This is a partially reconstructed point-cloud of a peanut field. When completed, UGA scientists will be able to tell the height, width, leaf cover, growth and disease anomolies for individual plants and track it through the season. Currently the research project is working to make the 3-d reconstruction accurate to within 1 mm.
This is a partially reconstructed point-cloud of a peanut field. When completed, UGA scientists will be able to tell the height, width, leaf cover, growth and disease anomolies for individual plants and track it through the season. Currently the research project is working to make the 3-d reconstruction accurate to within 1 mm.

University of Georgia scientist Glen Rains is combining 3-D images and robotics to help farmers identify crop problems before they become an issue that will affect  potential yields.

"We can really detect the smallest changes in the field, and I think farmers will find this very useful," said Rains, an agricultural engineer in the UGA College of Agricultural and Environmental Sciences on the UGA Tifton Campus.

Detection of the beginning stages of a disease means the farmer can be prepared to properly treat and manage the pest. Being able to examine the tiniest particle or fragment on a plant for disease could be a game changer for farmers.

Rains is in his first year of research using 3-D images with assistance from the AGCO Corporation. The company donated a Massey Ferguson 2635 tractor equipped with GPS and automated steering capabilities. Rains uses a tractor-mounted camera to take precise photos, complete with GPS location, in several peanut crop fields.

By producing a library of 3-D images, Rains knows what the plant should look like at different points in the growing season.

"The camera takes thousands of images per field trial, and we will do this two times a week during the summer," Rains said. "The idea is, if we take images often enough, we will be able use software developed by Georgia Tech to create 3-D maps and then zoom in, rotate them and analyze their structure, color and other physical characteristics. We are trying to be accurate down to the millimeter."

After the images are collected and processed, Rains and his team will send an autonomous robot into the field to take leaf and soil samples of the plants with early signs of disease, nutrient imbalances and insect infestations.

In addition to identifying issues of leaf curling and other signs of disease, this technology will show farmers the growing rates of their crops with more accuracy, according to Rains.

"If a farmer can tell what is going on early enough in the season, he can correct it, then he can keep his yield optimized," he said.

Rains believes in the future, 3-D imaging will be an essential tool that farmers and consultants use in managing crops.

If progress is made, the UGA research team will expand the 3-D image production trial to vegetables, specifically bell peppers and watermelons. Rains anticipates these 3-D imaging products will be ready for commercial use within the next year or two.