A System of obtaining geometric shapes of three dimensional objects is developed. It is based on the method of calculating the coordinates of the intersection of an object to be measured and the light plane made through a cylindrical lens from a laser ray. Video cameras are used to capture the image of the intersecting curve. The object is rotated at predetermined intervals to cover all over the surface.
Three main procedures are identified and programmed: Camera calibration; Range data calculation; and Surface modeling. In the camera calibration stage, a simple technique is possible when the structured light and the camera are well oriented. For other demanding situations, a more elaborate technique based on two cameras is desirable to reduce the error. In the range data calculation stage, a series of coordinates of the intersecting stripes in the structured light system are calculated and stored. The equation of the inverse perspective view is used for the coordinates of each point. A thinning operation is adopted to get the screen coordinates from the lighted image. In the last stage, a Ferguson surface model is used for the representation of the surface. The magnitude of tangent vectors is obtained by the chord length method.
A software system is developed and interfaced with the video image capture system. Several objects such as twisted rectangular vase, Agrippa and torso of Venus bave been tested. The error has ranged within 1mm in the experimentations.
