Drones and AI have learned to determine the maturity of the crop with high accuracy

To breed new soybean varieties during critical periods of the growing season, agronomists manually search for plants with the necessary traits. For example, with the early maturity of the pods. New research is allowing this process to be automated using drones and artificial intelligence.

Scientists at the University of Illinois predict soybeans’ maturation in two days using drone imagery and artificial intelligence. This reduces the need for a manual plant selection method that is slower and more exhausting for breeders and prone to error.

“Pod maturity assessment is very time consuming and error-prone. It’s a pod color-based grading system, so it’s human-driven, explains Nicholas Martin, assistant professor of crop production in Illinois and co-author of the study. “Many research groups are trying to use drone imagery to assess maturity, but not scale up and automate the process. We came up with a more accurate way to do it.”

Rodrigo Trevisan, a doctoral student, working with Martin, trained computers to detect changes in pods’ color from drone images. They were collected in five trials, three growing seasons in two countries.

The expert used deep convolutional neural networks (CNN) – a type of artificial intelligence. CNN’s work is similar to how the human brain learns to interpret images’ components – color, shape, texture – received from human eyes.

The new method of analysis will allow breeders to decide the fate of crops on a larger scale than ever before, the scientists conclude.

Google News button