Study On Intelligent Mosaicing And Fusion Technique For Microscopic Images

In many important fields such as material detection，biology techniques，medical analysis，microscopic image analysis now is widely used because the image of this kind contain theinformation eye usually cannot see, and some minor problem can be discovered through thisway．At most time, photographic apparatus get the microscopic images, and then the informationare converted digital images which computer can recognize and analyze，then the information canbe stored in the computer. But due to the limit of photographic apparatus and ambient conditionwhen photos are taken, the view which camera can obtain is much smaller than that of microscopes,which makes all the information cannot be showed in one panorama. In order to restore thepanorama of the whole specimen，certain ways of image mosaics technique are provided, andacquisition of a big image can be achieved．According the reasons as demonstrated above，imagemosaics technique is of great values in research and applications． Because of some disturbancewhich always exist during taking microscopic image, the same view has various focal planes．Thispaper demonstrate the ways how images are mosaiced and fused, and discuss some image fusiontechniques which are commonly used, and integrates partly clear images with different focuspoints to a whole clear image to present a whole image which contains all the information．According to the feature of obvious differences in microscopic images of light and geometrytranslation larger, this paper provides a SIFT feature matching to image mosaic. The SIFT isinvariant with rotation, scale zoom, gray change, and is stable for visual angle change, affinetransformation and noise.Comparing to the results of mosaicing, some mismatch is found in certain areas, and a newmethod is provided. Using this method, several microsection image are mosaiced.