Multispectral Imaging | How Multispectral Differs From Hyperspectral Imaging

Hyperspectral Imaging: An Image Based On True Spectroscopic
Data Derived From Contiguous Wavelength Data Points

Multispectral images capture a fixed field of view (FOV) sequentially through a series of bandpass filters placed in front of a camera. Bandpass filters can be either dielectric, or can be generated electronically using an Acousto-Optic Tunable Filter (AOTF) or a Liquid Crystal Tunable Filter ( LCTF).

Multispectral imaging is not "spectroscopic" and cannot be used with samples that either move or change during time it takes to acquire the FOV through the first to the final filter in the series.

For example, if a field of view contains seven fixed objects, each with one color of the rainbow, then all you need to identify each object is seven wavelength data points (bandpass filters) ranging from violet to red.

If the task is to define the accurate color of each object, or detect and quantify changes or variations in color, then it is necessary to use analytical spectroscopic data based on contiguous wavelength data points.

The best option, in this case, is the PARISS, Wavelength Dispersive, Analytical, Hyperspectral Imaging system.

Multispectral Basics: Low resolution, does not require spectroscopic data to detect target objects, or conditions
Samples: Because Multispectral Images are usually acquired by sequentially changing filters, objects in the field of view must remain motionless and unchanging during the acquisition series.
Key Multispectral applications: Segmentation of distinct colors found in fixed objects
Multiple fluorophores or chromophores: Can sometimes segment multiple fluorophores or chromophores, and identify auto-fluorescence. Accuracy often depends on the simplicity of the sample and the sophistication of the algorithms used for segmentation or "unmixing"
Spectral Resolution: 10 to 20 nm

To watch a presentation that describes both Multispectral Imaging and Hyperspectral Imaging "click to play" the interactive presentation

This page describes compares multispectral with hyperspectral imaging

Multispectral Imaging: Typically a Filter-Based Technique that Acquires Multiple Non-Contiguous Wavelength Data Points

Hyperspectral Imaging: An Image Based On True Spectroscopic Data Derived From Contiguous Wavelength Data Points

Multispectral Basics: Low resolution, does not require spectroscopic data to detect target objects, or conditions

Samples: Because Multispectral Images are usually acquired by sequentially changing filters, objects in the field of view must remain motionless and unchanging during the acquisition series.

Key Multispectral applications: Segmentation of distinct colors found in fixed objects

Multiple fluorophores or chromophores: Can sometimes segment multiple fluorophores or chromophores, and identify auto-fluorescence. Accuracy often depends on the simplicity of the sample and the sophistication of the algorithms used for segmentation or "unmixing"

Hyperspectral Imaging: An Image Based On True Spectroscopic
Data Derived From Contiguous Wavelength Data Points