Fast, Interactive Analysis of Remote Sensing Data with the HARVIST System
Abstract
Remote sensing instruments in Earth orbit provide a rich source of information about current agricultural conditions. Observed over time, patterns emerge that can assist in the prediction of future conditions, such as the yield expected for a given crop at the end of the growing season. These predictions can be made more accurate by incorporting other sources of information, such as weather conditions, soil properties, etc. However, the tools required to access and combine large amounts of data from multiple sources, at different spatial resolutions, are not readily available to the science community. The HARVIST (Hetereogeneous Agricultural Research Via Interactive, Scalable Technology) project has addressed this lack by demonstrating highly efficient, scalable methods that combine data from orbital remote sensing instruments and ground-based weather stations. The three major components of the system are: 1) data clustering, to identify major trends; 2) data classification, e.g., to identify different crop types or ground cover; and 3) regression, to predict continuous values such as crop yield. We have developed a graphical interface that allows users to interactively run clustering and classification algorithms on multispectral remote sensing data. Our recent "progressive rendering" advance for support vector machine classifiers quickly provides the user with an initial estimate of the result and then iteratively refines it in real-time. Finally, we have obtained state-of-the-art crop yield prediction results using several different regression methods.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2006
- Bibcode:
- 2006AGUFMIN21A1204K
- Keywords:
-
- 0402 Agricultural systems;
- 0480 Remote sensing;
- 1631 Land/atmosphere interactions (1218;
- 1843;
- 3322);
- 1694 Instruments and techniques