The color feature is the crucial basis for segmenting a stratum image

but traditional segmentation algorithms based on color feature clustering cannot get desirable results because of colors mixture near the layers' boundaries and heavy noisy.A new image segmentation method is proposed

it adopts an interactive image segmentation algorithm based on Random Walks to improve the computing method of the membership functions of fuzzy C-mean(FCM) incorporating spatial information.It inserts labeled cluster-nodes into the graph formed with pixels and their 4-connectedness

and takes the probabilities that a random walker starting its walk at a pixel first reaches a labeled cluster-node as the membership it belongs to this cluster.The experimental results show that pixels in the area of mixed colors near the layers' boundary are classified more accurately;the method decreases the noise sensitivity of FCM

and can effectively segment the structure physical modeling images.