| In remote sensing technique, scale transformation is a very effective and new method to extract remote sensing image with different resolutions, which has become a concern in the study of remote sensing technique, especially the information extraction. Hence, the study on scale transformation is of great significance on both theories and applications.The development of remote sensing in the past decades identifies that fractal theory which plays a useful role in this subject. Research shows that most of the natural world has the approximately same fractal characteristics with the gray image getting from the same field, and more importantly the majority of the nature phenomenon in gray image met with isotropic fractal brown random airport (FBR). Fractal dimension with the objects in image matches with fractal dimension with the objects on the surface, but the different textures in nature are different, hence, fractal dimension is regarded as one of the effective parameters which is used to indentify different f texture type. Fractal theory is important in spatial and spectral of remote sensing images.Scale transformation is the concern of this paper. First, we analyzed the fractal theory and geometric characteristics with gray image and identified the fractal dimension which was used as theory basis of our study. Then, we studied the character of the scale transformation models which fit calculated fractal dimension of the whole image rather, however these models don' t use to get fractal dimension of the pixel. We identified a new method of the calculation of fractal dimension of pixel. Then based on this new method, we analyzed fractal characteristics of different resolution images, studied the interdependent relationship between spatial resolution and fractal dimension. This paper introduceda new scale transformation methods based on non-variable-the dimension.Finally, using of image quality evaluation criteria, the new scale transformation methods were evaluated.
|
PDF Full Text Request