Large–scale digitization of herbarium specimens: Development and usage of an automated, high–throughput conveyor system

• DOI: 10.12705/671.10
• Publication Date: 2018-03-06T04:36:22Z
• Authors: Patrick W. Sweeney; Binil Starly; Paul J. Morris; Yiming Xu; Aimee Jones; Sridhar Radhakris...; Christopher J. Gr...; Charles C. Davis
• Keywords: Digitization; Herbarium

Abstract The billions of specimens housed in natural science collections provide a tremendous source under–utilized data that are useful for scientific research, conservation, commerce, and education. Digitization mobilization specimen images promises to greatly accelerate their utilization. While digitization collection has been occurring decades, the vast majority remain un–digitized. If task is be completed near future, innovative, high–throughput approaches needed. To create dataset study global change New England, we designed implemented an industrial–scale, conveyor–based workflow herbarium sheets. variation object–to–image–to–data prioritizes imaging capture storage container–level data. utilizes novel conveyor system developed specifically flattened specimens. Using our workflow, imaged transcribed specimen–level almost 350,000 over 131–week period; additional 56 weeks was required capture. Our project demonstrated it possible both image core database record 35 seconds per sheet (for intervals between 30 minutes or less) plus some overhead This rate line with pre–project expectations approach. throughput rates comparable other similar, focused on digitizing sheets as much three times faster than achieved more conventional non–automated used during project. We report challenges encountered development use discuss ways which could improved. apparatus software, schema, configuration files, hardware list, schematics available download GitHub.