Affine registration of multispectral images of historical documents for optimized feature recovery

Abstract

Multispectral (MSI) imaging of historical documents can recover lost features, such as text or drawings. This technique involves capturing multiple images of a document illuminated using different wavelengths of light. The images created must be registered in order to ensure optimal results are produced from any subsequent image processing techniques. However, the images may be misaligned due to the presence of optical elements such as filters, or because they were acquired at different times or because the images were captured from different copies of the documents . There is little prior work or information available about which image registration techniques are most appropriate. Image registration of multispectral images is challenging as the illumination changes for each image and the features visible in images captured at different wavelengths may not appear consistently throughout the image sequence. Here, we compare three image registration techniques: two based on similarity measures and a method based on phase correlation. These methods are characterized by applying them to realistic surrogate images and then assessed on three different sets of real multispectral images. Mutual information is recommended as a measure for affine image registration when working with multispectral images of documentary material as it was proven to be more robust than the other techniques tested.

Understanding multispectral imaging of cultural heritage: Determining best practice in MSI analysis of historical artefacts

“…It is an established method in Information Studies to involve researchers within methodological processes to understand and document their nuances ( [61], p1 and [62][63][64][65]). For our purposes, this included gathering information in the following ways: a review of published literature using MSI within cultural heritage; unstructured interviews and dialogues with practitioners who regularly operate and supply MSI systems for cultural heritage; and reflective practice using the system ourselves (for example, see [52], [66][67][68]), with diaries, case studies, transcripts, and reports being gathered before analysis and synthesis.…”

“…To detect near infrared light with a commercial camera, the infrared filter must be removed from the sensor ( [19][20], [41], [45], [86][87]), however, some cameras are now designed specifically without an infrared filter for MSI [68], [88][89][90]. Monochrome camera sensors ( [35], [57], [68], [91][92]) are more common than colour sensors [93][94], although the colour filters on the sensor used to image the Archimedes palimpsest enchanced the writing by revealing fluorescence [76].…”

“…To detect near infrared light with a commercial camera, the infrared filter must be removed from the sensor ( [19][20], [41], [45], [86][87]), however, some cameras are now designed specifically without an infrared filter for MSI [68], [88][89][90]. Monochrome camera sensors ( [35], [57], [68], [91][92]) are more common than colour sensors [93][94], although the colour filters on the sensor used to image the Archimedes palimpsest enchanced the writing by revealing fluorescence [76]. Silicon Charged Coupled Devices (CCD) or Complementary Metal-Oxide Semiconductors (CMOS) sensors are commonly used for MSI [20], [60], [87], [95][96][97], giving wavelength sensitivity from approximately 350 nm (ultraviolet) to 1100 nm (near infrared) [19][20], [28], [41], [60], [74], [87], [98], although the CHARISMA manual states that the sensitivity is actually wider than this range but is restricted by the choice of lens [19].…”

“…Software used to acquire the multispectral images includes Spectral XV (produced by Equipoise Imaging LLC [66], [68]) and Photoshoot (produced by MegaVision USA [45], [99]). Such software often automatically records relevant metadata, such as wavelengths, filters, and camera settings [125].…”

Image Similarity Metrics Suitable for Infrared Video Stabilization during Active Wildfire Monitoring: A Comparative Analysis

“…Translating gray level values into the more general measure of information content provides enormous flexibility. For this reason, MI is vastly employed not only for image similarity measurement but also to fuse multispectral [39,43,44] and multi-modal [45][46][47] information. However, this flexibility entails an elevated computational cost that may become prohibitive under certain circumstances.…”