An Inproved Up-scaling Algorithm Combined TSA and PSF

Abstract

Abstract Taking the Guanzhong Plain of Shanxi Province as the research area, the combine method of trend surface analysis method (TSA) and point spread function (PSF) (TSA+PSF) were used to up-scale the vegetation temperature condition index (VTCI) retrieved from Landsat 8 images (Landsat-VTCI) from a finer resolution to a coarser resolution. The up-scaled results were compared with VTCI images retrieved from Aqua MODIS (MODIS-VTCI) to provide technical support for the comprehensive application of drought monitoring results on two spatial scales. Meanwhile, a range of indicators, such as the semivariogram function (SVF), the structural similarity (SSIM), the correlation coefficients (r), root mean square errors (RMSE) were used to systematically compared the up-scaled methods. The results show that TSA+PSF performed better than TSA in terms of SSIM, the correlation and RMSE, the up-scaling model TSA+PSF has the higher accuracy, and it is more effective and robust than TSA. The model that uses PSF to analyze trend surface constructed by TSA is an improvement for up-scaling Landsat- VTCI images from a finer resolutions to a coarser resolutions.