Statistics Colloquium

Machine learning and signal processing techniques are gaining popularity in the areas of application related to remote sensing that is used in precision agriculture. This talk will introduce the results of using multivariate relevance vector machine models combined with wavelet analysis techniques to estimate, model, and forecast evapotranspiration (ET) from the local weather station variables and Landsat satellite images, and chlorophyll levels from unmanned aerial system called AggieAirTM. With the development of surface energy balance analyses, remote sensing has become a spatially explicit and quantitative methodology for understanding evapotranspiration (ET), a critical requirement for water resources planning and management. Limited temporal resolution of satellite images and cloudy skies present major limitations that impede continuous estimates of ET. First, a methodology combining wavelet multiresolution analysis with a machine learning algorithm, the multivariate relevance vector machine, in order to predict 16 days of future daily reference evapotranspiration will be introduced. This methodology laid the ground for forecasting the spatial distribution of ET using Landsat satellite imagery. Next, the results of using satellite images and relevance vector machine (RVM) to model ET will be presented. The second part of the talk deals with estimating chlorophyll levels from data collected by AggieAir at 15-cm resolution needed for precision agriculture applications. Simultaneously with the AggieAir ights, intensive ground sampling for plant chlorophyll was conducted at precisely determined locations. The results indicate that an RVM having LAI, NDVI, thermal and red bands as the selected set of inputs, can be used to spatially estimate chlorophyll concentration.