Abstract
The objectives of this study are to extend analysis functions of the multi-Temporal Analysis model with the Fractal theory (termed TAF model) and to improve the accuracy in analysis of land cover change applying the synthetic aperture radar (SAR) data. Through the box counting procedure, the fractal dimensions of the changed areas for the residential and the road-network were calculated, respectively. The high correlation between these two fractal dimensions was confirmed in twelve test areas. Especially, in case of using SAR data, the separability of the four typical types of the residential form and growth; the mono-focal, the ring/radial, the multi-focal and the twin type, was improved. These results suggest that SAR data could be substituted for the optical sensor data in the multi-temporal analysis with the fractal dimension.
