中文摘要：

寄生虫卵的自动识别是当今寄生虫医学图像处理的一个重要课题，目前已有算法一般都要求寄生虫标本杂质含量较少。提出一种基于形态学滤波的混合分割算法，首先采用B颜色信号提取有用信息，接着利用改进的形态学操作进行滤波，除去大量杂质及虫卵边缘粘合物，最后结合凸包运算并定义两个图像特征参数，即边界光滑度和区域填充度，做进一步选择。实验结果表明，该算法明显优于当前的一些虫卵分割算法，能充分利用虫卵的有用信息，有效剔除杂质，大大降低虚假目标。虫卵边界保留完整清晰，为寄生虫卵的自动识别打下了良好基础。

英文摘要：

Automatic recognition of parasite egg is an important topic in the field of parasite medical image processing. At present, existing algorithms always require fewer impurities in parasite specimens. In this paper, a hybrid segmentation algorithm based on morphological filtering was presented. Firstly, a B color signal was used to extract useful information, and then an improved morphological operation was applied to remove a large number of impurities and the egg adhesive materials. Finally, the parameters of image feature were acquired, such as area filling degree and boundary smoothness degree through convex hull operation for further selection. Experimental results have proved that the proposed algorithm is superior to the other existing segmentation algorithms. It can make full use of eggs＇ useful information, effectively remove impurities, significantly reduce the false targets, and keep the eggs＇ borders intact and clear. In conclusion, it lays a good foundation for the automatic recognition.