Research And Development On Auto Adjusting System For Cutting Plates With Preprinted Images

Preprinted plate cutting technique is a kind of technique that is intended to serve our life. This technique refers to cutting the plate along the contour of the image which is preprinted on the plate so the product with the same contour of the image is made. Being practical and directviewing, the technique is required and applied in many different areas. Currently, general NC machines don't have adjusting tools aiming at this kind of product yet, which results in low machining accuracy and high reject rate. Manual adjusting means can improve the accuracy in some degree. However, if there is rotation transformation between image coordinate system and machine coordinate system, the plate needs clamping again, resulting in a low efficiency.Based on knowledge of both image processing and computer vision, a new technique of auto adjusting for preprinted plate cutting is proposed, which realizes that image contours on the plate can be on-line located and then cut accurately without the need of clamping for a second time. The technique includes three important subordinate techniques namely contour vectorization technique, contour NC programming technique and image contour adjusting with marked points technique. The contour vectorization technique detects and fits the contour of the original image and adds marked points to original image for later adjusting. The contour NC programming technique is to figure out initial NC code which is based on image coordinate system. In the image contour adjusting with marked points technique, CCD camera is used to recognize the marked points on the plate. After that, the matrix transformation between image coordinate system and machine coordinate system can be installed which could be used to convert initial NC code to NC code that can be used to actual machining. In this dissertation, the contour vectorization technique and image contour adjusting technique will be investigated. The main contributions of this dissertation are summarized as follows:Firstly, based on the analysis of the characteristics of preprinted image, a rapid and effective flow of image segmentation is planed. Algorithms that are proper to the system are selected via analysis of many current algorithms and comparison between the experimental results of them. These algorithms can detect the image contours accurately. Furthermore, a contour optimization method is applied to get target contour closer to true edge, which improves the effect of image segmentation in some degree.Secondly, a method of piecewise fitting based on curvature is proposed, the idea of which is that find the cusp of the contour through curvature first and then fit the contour between two adjacent cusps with lines. Compared with methods without piecewise fitting, the efficiency is improved greatly.After the analysis of the experimental result of actual cutting, a new parameter, which is used to control the length of fitting, is added to the piecewise fitting algorithm, which greatly improves the smoothness of the contour after cutting and has no big effect to the efficiency on the same time. Thirdly, an image contour adjusting method based on measuring with CCD is proposed. The CCD camera, which can move with machinery move, can take pictures of marked points and compute the machine coordinate system of their circle center orderly. And then the linear transformation matrix between image coordinate system and machine coordinate system can be figured out. Given the transformation matrix, the initial NC code based on image coordinate system can be converted to the NC code based on machine coordinate system, by which the image contour adjusting is done.Based on theoretical research, an auto adjusting system for the cutting along the image contour preprinted on plate is developed. Experiments of a number of typical preprinted plates cutting along image contour have been carried out and the result proves that the NC code generated by this system has good accuracy and efficiency. Currently, the NC machine having installed this system was in 2009 Shanghai Machine Tool Exhibition (EASTPO 2009) in July, and achieved a good response.