Hyperspectral imaging technology

Hyperspectral imaging technology is designed to make images that reveal a maximum of spectral information for each pixel of the image.

Another way of looking at a hyperspectral image is as a set of image layers, each in another wavelength of the electromagnetic spectrum. The combination of those layers is called a hyperspectral data cube.

Want to see how it works? Download our software and a reference datacube.

There are multiple hyperspectral imaging technologies. Each applies different techniques to filter the light and capture the image data. The key differentiators you need to keep an eye on are:

1. acquisition speed – the time required to capture the hyperspectral data cube
2. snapshot capability – the ability to acquire a hyperspectral image, without scanning
3. spatial resolution – the size of the pixel array, similar to regular photography
4. spectral resolution – the number of frequency bands (or layers) each hyperspectral image contains.

Selecting the required properties is a tradeoff, driven by the requirements of the application. If you know which spectral bands to look for, there’s no need to capture the full spectral range. Especially because a lower resolution will enable faster acquisition speeds and open new possibilities with hyperspectral video imaging.

Read this 2023 press release about imec's combined VIS & NIR spectral camera system, complemented with hi-res RGB imaging, for data acquisition at video rate – allowing flexible assessment of the pros and cons of different spectral resolutions and ranges.

Imec can assist you in tailoring its hyperspectral imaging technology to your application. For instance, why not use one of our high-resolution evaluation systems to define which frequency bands are relevant for your needs? Afterwards, we can help you to build your dedicated filters or camera system.

Get in touch

On-chip hyperspectral imaging technology

Traditional hyperspectral imaging scanners contain a lot of precision optics to select and diffract the light. This makes them relatively heavy, expensive, slow and delicate – with a need for frequent recalibration.

Imec’s hyperspectral imaging technology takes a different approach. As a world-leading R&D hub in nanotechnology, we have developed a wafer-level CMOS process to integrate thin-film spectral filters directly on the pixels of the image sensor.

Our filters eliminate the complex optical camera design and can be deposited on a commercial CMOS imager, like the one in your smartphone or a model with scientific imaging capabilities.

The result is that our hyperspectral image sensors enable imaging systems that are:

• capable of operating in video mode
• highly customizable
• compact
• robust to environmental circumstances like shocks and vibrations, which eliminates the need for recalibration
• mass-producible and therefore less expensive

This leads to the increased adoption of hyperspectral imaging in a range of applications.

Flexible technology platform for high-res and real-time hyperspectral imaging

Our unique on-chip hyperspectral imaging technology allows us to make different spectral filter patterns. Those fall into two categories:

1. A mosaic pattern (Bayer-like) on top of a group of 3X3, 4X4 or 5X5 pixels. This enables real-time hyperspectral imaging, which is crucial for every application with a moving camera or ‘target’.
2. A striped pattern on top of each row of pixels. This results in high-resolution hyperspectral imaging – comparable to so-called linescan or push-broom cameras, only more compact, faster and easier to use.

This website offers you an overview of our ready-to-use evaluation kits, including remote support, and the easy-to-use HSI STUDIO, HSI MOSAIC and HSI SNAPSCAN software suites to make hyperspectral imaging accessible to anyone.

Of course, we can also develop custom spectral image sensors, with your desired pattern lay-out, filter specification, imager chip, ... All the way up to the design of the complete camera system.

Technology publications