SPATIAL SCALING TRANSFORMATION MODELING OF VEGETATION LEAF AREA INDEX RETRIEVED BY REMOTE SENSING IMAGE BASED ON FRACTAL

摘要

Scaling transformation is one of the basic and important scientific questions in quantitative remote sensing. This study proposed a leaf area index (LAI) scaling transfer model based on fractal theory for computing LAIs at different scales (spatial resolution). Based on scale invariance and self-similarity of remote sensing images in a statistical sense, the LAI scaling transfer model was developed by establishing the double logarithmic linear relationship between the scale n and the average LAIs of the image at different scales. The influence of the standard deviation of the image on the scaling transfer model was also analyzed. The results showed that the average LAIs of the image at different scales were well calculated by the scaling transfer model with an absolute percent error (APE) value of 0.27% and a root mean square error (RMSE) value of 0.0129. The fractal dimension of the image, the parameter of the scaling transfer model, increased as the standard deviation increased. This study suggests that the proposed method of LAI spatial scaling transformation based on fractal theory is feasible.