粘连蚕卵图像分离算法研究  

Study on Separating Algorithm for Overlapping Silkworm Ovum Image

作　　者：魏瑾[1] 韩斌[1] 张其亮[1] 

机构地区：[1]江苏科技大学电子信息学院,江苏镇江212003

出　　处：《江苏科技大学学报（自然科学版）》2006年第1期46-50,共5页Journal of Jiangsu University of Science and Technology:Natural Science Edition

基　　金：民品技术攻关课题(校编540126)

摘　　要：针对蚕卵图像中存在的粘连或者部分重叠现象,本文提出了一种新的分离方法。该方法先采用模糊C均值聚类算法将蚕卵图像从背景中分割,然后采用模糊距离变换得到模糊距离图像,最后使用分水岭算法将粘连或者部分重叠蚕卵分离。实验证明,此法能够将粘连颗粒物图像有效分离。In the actual silkworm ovum image, there are always conglutination or overlapping. Aimed at this phenomenon,a new algorithm used to separate the overlapping silkworm ovum is introduced. Firstly,a fuzzy C-means algorithm is developed to divide the silkworm ovum from the image. Then the fuzzy distance transform is applied to obtain the fuzzy distance image. Finally, the watershed algorithm is used to separate the conglutinant or overlapped silkworm ovum. Experimental results show that this method has a high performance in segmenting the conglutinant silkworm ovum. It is also applicable for the separation of all kinds of overlapping granule images.

关 键 词：图像分割 模糊C均值聚类算法 模糊距离变换 分水岭算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]