An Defect Detection Method Based on Zernike Moments for Industrial CT Image

摘要

Industrial comput fault scan imaging is able to provide intuitive and clear images, without any overlapped pictures or geometrical distortion and it is widely used in nondestructive test. When we use industrial CT to conduct nondestructive test on defect and assembly error of batches of complicated workpiece, we meet some problems about distinguishing CT image with manual method, thus assembly defect detection method based on Zernike moments is put forward, which is identification method in two steps. Firstly, it uses rapid arithmetic to select a possible matched location through reliable method with a low amount of calculation. Secondly, it will conduct further estimation on matched location using Zernike Moments and then select accurate matched location. The test has indicated that selection method we adopted is feasible and it is effective to use Zernike Moments for the final judgment. This method is fairly helpful for detecting neglected-loaded accessories and error assembly location.