Hyperspectral and multispectral imaging

Hyperspectral and multispectral imaging

Hyperspectral and multispectral imaging

Hyperspectral and multispectral imaging

Hyperspectral and multispectral imaging

Hyperspectral and multispectral imaging have gained importance in the last years in the world of spectroscopy, but do you really know in what they consist? In this post, we will define them, explain the different types of acquisition systems that are employed and describe their main elements, as well as mentioning their applications.


Hyperspectral imaging (HSI) and multispectral imaging (MSI) are optical spectroscopy imaging techniques [1]. Spectral images have three dimensions, where two are spatial dimensions (x,y) and one is a spectral dimension (λ). The difference between HSI and MSI is purely based on the spectral dimension, see Figure 1. HSI works with tens or hundreds of images that correspond to continuous wavebands, while MSI employs a few images (usually <10) that correspond to certain discrete wavebands [2,3].


Hyperspectral imaging obtains images with high spatial and spectral resolution at the cost of increasing the acquisition and processing time, while multispectral imaging provides images with lower resolution at a few important wavelengths, which enable faster image acquisition and its use in real time applications [2]. HSI helps establishing the optimal wavelengths for a certain application that have to be used by a MSI system.

Comparison between hyperspectral (HSI) and multispectral imaging (MSI).
Figure 1. Comparison between hyperspectral (HSI) and multispectral imaging (MSI). Adapted from [4].

The acquisition system employed in HSI/MSI has three main characteristics: spectral range, spectral resolution and spatial resolution [1]. The spectral range is the range of wavelengths covered by the spectral camera: the visible and near infrared (VNIR) spectrum (400 to 1000 nm), the NIR spectrum (from 900 to 1700 nm) or the short wavelength infrared (SWIR) spectrum (from 1000 to 2500 nm) are common spectral ranges. The spectral resolution is the difference between consecutive wavelengths. The spatial resolution is linked to the actual pixel size, which depends on the number of pixels in the x and y directions as well as the size of the image taken by the camera [3]. These three characteristics will have to be chosen depending on the final application.


Different types of acquisition systems can be distinguished depending on how the spectral (λ) and the spatial (x,y) information is obtained [1,5], see Figure 2. Spatial scanning techniques consist in obtaining the spectrum for a single point (point scanning or whiskbroom mode) or a single line (line scanning or pushbroom mode) and doing a spatial sweep (in (x,y) for point scanning and in x for line scanning). Spectral scanning methods consist in obtaining the spectrum for all the space (x,y) and a single wavelength and doing a spectral sweep in λ. This method is also known as plane scanning or area scanning mode. Finally, a single shot imager or snapshot sensor acquires both the spatial (x,y) and the spectral (λ) information on a single shot.

Acquisition systems employed in hyperspectral (HSI) and multispectral imaging (MSI).
Figure 2. Acquisition systems employed in hyperspectral (HSI) and multispectral imaging (MSI).


A spectral imaging system has three elements: a light source, a wavelength dispersive device, and an area detector [2]. Light sources can be illumination or excitation sources. Illuminations sources are broadband light sources and can be halogen sources (such as our TAKHI series Halogen light sources) or broadband LED sources (such as our COB series LED light sources or our BRETON series MultiLED light sources). Excitations sources are monochromatic light sources that excite a sample, which emits low-intensity light in a broadband wavelength range due to fluorescence or Raman scattering. They include lasers, xenon or mercury lamps among others [2].


Wavelength dispersive devices separate broadband light into different wavelengths and project the dispersed light to the area detectors. Imaging spectrographs, electronically tunable filters and beam splitting devices are examples of wavelength dispersive devices. Imaging spectrographs are based on diffraction grating, which can be transmission or reflection gratings. An electronically tunable filter modifies the bandpass wavelength by means of electronic devices. The two main types are acousto-optic tunable filters (AOTFs, wavelength isolation based on light-sound interaction) and liquid crystal tunable filter (LCTFs, wavelength isolation controlled by liquid crystal cells). Beam splitting devices acquire images at two or more wavelengths at the same time by dividing the light into several parts [2].


Area detectors are in charge of collecting the light. CCD (charge-coupled device) cameras are the most common type of area detectors. A CCD sensor contains several small photodiodes (also know as pixels) with are made of light-sensitive materials such as silicon (Si, used in silicon CCD cameras) or indium gallium arsenide (InGaAs, employed in InGaAs CCD cameras). Silicon CCD cameras are utilized in the visible and SWIR regions while InGaAs cameras are employed in NIR applications.


Applications of hyperspectral and multispectral imaging include agriculture (monitoring of the development and health of the crops), food processing (sorting, identification of foreign bodies, quality evaluation), medicine (non-invasive scans of the skin) or geology (identification of minerals) [5,6]. Click to know more about spectroscopic solutions for industry.


However, in spite of the increasing number of HSI and MSI applications, these technologies have not been widely adopted by the industry yet. The main reason is their high price when compared with other systems. Therefore, right now, the main challenge is to develop HSI and MSI systems that are more economic, compact and easy to handle.


Written by J.J. Imas




[1] Ortega, S.; Halicek, M.; Fabelo, H.; Callico, G.M.; Fei, B. Hyperspectral and multispectral imaging in digital and computational pathology: a systematic review. Biomed. Opt. Express 2020, 11, 3195, doi:10.1364/boe.386338.


[2] Qin, J.; Chao, K.; Kim, M.S.; Lu, R.; Burks, T.F. Hyperspectral and multispectral imaging for evaluating food safety and quality. J. Food Eng. 2013, 118, 157–171.


[3] Amigo, J.M. Hyperspectral and multispectral imaging: setting the scene. In Data Handling in Science and Technology; Elsevier Ltd, 2020; Vol. 32, pp. 3–16.


[4] Giannoni, L.; Lange, F.; Tachtsidis, I. Hyperspectral imaging solutions for brain tissue metabolic and hemodynamic monitoring: Past, current and future developments. J. Opt. (United Kingdom) 2018, 20, 044009.


[5] Hyperspectral and Multispectral Imaging | Edmund Optics Available online: https://www.edmundoptics.es/knowledge-center/application-notes/imaging/hyperspectral-and-multispectral-imaging/


[6] Hyperspectral imaging – Wikipedia Available online: https://en.wikipedia.org/wiki/Hyperspectral_imaging 

Subscribe To Our Newsletter

Get updates and learn from the best