Fuel Estimation

Just left the beautiful and productive Golden Oak Ranch. Sure was a disappointing sight seeing the burn scars, smoldering and devastated habitat. In OCT of 2016 we identified the need to do periodic wildfire fuels estimation at this ranch, and in MAR-APR of 2017 a baseline was collected. Then the fire at the end of JUN 2017 took over 400 acres of the area including almost 200 acres of the area we baselined as part of our LASR-Air Internal Research & Development Program to validate the use of drones for fuels estimation. Originally funded by a grant from a company and the fire department in Anchorage, AK, the program has matured to the point of validation in various fuels associated with the Wildland-Urban Interface (WUI).

We used off the shelf processing "software as a service" providers to get the baseline images, and then ran the test algorithms against the pixel values obtained from each of the vendors images. Results were mixed to not useful, but we are improving on those with promise. Below is the best result. It shows the color coded areas with about 60 percent accuracy for all categories, but is 88 percent accurate for low moisture, bulk plant material (the drone camera did the collection, and no ground truthing was required to get this result). Bright Red is soil; Orange is low moisture fuel; Yellow is high moisture fuel; Green is plant material in photosynthesis active stage.