Skip to main content
ARS Home » Research » Publications at this Location » Publication #309939
Title: Processed multispectral imagery differentiates wheat crop stress caused by greenbug from other causes

Author
item BACKOULOU, GEORGES - Oklahoma State University
item Elliott, Norman - Norm
item GILES, KRISTOPHER - Oklahoma State University
item MIRIK, MUSTAPHA - Texas Agrilife Research

Submitted to: Computers and Electronics in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/11/2015
Publication Date: 6/1/2015
Publication URL: http://handle.nal.usda.gov/10113/61768
Citation: Backoulou, G.F., Elliott, N.C., Giles, K.L., Mirik, M. 2015. Processed multispectral imagery differentiates wheat crop stress caused by greenbug from other causes. Computers and Electronics in Agriculture. 115:34-39.

Interpretive Summary: The greenbug is an important pest of wheat and other small grains. Economic losses in wheat due to greenbug in Oklahoma, Texas, and Kansas average more than $100 million annually. The objective of this study was to determine the potential for multispectral imagery acquired in specific visible and near infrared wavelength bands of light combined with mathematical modeling can be used to detect greenbug infestations in wheat fields. Multispectral images of wheat fields were acquired using a Duncantech MS3100-CIR multispectral camera mounted in an aircraft. Stress observed to wheat plants in wheat fields was grouped into categories: greenbug, drought and agricultural conditions. Spatial pattern metrics quantify attributes of the shape, area, and geographic distribution of spatial objects identified in imagery. A set of 10 spatial pattern metrics were computed for each stress factor. The combination of multispectral data and spatial pattern metrics subjected to statistical modeling produced a model that accurately differentiated patches of injured wheat infested by greenbug from patches caused by drought and agronomic conditions. The detection and differentiation of stressed patches will help in mapping stress within fields for the purpose of site-specific pest management and for monitoring systems to identify greenbug infestations at individual field and regional scales.

Technical Abstract: The greenbug, Schizaphis graminum (Rondani) (Hemiptera:Aphididae) is an important pest of small grains such as winter wheat (Triticum aestivum). Economic losses in wheat due to greenbug in Oklahoma, Texas, and Kansas were estimated at $100 million annually in the 1990's. The objective of this study was to determine the potential for multispectral imagery analyzed using spatial pattern metrics subjected to discriminant function analysis to differentiate patches of wheat plants within wheat fields infested by greenbug from stressed patches caused by other factors. Multispectral images of wheat fields were acquired using a Duncantech MS3100-CIR multispectral camera. Stress observed to wheat plants in wheat fields was grouped into categories: greenbug, drought and agricultural conditions. ERDAS Imagine software was used to process and analyze images, and FRAGSTATS was used to quantify spatial pattern. A set of 10 spatial pattern metrics were computed at the patch level for each stress factor. The analysis of spatial pattern metrics by discriminant function analysis revealed that the three types of stress could be reliably differentiated. The combination of multispectral data and spatial pattern metrics made it possible to differentiate patches in wheat fields infested by greenbug from patches caused by drought and agronomic conditions. The detection and differentiation of stressed patches may help in mapping stress within fields for the purpose of site-specific pest management and for monitoring systems to identify greenbug infestations at individual field and regional scales.