Virtual Field Reference Database (VFRDB) description

For the purpose of this study, a VFRDB was constructed consisting of 3024 georeferenced digital photographs taken throughout the state from May 2002 to February 2004. Photographs were taken with a Canon S200 digital camera and georeferenced to GPS data from a Garmin etrex GPS unit. The time stamp from the track log on the GPS unit was downloaded and compared to the time stamp embedded in EXIF header information.

The result is a rich web-based data repository that allows for classification of fire fuel models based on the in situ photographic images as well as the Landsat ETM+ image and supporting orthophoto and RIGIS data. These data are spatially linked to each georeferenced digital image via a custom web based SQL interface which allows users of the data repository to view the in situ digital image along with all supporting remote sensing and RIGIS data for that image.

Additionally, each image was assigned a fuel model category and assigned rankings of Low, Medium, High fuel loadings per 1, 10 and 100-hour time lag fuel classes as well as herbaceous and 1-hour live fuel classes. Of the 3024 images taken, 2729 were useable for assignment of fuel model, with the remainder suffering from blur, shake or otherwise flawed.

Schematic of VFRDB point. A VFRDB can provide a robust source of reference data to support quantitative map evaluation.

VFRDB points provide in-situ measurement data that can be applied to many classification systems and can serve as a baseline data library to support long-term development and evaluation.