Remote sensing data is the essential source of information for enabling monitoring and quantification of crop state at global and regional scales. Crop mapping, state assessment, area estimation and yield forecasting are the main tasks that are being addressed within GEOGLAM. Efficiency of agriculture monitoring can be improved when heterogeneous multi-source remote sensing datasets are integrated. The research activities are directed to the integrated exploitation of MODIS, Landsat-8, Sentinel-2 and Sentinel-1 data along with meteorological data for crop yield forecasting, mapping and area estimation. Archived coarse spatial resolution data, such as MODIS, VIIRS and AVHRR, can provide daily global observations that coupled with statistical data on crop yield can enable the development of empirical models for timely yield forecasting at national level. With the availability of high-temporal and moderate spatial resolution Landsat-8 and Sentinel-2A/B imagery (at 10-30 m), course resolution empirical yield models can be downscaled to provide yield estimates at regional and field scale. Since the yield model requires corresponding in season crop masks, we are developing automatic approaches for crop mapping integrating satellite imagery, meteorological information, and information on crop calendar. Different machine learning algorithms are extensively applied to solve tasks of regression and classification. Coarse resolution crop maps are further used for stratification purposes within area frame sampling for crop area estimation. This approach is successfully applied for estimating the area of winter crops in Ukraine for 2016 using Landsat-8 and Sentinel-2A images. Thanks to combination of statistical approaches and machine learning techniques, the derived values will also include the corresponding uncertainties of the estimates.
Roger Eric Vermote