基于模糊C均值聚类和数学形态学的图像分割

Image segmentation based on fuzzy C-means clustering and mathematical morphology

心肌细胞钙离子实时激光扫描共聚焦光学切片呈现为点状分布的荧光图像并且受到噪声的严重干扰,单独利用模糊C均值聚类不能对这种图形进行有效分割。针对这种特定的图像提出了一种基于模糊C均值聚类和数学形态学的图像分割算法。首先利用邻域平均对图像预处理,然后利用模糊C均值聚类做分割,最后利用数学形态学的方法对图像做了平滑、连通和去噪处理。这种方法,不但有效地抑制了噪声,而且分割出的图像边缘连续、清晰。

The fluorescent images on the laser scanning optical biopsy of the myocardial cell calcium-confocal have the points-like distribution and interfered seriously by the noise. Using alone the fuzzy C-means clustering such graphics cannot be effectively divided. Hence an image segmentation algorithm based on the fuzzy C-means clustering and mathematical morphology is presented. This algorithm uses an average field of the image segmentation to preprocess, then uses the fuzzy C-means clustering to divide and finally uses the mathematical morphology of the image a smooth, connectively and de-noising treatment is made. This approach effectively curbs the noise so that the split edges are continuous and clear.