Skip to main content
ARS Home » Northeast Area » Beltsville, Maryland (BARC) » Beltsville Agricultural Research Center » Hydrology and Remote Sensing Laboratory » Research » Publications at this Location » Publication #323265
Title: Rapid detection of Colorado potato beetle damage using small unmanned aircraft

Author
item Hunt Jr, Earle
item RONDON, S.I. - Oregon State University
item HAMM, P. - Oregon State University
item TURNER, R. - Collaborator
item BRUCE, A. - Collaborator

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 7/28/2015
Publication Date: 8/18/2015
Citation: Hunt Jr, E.R., Rondon, S., Hamm, P., Turner, R., Bruce, A. 2015. Rapid detection of Colorado potato beetle damage using small unmanned aircraft. Meeting Abstract. https://scisoc.confex.com/scisoc/2015am/webprogram/Paper95673.html.

Interpretive Summary:

Technical Abstract: Remote sensing with small unmanned aircraft systems (sUAS) has potential applications in agriculture because low flight altitudes allow image acquisition at very high spatial resolution. Damage to potato fields by the Colorado potato beetle (Leptinotarsa decemlineata) rapidly increases from initial infestations; early detection allows more options for integrated pest management and biological control. We conducted an experiment on Ranger Russet potatoes at the Oregon State University Hermiston Agricultural Research and Extension Center (HAREC) with 4 levels of infestation (0, 1, 3, and 5 beetles per plant) and 4 replications arranged in a randomized block design. Daily, we flew a quadcopter at three altitudes with a Tetracam, Inc. (Chatham, CA) Multiple Camera Array with 5 bands (blue, green, red, red-edge, and near-infrared) and one up-looking calibration channel. Over three days, damage in some plots increased from 0 to 29%, with almost all of the damage becoming visible on the third day. Damage was correlated with the total number of beetles (sum of released and immigrants) beetles). Plot-scale spectral vegetation indices were not well correlated with initial levels of damage, whereas plot heterogeneity of NDVI was best correlated. Simple algorithms of heterogeneity may be automated and do not require image mosaicing. False positive detections will be frequent but manageable.