Defects detection for machined part surface texture image can obviously enhance the entire efficiency for the automatic manufacturing.Because of the particularity of texture image for the machined part surface

the normal texture image segmentation is useless in this kind of defects detection.Since the infomation of defects on machined part surface don’t be got in advance

it is a kind of unsupervised texture segmentation.A modified fuzzy clustering method which can detect the defects on machined part surface automatically is presented.The results of experiment demonstrate the feasibility of the algorithm.