About the project

This heritage project focuses on digitization, enrichment and cross collection linking of bound historical herbaria. The herbaria pages were carefully scanned with a color scale and barcode. Scripts were developed that can detect / analyze the page, control strip and barcodes and extract them automatically from the RAW scans. The herbarium sheet was automatically straightened so that it could be offered digitally under optimal conditions for further investigation / consultation. The pages were also optionally de-warped which was found to improve OCR results. Finally, the link with the inventory was also made using the barcodes, so that existing metadata could be linked with the scan.

IDLab role

IDLab has the following tasks within the Flore De Gand project

1. Developed a novel page detection algorithm for preprocessing of the scanned pages.
2. Developed an interactive smart table to crowd source scanning, enrichment, and geolocalization of historical images in today’s maps.
3. Created a Demonstrator on Van Hyck’s The Ghent Alterpiece to demonstrate the power of cross collection linking.