Rearch And Improvment Of Two Dimensional Industrial CT Image Vectorization System

Industrial computer tomography (ICT) is one of the most advanced non-contact measurement methods. It can show the internal structure,composition and defect of the detected objects in the form of slice images while the structure and material of the workpiece is not damaged. Therefore, as a practical non-destructive testing technology, it has broad prospects in its application to reverse engineering. The section slice obtained by ICT is gray image, which can not be used in reverse engineering until it is vectorized. As the developed ICT image vector system is with poor robustness, inadequately friendly human-computer interaction and low efficiency in vectoring images, the optimization research is made in this article.Firsrly, in the image preprocessing module, an enhance-edge-itself detection methods based on Canny algorithm is used to search the edge, then, the edge is extracted according to the threshold. The extracted edge is edited to obtain a closed single-pixel image contour.Then, in the vectorizaton module, the contour is segmented according to the vector changing rate of the contour point. The initial vectorized image edge is obtained by fitting these segments with line and curve. The initial vectorized image edge is edited to reduce the number of feature point and improve the accuracy of vectorized image. By optimizing the entire circle fitting, two point linear fitting and three point arc fitting functions, the stability and operating friendliness of the software is improved.Finally, compared the image vectorized by commercial software with the software mentioned in this paper, as the experiment results show that the improvement is effective.