Optimization Method Of Industrial CT Image Processing System

Industrial CT technology is one of the most advanced measure methods in the world. Without the limitation of object size, structure and material, the object can be scanned non-destructively in tomography. As the scanned results show the internal and external structure of the tested compounds accurately in the image, this technology could be applied broadly in reverse engineering. The developed industrial CT systems can transform bitmap slice images into vector data and output the results with the common CAD data formats. However, the system is low efficiency, bad precision and inconvenient human-machine interaction in processing, which effects negatively on following tasks. In order to solve these problems, the optimization research was carried out in this paper:The author studied the classical edge detection operator, Canny operator and C-V methods. Based on the characteristics of industrial CT images, the author selected both the Canny operator and an improved Canny edge detection algorithm as the detection method adopted in the system according to the experiment results of each method.According to the gap edge linking method and fuzzy decision based edge connection method, a new method is given during CT image processing. This method can connect the margins automatically to some extent, which will reduce the labor of human-computer interaction. It can meet the requirements of edge connection in processing CT images.A modified method of intersecting line identification is applied in fitting lines of industrial CT images to improve the accuracy of line identification and fitting.User observation test is applied to assess the availability of system. Base on the investigation and research, problems existed in human-computer interaction was found out and solved, which would improve the convenience of human-computer interaction.Finally, compare with business software, the optimization results was validated. Also, some shortcomings were discovered which would help guide the next step of optimization.