Methods for removing 'honeycomb' noise of fiber-endoscopic images

摘要

The fiber-endoscope has been widely used in medicine. The image fiber bundle usually has pixels from several thousands to tens of thousand. Because of the non-transparent wall cladding of individual fibers, the images putout by the image fiber bundle present "honeycomb" pattern (noise). It will influence the image visual effect, so it is very important to find methods to remove these "honeycomb" noise and improve the image quality. In this paper, three methods were used to process the fiber-endoscopic images for removing the honeycomb noise. First, low-pass spatial filtering mask was used to process the image. Secondly, the image special frequency was gotten by Fourier transform, and the honeycomb pattern frequency is separated from the image message. It's possible to remove these honeycomb pattern frequency without degrade the iamge quality. Finally, the linear interpolation method was used to process the image. We compared the processing results of these methods. These methods can be used in real color images as well as gray level images.