Hyperspectral imaging is an innovative and exciting technology that holds incredible diagnostic, scientific and categorization power. Current industry innovation is a testament to the creative power and imagination of the diverse community seeking to optimize this technology. However, fundamental instrument performance is not consistently well characterized, well understood or well represented to suit distinct application endeavors or commercial market expectations. Establishing a common language, technical specification, testing criteria, task-specific recommendations and common data formats are essential to allowing this technology to achieve its true altruistic and economic market potential. In 2018 the IEEE P4001 was formed to facilitate consistent use of terminology, characterization methods and data structures. This talk is a progress report to inform the hyperspectral community of the status of the work to date, the interconnection with other standards and outline the roadmap.
Uniformity from Lambertian optical sources such as integrating spheres is often trusted as absolute at levels of 98% (+/- 1%) or greater levels. In the progression of today’s sensors and imaging system technology that 98% uniformity level is good, but not good enough to truly optimize pixel-to-pixel and sensor image response. The demands from industry are often for “perfect” uniformity (100%) which is not physically possible, however, perfectly understood non-uniformity is possible. A barrier to this concept is that the definition and measurement equipment of uniformity measurements often need to be very specific to the optical prescription of the unit under test. Additionally, the resulting data are often a relativistic data set, assigned to an arbitrary reference, but not actually given an expression of uncertainty with a coverage factor. This paper discusses several optical measurement methods and numerical methods that can be used to quantify and express uniformity so that it has meaning to the optical systems that will be tested, and ultimately, that can be related to the Guide to the Expression of Uncertainty in Measurement (GUM) to provide an estimated uncertainty. The resulting measurements can then be used to realize very accurate flat field image corrections and sensor characterizations.
Sensor fusion and novel “multi-image” systems that have several different spectral ranges are proliferating in tactical and commercial applications. Calibrating these devices requires a variety of sources from quartz-tungsten halogen to blackbodies to more selectable band sources such as LEDs. Usually these sources are used independently in discrete spectral regions, but real reflective and emissive targets often have signatures that make combining these sources necessary if one is to emulate these real spectrums for testing in either image (collimator) or flood (sphere) configurations. A novel approach to combine LED and broadband emitters has been developed to effect stable, calibrated, traceable sources that can match real target spectral signatures.
Many existing and emerging remote sensing applications in the UV, Visible, NIR, SWIR, MWIR and LWIR regions are challenging the conventional thinking of radiance and temperature calibration techniques. While the relationship between blackbody temperature and optical radiation is well understood, often there is an “invisible” dividing line between treatments of these values as either optical radiance or temperature. It is difficult to perform seamless temperature and radiance calibrations across the point of 2.5um. Spectrum above 2.5um is typically related in temperature terms and below 2.5um may be either spoken of in terms of temperature or optical radiance. There is also a natural unit “convergence” issue at 2.5um, due to the crossover of significant levels of emissivity, reflectance and temperature at this point. NMI traceability in the spectral region of 2.5-14.0um can also be a problem especially for spectral radiance. This paper will outline a possible turn-key test bench solution that provides traceable solutions for both temperature and radiance value in these regimes. The intent of this paper is to offer a possible solution and challenge the infrastructure that exists today over the 0.3-14um range in order to obtain a valid spectral radiance or temperature value, or both, to support emerging sensor fusion technology.
Hyperspectral imaging (HSI) is an exciting and rapidly expanding area of instruments and technology in passive remote sensing. Due to quickly changing applications, the instruments are evolving to suit new uses and there is a need for consistent definition, testing, characterization and calibration. This paper seeks to outline a broad prescription and recommendations for basic specification, testing and characterization that must be done on Visible Near Infra-Red grating-based sensors in order to provide calibrated absolute output and performance or at least relative performance that will suit the user’s task. The primary goal of this paper is to provide awareness of the issues with performance of this technology and make recommendations towards standards and protocols that could be used for further efforts in emerging procedures for national laboratory and standards groups.
Sintered PTFE is an extremely stable, near-perfect Lambertian reflecting diffuser and calibration standard material that has been used by national labs, space, aerospace and commercial sectors for over two decades. New uncertainty targets of 2% on-orbit absolute validation in the Earth Observing Systems community have challenged the industry to improve is characterization and knowledge of almost every aspect of radiometric performance (space and ground). Assuming “near perfect” reflectance for angular dependent measurements is no longer going to suffice for many program needs. The total hemispherical spectral reflectance provides a good mark of general performance; but, without the angular characterization of bidirectional reflectance distribution function (BRDF) measurements, critical data is missing from many applications and uncertainty budgets. Therefore, traceable BRDF measurement capability is needed to characterize sintered PTFE’s angular response and provide a full uncertainty profile to users. This paper presents preliminary comparison measurements of the BRDF of sintered PTFE from several laboratories to better quantify the BRDF of sintered PTFE, assess the BRDF measurement comparability between laboratories, and improve estimates of measurement uncertainties under laboratory conditions.
Integrating sphere (IS) based uniform sources are a primary tool for ground based calibration, characterization and testing of flight radiometric equipment. The idea of a Lambertian field of energy is a very useful tool in radiometric testing, but this concept is being checked in many ways by newly lowered uncertainty goals. At an uncertainty goal of 2% one needs to assess carefully uniformity in addition to calibration uncertainties, as even sources with a 0.5% uniformity are now substantial proportions of uncertainty budgets. The paper explores integrating sphere design options for achieving 99.5% and better uniformity of exit port radiance and spectral irradiance created by an integrating sphere. Uniformity in broad spectrum and spectral bands are explored. We discuss mapping techniques and results as a function of observed uniformity as well as laboratory testing results customized to match with customer’s instrumentation field of view. We will also discuss recommendations with basic commercial instrumentation, we have used to validate, inspect, and improve correlation of uniformity measurements with the intended application.
Integrating spheres for optical calibration of remote sensing cameras have traditionally been made with Quartz
Tungsten Halogen (QTH) lamps because of their stability. However, QTH lamps have the spectrum of a blackbody
at approximately 3000K, while remote sensing cameras are designed to view a sun-illuminated scene. This presents
a severe significant mismatch in the blue end of the spectrum. Attempts to compensate for this spectral mismatch
have primarily used Xenon lamps to augment the QTH lamps. However, Xenon lamps suffer from temporal
instability that is not desirable in many applications. This paper investigates the possibility of using RF-excited
plasma lamps to augment QTH lamps. These plasma lamps have a somewhat smoother spectrum than Xenon. Like
Xenon, they have more fluctuation than QTH lamps, but the fluctuations are slower and may be able to be tracked in
an actual OGSE light source. The paper presents measurements of spectra and stability. The spectrum is measured
from 320 nm to 2500 nm and the temporal stability from DC to 10 MHz. The RF-excited plasma lamps are quite
small, less than 10mm in diameter and about 15 mm in length. This makes them suitable for designing reasonably
sized reflective optics for directing their light into a small port on an integrating sphere. The concludes with a
roadmap for further testing.
Application-specific integrating sphere-based, integral veiling glare measurement systems are described. The sources use
the integral method for measuring the veiling glare (VG) index of various lens-based imaging systems. The calibration
source has provisions in the form of a collimating lens holder to simulate a situation where the black target and bright
surround are at a sufficiently great distance to give measurements of VG index which are the same as that which would
result if the distance where infinite. The design criteria for the integral VG test source are presented. Included is a
summary of the end-user specifications in regards to spectral radiance, levels of attenuation, irradiance stability, and
aperture uniformity and contrast. Spectral radiometric predictions and actual output levels are compared.