基于优化模糊增强的顺序形态学细胞图像边缘检测算法  

Edge Detection Algorithm of Cell Images Based on Optimized Fuzzy Enhancement and Sequential Morphological

作　　者：张瑞华[1] ZHANG Ruihua(School of Physics and Information Engineering,Jianghan University,Wuhan 430000,China)

机构地区：[1]江汉大学物理与信息工程学院,武汉430000

出　　处：《湖北民族大学学报（自然科学版）》2021年第1期86-90,96,共6页Journal of Hubei Minzu University:Natural Science Edition

基　　金：国家自然科学基金项目(61575085);湖北省教育厅教学研究项目(2016281);武汉市教育局教学研究项目(2017086).

摘　　要：为实现对医学细胞图像的边缘检测,提出了一种基于优化模糊增强的顺序形态学边缘检测算法.首先利用优化模糊增强算法增强细胞图像边缘,优化算法采用二维直方图斜分法初步定位边缘区域并结合自适应隶属度阈值进行模糊增强.然后使用尺度为5或4个不同方向的结构元素,百分位(p,q)为(0,1/2)和(1/4,3/4)两种百分位对增强后的细胞图进行复合顺序形态学边缘检测,得到多张边缘子图,最后融合多幅子图,细化后得到最终的细胞边缘.对多幅细胞图像进行边缘检测试验,结果表明该算法在增强边缘的同时可消除部分边缘模糊和粘连,检测到的边缘连续、清晰、完整,效果明显优于canny算子、水平集c-v算法和一般形态学检测算法.For edge detection of medical cell images,an edge detection algorithm of cell images based on fuzzy enhancement and sequential morphological is proposed.Firstly,the edge of cell image is enhanced by using optimized fuzzy enhancement algorithm.The optimization algorithm uses the two-dimensional histogram oblique method to locate the edge area and uses adaptive membership threshold to realize fuzzy enhancement.Then the enhanced image is divided into several edge sub-images by using different direction elements and different size percentiles.Multiple sub-images are fused based on information entropy.Finally the edge detection results are obtained by morphological refinement.Experimental results show that the algorithm can eliminate some edge blur and adhesion while enhancing edge,and the detected edges are continuous,clear and complete,whose effect is apparently superior to canny operator,level set c-v algorithm and general morphological detection algorithm.

关 键 词：边缘检测 模糊增强 细胞粘连 水平集 阈值 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TP37[自动化与计算机技术—计算机科学与技术]