This is a draft schedule. Presentation dates, times and locations may be subject to change.

661
Comparison of High Resolution Aerial Photography to Manual Field Collection in Assessing the Control of Red Cedar Using Goats

Monday, July 10, 2017
Exhibit Hall (Baltimore Convention Center)
Raquel V Lourencon, American Institute for Goat Research, Langston University, Langston, OK
Steven P. Hart, American Institute for Goat Research, Langston University, Langston, OK
Terry A. Gipson, American Institute for Goat Research, Langston University, Langston, OK
Marcio White, Department of Agriculture and Natural Resources, Langston University, Langston, OK
In a research study using goats to control red cedar (Juniperus virgiana), it was necessary to catalog trees in the study area by GPS coordinates and basal diameter in order to monitor their control by goats. The purpose of this study was to compare the accuracy of high resolution aerial photography to manual field GPS collection. A X8-M multicopter drone with a 12-megapixel Canon® camera was used to take high-resolution aerial pictures (2 cm/pixel). The images were taken in the winter when red cedar is usually the only green plant in the field. An onboard GPS unit facilitated assembly of individual pictures into an orthomosaic using Pix4Dmapper® software. The orthomosaic and the manual field GPS data were brought into ArcMap® as layers and tree diameter was calculated as shapefiles on the orthomosaic layer of the selected tree GPS data points. For field measurements, a Trimble® GEO 7X was used to determine tree GPS coordinates and the radius was measured and doubled for diameter. The difference (DIFF) between the diameter measured by the GPS field data and that diameter calculated based on the orthomosaic, the distance (DIST) between the GPS coordinate in the field and the centroid of the shapefile, and the compass point position (BEAR) of the shapefile centroid relative to the GPS point were calculated. Regression analysis was used to determine linear relationships between DIFF, DIST, and BEAR using R (R Core Team, 2013). There was no significant relationship between DIFF and DIST or between DIFF and BEAR (P > 0.10). Therefore, no discernible bias exists between the two methods. The diameter measured by aerial photography showed a positive correlation (R2 = 0.59; P < 0.05) with the diameter calculated from the manual field GPS collection. However, since some dispersion of the shapefiles on the orthomosaic was observed, aerial photography, at this time, cannot replace the manual field collection of the data.