Professor Lee Kyung-hwan’s research team at CNU developed technology to identify the characteristics of fruits in orchards using drones and AI.
Professor Lee‘s team at CNU’s Department of Convergence Bio-systems Mechanical Engineering succeeded in developing technology to create 3D images of actual orchards using drones and accurately identify the number, size, and location of fruits in each tree. As a result, it is drawing attention to the potential development of orchard digital twin technology, allowing precise management of orchard environments in the digital space.
The research team used drones equipped with multiple cameras to capture the entire orchard from various angles, connecting the key points of each video to create a 3D image of the orchard. They also included GPS reference points when capturing the orchard to coordinate all points in the 3D image with GPS coordinates. Using deep learning instance segmentation in the 3D orchard image, they identified and measured the size of fruits and then mapped the location of each fruit using GPS coordinates. This information allows them to identify the number and size of fruits on each tree and even map their positions.
The research team filmed the entire orchard with a drone equipped with multiple cameras, then connected the feature points of each image to create a 3D image of the orchard, which allowed them to closely observe the characteristics of the orchard from various angles.
Additionally, when filming an orchard, all points in the 3D image, including GPS reference points, can be converted to GPS coordinates. The research team used a deep learning instance segmentation method on a 3D image of an orchard to recognize fruit, measure its size, and then use GPS coordinates for the location of the fruit. Through the information obtained in this way, not only can the number and size of fruits be identified by the height of each fruit tree, but even their location can be mapped.
Previously, this technology received great attention after being announced at the U.S. Agricultural Robot Symposium and Exhibition (2022 FIRA USA) last year, and received international cooperation proposals from Israel’s advanced robot companies and large agricultural companies.
Currently, the research team has submitted an international collaborative research proposal to the Korea-Israel Industrial Research and Development Foundation and is awaiting project selection results. They are also pursuing collaboration with Silicon Valley companies in the United States to commercialize and export this technology.
This research project was conducted with support from the Advanced Agricultural Machinery Industry Technology Development Program of the National Institute of Agricultural Sciences and the BK21 Stage 4 IT-Bio Convergence System Agriculture Education and Research Group. The research results were pre-published in the online edition of the leading international journal in the field of convergence agriculture, Computers and Electronics in Agriculture (IF: 8.3, top 0.9%), in the September issue.
