At present, almost all algorithms are developed for specific images. So the paper puts forward the algorithms of removing noise and segmenting images according to the previous algorithms and the characters of digital images acquired from Animal Science and Medicine College of Inner Mongolia Agricultural University.In removing noise, the paper puts forward a new automatically adapting median filtering algorithm, the algorithm makes the 3×3 median filter, 5×5 median filter and multilevel weight median filter as the candidates, automatically chooses a optimum filter from them to removing noise following the least variance rule, and the effect is better. In karyons segmentation, the paper puts forward a new genetic clustering algorithm basing on the entropy of the image, the algorithm takes the entropy of the image as the applicability function of genetic clustering algorithm and realizes more perfect karyons segmentation. The preprocessing results establish the foundation for farther processing.Furthermore, the paper finishes some interrelated preprocessing. All algorithms are compiled and run with Visual C++6.0 on Windows2000, with the better flexibility and convenient operation.
|