Rapid urbanization has many social and environmental impacts, so it is important to be able to monitor its evolution and try to limit its extent and effects. In countries without geographic or statistical information system, satellite images are the main source of data to observe and quantify this phenomenon at the scale of a metropolitan area. This paper introduces a methodology that utilizes satellite images to extract vital information for analyzing urban growth. The method focuses on determining the density of built-up areas, the homogeneity of these built-up areas and their boundaries, and the structuring of their agglomeration into centers and subcenters. This enables the assessment of phenomena such as urban sprawl, leapfrog development, and area saturation. The approach combines local fractal analysis and mathematical morphology techniques to derive the desired information. Their usefulness is illustrated by a qualitative analysis of the growth of the Greater Accra Metropolitan Area (GAMA), Ghana, as well as in a preliminary neural network modeling of the urbanization process of this same area.