Unlike traditional color imaging, which records light in just three color bands (red, green, and blue), spectral imaging divides light into hundreds of spectral bands og collects detailed information beyond what is visible to the eye.
Two pills might look identical, but their chemical composition could be different.
An apple could look fine on the surface but have a hidden bruise.
A drop of water might not show up on a normal picture, but by using spectral imaging you can record wavelengths in the entire electromagnetic spectrum and collect detailed information about the inspected material.
Advanced algorithms process this data, analyzing the spectrum of each pixel to detect variations, which enables the identification of specific materials or features, like subtle differences in the chemical composition, material properties, and biological characteristics of the inspected items.
Hyperspectral imaging uses hundreds of narrow wavebands, allowing for great detail and the ability to distinguish between materials in a way that is not possible with multispectral imaging.
The downside is the complexity of data processing and interpretation.
%20(1).jpg?width=1000&height=700&name=multispectral%20(1000%20x%20700%20px)%20(1).jpg)
Multispectral imaging uses fewer, broader bands (often 3-10).
This sacrifices the very detailed spectral resolution for a faster and more cost-effective solution.
Using specific wavebands beyond the visual spectrum is another form of spectral imaging.
Examples could be using ultraviolet (UV), near-infrared (NIR), or short-wave infrared (SWIR) light to detect certain materials.
.jpg?width=490&height=440&name=Water%20on%20steel%202%20(490%20x%20440%20px).jpg)
At JLI, we have used hyperspectral imaging to detect unwanted water droplets on the surface of steel tools that have gone through an autoclave process for sterilization.
Hyperspectral imaging can be used to detect the invisible droplets because they absorb different electromagnetic wavelengths than the steel. It is even possible to detect water on the surface of the steel when the tool is packed in a bag.
Hyperspectral imaging will be useful for e.g. sorting plastic into specific plastic types.
Although the plastics may look the same visually, their chemical composition - their polymer basis, filler, and other additives like flame retardants, softeners, etc. - is different, and this can be detected with a hyperspectral camera.
Hyperspectral imaging is well suited for quality control in the pharmaceutical industry.
This could be anything from detecting if the chemical composition of a medicine differs from the standard to measuring the thickness of a pill's coating.
.jpg)
.png)
Just because we can't see it doesn't mean it isn't there.
In this video, Machine Vision Engineer Martin Falk Jensen explains how you can use hyperspectral imaging to see the invisible.
He goes through a number of use cases for hyperspectral imaging, like finding knots in wooden boards.
Book a meeting to learn more:
.png)
Send me an email at hb@jlivision.com
or book a meeting, and let's find out how we can help you.
