The Shiu-Ying Hu Herbarium of the School of Life Sciences at The Chinese University of Hong Kong (CUHK) has established an innovative database, the Virtual Carpological Herbarium of Fruits and Seeds. By combining plant taxonomy and photogrammetry, researchers reconstructed more than 300 plant specimens into 3D models that reflect their genuine structures, sizes and colours.
The database will provide an insightful reference for native plant identification, ecological survey, and conservation of rare and endangered species. The research methodology and results were published in the scientific journal PLOS ONE.
Dr David Lau Tai-wai, Curator of the Shiu-Ying Hu Herbarium, alongside his research team spent two years building 3D models of more than 300 plant species, including about 200 native species. Each 3D model was reconstructed according to an authenticated voucher plant specimen using the “structure-from-motion” technique. This required taking 90 to 100 photographs per specimen and conducting rigorous comparisons with identified species to ensure accuracy.
The database is now available online for researchers and the public to view the 3D plant models on computers and mobile devices without additional software. The 3D models can be rotated through 360° and magnified to observe the structures and fine details. The high resolution allows users to inspect macroscopic features smaller than 1 mm such as close-up textures of leaves and seeds, much like seeing the real specimens.
Dr Lau stated that 3D models are extremely valuable for scientific research and citation. The authenticated voucher plant specimens will be permanently stored in the archives of the Shiu-Ying Hu Herbarium. In the future, more native plant specimens will be included in the database.
This project demonstrated the feasibility of applying advanced technologies in botanical research, not only facilitating the identification of plant varieties and endangered species but also contributing to STEAM education. The team has published the methodology for creating these 3D models and hopes to set an example for other botanists and herbaria to establish their own 3D databases.
Other team members from the Shiu-Ying Hu Herbarium include the first author of the publication and researcher Mr Wang Ho-lam, education manager Mr Wong Tin-hang, and researchers Mr Chan Yiu-man and Mr Cheng Yat-sum.
The paper states that mass digitisation is a trend that has grown in the past years to document natural history collections and herbarium archives which facilitated the knowledge transfer of biodiversity information. A major digitisation method is 2D scanning. This method, used in many herbaria and museums, involves specimens being captured via high-quality images. However, 2D images are usually plain and fixed – a specimen cannot, thus, be viewed from all angles. In addition, due to overlapping, some structures cannot be observed from the images.
Technological advancement in the development of 3D digitisation of natural history collections has garnered much attention. The technique assists in the transformation of natural history collection into 3D models which provides more information than those from the 2D counterpart. Laser scanning is one of the common methods to create 3D models of specimens with detailed external structures.
Another method is via computed tomography (CT) scanning where an X-ray is used to penetrate specimens and document the internal structures in a pile of images, which can be stacked into 3D models. However, the genuine colour of the specimens cannot be reflected on the 3D models reconstructed by laser or CT scanning, so this important phenotypic character is missing. Hence, the team’s work aims to remedy these drawbacks in one solution.
Other universities and institutes are keen to shore up the nation’s sustainability ambitions and quality of life. Recently, HKUST’s Institute for the Environment and the HKSAR Environmental Protection Department said they would collaborate on a large scale to study to develop science-based regional ozone and photochemical smog control strategies.