Working with spectral data requires software that can read and process data other than the standard digital photo (i.e., three spectral channels – red, green, blue). There are a number of applications that is being used in the field for image processing, but most are expensive and not intuitive (e.g., ENVI). Below are two open source software applications that are being developed for the Cultural Heritage field for image processing of spectral data (both multi-and hyperspectral data).

Hoku Image Processing Software

Hoku is an image processing package for multispectral images. This software is distributed free of charge and is intended for the recovery of erased or damaged writings on ancient manuscripts.

Hoku is a batch processing software package. Jobs are defined and parameters are set interactively with a graphical user interface, but jobs are run in batch mode until they conclude. There are no interactive image processing methods available. You cannot touch-up or draw on an image with this software, but you can apply any of several image processing methods to enhance image details that otherwise are difficult to see.  The software is written in Java and requires having installed 64-bit Java, version 1.8.

NOTE: As of June 23, 2023, a new version of Hoku is available. See here.

Please send any comments, suggestions, descriptions of problems to Keith Knox at knox@cis.rit.edu.

MISHA Initiative: Capture and Processing Software

A government-funded grant from the National Endowment for the Humanities (NEH) led to the development of a low-cost, low barrier-to-entry multispectral imaging system, PR-268783-20 Low-Cost End-to-End Spectral Imaging System for Historical Document Discovery. This imaging system is known as MISHA (Multispectral Imaging for Historical Artifacts).

Directed (from 2020- present) by David Messinger and Juilee Decker at Rochester Institute of Technology, the goal of the MISHA initiative was to produce a system and a companion, free, standalone application that has a low barrier-to-entry and supports the basic functionality needed for multi- and hyperspectral image analysis for Cultural Heritage objects.

Image capture and processing software is written in Python. Capture software is only available for Windows platforms; processing software can be run on either Windows or Mac platforms.

The MISHA software, and the associated parts list for building a low-cost, low barrier-to-entry multispectral imaging system were launched on April 9, 2024 at the National Archives in Washington, D.C. Information about the steps to receive the plans and software for the MISHA system and registration for our MISHA Users Group may be found here.

Or, you may email us: misha@rit.edu