Related Experiment Video
Updated: Jun 12, 2026

Computer Vision-Based Biomass Estimation for Invasive Plants
Published on: February 9, 2024
A Manually Annotated and Curated Mediterranean Plant Image Dataset of Native Species from Lebanon
Yasmine Taki1, Dany Abou Jaoude1,2, Ibrahim Issa1,3
1Saving Lebanon's Biodiversity Center Incubator, Maroun Semaan Faculty of Engineering and Architecture, American University of Beirut, Beirut, Lebanon.
Abstract:
The FloraLebanon (FL) is a curated field dataset of 24,944 plant images from 102 Mediterranean species native to Lebanon. University student volunteers took images during field expeditions to the Al Shouf Biosphere Reserve in 2024. The volunteers followed a structured image-capture protocol developed for the project to document whole plants, leaves, stems, flowers, fruits, and habitats. Image pre-processing involved the manual removal of redundant and poor-quality images, followed by a second cleaning technical validation. Species identification was validated by examining voucher specimens taken from the field and confirmed by two independent taxonomic experts. The FloraLebanon (FL) dataset includes plant images, a metadata sheet of plant nomenclature, and scans of voucher specimens. The FloraLebanon (FL) dataset is open access, collected and curated to support machine learning and biodiversity monitoring. Our goal is to expand the dataset to cover Lebanon's flora, estimated at 2,600 species.
Related Concept Videos
Light Acquisition
Non-vascular Seedless Plants

