Location:Home > Engineering science > Mechanical Engineering > Mechanical and Electronic Engineering > Research on Graph Element Self-identification Focusing on Industrial CT Image Vectorization and Accu

Research on Graph Element Self-identification Focusing on Industrial CT Image Vectorization and Accu

Downloads: []
Tutor: LiuFengLin
School: Chongqing University
Course: Mechanical and Electronic Engineering
Keywords: Computed Tomography,Vectorization,Reverse Engineering,Image Measurement
CLC: TP391.41
Type: Master's thesis
Year:  2008
Facebook Google+ Email Gmail Evernote LinkedIn Twitter Addthis

not access Image Error Other errors

Industrial Computerized Tomography technology is using ray to transmit the dislocation of objects to be test with the scan way from various directions under the objects undamaged. Next, it applies special detector to gather information decayed through the tested object ray, and then to get the workpiece¡¯s dislocation image using the computer special image reconstructive calculation. Despite of shape or quality of the object to be test, Industrial Computerized Tomography not only can detect the internal construction and identified flaws in it, but also can carry out accurate measurement on the internal and external surface of the object. Especially, the technology plays an unreplaceable role in the Non-destructive Testing of closed-parts¡¯ internal part. Therefore, it acts a great part in the Non-destructive Testing and Reverse Engineering.However, bitmap format slice picture obtained from the Industrial Computerized Tomography scanning cann¡¯t be applied directly to Reverse Engineering. Considering it, the paper made a research on Industrial CT image vectorization and graph element self-identification and developed measurement system which laid the foundation for its application to Reverse Engineering.Firstly, analyzed the characteristics of Industrial CT image and denoised, enhanced, binarized and contour extracted it etc, next, to track the edge contour of the image using Freeman chain code and then to store the information of the edge point in the chain for getting ready for the following graph element self-identification and Industrial CT image vectorizationThe edge of image was identified as line segment, circle and circular arc etc graph element. For the identification of line segment, the study applied Bresenham-based integrated efficient line segment-generating algorithm to get line segment fitted. While for the circle, the author used the circular measurement method based on probability of existence and modified it. The detailed procedures are as follows: the circle range was chosen before ahead, and a lot of probability of existence which is impossible as the center of circle was not calculated. Thus, the calculating task is reduced and the testing efficiency is also greatly improved, and that the circle¡¯s parameter stored in the chain reduces the EMS memory¡¯s expenditure. The result shows the calculation expericing modification is faster 33-178 times than before. Combining the vertical bisectrix method, circular arc searching algorithm and method of least squares function to fit the circular arc and store distinctively the acquired graph element in the line segment, circle and circle arc chain to produce DXF file and transmit them into CAD system to edit and perfect them.The paper made a research on the circle measurement in Industrial Computerized Tomography image. Using the circle testing measurement based on probability of existence detecting the circle and appling the method of least squares function to obtain the pixel size. According to the arithmetic product obtained multiplying the pixel size of circle by the actual size of the pixel equivalent, which is the actual size of each pixel represented in the Industrial Computerized Tomography image, to get the acual size. Comparing the measurement result to true value, the result shows that the abosolute error is less than 0.1mm and the relative error of the measuring circle semidiameter size and the actual size is less than 0.5%.
Related Dissertations
Last updated
Sponsored Links
Home |About Us| Contact Us| Feedback| Privacy | copyright | Back to top