This paper aims to present our research and discussion on establishing a practical system which integrates the technologies of Image Processing, Photogrammetry and GIS for the purpose of middle and large scale map making and revision. As it is very difficult for a computer to understand an image automatically and the low effiency of the data aquisition is still the big limitation of the application of GISs now, an idea of operator-guided semi-automatic processing for geographic data acquisition was proposed and studied in this paper.