Real-time Hyperspectral Imaging for In-field Data Collection 

Examples include: 

• Troughs in the reflected spectrum associated with specific molecular and atomic absorption wavelengths
• Variations in reflectivity across wavelength bands beyond the visible range
• Small spectral variations enabling observers to distinguish between objects of the same colour.

This kind of spectral information is invisible to the naked eye and undetectable using traditional RGB photography. To reveal it, we need to look at the world in more detail. We need to use hyperspectral imaging technology.

Light captured by a traditional RGB camera is grouped into three broad channels (red, green, or blue). In contrast, hyperspectral imaging systems capture light using tens or hundreds of channels to record its characteristics  more accurately. This means that hyperspectral technology can recreate the spectrum of light hitting the camera, outputting both spectral and spatial data in a single image.

Adding a new layer of hyperspectral data to an image enables many exciting use cases across a range of industries. However, maximising the potential of hyperspectral imaging requires next-generation cameras that are more robust and flexible, easy to use, and can be quickly and effectively operated in the field, away from controlled laboratory conditions or fixed single-purpose systems.

The Living Optics camera represents a technological leap in hyperspectral imaging, enabling advanced portable and mobile data collection while also delivering real-time spectral information. Seeing the world in more detail is more powerful when you can actually go out and see the world.

Portable hyperspectral imaging systems

A lot of studies demonstrate the potential of hyperspectral imaging technology in the lab, using static systems under controlled conditions. While this has value, to enable new use cases, we need portable, flexible, and robust instrumentation capable of performing in-field hyperspectral imaging regardless of the environment.

Portable hyperspectral imaging systems are made possible by the miniaturisation of hyperspectral imaging components, improved edge computing systems, and new hardware designs.

Hyperspectral cameras generate significantly more data than traditional RGB sensors, which compress reflectance spectra into only three channels. Therefore, we need a camera with significant local compute and innovative compression technology to rapidly output 3-dimensional hypercubes of data.

Previous systems also relied upon scanning techniques, building up a hyperspectral image line by line, requiring longer exposure times. This makes them impractical for use in the field, extending the time it takes to capture a single hyperspectral frame and complicating the development of mobile systems that need close to instantaneous image capture to avoid image blur.

Finally, the hardware must be sturdy and flexible enough to operate in the field. This requires:

• A robust design to house delicate optical devices while withstanding transport and use in the field without breaking
• A device that is easy to use while operating in various setups
• A sensor that can capture meaningful hyperspectral data in different lighting conditions.

Taking hyperspectral technology from the lab to the real world with Living Optics 

The Living Optics camera is a portable, visible/near-infrared imager that uses snapshot spectral technology to output video-rate hyperspectral data. The camera generates a 2048 x 2432 pixel RGB image with 4384 evenly spaced hyperspectral sampling points, each consisting of 96 bands across the visible and near-infrared spectral range (440 – 990nm). These two data streams can be output at up to 30 frames per second for real-time data analysis.

The first mass-produced, affordable, portable hyperspectral system, the Living Optics camera aims to take hyperspectral imaging out of the lab and out into the world, powering exciting new use cases.

Weighing 1kg, the compact device (203mm x 148mm x 70mm) is easy to transport and assemble in the field, requiring just a few steps and minimal training to start capturing images. It can be used as a handheld device or mounted in various orientations using multiple attachment points, including on a tripod or on a mobile vehicle.

The camera comes with interchangeable lenses to adapt its field of view and a software development kit with a UI for camera control and data analysis. The system can also be viewed and controlled wirelessly via any mobile device.

The Living Optics solution enables fast and simple on-site hyperspectral data collection regardless of the environment. One of the most exciting fields for portable systems is hyperspectral imaging in agriculture.

Mobile hyperspectral imaging for agriculture

Hyperspectral imaging has found a home in agriculture, with flight instruments (satellites, aircraft, balloons, etc.) capturing large-scale agricultural data to measure environmental conditions and identify crop health trends.

The technology is also being applied to controlled environment agriculture (CEA), where crops are grown in almost clean-room conditions to maximise crop yield. In CEA, hyperspectral data is used for plant monitoring (including identifying disease or stress factors) and nutrient management (optimising the use of irrigation, fertiliser, etc.).

Hyperspectral remote sensing provides valuable data for large-scale agricultural trends, and fixed CEA hyperspectral systems enable focused plant health analysis. But what about hyperspectral use cases for a typical in-field farming operation?

With truly portable devices, like the Living Optics camera, farmers can implement mobile ground-based systems that deliver high-quality and high spatial resolution hyperspectral data.

Whether it is cameras mounted on agricultural vehicles for extensive coverage or images taken by hand to sample spectral information at different points around the field, it is possible to estimate a range of plant quality parameters at the ground level using real-time hyperspectral data.

Parameters that could be monitored include:

• Phenotypic analysis – the variety of crop genetics present and how their growth is affected by different environmental conditions
• Fruit ripeness – factors related to ripeness, such as the accumulation of sugars and starch, to determine the best time for harvesting fruit
• Chlorophyll content – for plant monitoring during growth and improved yield estimates
• Nitrogen content – a non-destructive method of assessing nitrogen and improve plant growth
• Fungal diseases – high spatial resolution data to ensure fungal diseases are identified early, limiting the potential impact.

Get in touch today

From precise orchard monitoring to identifying fungal diseases in cereal production and much more, with portable devices designed to work out in the field, farmers can find countless ways to incorporate hyperspectral data and improve their operations.

Contact Living Optics to learn more.