Project Number: 6217-22000-013-09
Project Type: Reimbursable

Start Date: Jul 01, 2008
End Date: Jun 30, 2010

Objective:
Demonstrate that an image analysis approach based on multi-spectral imagery combined with spatial pattern analysis can work to identify Russian wheat aphid-infested wheat fields.

Approach:
We will identify study fields infested with Russian wheat aphids and fields not infested from which to acquire multi-spectral imagery using a Terraverde multi-spectral imaging system. Ground-based sampling will be accomplished to ground-truth fields as infested or non-infested and to characterize the spatial pattern of the infestation in each field. Fields will be imaged from an altitude of 5000 ft above ground level using from a Cessna 172 aircraft. We will evaluate reflectance patterns in multi-spectral imagery using image analysis software. Previous research has shown that a strong, repeatable relationship between Russian wheat aphid density and/or plant injury can be detected. The analytical approach will involve imagery using a filtering operation whereby GIS operations are used to adjust the image for variation in reflectance expected based on topography and soil type (and perhaps other factors) which affect the growth and vigor of wheat plants, and thus affect light reflectance patterns from the plant canopy. Once an acceptable statistical model is developed it will be tested by categorizing numerous wheat fields with respect to overall level of infestation by Russian wheat aphids. Testing will be accomplished by imaging many wheat fields and simultaneously sampling each field for infestation. The value of the model will be assessed by its ability to correctly classify fields as infested versus not infested (or lightly infested).