The main difference between multispectral and hyperspectral imaging is the number of wavebands being imaged and how narrow the bands are.

Multispectral imagery generally refers to 3 to 10 discrete “broader” bands .

Hyperspectral imagery consists of much narrower bands (10-20 nm). A hyperspectral image could have hundreds of thousands of bands.

Hyperspectral and multispectral images have many real world applications. For example, hyperspectral imagery has been used to map invasive species and help inmineral exploration.

An example of a multispectral sensor is the Landsat-8 satellite. Landsat-8 produces 11 images with the following bands:

Band 1: Coastal aerosol (0.43-0.45 um)
Band 2: Blue (0.45-0.51 um)
Band 3: Green (0.53-0.59 um)
Band 4: Red (0.64-0.67 um)
Band 5: Near infrared NIR (0.85-0.88 um)
Band 6: Short-wave Infrared SWIR 1 (1.57-1.65 um)
Band 7: Short-wave Infrared SWIR 2 (2.11-2.29 um)
Band 8: Panchromatic (0.50-0.68 um)
Band 9: Cirrus (1.36-1.38 um)
Band 10: Thermal Infrared TIRS 1 (10.60-11.19 um)
Band 11: Thermal Infrared TIRS 2 (11.50-12.51 um)

Each band has a spatial resolution of 30 meters with the exception of band 8, 10 and 11. Band 8 has a spatial resolution of 15 meters. Band 10 and 11 have spatial resolutions of 100 meters.

NASA’s Airborne Visible / Infrared Imaging Spectrometer (AVIRIS) is an example of a hyperspectral imager. AVIRIS delivers 224 contiguous channels with wavelengths from 0.4-2.5 um.

Having a higher level of spectral detail in hyperspectral images gives better capability to see the unseen. By comparison, hyperspectral remote sensing was able to distinguish between 3 minerals because of its high spectral resolution. The multispectral Landsat Thematics Mapper could not distinguish between the 3 minerals.

However hyperspectral imaging  also adds a level of complexity. 200 narrow bands can be difficult to work with at times.

Hyperspectral cameras were first developed in the 1980s by military and government agencies like NASA, which used the cameras for remote sensing. Mounted on light aircraft or satellites, detectors mapped the Earth’s surface not just in the visible spectrum, but also in the infrared spectrum to monitor different types of vegetation and minerals from the sky.

Human eyes are sensitive only to broad overlapping ranges of frequencies that peak around red, blue and green, three colours that cover all the wavelengths in the visible spectrum. Similarly, a regular camera records three spectral channels in every pixel (red, green, and blue).

Hyperspectral cameras, on the other hand, can detect many different wavelengths separately. They can also see across a wider spectrum than humans can, extending into infrared and ultraviolet. Hyperspectral imaging is the technique of acquiring a 2D image where every pixel in the image contains a continuous spectrum.

Hyperspectral imaging, provides a digital image with far more spectral (colour) information for each pixel than traditional colour cameras. The raw data output is often visualized as a “datacube.” This can be thought of as a stack of tens to hundreds of pictures with each successive image representing its own specific colour (spectral band), or equivalently, as a detailed spectral curve for each pixel.

When light enters the hyperspectral camera, a prism or spectral filter splits the light into its constituent wavelengths. Detectors measure the light in each of these hundreds of narrow bands to see how the material reflects and absorbs light over the full spectrum.

This can be used to learn about the material: what kind of atoms it is made of and how they are bonded.

This division of the light spectrum into many small wavelength bands, captures a unique spectral fingerprint or signature of an object. This spectral signature gives very detailed information about the material constitution of the imaged object, considerably improving its identification and classification and is today recognized as a key enabling technology for next generation industrial inspection, medical diagnosis and security applications.

Early hyperspectral cameras were relatively large because of the arrays of intricate optical components, and consequently extremely expensive and thus the province of satellite platforms and advanced laboratory use. Many cameras usually had to be cooled to -200°C, making them bulky and expensive.

In addition, the speed at which satellites could function was hampered by having to send huge amounts of data back to Earth for processing. This is because traditional hyperspectral cameras scan an image line-by-line, like a printer churning out text. As each pixel comes with a tower of data, every line comes with its slice of a “data cube”. Once the full image is captured, each horizontal cross- section shows the entire image in a single wavelength.

Innovations in “snapshot” imaging are now enabling some cameras to capture both the spectrum and spatial position of each pixel at once, allowing hyperspectral cameras to take many images per second and making hyperspectral movies, and greatly reducing the data payload in the process.

As a result, hyperspectral imagers are not only becoming quicker and more sensitive, but also ever smaller. The latest advances in silicon technology promise to reduce the weight of a hyperspectral camera from kilos to a few grams. This means that hyperspectral sensors can now be made small enough to be accommodated in small unmanned UAVs and small enough to be integrated into mobile and handheld devices. By as early as 2018, handheld versions will allow everyday professionals – farmers, doctors, police officers and environmental engineers – to instantly access this invisible world, which could open up a host of new opportunities. It gives mankind a whole new set of eyes.

Blackroc are currently working with leading research and specialist optics groups to create a mobile imaging handheld product utilising a hyperspectral imaging handheld system based on an image sensor, at the level of the chip itself, removing the need for expensive, bulky and complex optics that are used on traditional systems today. This will be small enough for the mobile world and at a cost which increasingly, with the scale of volume, will make this exciting technology available to many more applications.

Blackroc are also working with the Centre for Precision Agriculture in UK, assessing both multispectral and hyperspectral handheld solutions which will assist in the identification of disease and nutritional deficiency in crops, and effects due to drought and insect infestation.

Hyperspectral cameras are already having an impact across a broad variety of industries. One is food and drink, where infrared images can detect the tenderness of meat or the taste of a sponge cake by analysing sugar, fat and moisture content.
Another is medicine, where chemical and structural changes in tissue can point to skin cancer.  Eventually, compact Hyperspectral systems could be adapted for much wider use such as in outpatient medical clinics for point of care diagnostics.

It has been described as the ultimate non-destructive testing.