| Enough contrast and clearness of X-ray image are the pre-condition of detecting defects and industry CT reconstruction. Due to the properties of the X-ray system and the disturbance of the external circumstance, the actual X-ray images have large noise, leading to be difficult in the following image processing.Image edge detection plays an important role in the image processing. The results of the conventional edge detection operators are somewhat non-continuous and missing. Furthermore they are sensitive to noise. For noise image, filtering followed by edge detection is generally accepted. But this procedure will blur the image edge, while eliminating noise. It is difficult to hold image edge and eliminate noise simultaneously.Mathematical morphology, a methodology of nonlinear filters, offers a better method to solve the problem mentioned above. It is a new developing science based on strict mathematical theory. It influences the theory and technology of image processing deeply, and has been widely used in the area of image processing such as noise suppression, edge detection, image segmentation, feature extraction and so on.The following research is mainly done, on the basis of studying the theory of the mathematical morphology in depth.1. There are drawbacks such as low detection precision, being sensitive to noise, in conventional edge detection operator. A new morphological gradient based on opening and closing is proposed. The results indicate that this method have good effect for ideal image and noise image.2. A method to enhance the image contrast is studied, which is based on Toggle mapping. From the results, it can give a good enhancement to low contrast image.3. According to different noises in the X-ray image from factory, we have designed three different morphological filters. Eliminating noise and holding image structure information are simultaneously obtained by using these filters, which is almost impossible for linear filters. |
PDF Full Text Request