Herbarium collections, which contain more than 395 million specimens around the world, are critical libraries of plant and fungal life. These collections are the foundation for our understanding of Earth’s biodiversity, evolution, and ecology. Recent developments in artificial intelligence (AI) are opening new doors that allow researchers to interact with specimens in unprecedented ways. This type of innovation is opening up a treasure trove of previously untapped data.
Down Under, the National Herbarium of New South Wales (NSW) has gone even further. In the past three years, they’ve been able to digitize over 1.15 million specimens. This large collaborative effort is a huge step towards providing herbarium data to the scientific community for use in a wide range of research. The digitization process allows scientists to rapidly study plant evolution and ecological changes, which is increasingly important in the face of climate change and biodiversity loss.
AI has been central to this transformation. AI technology helps improve research efficiency by automating the extraction of detailed leaf measurements and other traits from digitized herbarium specimens. This innovation enables botanists and ecologists to gather information quickly and accurately, making it easier to analyze trends in plant life over time.
The Scope of Herbarium Collections
Herbarium collections around the world thus provide an unparalleled, readily available archive of our planet’s plant and fungal diversity. Researchers investigating plant evolution, species interactions, and ecological dynamics depend on these collections. With a staggering 395 million specimens cataloged around the world, they represent one of the greatest treasures. As such, each specimen represents a snapshot of biodiversity at the time and place it was collected. This makes it possible for scientists to monitor shifts within and across generations.
In Australia, the University of Melbourne and other institutions keep large herbarium collections. The University of Melbourne’s herbarium holds about 170,000 specimens. Just like our museum’s nationwide collection, this collection is vital to documenting and understanding biodiversity across the nation. Especially as these collections continue to grow, the need for effective, scalable digitization and data extraction is paramount.
And yet, despite their critical importance, many herbarium collections could stay just partially digitized for decades at the current rate. We know that digitization is not just time-consuming—it’s resource-intensive. Consequently, much of this data is still trapped on paper forms, creating an inaccessible barrier for researchers who could benefit from it. This constraint has generated significant interest in using AI to speed up the digitization process and make data more accessible.
Advancements in AI and Digitization
Recent advances in AI technology are starting to meet the unique challenges of herbarium digitization. One such innovation is Hespi, an open-source software to automatically digitize information contained in herbarium specimens. Using natural language processing, Hespi can identify the main specimen label on a herbarium sheet in an instant. Then, in a few seconds, it scans all the information that’s in there.
The accuracy achieved by Hespi is impressive. Researchers have seen incredible accuracy in pulling information out of herbarium specimens. They’re able to provide precise species identification, specific location coordinates, and detailed collection dates. This unprecedented level of efficiency is currently saving time and increasing the reliability of data generated from herbarium collections.
AI integration extends beyond technological innovation. It catalyzes a true paradigm shift, enabling scientists to tap into botanical data in ways that were previously unimaginable. By feeding digitized specimens into large biodiversity portals such as the Australasian Virtual Herbarium, the Atlas of Living Australia, and the Global Biodiversity Information Facility, researchers can share valuable information across disciplines and geographical boundaries.
The Future of Biodiversity Research
As you can see, the future of biodiversity research is deeply connected with AI technology and digitization efforts. By unlocking the treasure trove of data held in herbarium specimens, researchers can conduct rapid investigations into plant evolution and ecological shifts. This work is important foundational work for developing plans to combat the current crisis of biodiversity loss. It’s improving our understanding of the impacts of environmental change on ecosystems.
As AI advances, we can expect to see its use on herbarium research go even further. The promise for real-time data extraction and analysis to accelerate scientific discovery and improve collaboration between researchers across the globe is huge. It is still important to make sure that digitization stays at the top of the funnel to get the most out of these new technological advances.