摘要：

针对生物切片图像颜色相近、边缘复杂、细节模糊等问题，提出一种基于形态学的图像分割方法。将彩色图像转换到HSV颜色空间后进行二值化，经形态学填充、去噪、膨胀得到图像分割模板，运用到原图各分量上，实现目标的批量分割。将该方法应用于2组小鼠舌头切片图像的目标分割中，实验结果验证了该方法的有效性和鲁棒性。

关键词: HSV颜色空间, 最大类间方差法, 形态学, 结构元素, 阈值分割

Abstract:

The biopsy image is difficult to segment, since it usually contains similar color, complex edge, fuzzy boundary and so on. To solve this problem, this paper proposes a segmentation method of slice image based on morphology. It transforms the color space from RGB to HSV for binarization. Then it gets segmentation template after operations of filling, denoising, and expanding the image with morphology. The template can be applied to each component of original image, in order to realize the batch segmentation of objects. Experimental results verify the validity and robustness of this algorithm after two groups of experiments on mice tougue.

Key words: HSV colorspace, Otsu method, morphology, structural element, threshold segmentation

中图分类号: 

• TP391