Abstract
Describing the properties and performance of a hyperspectral camera is more complex than for a conventional camera. Still, up to now, hyperspectral cameras have been described largely in the same terms as conventional cameras, which fall short of what is needed to convey the actual performance to users and potential buyers. Responding to this situation, the IEEE Standards Association established a working group in 2018 to develop a new standard for hyperspectral imaging. More than 200 members of the hyperspectral community have taken part in the work, representing users, manufacturers, government labs and academia. This extensive work has resulted in the IEEE 4001 standard, which as of April 2025 is in the final stages of approval by IEEE. This standard defines a comprehensive set of characteristics to describe the performance of a hyperspectral camera as a “black box.” As a result, the standard provides for comparison of performance across several different internal camera architectures. This has required adoption and development of several new characteristics, but only where it has been necessary in order to arrive at a minimal (necessary and sufficient) set of performance metrics. Examples include quantities describing dynamic range, spatial co-registration of bands, spectral co-registration of pixels, actual spatial resolution, actual spectral resolution, light collection, noise floor, and stray light.
Arguably, image exploitation that is not informed about camera properties risks producing suboptimal results. Therefore, the standard also defines a set of metadata needed to capture camera-related properties of a hyperspectral image. With this information, it is, for example, possible to estimate the level of signal-dependent physical noise in individual pixel values. It is also possible to estimate the magnitude of crosstalk from spatial to spectral contrast that results from an error in the pixel-level co-registration between bands. An interesting aspect of the standard can be to study what benefit this extra information brings into different applications. The standard is expected to become an important tool for the hyperspectral imaging community in several ways, ranging from camera design decisions to procurement processes. The novel elements introduced thanks to the “black box” approach also have clear potential for use outside the hyperspectral field, in conventional imaging as well as in spectroscopy. Annexes in the standard outline procedures for testing cameras according to the standard, using commonly available test equipment. It is now a good time for the community to become familiar with the standard and start using it.
The IEEE 4001 Standard for Hyperspectral Imaging
Describing the properties and performance of a hyperspectral camera is more complex than for a conventional camera. Still, up to now, hyperspectral cameras have been described largely in the same terms as conventional cameras, which fall short of what is needed to convey the actual performance to users and potential buyers. Responding to this situation, the IEEE Standards Association established a working group in 2018 to develop a new standard for hyperspectral imaging. More than 200 members of the hyperspectral community have taken part in the work, representing users, manufacturers, government labs and academia. This extensive work has resulted in the IEEE 4001 standard, which as of April 2025 is in the final stages of approval by IEEE. This standard defines a comprehensive set of characteristics to describe the performance of a hyperspectral camera as a “black box.” As a result, the standard provides for comparison of performance across several different internal camera architectures. This has required adoption and development of several new characteristics, but only where it has been necessary in order to arrive at a minimal (necessary and sufficient) set of performance metrics. Examples include quantities describing dynamic range, spatial co-registration of bands, spectral co-registration of pixels, actual spatial resolution, actual spectral resolution, light collection, noise floor, and stray light.
Arguably, image exploitation that is not informed about camera properties risks producing suboptimal results. Therefore, the standard also defines a set of metadata needed to capture camera-related properties of a hyperspectral image. With this information, it is, for example, possible to estimate the level of signal-dependent physical noise in individual pixel values. It is also possible to estimate the magnitude of crosstalk from spatial to spectral contrast that results from an error in the pixel-level co-registration between bands. An interesting aspect of the standard can be to study what benefit this extra information brings into different applications. The standard is expected to become an important tool for the hyperspectral imaging community in several ways, ranging from camera design decisions to procurement processes. The novel elements introduced thanks to the “black box” approach also have clear potential for use outside the hyperspectral field, in conventional imaging as well as in spectroscopy. Annexes in the standard outline procedures for testing cameras according to the standard, using commonly available test equipment. It is now a good time for the community to become familiar with the standard and start using it.