Hyperspectral imaging is a hybrid technique, combining spectroscopy and 2D imaging, and most recently 3D imaging. A hyperspectral image is generated by collecting and processing images of the sample within narrow spectral bands (typically between 10 and 100) covering a continuous spectral range.
This method produces spectral data for each individual pixel in the field of view. This provides valuable information about the emission, reflection, and absorption of electromagnetic radiation, which can be used to determine the chemical composition at different areas of the sample. No prior knowledge of the sample is required, which makes it a highly valuable technique for many different users.
Hyperspectral imaging has been used in the past for applications such as air reconnaissance, satellite imaging and other markets that are not overly price sensitive. However, the emergence of alternative methods provides hyperspectral imaging with potential in volume markets and even consumer applications and products.
Precision farming, cancer detection, and food testing in supermarkets are just some examples of where the new approaches could be employed.
Standard Hyperspectral Imaging Cameras
Standard hyperspectral imaging cameras use gratings or prisms as dispersive elements. A large distance is required between the sensor and diffractive element when gratings and prisms are employed, which makes the instrument large, taking up a lot of laboratory space or makes the instrument too heavy for the increasingly popular drones or UAVs.
These cameras are often expensive and are sensitive to misalignment due to mechanical influences, and they also require a slit to obtain high spectral resolution, which significantly restricts the light efficiency.
Bandpass Filters for Hyperspectral Imaging
Bandpass filters could be used to overcome the disadvantages of standard hyperspectral imaging cameras, these filters can be placed directly on top of the sensor, by either replacing the cover glass with the filter or a mechanical holder or the filter can be glued on the surface of the sensor. This considerably reduces the size and weight of the detector, bringing high resolution hyperspectral imaging to UAVs and small cube satellites. A slit is not needed by this optical design, as light is gathered through the entire aperture of the lens, and this, in addition to the optical bandwidth of the filter and the high transmission, results in hyperspectral imaging cameras that are extremely light efficient.
The full scene is displayed in each image acquired, making it possible to use the “step-and-stare” method. This enables the scene to be imaged from different positions, and the hyperspectral data cube to be constructed using image pattern recognition methods. It also enables stereoscopic 3D reconstruction of the scene, adding height information to the hyperspectral data cube.
The filters also provide excellent signal to noise ratio, a short measurement time, and high stray light suppression. Compared to gratings and prisms, the filters have a bigger aperture and generate higher transmission. Please e-mail to [email protected] for more information.