The James Webb Space Telescope (JWST) launched in December 2022, first started collecting data in June 2022, and is already revolutionizing our understanding of the distant Universe. With its large segmented mirror, and optimization for infrared wavelengths, JWST was designed to detect and characterize some of the first galaxies to form in our universe and investigate how galaxies then evolve over the age of the Universe to the present day. In this talk, I will present how JWST has pushed our cosmic frontier beyond what was possible with Hubble and share some early results from extragalactic deep surveys and their implications for our understanding of the early universe.

The IEEE Standards Association Project 4001 (P4001) has worked since 2018 to create a standard that covers a range of unmet needs in the field of hyperspectral imaging. Now in its final stages of review, the P4001 standard is the result of the collective work of a large group comprising a significant fraction of the hyperspectral community. The formal scope of the P4001 standard is hyperspectral imaging in the reflective domain (0.25 to 2.5 micrometer wavelength) using cameras which record at least 30 bands, based on four of the most common camera architectures. A main part of the standard is a set of characteristics covering the spatial, spectral, radiometric, temporal, coregistration, stray light, and imperfection aspects of camera performance. The set of characteristics covers several aspects where current practices have been inadequate including resolution, light collection, and coregistration. The characteristics are defined as physical quantities that are thus not tied to particular test procedures. However, the P4001 standard will incorporate a set of recommended testing procedures based on commonly available test equipment. The P4001 standard also defines a set of camera-related metadata aiming to provide information that can support extraction of reliable and accurate information from images, including information about noise, distortions, and uncertainties. Three notional use cases are defined in the standard representing machine vision, laboratory, and geoscience applications. The standard mandates different sets of characteristics and metadata for P4001-compliant specifications and image data respectively. The P4001 standard will undergo formal reviews during 2023 and is expected to be published in its first version at the end of the year.

Hyperspectral imagers provide spectral reflectance values representative of the material being observed. Methods to evaluate uncertainties of hyperspectral imagers remains elusive. All spectrometers are subject to uncertainties attributed to various factors such as wavelength, spectral bandwidth, and linearity. Measuring a set of color tile standards with a variety of known spectral reflectance curves provides a means to quantify the differences between the measured and the known reference values. In effect, this takes a bulk of the uncertainty factors into account. Here we report on the development and application of this method to reduce the measurement errors.

Hyperspectral imaging cameras must be characterized and calibrated prior to data collection. Efforts are underway to standardize the protocols to characterize and calibrate these systems. The goal of this research is to create an approach using the existing NVLabCAP software produced by the NVESD Advanced Sensors Evaluation Facility (ASEF) to aid in characterizing and calibrating specifically visible near-infrared (VNIR) pushbroom style hyperspectral imagers. NVLabCAP has been established as the standard for characterizing and calibrating electro-optical (EO) and infrared (IR) sensors for the Department of Defense and concepts used for EO/IR sensors can be carried over to hyperspectral systems. This paper will cover the theory behind the hyperspectral imager used, valid methods of characterizing VNIR pushbroom imagers, what methods from NVLabCAP can be used for this purpose, and the transformations needed to fit the data into a structure NVLabCAP can utilize.

The work presented introduces a deep-learning based technique to spectrally unmix data containing more than one endmember. It uses a novel loss function with a soft-attenuation mechanism leading the neural network to focus on visual features of the input spectra. A Deep Neural Network was developed to detect Ammonium Nitrate in Visible to Near Infrared (VNIR) and Short Wave Infrared (SWIR) co-aligned aerial hyperspectral imagery. We compare the target detection accuracy of our method, against a well-known classical method referred to as the Adaptive Cosine Estimation (ACE). We show that our DNN based method outperforms ACE by two-orders of magnitude.

Accurate retrieval of surface emissivity from longwave infrared hyperspectral imaging data is necessary for many scientific and defense applications. Emissivity retrieval requires an atmospheric scene model and consists of two steps: atmospheric compensation (AC) and temperature-emissivity separation (TES). AC converts the at-aperture radiance to ground radiance, and TES uses the ground radiance to produce a temperature and emissivity estimate. TES assumes smooth emissivity spectra for solids, compared to atmospheric features. These model-based techniques find a high-resolution atmospheric model which provides the smoothest corrected target emissivity. The high-resolution model is then band-averaged to match the sensor’s spectral response function (SRF), which is difficult to characterize and maintain. Any SRF errors are propagated to the atmospheric spectra, emissivity estimates, and lead to an incorrect atmospheric retrieval. The proposed technique improves the quality of the retrieved atmospheric model and emissivity by correcting SRF errors concurrently with model retrieval. High emissivity pixels are found using an in-scene AC technique. The emissivity and temperature of each pixel are identified using a chosen atmospheric model, material library, and featureless atmospheric bands. At-aperture radiances are then generated for reference. The linear relationship between the generated reference and measured at-aperture radiances is used to determine model correction factors. The correction magnitude is dependent on the calibration and model error. The correct atmospheric scene model is determined as the model requiring the smallest correction factor. Using simulated data, we demonstrate the capability of this technique to substantially reduce the effects of calibration error on atmospheric model and emissivity retrieval.

This conference presentation was prepared for SPIE Defense + Commercial Sensing, 2023.

HyperField, a constellation of polar sun-synchronous nanosatellites, is being developed by the Finnish company Kuva Space. The constellation that will be launched in 2023 consists of 100 CubeSats with hyperspectral systems operating in the visible to near-infrared (VIS-NIR, 450-1100 nm) or Visible-to-shortwave infrared (VIS-SWIR, 450-2500 nm) ranges and provides two to three times daily images of every location on Earth. This paper presents the first and second generations of the Hyperfield satellites. It reviews along with their innovative platform and detector technology, the optical modes, planned mission operations, advanced AI-based processing architecture and novel algorithms are developed to improve the acquisition, enhance the image quality and produce tailored-based services.

The use of multispectral imaging of the study of medieval manuscripts, maps, and other artifacts has long be known to be effective, particularly in the study of palimpsests, manuscripts on which the original writing was erased or scraped off, and the parchment was recycled for another layer of writing to be applied. Commercial solutions exist, providing very high quality imagery, however they are generally expensive, a challenge to transport, and technically difficult to operate. Under funding from the National Endowment for the Humanities we have developed a low-barrier-to-entry imaging system called MISHA: Multispectral Imaging System for the study of Historical Artifacts. The camera system is largely COTS equipment and uses two open-source software packages, one for image capture and one for image processing. The system has been intentionally designed to be used in a wide range of institutions by non-technical staff, with minimal training. To date the systems have been deployed to several libraries, archives, and museums in the US and Europe, and have imaged well over 200 artifacts. Here, we will present the basic system design and operating concept, the capture and processing software packages, and show examples across a wide range of artifacts, time periods, and materials.

Size, weight and power are traditionally limiting factors for unmanned airborne deployment of long wave infrared (LWIR) hyperspectral imaging systems. Spectrum Photonics’ novel HyCARS technology has resulted in the development of a near optimally sized cooled LWIR hyperspectral imaging system by integrating the spectral optics directly within the camera dewar. The completed HyCARS system, which includes an integrated zoom lens, roll correction, calibration and on-board processing, recently completed an airborne demonstration. The HyCARS system design approach, performance specifications, airborne platform integration, results from recent flight testing, and pathways for future development will be described.

The Survey3 by MAPIR is a multispectral camera that can be mounted to a drone to collect imagery that when processed can give insights into the wellbeing of crops. The Survey3 offers several triple-band filter options, but the most popular is the RGN filter which allows agriculturalists to map various vegetation indices like the NDVI. Obtaining accurate NDVI values requires that each RGN band be isolated from the other to achieve accurate ratios when applying a vegetation index. We have found that the Survey3’s RGN bands have significant leakage, especially in the NIR band, and the calibration performed in their software is inadequate to resolve this discrepancy. In this paper, we characterize the amount of overlap in the bands and use this information to develop methods for correcting the results.

The drawbacks of traditional multispectral imaging systems are their slow scanning speed and requirement for extremely accurate optical alignment. A bayer mosaic pixel type image sensor was manufactured by monolithically combining an infrared multi-band plasmonic microfilter with an InAs/GaSb type-II superlattice (T2SL) photodetector for each mosaic pixel. The quasi-three-dimensional (quasi-3D) plasmonic nanostructures as plasmonic microfilter had a period of 1.2 to 2.6 μm with 0.2 μm steps and were fabricated using nanoimprint lithography. The quasi-3D plasmon nanostructure has characteristics such as high optical transmission, wavelength selectivity, and wide incidence angle independence due to the effects of propagating and localized surface plasmon resonance of structure. Our research results suggest that it will be used as a core device in various fields by converging with IoT and big data technology because it does not require a separate optical system and can implement information for each infrared spectrum in real time.