University of Tartu researchers create plant digital twins with ordinary cameras
Researchers at the University of Tartu published a study showing how everyday cameras can build detailed 3D models of plants for botanical gardens.
Study combines two scanning techniques
The team led by Toivo Aukusti Ylinampa and Urmas Kõljalg used photogrammetry and 3D Gaussian splatting to capture plant growth. Their primary tool was a small Olympus TG-6 camera, which they paired with a macro lens device, a phone, and a drone. Researchers converted digital records of specimens that included one orchid bloom and a big tree within the garden.
Two datasets show different results
In the first dataset, about 8,500 photographs of a fern were taken on 12 days across a 15-day period. Every daily model maintains identical scale and orientation so that fern growth can be compared directly across time. In the second dataset, about 20,000 photographs of the tropical greenhouse were processed in 14 overlapping sections for an interactive walkthrough.
Limitations and future steps noted
Although a mild breeze clearly lowered model performance, the greenhouse segments were put together by hand to allow visual inspection instead of exact measurement. A macro test showed a model measurement of 13 mm closely matched a caliper reading of 13.2 mm. According to the authors, future work will focus on automating section stitching and performing separate quantitative validation within their workflow.
Reported by one outlet
Only one outlet has published this. Nothing here has been checked against a second report, so read it as that outlet's account and follow the link below for the original.
Reported by
1 independent outlet. Headline as published. Links open the original report.