Spatial scaling transformation modeling based on fractal theory for the leaf area index retrieved from remote sensing imagery

Abstract

This paper proposes a scaling transfer model based on fractal theory to retrieve the leaf area index (LAI) at different spatial resolutions and to evaluate the scaling bias on the LAI retrieved from coarse resolution images. The LAI scaling transfer model was developed by establishing the double logarithmic linear relationship between the scale n (spatial resolution) and average LAIs of the image at different scales. Thereafter, the influences of four factors, namely, coefficients of LAI retrieval model, image size, spatial resolution, and image standard deviation, which may have impact on the scaling transfer model were analyzed. The results indicated that the scaling transfer model performed well in estimating LAI with a determination coefficient ( R 2 ) value of 96.99% and in evaluating the scaling bias with a root-mean-square error of 0.0188. The scaling transfer model was considerably influenced by the image standard deviation. As the model parameter, the fractal dimension of image was highly correlated with the standard deviation of the normalized difference vegetation index image. Results indicated that the proposed method based on fractal theory is feasible for LAI spatial scaling transformation.