Poster Open Access
C Willard; M Strlic; J R Gilchrist; R van Langh; T Tuckett; A Gibson
Many hyperspectral imaging systems have a limited spatial resolution, making it challenging to capture the details of large heritage surfaces. To create a detailed high-resolution image, multiple images can be stitched together using image mosaicking techniques. This work develops an automated mosaicking technique for hyperspectral images taken using a scan mirror assembly, aiming to ensure accurate alignment and stitching of hyperspectral images. Seam line cutting is investigated as an approach to preserve the accuracy of spectral reflectance data across image boundaries whilst minimising visible seams. A case study has been conducted on a genealogical roll held at UCL Special Collections, the roll is 6 metres long and constructed from joined parchment sections. The roll traces the lineage of England’s Plantagenent rulers back to Adam and Eve (https://ucldigitalpress.co.uk/Book/Article/2/9/18/). It is suspected that the roll was completed in different sections, spectral analysis of the inks can help to determine whether there are distinct segments within the roll. This work will follow on to a research project using a new hyperspectral scanning platform capable of imaging large heritage surfaces up to 2 m x 2 m in high resolution.
Images courtesy of Special Collections, UCL Library Services.