Project Number: 8042-22000-167-034-R
Project Type: Reimbursable Cooperative Agreement
Start Date: Jan 1, 2022
End Date: Dec 31, 2022
The objectives of this agreement are to refine precision agriculture technology (computer vision and artificial intelligence tools), to estimate and map weed density and biomass in U.S. soybean production regions, and build a web-based application that automates data analysis and visualization for both farmers and researchers.
The research will be carried out at two scales: in small-plots using a hand-held camera system, and on-farm using tractor-mounted cameras. It is anticipated that work will be expanded to include drones in years two and three. The team working on this project has developed two systems that are both inexpensive and efficient in collecting weed density and biomass. The first system uses a 3-D reconstruction technique (monocular camera and structure from motion analysis of digital videos) using a hand-held pole-mounted GoPro camera. The second system is a tractor-mounted sensor box consisting of multi-spectral, LiDAR, and ultrasonic sensors. Both systems coupled with already develop data flow systems are expected to accelerate efficient data collection throughout the growing season and across locations. A mobile app has also been developed for collecting pertinent meta-data that accompanies the collected video data. These technologies have been beta-tested at select locations in the U.S. This project will greatly expand the coverage across important U.S. soybean production regions and calibration for more weed species. Further, a web-based user interface that automates data analysis, visualization, and recommendation systems will be developed.