This work describes a novel hyperspectral imaging system for the long-wave infrared domain. Being a spectral scanner, it is more versatile compared to single-point and push-broom techniques, which require relative movement between instrument and sample to acquire an image. The imager is equipped with a 1024×768-pixel uncooled microbolometer sensor, which is sensitive to wavelengths in the range from 8 to 14 µm. A scanning Fabry-Pérot interferometer is placed in front of the collecting optics. The distance between its two mirrors determines the distribution of wavelengths entering the system. By capturing images at different mirror separation distances, a hyperspectral data cube is formed. The spectral information provided in such a hyperspectral data cube can i.a. be used for material classification, as here demonstrated with an example of different gemstones.
|