Ensuring Quality in Large-Scale Herbarium Digitisation: Methods, Strategies, and Lessons from Kew
Year: 2025
Type: article
Abstract: The Herbarium and Fungarium collections at the Royal Botanic Gardens, Kew, house an estimated eight million specimens. While previous digitisation efforts at Kew were conducted internally or in collaboration with institutional partners, the current project, funded by the Department for Environment, Food and Rural Affairs, marks the first large-scale digitisation effort involving an external supplier. The supplier is responsible for producing approximately 5.9 million transcriptions and 7.4 million images, a task that requires balancing speed with accuracy while working with complex and often inconsistent botanical specimens. Achieving high-quality outputs at this scale presents significant challenges, particularly when working with supplier staff who may lack prior experience with herbarium collections. While the supplier provides staff with imaging training and equipment, Kew is responsible for ensuring that specimen handling, data capture, and transcription meet the necessary standards. This presentation focuses on Kew’s quality assurance framework, highlighting the methods and strategies used to maintain accuracy and consistency across millions of records. Key aspects include iterative transcription and imaging protocols, spot checking, feedback cycles, systemic error detection, and collaboration between Kew staff and supplier teams. We will share lessons learned, particularly in balancing efficiency with precision and adapting workflows to accommodate the complexities of herbarium specimens. By outlining our approach to quality assurance in large-scale digitisation, this presentation aims to provide practical insights for institutions planning similar projects, whether working with external suppliers or managing digitisation in-house.
Authors: Korpak, Piotr, Bowles, Benjamin
Institution: Royal Botanic Gardens, Kew
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7084414922
FWCI:
Cites:
Cited by:
Related to: