基于结构重建距离分水岭的细胞图像目标提取  

Cell image target extraction based on structural reconstruction distance watershed

作　　者：李凯臣 李俊芳[1,2] 于晓[1,2] 许晓刚 姜艳艳 LI Kaichen;LI Junfang;YU Xiao;XU Xiaogang;JIANG Yanyan(School of Electrical Engineering and Automation,Tianjin University of Technology,Tianjin 300384,China;Tianjin Key Laboratory for Control Theory and Application in Complicated Systems,Tianjin University of Technology,Tianjin 300384,China)

机构地区：[1]天津理工大学电气工程与自动化学院,天津300384 [2]天津理工大学,天津市复杂系统控制理论与应用重点试验室,天津300384

出　　处：《天津理工大学学报》2023年第4期39-45,共7页Journal of Tianjin University of Technology

基　　金：国家自然科学基金(61502340);天津市自然科学基金(18JCQNJC01000);天津市教委科研计划项目(2018KJ133);天津理工大学教学基金(YB20-05);天津理工大学教材建设基金(JC20-11);天津市复杂系统控制理论及应用重点实验室开放基金(TJKL-CATCS-201907)。

摘　　要：细胞形态不一与粘连性严重是细胞图像的一类表现特征,准确提取出各个细胞是细胞图像目标提取的重点。通过分水岭算法的图像目标提取方法,可得到连续、封闭及单像素宽的精确边缘,准确定位出目标细胞。通过形态学和距离法对图像进行细节处理,提出基于结构重建距离分水岭的细胞图像目标提取方法。运用形态学重构法解决噪声和孔洞等杂质影响,采用距离法处理细胞间的粘连性。试验结果表明:该方法对于粘连性严重的细胞同样能实现较准确的提取,且相对于其他图像目标提取算法,提取速度更快。Cell morphology is different,and serious adhesion is a kind of cell image performance characteristics,accurately extracted each cell is the focus of cell image target extraction.Based on the watershed algorithm,the image target extraction method can obtain precise edges of continuous,closed and single pixel width and accurately locate target cells.Detailed images are processed by combining morphology and distance method and a method of image target extraction based on structural reconstruction distance watershed cells is processed.Aiming at problems of cell image noise,impurities and complex cell adhesion,the algorithm solves the influence of noise,holes and other impurities by morphological reconstruction and processes the adhesion between cells by distance method to achieve accurate and rapid extraction cells in cell images.At the same time,more accurate extraction can also be achieved for cells with severe adhesion.The experimental results show that the method can extract cells more accurately and quickly than some other image target extraction algorithms.

关 键 词：细胞图像目标提取 粘连性处理 形态学重建 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]