FUSION OF REAL AND SYNTHETIC-IMAGES FOR REMOTE-SENSING SCENE UNDERSTANDING

The high complexity of remotely sensed images and measurements, provided by the last generation of sensors, demands new techniques for scene understanding and analysis. The paper introduces topics in multisensor and synthetic images fusion. The fractal geometry is applied for unknown information modelling. Integrated techniques in computer graphics and computer vision are used. A new method is introduced for more accurate representation and visualisation of the fractal surfaces. A multiresolution approach is considered for the accurate description of surface radiometry and geometry. The image synthesis is based on the knowledge of surface geometry, on radiation source, and sensor characteristics, and on the radiation scattering process for different cover types. In the final part, results of the developed techniques are applied in satellite imagery understanding.