【摘要】 研究了显微图象中目标对象的识别问题。首先比较了显微图象和宏观图象之间的区别,总结了大部分显微图象具有的一般特征,讨论了宏观图象处理方法用于显微图象时存在的缺陷。然后提出了一种目标对象识别的新方法,该方法利用最优颜色通道识别显微图象中的目标对象,动态地调节识别阈值,减少了误识别象素的数量。对残存的误识别象素和脏点象素采用面积滤波的方式去除,进一步提高识别的效果。该文通过实验,识别了几种显微图象中的目标对象,实验发现,使用这一方法不仅能够提取背景中的目标对象,还可以大量减少误识别象素和脏点象素,说明该方法用于显微图象中目标对象的识别具有一定的优点。更多还原

【Abstract】 The recongnition of objects in microscopic images is studied in this paper.Firstly,The common characters are summarized by comparing microscopic images with macroscopic images.And the limitation of macro image processing algorithms is also discussed.Secondly,a new method of recognition which uses the best color channel and dynamicly adjusts the recongnized threshold is developed.This method can reduce the number of erroneous and dirty pixels.The quality of recognition will be improved ulteriorly through erasing the remained erroneous and dirty pixels.The objects in different microscopic images are recongnized through experiments.Experiments show that the method of recongnition can not only acquire the objects from background but also reduce the number of erroneous and dirty pixels greatly.And the method is suitably applied to recongnize the objects in microscopic images更多还原