Spectral imaging in real-time—Imaging principles and applications

摘要

The journal of Real-Time Imaging has reported for a decade recent developments in real-time image processing in a very large range of applications. This demonstrates the maturity of the technology and its spread into many industrial fields. "Classical Image Processing" grew mainly out of machine vision applications by exploiting the luminance information in industrial scenes, investigating either single shots of grey value images as still images, extracting additional stereo information from multiple camera positions, or grabbing continuously changing spatial luminance in image sequences. Colour was the supplementary dimension for simple discrimination via selective spectral features. Nowadays, many methods originally developed for grey value image processing are applied to multi-colour data arrays, although special considerations need to be taken into account for the sensors capturing data (e.g. single or multi-chip CCD or CMOS cameras), the illumination (by different colour temperature and spectral characteristics), and the colour system used (R G B, I H S, etc.) in which the colour spectrum is decomposed.