提出了一种新的用于图像平滑处理的滤波器自适应聚类滤波．它具有以下特点：（１）滤波器对像素点邻域内的灰度进行聚类，将聚类中心作为滤波器的输出；（２）滤波器能够根据像素点周围的拓扑特征，自适应地确定平滑窗口和窗口内平滑权重系数的大小，具有输入空间和输出空间两方面的自适应性．自适应聚类滤波用于图像的平滑处理，对滤除冲激噪声特别有效，同时较好地保持了边缘信息．实验结果证明了算子的有效性 A new filter adaptive clustering filter for image smoothing is proposed. It has the following characteristics: (1) The filter classifies the gray level in the neighborhood of pixels, and takes the cluster center as the output of the filter; (2) The filter can adaptively determine according to the topological features around pixels The smoothing window and the size of the smoothing weight coefficient in the window have the adaptability of both the input space and the output space. Adaptive clustering filtering is used to smooth the image and is especially effective in filtering impulse noise while keeping the edge information well. Experimental results show the validity of the operator