News
New publication with Prof. Tobias Hank
17 Dec 2024
"Plant trait retrieval from hyperspectral data: Collective efforts in scientific data curation outperform simulated data derived from the PROSAIL model."
Our team member, Prof. Tobias Hank, contributed to a new study led by Universität Leipzig, published in the ISPRS - International Society for Photogrammetry and Remote Sensing. This collaborative research explores the effectiveness of hyperspectral data and machine learning for retrieving plant traits, comparing combined real-world and simulated data.
The study highlights that combining real-world data is far more effective than relying on simulated data. It underscores the importance of collective efforts within the scientific community and open data-sharing to build robust, transferable models for ecosystem monitoring.
While simulated data can help fill gaps for traits that are poorly represented in existing datasets, for most traits, adding simulated data has little to no benefit—and in some cases, it can even reduce accuracy.