Research news on multispectral imaging

Multispectral imaging is an optical acquisition method that records scene radiance or reflectance in multiple discrete spectral bands, typically spanning the visible, near-infrared, and sometimes shortwave infrared regions. Using bandpass filters, tunable filters, or sensor arrays with spectrally selective elements, it generates co-registered images where each band captures wavelength-dependent material properties. Quantitative analysis of these bands enables spectral feature extraction, classification, and segmentation based on differences in absorption, scattering, or fluorescence. In research, multispectral imaging is applied to non-destructive assessment, tissue characterization, remote sensing, and cultural heritage analysis, often serving as a lower-complexity, higher-throughput alternative to full hyperspectral imaging.

Observing oscillations, flares and tornados on the sun

For six and a half days in July 2024, the balloon-borne solar observatory Sunrise III kept its gaze fixed on the sun. The stratospheric flight, which stretched from the northernmost tip of Sweden to Canada's Northwest Territories, ...

Food industries embrace AI sensors to improve efficiencies

Food waste is a nagging problem that weighs heavily on global food production, distribution and sales industries—but an emerging generation of AI sensors is providing a raft of fresh solutions. The embrace of AI in food industries ...

page 1 from 8