Evaluation of Unmanned Aerial Systems Technology for Pine Beetle Detection

This Phase II effort supports an AFIMSC/AFWERX pilot study to test the feasibility of utilizing small-scale unmanned aircraft system (UAS) platforms outfitted with image capturing sensors to perform early detection of pine beetle infestations within a pine forest. The pilot study will test the hypothesis that remote controlled UAS can be a more efficient and more cost-effective means to detect and control an incipient pine beetle attacks before infestations expand to epidemic proportions. Machine learning algorithms will be used to: fuse visible(EO), long-wave infrared (LWIR), hyperspectral, and 3D laser-range finding (LIDAR) imaging data; discriminate between tree species and other ground cover; predict "health indices" that measure tree stress from aerial views of forests; deliver these results visually within a GIS software application to support planning by USAF and US Fish & Wildlife personnel. Imaging data will be obtained from ponderosa pine-dominated forests outside Austin, Texas and at the United States Air Force Academy (USAFA) in Colorado.