A framework for storage, visualization and analysis of multispectral data

In this paper we describe our database centered work-flow for acquisition, enrichment, long-term storage and webbased analysis of multi- and hyperspectral image data and measurement metadata. We propose a standardized way of storing large amounts of measurement data using a hybrid approach consisting of a relational database and direct file access through a common data access layer. Data import and export is performed by either using proprietary file formats like ENVI or by using an universal XML-based data format. To allow preview and analysis of image data, a web-based application has been developed which supersedes the need for client-side installation of tools like MATLAB to perform spectral or spatial analysis of the data while still allowing third-party applications to retrieve data from the database to perform in-depth analysis like automated filter design. Successful implementation of the workflow is demonstrated by the example of rapid application development for mineral sorting.