Plenty of researches have been presented to combine satellite remote sensing observations with ground-level air quality monitoring sites measurements together to accomplish spatially continuous observation of air pollutant, such as using Aerosol Optical Depth (AOD) derived from Moderate Resolution Imaging Spectroradiometer (MODIS) and ground observed fine particulate matter (PM2.5) to estimate the PM2.5 concentrations that has the same spatial coverage and continuity with MODIS AOD. However, because the temporal resolution of MODIS, i. e. twice in daytime, is too limited to accomplish real-time air quality monitoring. Hence, geostationary satellite with higher temporal resolution is introduced in this study, e. g., Multi-functional Transport Satellite 2 (MTSAT-2), which provides Earth observations for every half an hour or even more frequently. This paper presents an integrated spatial-temporal-spectral image fusion model to blend the spatial, temporal and spectral resolution of MODIS and MTSAT-2 images to generate the synthetic data with high spatial, temporal and spectral resolution simultaneously, which can provide more comprehensive data for real-time air quality monitoring. The fusion process involves MTSAT-2 visual band sharpening, spatial-temporal fusion, and temporal-spectral fusion, which aims to improve spatial, temporal, and spectral resolution respectively. The proposed image fusion model was tested on one set of MODIS and MTSAT-2 images within the Pearl River Delta (PRD) region, China. Experimental results indicate that, the advantages of the spatial, temporal, and spectral resolution of these two satellites are merged quite well, which can produce synthetic half-hourly MODIS-like images for real-time air quality monitoring.
