NDSU Using Sensors to Identify Weed Infestations

North Dakota State University researchers have received a grant to demonstrate the use of electronic sensors to identify weed infestations in cropland remotely.

The project will identify which weeds are infesting specific crops, which parts of fields the weeds are infesting and the degree of the weed problem.

“Crop producers and crop consultants need to identify specific weed infestations growing in various crop fields to implement effective weed management regimes,” says NDSU Extension Service agricultural machine systems specialist John Nowatzki, the lead investigator on this project.

Nowatzki, Agricultural and Biosystems Engineering Department Chair Sreekala Bajwa and Extension weed specialist Rich Zollinger have teamed up with Kris Poulson, Casselton-area agricultural producer and consultant, and agricultural lead for Sentera LLC in North Dakota, for the project.

Sentera, a company based in Shakopee, Minn., that designs sensors, unmanned aerial vehicles and software technologies for a variety of uses, is the private-sector partner on this project. Company personnel will provide hardware and software engineering support for data collection and imagery processing.

The North Dakota Department of Commerce awarded an $85,500 Research ND grant for the weed project.

Project personnel are using an NDSU hand-held radio spectrometer and color, thermal and infrared sensors to measure light reflectance, intensity and color of weeds growing in NDSU greenhouses. They’ll use the same hand-held sensor to collect data throughout the 2015 growing season in outdoor plot and field weed patches at NDSU’s Carrington Research Extension Center and Prosper Research Site.

Researchers also are using commercial sensors designed for manned and unmanned aircraft to collect reflectance values on the same greenhouse, plot and field weeds. The researchers will correlate the data from the commercial sensors and hand-held spectrometer, then publish and share the results with commercial agribusinesses and crop producers for use in their operations.

The weeds they are identifying are kochia, wild buckwheat, green foxtail, wild oats and Canada thistle.

The researchers also are monitoring NDSU research fields and commercial crop fields on the ground to identify the weeds growing in each one. They’ll monitor each field every other week with sensors and cameras to demonstrate the effectiveness of imagery in weed identification. They will use computer analysis software to establish positive identification.

This research will help producers spot weed problems early in the growing season so they can select the proper herbicide and get it applied on the crop.

“Earlier herbicide application results in reduced weed competition with the growing crops and correlates with higher yields,” Nowatzki says.

Project results will be incorporated into NDSU Extension programs focused on helping producers and crop consultants use sensors to enhance their business activities. The information learned from this project also will be shared at a project field day.

Source: John Nowatzki and Ellen Crawford, North Dakota State University