中文说明：对于细胞图像序列中多目标的追踪是细胞运动研究中的难点，针对高密度细胞图像中细胞运动的复杂性，本文提出一个细胞分割和追踪的系统。在细胞分割部分，针对实验所用细胞图像序列的特点，本文分别采用了不同的分割方法。在基本的细胞分割后，由于得到的分割图像存在着一些粘连细胞，为了将之分离，采用了基于Freeman code法对细胞轮廓进行跟踪编码。根据编码所得的链码特征分析细胞的轮廓形态，找出粘连细胞图像中的凹角点，再将凹角点进行分组配对完成粘连细胞的分离。在追踪部分，针对细胞的运动特性，将细胞分为惰性细胞和非惰性细胞，分别采用区域重叠法和拓扑约束法进行追踪。针对高密度细胞追踪中由于细胞密集加上细胞分裂很容易导致错误追踪的问题，本文提出将面积项引入实现追踪的拓扑约束条件中，对原方法进行改进，有效的解决了这一问题，有助于提高追踪效率，特别是对于细胞密度比较大，且分裂频繁的细胞图像效果明显。另外，本文提出根据细胞数量设置产生图的距离门限，使算法的适用性更强 。
English Description：
Tracking multi-target in cell image sequence is a difficult problem in cell motion research. Aiming at the complexity of cell motion in high-density cell image, this paper proposes a cell segmentation and tracking system. In the cell segmentation part, according to the characteristics of the cell image sequence used in the experiment, this paper adopts different segmentation methods. After the basic cell segmentation, because there are some adhesion cells in the segmentation image, in order to separate them, the Freeman code method is used to track and code the cell contour. According to the chain code features, the contour of cells was analyzed to find out the concave corners in the image of adhesive cells, and then the concave corners were grouped and paired to complete the separation of adhesive cells. In the tracking part, according to the movement characteristics of cells, the cells are divided into inert cells and non inert cells, and the region overlapping method and topological constraint method are used to track the cells respectively. In high-density cell tracking, it is easy to lead to wrong tracking due to dense cells and cell division. In this paper, the area term is introduced into the topological constraints to realize the tracking, and the original method is improved, which effectively solves this problem and helps to improve the tracking efficiency, especially for the cell image with large cell density and frequent division. In addition, the distance threshold is set according to the number of cells to make the algorithm more applicable.