Factors to Consider when Selecting Hyperspectral Cameras
If you're starting your search for a hyperspectral camera and can't send samples, you'll likely begin by examining datasheets. Like other manufacturers, Resonon offers detailed datasheets for each of our hyperspectral cameras. While datasheets are intended to help customers compare products, they can sometimes be misleading. Ultimately, the key is determining if the chosen hyperspectral camera will meet your specific needs.
Figure 1: Important attributes of hyperspectral cameras for most hyperspectral imaging customers.
The image above showcases the key attributes most relevant to our customers. Each attribute is a part of our design trade space. The ideal camera for you will depend on the specific problem you're trying to solve or the opportunity you want to seize, typically determined by a subset of these attributes.
Image Quality
Figure 2: The image on the left was captured with a competitor’s hyperspectral camera, while the image on the right was taken with a Resonon Pika XC2 hyperspectral camera.
Image quality plays a varying role across different applications. When objects of interest are large and occupy a significant part of the field of view, ultra-high resolution might be less critical. However, for capturing small details, image quality becomes paramount.
Assessing image quality often involves understanding the Modulation Transfer Function (MTF) of the imaging instrument, which isn't commonly included in datasheets and can be complex to interpret. To help visualize the capabilities of our hyperspectral cameras, sample images, as detailed in our blog, offer a straightforward way to gauge their image quality.
Radiometry
Figure 3: Mineral classification maps generated from NASA AVIRIS hyperspectral data. The image on the right, following an instrument upgrade, demonstrates a fourfold improvement in Signal-to-Noise Ratio (SNR) compared to the image on the left. Image courtesy of USGS.
Radiometry refers to the effectiveness of light collection by an instrument, directly influencing the Signal-to-Noise Ratio (SNR).
While most datasheets typically provide a single SNR value, SNR actually varies across the wavelength range of the camera. To support informed decision-making, Resonon offers SNR plots for all our cameras.
Radiometry is often an overlooked yet crucial aspect, especially in the context of hyperspectral cameras. Unlike conventional cameras, hyperspectral cameras split the signal into numerous spectral channels, which can reduce signal levels and increase noise in data. Insufficient light can exacerbate this issue, compromising the benefits of additional spectral information.
Increasing illumination or adjusting scan speeds can enhance SNR. However, in scenarios like crewed airborne acquisitions, factors such as fixed illumination and scan rates are non-negotiable due to practical constraints. Our feasibility studies assist in determining the necessary illumination levels and scan speeds to achieve the required SNR for your specific application.
Speed
Figure 4: Some hyperspectral applications necessitate rapid processing speeds akin to a cheetah, while others do not prioritize speed as significantly.
In our lineup of hyperspectral line-scan cameras, the camera speed, also known as the maximum frame rate, directly influences image quality. If an object is scanned with a camera that operates too slowly, the resulting image may appear blurred or stretched in one dimension.
In certain scenarios, adjusting the scanning speed can mitigate these issues. However, for applications such as scanning materials on high-speed conveyor belts, slowing down is not feasible, making speed a critical consideration. Conversely, in airborne imaging and similar applications, excessive camera speed often goes unused because the available amount of light dictates the maximum data acquisition rate.
Size / Mass and Cost
Figure 5: Size, weight, and cost information are readily accessible in camera datasheets and price quotations.
These details are typically found in datasheets and quotations, making them straightforward to grasp. Size and weight are particularly crucial for applications such as deploying airborne drones, whereas they carry less significance for machine vision or laboratory settings. Naturally, cost is almost always a factor to consider.
Ease of Use
Ease of use encompasses hardware, software, and customer support, and is consistently paramount.
Our aim is to enable you to swiftly gather and utilize hyperspectral data. This involves supplying all essential components, detailed assembly instructions, and quick-start guides, in addition to hardware crafted for easy setup and adjustment. We also offer fully integrated airborne systems that include drones, eliminating integration challenges for end-users.
Software usability, though not easily quantifiable on a datasheet, is addressed through our free Spectronon software and example datacubes. We provide detailed user manuals, quick-start guides, and tutorial videos to facilitate a smooth start. For industrial clients and teams managing substantial data volumes, our RVS page offers videos showcasing the user-friendly interface.
Figure 6: A training screen in Resonon's Real-time Vision System (RVS) software, together with intuitive tools, video tutorials, and user manuals, allows customers to quickly become adept with the software.
What About Other Specifications?
You may have noticed that certain commonly mentioned hyperspectral camera specifications, such as distortions and spot sizes, were not listed as key attributes. While these specifications significantly impact instrument performance, we do not want to diminish their importance.
However, for most customers selecting the right hyperspectral camera for their application, directly comparing these specifications is less relevant than focusing on the six key attributes we've highlighted here.
For further information on how we can address your specific application, please contact us.