Loading…
Thursday, April 19 • 11:00am - 11:30am
Use of High resolution UAV imagery and ground-based data in a Montana winter wheat (Triticum aestivum) field to support landcover classification and site-specific weed management

Sign up or log in to save this to your schedule and see who's attending!

Precision agriculture technologies enable site-specific management of crops by allowing a farmer to vary seed, fertilizer, herbicide, and pesticide application rates within a field. One application of precision agriculture technology is site-specific management of weed populations. In this context, weed control is based on economic thresholds determined by weed density. Weed patches with densities that are perceived by the farmer to be at or above an economic threshold are controlled, while lower density patches and areas with no weeds are left untreated. In an attempt to map the distribution and extent of an invasive cheat grass (Bromus tectorum) population in a winter wheat (Triticum aestivum) field near Sun River, MT we captured centimeter-scale, red, green, and blue digital aerial imagery during the green-up growth phase in the spring of 2017. At the same time, we also collected ground-based data describing winter wheat, cheat grass, and bare soil proportions with the extent of field plots. We randomly selected a subset of our plots and related their location and ground characteristics to spectral components of the collected imagery. With this relationship we developed a classification of wheat crop, cheat grass, and bare ground using the Random Forest algorithm. Based on a comparison to the subset of plots not used in the classification, we found that we were able to identify and map cheat grass patches within the wheat field with an overall map classification accuracy of 88%. Based on these results, we suggest that the combination of ground and remotely sensed imagery collected by UAV devices has good potential for effectively characterizing crop versus non-crop vegetation within agricultural settings. Knowledge of favorable and unfavorable plant distributions can further be used to establish efficient vegetation management strategies that have the potential to reduce costs, environmental contamination, and save time.

Speakers
PD

Philip Davis

Philip is an agricultural research specialist. He works with scientists at Montana State University to assess, and prescribe effective agricultural practices, specifically related to wheat production.


Thursday April 19, 2018 11:00am - 11:30am
Governor Room