中文摘要：

悬浮细胞的培养过程需要对活细胞计数．针对人工计数方法存在的缺陷，利用图像分析技术对拍摄的细胞图片进行处理，实现活细胞自动识别及计数．要实现细胞识别，首先要对图像进行正确分割．对于前景和背景对比度较小且细胞很稀疏的图像，采用背景拟合设定阂值进行分割；针对前景和背景对比度较大的图像，采用一种基于局部区域直方图的方法进行分割．对分割后的图像进行链码跟踪并测量细胞参数，最终实现细胞识别及计数．实验结果表明上述方法能快速有效地计算出活细胞个数．

英文摘要：

It is needed to count live cells after suspension ceils being cultured. To solve problems caused by counting ceils manually, live cells are recognized and counted automatically by image-analyzing technique. Two methods are used to segment two different kinds of ceil images respectively. One method is based on background fitting to the gray of original cell image and the other on a local histogram. Then the parameters of every cell can be obtained by tracking a segmented image by chain code. With these parameters, all cells in a cell image are tracked and counted quickly and efficiently.