Simulate TM Image from MODIS Data, Using Linear Mixture Model

摘要

Satellites sensors, e.g., MODIS, NOAA-AVHRR, SEAWIFS, are widely used satellite data. However, the coarse spatial resolution of these data limited the application. In this paper, we propose a novel approach for the simulation of high spatial resolution image from mutil-temporal coarser resolution satellite data that will has both high spatial resolution and temporal resolution. Through detecting temporal change information from time series MODIS images, the algorithm produces high temporal synthetic Landsat data from a previous Landsat image. The algorithm was tested over a 1400km2 area located in Arkansas, USA. High correlation coefficient r of 0.81 between synthetic imageries and actual observations was acquired.