Why Hyperspectral Imaging Matters in a Data-Driven World? 

To make the most of richer, more nuanced datasets, you need analysis techniques that can keep up – Sophisticated software that is capable of low-level analysis and getting stuck into the fine details of the data. But software that can also abstract away the complexities of its analysis, making it easier for users to extract meaningful insights and achieve their high-level goals. 

Incorporating more intricate and advanced data into new applications requires overcoming this gap in software and analysis. Simply plugging in-depth datasets into pre-existing software barely scratches the surface of what is possible with the actual information it contains.  

A great example of a technology overcoming this gap is Hyperspectral Imaging (HSI). 

Innovating Visual Tasks with Hyperspectral Imaging Datasets

Hyperspectral imaging cameras capture and process light in significantly greater detail than the human eye or standard cameras. While photoreceptor cells and RGB sensors record light using just three broad wavelength bands (red, green, and blue), hyperspectral imaging cameras use tens or hundreds.  

With many more narrow wavelength bands, hyperspectral data makes it possible to accurately recreate the spectra of light hitting the sensor while retaining spatial information. This enables spectral analysis across the image to reveal new information and identify unique signatures related to physical and chemical properties.  

Beyond the dramatic increase in spectral resolution, hyperspectral datasets often extend the spectral range into “invisible” parts of the light spectrum, such as near-infrared (NIR). Additional data from the NIR offers critical information when identifying chemical properties, in particular organic compounds and some minerals. 

Users can then leverage these hyperspectral datasets to uniquely detect, identify, segment, and track materials and objects for their specific application. Using enhanced spatial-spectral information to empower computer vision analysis rather than restrict it to simple RGB data. 

By offering high spectral resolution datasets covering VIS-NIR light, hyperspectral imaging unlocks richer, more precise analysis to enhance existing use cases and enable new ones: 

• Agriculture and Environmental Monitoring: tracking plant health and environmental metrics to identify stress factors, enhance farming practices, and simplify mapping land usage. 

• Healthcare: Use cases across diagnostics (e.g., skin cancer triage, burn assessments), patient monitoring (e.g., blood flow and oxygenation levels), and surgery (e.g., distinguishing between different organs and tissues).  

• Manufacturing and Quality Control: Monitoring production lines and processes, ensuring high standards and cleanliness, and inspecting raw materials. 

• Defence and Security: Advanced surveillance and scene understanding to detect visually camouflaged objects through spectral analysis, identifying new threats. 

To make these applications of hyperspectral imaging a reality and move the technology from research to the real world, we need a new and more advanced camera matched with an advanced software tool. 

The Living Optics Camera – Bringing Hyperspectral Data to Real-World Applications

Launched in January 2024, the Living Optics Camera is a technological leap in hyperspectral imaging and analysis, delivering video-rate frames in a small, portable, easy-to-use device.  

Previously, most commercial devices used line-scan methodologies to slowly gather hyperspectral data and create a full frame. They scan the camera across the scene, capturing high spectral resolution data line by line, and then combine all of this data into a final completed image. This process is inherently slow and incapable of delivering video-rate frames for real-time analysis. 

The Living Optics Camera relies on next-generation snapshot technology, capturing all of the data needed in a single exposure. The camera utilizes two sensors, capturing a full RGB image with one and detailed spectral information with the other. There is no slow build-up of data scanning across the scene. The Living Optics Camera pumps out 30 frames of hyperspectral data every second. 

This solution aims to deliver on the promise of widespread hyperspectral imaging, getting new insights and high spectral resolution data into the hands of real users in the field who can make the most of it. 

While having a small, compact, and portable device solves the hardware challenges of widespread hyperspectral imaging, significant data processing challenges related to outputting and analysing so much data in real-world conditions remain. 

To get into the fine details of a much more detailed image, we need advanced hyperspectral software solutions that can process, manipulate, and analyse vast datasets efficiently and effectively. 

Overcoming the Challenges with Hyperspectral Data with Living Optics Software

With over ten times more data per pixel, there is a gap between capturing hyperspectral data in real-time and extracting meaningful insights in real-time. A tremendous amount of data needs to be quickly extracted from the camera and processed using advanced hyperspectral imaging software. 

In the year since its launch, the value and versatility of the Living Optics Camera has been repeatedly demonstrated. Teams around the world have deployed the Living Optics camera in a wide range of applications, including precision agriculture, food quality inspection, security, material sorting, and more.  

Large amounts of enhanced spatial-spectral data have been generated during this time, and computer vision deployments are already underway. Now, by working with customers and listening to feedback, we have upgraded our software tools to overcome the significant data mining challenges present when analysing hyperspectral datasets. 

With Living Optics’ new software package, real-time hyperspectral data analysis is more accurate and accessible than ever before. Advances in our hyperspectral imaging analysis include: 

• AI-powered spectral labelling to rapidly segment images, automate material classification, define objects, and detect anomalies from the outset 

• The ability to instantly visualise hyperspectral data for faster decision-making in the field or lab 

• Incorporating deep learning algorithms for enhanced spatial-spectral analysis during inference and at deployment 

• Enabling users to capture data, train, and deploy AI models at the edge based on hyperspectral data 

• Integrating with industry-standard APIs and third-party software for plug-and-play deployment 

• An upgraded and simplified interface to meet the diverse needs of our customers 

• Cloud-enabled capabilities for remote data access, collaborative research, and inspecting hyperspectral data in distributed environments 

Living Optics’ software enable local real-time hyperspectral data analysis for real-world applications. You can even run AI and machine learning algorithms locally without the need for external compute or separate model integration.  

These development tools are also hardware agnostic and can be fed many types of hyperspectral data. This means you can purchase the Living Optics Camera to access our advanced hyperspectral analysis capabilities or utilise them as a standalone solution that plugs into your existing systems. 

The Future of Hyperspectral Imaging in a Data-Centric World

All these capabilities combine to streamline hyperspectral data analysis and make the most out of the Living Optics hardware regardless of the environment. By simplifying how you process hyperspectral data and train your models, we aim to reduce the time to deployment for new computer vision systems across different industries.  

From maximising crop yield through automated intervention powered by local analysis to systems that immediately identify manufacturing problems causing a drop in product quality, hyperspectral imaging software has caught up with the data processing challenges of video-rate output in the field. 

Plus, with the increasing role of AI and cloud-based analysis, the Living Optics camera offers the perfect tool kit to incorporate next-generation algorithms and collaborate with colleagues in real-time, no matter where they are based. 

No longer capture hyperspectral datasets and wait to understand what you’ve seen. Incorporate advanced hyperspectral analysis and tooling techniques to gain immediate, usable insights. While hardware advances opened the door to robust, video-rate hyperspectral imaging, software innovations will unlock the full potential of the technology. 

So, whether you are interested in the Living Optics Camera and analysis tools or want to see what our software could do with your own hyperspectral datasets. By contacting our sales team, you will be on the path to enhancing your computer vision capabilities through advanced data and analysis. Also stay tuned because we have more plans for the future, including translating our software tools for line-scanning systems using 3^rd party APIs and integrating geographic information systems to track the location of hyperspectral measurements. Critical for a number of use cases in AgriTech and beyond. sets taken in the field for detecting, segmenting, counting, and sizing fruit in orchards to estimate fruit ripeness and potential crop yields.

Watch our software experts as the explain the benefits of the Living Optics Software tool.