Automating Plant ID with Laser Scanning and AI

photo of Saket Navlakha is Automating Plant ID with 3D Laser Scanning
Saket Navlakha is Automating Plant ID with 3D Laser Scanning

Automated camera systems have enabled scientists to quickly track traits of thousands of plants. Now researchers at the Salk Institute are automating plant identification with 3D laser scanning and machine-learning algorithms. The algorithms teach a computer to analyze three-dimensional shapes of plant branches and leaves. They also may help scientists better quantify how plants respond to climate change, genetic mutations or other factors. They are

Salk researchers are using 3D laser scanning to capture the structure of plant architectures, thereby automating the process. They shine a laser at each plant to “paint” its surface with a beam. Resulting data – called a 3D point cloud – show fine details of a plant’s surface. But quantitatively analyzing point clouds can be challenging because the technology is so new and the datasets so large.

“The resolution and accuracy of this data is much better,” said Saket Navlakha, an associate professor in Salk’s Integrative Biology Laboratory. “But methods that have been developed for analyzing leaves and branches in 2D images don’t work as well for 3D point clouds.”

Navlakha and Illia Ziamtsov, a graduate student at the University of California-San Diego, used a 3D laser scanner to scan 54 tomato and tobacco plants grown in a variety of conditions. They inputted the resulting 3D point clouds into machine-learning algorithms. That enabled them to teach the program how to phenotype plants. The technique involved the researchers manually showing the location of leaves and shoots. The software then began to automatically recognize the features.

The researchers focused on teaching the program to make three phenotype measurements – separating stems from leaves, counting leaves and their size, and outlining a plant’s branching patterns. The program had a 97.8 percent accuracy at identifying stems and leaves.

The Salk researchers plan to fine-tune the approach – differentiating close-together leaves can still be challenging, for instance. And the current software version may not work on all types of plants. The researchers hope to generalize the software to work on plants from vines to trees, and also to analyze roots.

“There are a lot of challenges in agriculture to increase crop production and better sequester carbon,” Navlakha said. “We hope our tool can help biologists address some of these challenges.”

The Salk researchers will release their software as open-source for other researchers to use. The study recently was published in “Plant Physiology.”

For more information visit plantphysiol.org and search for “Saket Navlakha” for more information.

 

Note – If you liked this post click here to stay informed of all of the 3D laser scanning, geomatics, UAS, autonomous vehicle, Lidar News and more. If you have an informative 3D video that you would like us to promote, please forward to editor@lidarnews.com and if you would like to join the Younger Geospatial Professional movement click here.

 

This entry was posted in 3D Modeling, Agriculture, AI, artificial intelligence, Data, Environmental, Forestry, Laser Scanning, Lidar, machine learning, point clouds, Research, Young Geospatial Professional and tagged , , , . Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *

*

This site uses Akismet to reduce spam. Learn how your comment data is processed.