Dimensional Control of Industrial Parts using Image Processing and CAD Algorithm, Reza Nejatpour, 2008

Abstract:

Measuring and dimensional control is one of the most important sections of manufacturing processes. Nowdays regarding the versatility, high quality and numeral complexity of products, the traditional contact measuring methods are not adequate. In order to solve these problems we need a method with which dimensions and charactristics of a product with complex curves are calculatable in a short time with the highest required accuracy. In this research a machine equipped with a CNC table and a CCD camera is designed. Using image processing and CAD algorithms this machine can automatically calculate the dimention of 2D industrial products with good accuracy. In the first part of this paper the system hardware such as CCD camera, lens and CNC table are introduced and the distinct features of each of them is described. Afterwards, two main methods of camera calibration are explained. Then the necessity of image stitching is described and a new method based on camera calibration is proposed. In the next part, the image processing algorithms which are used for edge detection, gage feature extraction and line and edge detection are explained. How to interpolate a BSpline on the edges of the images is also illustrated and a method for decreasing the errors in sharp comers is applied which can easily model free and complex shapes. Using this method measurment process is shortened and its error is decreased. Automation of the measurment has also omitted human mistakes and errors due to uncalibrated tools. ­

Keywords:

Image Processing, Dimensional Control, Camera Calibration, BSpline