As we know, a main problem in image classification and explanations the segmentation of image. Use of the boundary contour rather than most of other preprocessed data help in simpler image representation, registration, and identification. Thus the problems of edge detection and contour following have been widely addressed with proposed solutions ranging over simple thresholding, signal processing, filtering techniques, and definition of admissible edge profile function, etc. Most of these techniques seem to be complex and time consuming. To solve the conflict between noise and time consume, this paper presents a fast method for detecting the contour of images. It consists of three steps: first, reduce the image noise by using a high frequency noise sensitive smoothing operator; second, detect the edge by using an algorithm based on modified Sobel operator; and third, follow the boundary contour and determine the area of interesting in the image. The results of experiment show that the method is fast and accurate.