by Enrico Villamaino
The University of Illinois and University of Oregon teamed up with an aerial imagery analytics company to host a workshop to discuss the challenges and opportunities for computer vision (CV) in agriculture.
The Agriculture-Vision workshop is a new addition to the annual Computer Vision Pattern Recognition (CVPR) conference. CVPR is the premier annual event for professionals working in the field of CV. Agriculture-Vision was presented online on June 14. A major portion of the workshop was dedicated to the progress of Intelinair, the aerial imagery analytics firm co-sponsoring the workshop.
The workshop was an opportunity to increase the exposure of the work they’ve been doing at Intelinair, said Dr. Naira Hovakimyan, the University of Illinois’ W. Grafton and Lillian B. Wilkins Professor of Mechanical Science and Engineering. Hovakimyan founded Intelinair in 2014.
Intelinair offers its AgMRI service to farmers in Illinois, Indiana and Iowa. Farmers subscribe to the AgMRI service and receive its timely alerts through a downloadable app.
“People go to the doctor once a year for an annual checkup to make sure they’re healthy and to catch any potential issues early,” Hovakimyan explained. “The AgMRI is the same idea. It’s a regularly scheduled checkup, every two weeks, for your fields. We deliver what we call ‘crop intelligence.’”
Over a six-month period (typically April – October), AgMRI subscribers receive 13 flyovers by airplanes outfitted with five different types of sensors. Using its patented technology, Intelinair collects and analyzes data from these various sources. Aerial images, temperature readings, humidity measurements, rainfall, soil samples, terrain type, equipment utilized, planting rates and applications are among the variables scrutinized. Hyperspectral analysis, CV and deep learning (DL) are used to identify patterns and anomalies of every monitored field for the growing season. AgMRI tracks a farm’s progress from week to week. The newer information is continuously aggregated and correlated with what has already been recorded, relating and connecting past and present situations. Consequently, AgMRI is able to identify abnormal crop conditions long before the human eye can.
“We are currently able to notify farmers of over 40 different anomalies affecting their fields,” Hovakimyan said. Such anomalies can include weeds, sectional needs to replant and the first signs of emerging diseases. Hovakimyan noted, “Our clients receive any updates to their farm assessments within 24 to 48 hours of their most recent flyover.”
Intelinair’s quick feedback, with up to date insights into the performance of their crops, have granted growers improved decision-making and maximized crop yields.
Agriculture-Vision is expected to have over 700 participants, including representatives from Google, Microsoft, IBM, Stanford University and the Massachusetts Institute of Technology. Hovakimyan hopes that the participation of this diverse group of technological leaders and academics will spur on further advances in the field of crop intelligence.
“Next-gen farming has to be data-driven,” she stated.