such as inefficient manual inspection, not high accuracy, damage of workpiece surface and so on, while needing to be online to measure a large number of workpieces’ size. For the common machine measure instrument, although problems of human vision measure rarely happened, it needs high accuracy of workpiece’s location because of some fixed formula controlling the measure. Otherwise, the measure will not be accomplished. It is not fully intelligent to for the workpiece’s measure. However, non-contact binocular stereo vision system not only has the advantages of high accuracy and high efficiency, but also realize the intelligence for measuring the three-dimensional size of the workpiece. So, it is favored by majority of manufacturing enterprises. But because the technology has not yet reached a very mature stage, the practical applications based on binocular three-dimensional measure system are not many. So, three-dimensional measure of the workpiece’s size the paper researched based on binocular vision can make a contribution to industrial automation more.Based on HALCON’s powerful visual algorithms library and the.NET’s C#development environment, the paper realize the three-dimensional measure for the study of the workpiece. The binocular camera tilted inward is calibrated for stereo vision system to acquire cameras’internal and external parameters. The left and right images of the workpiece are corrected by the epipolar, which makes the corresponding points of the workpiece not have the level difference. The workpiece is automatically extracted by image preprocessing. What’s more, the disparity image and three-dimensional point cloud image are obtained by searching for the corresponding matching point making use of dense type stereo matching algorithm based on gray-related and the error matching points are excluded secondarily by three-dimensional point cloud. In the end, the three-dimensional size of the workpiece is determined by calculating that of the externally enclosing rectangle of the three-dimensional point cloud in the direction of the major axis inertia.A set of binocular measure system software has been accomplished according to this research content and methods. The fact that the three-dimensional measure has a high accuracy explains that the innovative methods like automatically extracting workpiece, secondarily excluding error matching points by three-dimensional point cloud and calculating outside the enclosing rectangle the paper used are practicable.