| Summary: | 碩士 === 國立中興大學 === 機械工程學系所 === 107 === In computer vision, camera obtained the two-dimensional image from an object. We could calculate the transformation of coordinate systems between camera and object by using Perspective N Points(PnP), then proceeding to know the movement which include translation and rotation. And this method could also use on the calibration of robot arm.
In this research, we have a non-contact measurement system established by three cameras and the standard part which composed of precision steel balls would be driven by linear guideway and rotated table. Three-dimensional coordinates of the standard part would be obtained at different positions, while three cameras shoot from various directions. After using edge detection to obtain the edge information of the image, the center position of the image could be detected. In this way, we would have the coordinate position of the object in two dimensions and three dimensions, and we ultimately use the weighted average trajectory improve measurement methods which was proposed, calculates a more accurate object movement trajectory.
|
|---|
