硬聚类算法隶属度只有两个值 0 和 1，即一个样本只能完全属于某一个类或者完全不属于某一个类，适用于背景和目标区别明确的图像。-Hard clustering algorithm membership only two values 0 and 1, that is, a sample can only belong to a class entirely or completely do not belong to a particular class for a clear distinction between background and target image