Abstract: Aiming at detecting regions with missing information in texture image, a coarse-to-fine method is proposed to achieve the automatic locating of defective regions.In the coarse locating phase, Gaussian filtering and threshold segmentation are employed to remove background texture and acquire the rough contour of defects.In the precise positioning stage, an improved Chan-Vese active contours model based on fuzzy energy is introduced and the rough contour is taken as the initial curve for the model.During implementation, level set method is used to find the precise edges of defective regions. Experimental results show that our method is able to achieve accurate and automatic locating of different defective regions with high computational efficiency and robustness.