Digital change detection using satellite data is a potentially powerful means for a variety of applications, such as determining the extent of flooding of large rivers, and assessing urban expansion. Where, the original digital images used in the change detection process should be registered to each other prior to use them in one of change detection methods. In this thesis, the change detection technique used is the image differencing, and two methods are applied to the original images prior to use this technique. The first is edge-based method and the second is a supervised classification method based on "minimum distance to means" decision rule followed by closed contour extraction process. In the first method, an algorithm for detecting the edges is used which comprises edge thinning and removing of isolated points, also, another algorithm for detecting straight lines is developed which comprises identifying the edge point directions and storing them in a direction file, this step is followed by edge thinning, edge linking according to its direction, and removing of-isolntcd points. The next steps of this developed method are edge chains extending to form line segments, and the final step is extending and connecting the resultant straight lines from the previous steps. In the second method, the classification maps which result from the classification process of the digital image are used to extract closed contours which represent the features of the area under study using a simple and effective procedure. All the programs used in this thesis are written in quick basic language using personal computer (PC) with main processor type 80486.