Building: Main Building 1st Floor
Room: Salone degli Oceani
Last modified: 2013-10-02
Abstract
Mass-digitisation of biological collections is becoming a well-understood industry. Conveyor belts coupled with scanners can be used to create images of thousands of herbarium sheets per day.
Digitarium is a service centre established in 2010 for outsourcing digitisation. Its mass-digitisation system is available both as service in Joensuu, Finland, and infrastructure that can be delivered to other locations. Speed of digitisation is 200-300 images per hour, employing two human operators. The cost is 1-2€ per sample, depending on how much manual transcription is wanted. Digitarium’s intelligent, automated digitisation ”assembly lines” come in different sizes and multiple camera setups, supporting diverse sample types, such as herbarium sheets, large insects with a separate image of the insect and the labels, and small insects that are imaged from multiple angles with labels intact. All samples receive a URI identifier written in a QR-(Quick Response) code.
The infrastructure has been used to digitise the fern collection of the Finnish Museum of Natural History. The work covered 31 000 sheets in 2½ months. Digitisation is underway for the Herbarium of the University of Oslo, which will process 167 000 sheets in 5 months. A Coleoptera collection of 10 000 beetles, endowed to the Finnish Museum of Natural History, is being processed in the new production line for small insects.
After imaging, further steps of the digitisation process take place from a backlog of images, which has been categorised by organism group and geography. Distance workers in different parts of the world use Digitarium’s web application for transcribing, while customers use it for quality control. Organising a community of expert transcribers is a current challenge. Voluntary work has been shown by ALA and the Notes from Nature project to work well for transcribing museum heritage collections, but it is not an option for the contract-based service industry. For this other avenues may need to be pursued, such as a micro-payment per transcribed image, and transcribing on-demand as required by ongoing research projects.
It is concluded that a network of service centres and well-organised transcriber communities may be able to digitise all major collections in the next 30 years.