As a key input parameter in many climate and land-atmosphere models, the validation of retrieved leaf area index (LAI) on regional scale from remote sensing data makes great senses. The problem of scale between the field experiments and the ground parameters retrieved from satellites is still one of the most difficult problems in the validation of satellite remote sensing data. The difficulty is twofold: First, the field measurements are not exhaustive; Secondly, the model is not linear and surface on satellite pixels is not homogenous. Therefore the objective of the scaling transform study is to estimate a non-linear function describing spatial distribution information of pixels from information on sub-pixels. The Computational Geometry Model is a general spatialization method which can realize the scaling of non-linear and discontinuous function. However it needs a large amount of computing time and a special algorithm to retrieve convex hull when facing a large number of input arguments. In this paper QuickHull algorithm is introduced to resolve the scaling problem of the bivariate LAI retrieval function. The scaling effect is analyzed through aggregating the high-resolution LAI (pixel size of 30 meters) retrieved from TM images by means of CGM method and directly aggregated method respectively. The CGM method is proved to have the capability of improving the scaling effect of LAI at larger aggregated scales. It is a prospect method to resolve the scaling problem and will take effect for the validation with limited field experiments.