Object-Oriented Raster-Based Data Model for Urban Growth Simulation System

The growth of urban areas into forested and agricultural lands is a local and global planning issue of great concern. The proposed model is hence based on the raster data model. The raster-based GM is implemented using object-oriented Microsoft Visual C++ programming under the Microsoft Windows environment. To satisfy more analysis methods expended for this system, such as CA and neural network models, a virtual base class (CPredictModel) is defined for their united interfaces. The vector data are integrated into the raster-based GM method by the vector-to-raster conversions. To validate the raster-based data model, time series of raster data are generated from Landsat TM for the Jilin City, China, in 1987, 1996, and 2000. The changes between any two dates can be analyzed efficiently, and the growth trends of Jilin City in the future are evaluated successfully by the presented raster-based data model.