最大类间方差法在图像分割中的应用

基于阈值分割的方法,在图像分割过程中,物体像素的灰度级与背景像素的灰度级有不同,可以将物体从背景中较好的分割出来,总能用封闭而且连通的边界定义不交叠的区域。本文针对细胞图像的特点,首先对细胞图像分别利用经典的Sobel算子和Log边缘检测进行阈值分割,此分割方法是基于最大类间方差准则的分割。然后利用最大类间方差法对图像进行阈值分割。分割效果较令人满意。经实验证实运算速度较快,实现起来较为简单。

Applications of Maximum Between-cluster to Image Segmentation

Based on Threshold Value Division method,in the processing of image segmentation the gray levels of object pixel and background pixel are different. So we can divide the object better from background and can always difine the non-overlapping area with closed and connected boundary.The text according to the features of cell image to divide.First,using classical Sobel operator and Log edge detection respectively.This method based on the criterion of maximum between-cluster.Then using the method of maximum between-cluster.The effect of division is satisfied.The operation speed is faster and it is easily come true by experiment.