This special issue of Applied Optics on Advanced Infrared Technology and Applications collects significantly expanded refereed papers presented at the conference of the same name, held in Quebec City, Canada, Sept. 27 to Sept. 30, 2017. All the authors who participated at the conference were contacted and invited to contribute to this special issue. Furthermore, the AO dedicated issue on AITA was open to contributions from other practitioners of IR, through direct contact and a call for papers published in AO.
© 2018 Optical Society of America
Dr. Laura Ronchi Abbozzo of the Ronchi Foundation, formerly with the University of Florence, has been organizing the AITA (Advanced Infrared Technology and Applications) meetings in different cities, usually in Italy, for the last 30 years with the help of a small group of prominent Italian scientists. This is the second time that the conference was organized outside Europe; it first took place exactly 10 years earlier in Leon, Mexico, in 2007. AO received 39 papers; of those 22 were accepted upon the strict review process involving two reviewers, as is customary for journals at The Optical Society. The submission deadline was December 31, 2017, with the accepted papers published only five months later (May 1, 2018). Most were available online within two to three months after submission.
Reasons for rapid advances in infrared (IR) technology are abundant, and quite interesting to contemplate, when compared with the advances in the visual interval . The IR interval includes a rather wide spectral band between the visible and the terahertz range. The edge of the visible is relatively easy to define: the typical person does not see any radiation beyond the red. The onset of the terahertz range is defined by the use of different sources and detectors and, even more, by specialized, advanced electronics.
Several manuscripts, subject to demanding reviewer observations, are still in the revision process, but expected to be published within the next few months. As IR technology and applications are moving from the novelty stage to real-life applications, we are finding that the algorithms of image processing techniques developed for the visible interval are much more difficult to implement in the noisy and ambiguous IR spectral range. This will lead into development of new algorithms to specifically address the idiosyncrasies of the invisible IR spectral range.
Still, three such image processing techniques and simulations relying primarily on some modification of neural network algorithms were accepted on time for inclusion in this compendium.
Highlights among the papers published include a group at Virginia Tech, USA, headed by Mahan, that determined limitations of the Monte Carlo ray-trace (MCRT) diffraction technique based on the Huygens–Fresnel principle. They conclude affirmatively that the MCRT method may indeed be applied effectively to accurately describe diffraction. However, they caution that careful attention must be paid to the interplay among the key parameters: the number of aperture points, the number of rays traced per aperture point, and the number of bins on the screen.
Guei and Akhloufi, of Perception, Robotics and Intelligent Machines (PRIME) group at the University of Moncton, Canada, evaluate deep learning enhancement of IR face images using generative adversarial networks. Deep learning framework, based on the use of deep convolutional generative adversarial networks (DCGAN), is developed for the IR face image super-resolution. The authors implement DCGAN for upscaling the images by a factor of , starting at a size of and obtaining a face image. Tests are conducted using different IR face datasets operating in the NIR and the LWIR spectral intervals. They demonstrate that the proposed framework performs well and preserves important facial details.
Zewei et al. at the State Key Laboratory of Fluid Power and Mechatronic Systems, Zhejiang University, China, and Christel-Löic Tisse, France, collaborated on single image based non-uniformity correction (NUC) of uncooled long-wave IR detectors. They propose a deep-learning approach to eliminate the fixed-pattern noise (FPN), caused by the non-uniform response of focal-plane-array and its associated optoelectronic system. They successfully demonstrate that a better single-image-based NUC operator can be learned from a large number of simulated training images. The training scheme is based on convolutional neural networks and a column FPN simulation module. The successful technique reconstructs a noise-free IR image from the corresponding noisy detected one. The authors demonstrate that the deep-learning-based approach delivers superior performance in terms of the FPN removal, detail preservation, and suppression of artifacts.
The materials that are used in IR imaging systems have different performance requirements. We include two technology development papers that report on the achievements of detector optimization.
The combination of high detectivity, wideband wavelength coverage, spectral flatness, and spatial uniformity make blocked impurity band (BIB) trap detectors an excellent reference detector for spectrally resolved measurements and radiometric calibrations over the near- to far-IR spectral range. Woods and Defibaugh of the Sensor Science Division at the National Institute of Standards and Technology, together with a group of collaborators at the Jung Research and Development Corp., all in the U.S., describe the achievements with the wideband IR trap detector based on doped silicon photocurrent device. The detectors are fabricated from arsenic-doped silicon (Si:As) BIB photodetectors, composed of two detectors in a wedge geometry, with an entrance aperture diameter of either 1 or 3 mm. At 10 K temperature, nearly ideal external quantum efficiency () may be attained over much of the range from 4 to 24 μm, with significant responsivity covering a broader interval from 2 to 30 μm. The traps exhibited maximum etalon oscillations of only 2%, which is about 10 times smaller than those of the single Si:As BIB detectors measured under similar conditions. Spatial non-uniformity across the entrance aperture of the traps is about 1%. These detectors promise to be highly suitable for the near- to far-IR applications.
Rogalski and coworkers at the Institute of Applied Physics of the Military University of Technology, Poland, describe the potential performance prediction of p-i-n HgCdTe long-wavelength IR HOT photodiodes. An advanced computer program was applied to explain in detail the influence of different recombination mechanisms. The photon recycling effect was determined to drastically limit the influence of radiative recombination on the performance of small pixel HgCdTe photodiodes. The authors offer an additional insight on the ultimate performance of long-wavelength IR, high-operation-temperature HgCdTe arrays with pixel densities that are fully consistent with background- and diffraction-limited performance of the imaging optical system.
One of the most attractive features of IR radiation for the non-destructive, non-obtrusive, non-interfering testing is that any body at finite temperature emits the radiation. Coupled with image processing techniques, the visible/IR spectroscopy offers means of rapid and non-invasive characterization of matter. We include four publications dealing with material studies, including one in-vivo tissue characterization.
Lin and collaborators at the Institute of Intelligent Machines, Chinese Academy of Sciences, and State Key Lab of Pulse Power Laser Technology, National University of Defense Technology, China, predict accurately and rapidly the detection of soil and fertilizer properties based on visible/near-IR spectroscopy. The authors demonstrate that the visible and near-IR diffuse reflectance spectroscopy is an effective tool for extensive investigation of soil and fertilizer properties. Both the principal component regression and genetic algorithm may sufficiently reduce the number of variables to pinpoint the characteristic ones. This results in enhanced speed and in greatly improved prediction accuracy.
One of the great advantages of living in the 21st century is that new materials are discovered every day. Even before we learn to utilize them well and efficiently, we may sometimes learn that they are toxic. Then we are forced to start searching again, looking for the advantages of employing new materials without the disadvantages of toxicity. Ullah and Wang of the Laboratory for Information Optics and Optoelectronic Technology, Shanghai Institute of Optics and Fine Mechanics, China, study bisphenol “S” and bisphenol “A” using FTIR spectroscopy. The first compound was discovered to be toxic, so the second one was developed in hopes of avoiding toxicity. The authors are comparing their spectra, suspecting that the second one will also prove toxic. They hope that their study will offer the explanation for their toxicity leading to the possibility of finding a hazard-free compound in the future.
Martan et al., from the New Technologies Research Center (NTC), University of West Bohemia, in the Czech Republic, present analysis of short wavelength IR radiation during laser welding of plastics. A new measurement system and a new approach for calculating the IR radiation are laid out. Detailed investigation in the quasi-simultaneous transmission of laser welding in plastics is presented. The new calculation is based on the simplification of the process to two positions and two temperatures (surface and molten interface), incorporating the knowledge of the spectral optical properties of the material, filters, and the camera response. The results of measurement and calculation for three different optical filters and polyoxymethylene (POM) samples with two thicknesses are given. Good agreement is obtained for the calculations using normal transmissivity of the semi-transparent polymer.
Verdel et al. of Physics Institute Jožef Stefan, Ljubljana, Slovenia, collaborated with the researchers at the Beckman Laser Institute of the University of California, Irvine, USA, to perform a feasibility study of skin in vivo. They combined photothermal radiometry (PPTR) and optical spectroscopy in the noninvasive assessment of skin structure. The analysis involves simultaneous multidimensional fitting of the measured PPTR signals and DRS spectra to the predictions of a numerical model of the Monte-Carlo radiation transport, implemented in a four-layer optical model of human skin. The measured values correspond well with the maximum epidermal thicknesses determined with the multiphoton microscopy images.
Much information may be made visible in the world beyond the visible (i.e., in the IR). In this research field, new terminology has been introduced for imaging, that we call “thermography”. This word denotes just about everything that may be accomplished when one is recording a two-dimensional incidance of radiation with the two-dimensional IR detector.
The first part of this word, “thermo”, implies that we are recording thermal radiation, same as that recorded by Lord Kelvin when he exposed a thermometer to the solar spectrum beyond red. The second part, “graphy”, is defined as a descriptive science, denoting (a method of) writing and recording. The introduction of this word was actually quite necessary because the IR image has to be recorded into some medium before it may be presented as information to a human. Also, in this aspect, IR is different from visible: detected thermal images require an additional transducer to make them visible, often some current-voltage combination that drives the display. Three applications of thermography are described next.
An attractive feature of IR radiation is that we can control the amount of emitted radiation by selectively supplying thermal energy to the tested piece for the purpose of generating the desired signal. This represents an interesting and complicated inverse problem. We know that the time-dependent inhomogeneous heat equation determines the temperature distribution; therefore, we may deposit an appropriate amount of thermal energy. Absorbed energy elevates the local temperature that increases the amount of the emitted radiation, according to Planck’s law. The relationship between temperature distribution induced in the sample and the thermal images that the sample and its internal and external geometry and physical properties create in the IR light is often used for defect localization.
Furthermore, in IR several options are available in generating the structured radiation by controlling and appropriately stimulating the temperature of the radiation-emitting object area. This flexibility resulted in the creation of a number of special techniques to measure the response of an object, and specifically its exposed surface, to the structured irradiation. In addition to the availability of a laser with precise timing control as a source, a pair of synchronized lamps may also be placed in a symmetrical configuration. These two techniques are often referred to as active thermography. The surface response may be characterized from a series of recorded frames as a pulse-phase or lock-in thermography, the former detecting phase and the latter detecting amplitude. Five teams discuss these techniques in diverse applications. Two authors present active thermography for robotic applications, two apply lock-in and pulsed-phase thermography in diverse novel applications, and one uses induction thermography to detect cracks.
Autonomous inspection of advanced components remains a challenging task for the industrial applications of robotics, even though this technology has appreciably advanced within the last few years. When the inspection needs to be implemented on the assembly line, with controlled motion and inspection intervals, a number of procedural and technological advancements have to be implemented for its successful operation. Peeters and collaborators from the University of Antwerp, Op3Mech and Laboratory of Biomedical Physics, and Vrije Universiteit Brussel, Belgium, describe an optimized robotic setup for automated active thermography. They implement algorithms using advanced path planning and visibility study. The automated methodology uses numerical simulations to determine the best set of experimental parameters to inspect the structure for defects, employing active thermography. The inspection is accomplished using a robotic arm and advanced path-planning tools to determine the optimal positions of the excitation and measurement loci. The optimization task implements a genetic algorithm and spline regression models to optimize the heat power, robot-arm speed, camera frame-rate, and excitation timing to achieve the automatic inspection.
Švantner and colleagues at the New Technologies Research Center, of the University of West Bohemia, Czech Republic, report on the quantitative evaluation of active thermography using contrast-to-noise ratio. Their contribution is focused on the assessment of the results, as impacted by the choice in parameters. Different interpretation procedures are introduced. The authors argue that the choice of working parameters impacts the reported results. They find that the contrast-to-noise ratio may be used for superior quantitative evaluation of the results.
A group of investigators at Arcelor Mittal Maizieres Research, led by Taram, and another one at UFR Sciences Exactes et Naturelles, directed by Bodnar, both in France, evaluated spot welds using eddy current thermography. They perform nondestructive testing of resistance in the spot welds. Their sample is a three-sheet stack-up, joined together by welding spots. An inductive Helmholtz coil is used for heating up the sample. The thermal response of each spot is monitored using a cooled indium antimonide camera. The sequence of captured images is processed using the pulsed-phase thermography technique. The phase images are augmented with a supplementary filtering operation to facilitate the identification of the defect outline.
Muzika et al. at the New Technologies Research Center of the University of West Bohemia, Czech Republic, apply lock-in and pulsed thermography to the testing of solar cells. LED-illuminated (LEDILIT) lock-in and flash-pulse thermographic techniques (FPT) are compared in this study. Traditionally, the lock-in methods have more frequently been used for the solar cell inspection. In this study, the LEDILIT has been found a highly appropriate technique for inspection of defects related to the photovoltaic effect. Local shunts, cracks, and artificial, laser-made defects are efficiently detected. This may be contrasted with the FPT method that is able to identify only a few of the most noticeable shunts and laser-made defects.
Genest and Li, from the Aerospace Research Centre, National Research Council Canada, evaluate the use of induction thermography of steel coupons with cracks. The feasibility of the induction thermography technique to detect cracks is assessed experimentally and numerically on notched steel coupons using two coil configurations: straight line and loop shape. The coupons had different lengths of fatigue cracks, varying from 0 (zero) to 3 mm. The loop coil resulted in the creation of higher temperatures than the straight line. The authors demonstrate that the induction thermography may detect a crack as short as 1 mm in the notched steel coupons.
Those of us who work in the IR like to think that IR is that part of the E&M spectrum, where all the characteristics that we know about the visible are still present, except the human eye has to be replaced, or better put, augmented by suitable detectors. Just transitioning the spectral band from the human-centered (visible) to machine-centered (IR), we observe that the number of problems that can be solved is significantly increased by the special characteristics of IR radiation. Three such works describe transitioning of techniques between the IR and other spectral regions.
Liu and Wang, in College of Physics and Optoelectronics, Taiyuan University of Technology, China, studied light-intensity- and FOV-controlled adaptive fluidic iris. A light intensity and field of view (FOV) controlled adaptive fluidic iris is created using a 90° twisted-nematic liquid crystal (TNLC) cell and two orthogonal polarizers, attached on the substrate. A black mask is furthermore placed in the middle of the chamber. When changing the hydraulic pressure from the inlet and outlet, the black mask moves within the chamber. This behavior is equivalent to changing the iris position along the optical axis. Thus, the iris may control the FOV of an optical system. The experiments show that the device can control the light intensity from 100% to 0% by applying a voltage of 9 V on the TNLC cell.
Amin and collaborators in the Department of Electrical and Computer Engineering, George Washington University, with colleagues at Johns Hopkins University and 3 Omega Optics, Inc., all from the USA, describe attojoule, efficient graphene optical modulators. Considering that the silicon-based modulators are bulky and inefficient, they describe elegant, heterogeneously integrated graphene-based devices. Furthermore, they present recent results on a graphene-based hybrid-photon-plasmon modulator on a silicon platform. They discuss electron beam lithography treatments for transferring graphene for the relevant Fermi level tuning. This, physically compact, 100 aJ/bit modulator opens the path toward a novel class of attojoule efficient opto-electronics components and devices.
In a collaborative effort between Spain and Canada, Usamentiaga from the Department of Computer Science and Engineering, University of Oviedo and the Maldague group in the Computer Vision and Systems Laboratory of Laval University perform a comparison among three calibration methods for IR cameras: a direct and an iterative estimation of the transformation between the image and world coordinates, and a complete camera calibration method using a specifically designed calibration target. The complete method clearly outperforms the two others methods with an average error of only 0.060 mm, representing 0.08% error in the measured distance of the setup.
In presenting the advancements in the field of IR in this dedicated issue, we can look with great deal of satisfaction at the breadth of applications and contributions of IR technology to the well-being of humanity. We will continue using the IR part of the E & M spectrum as a region where nondestructive testing may be performed to detect material defects and performance anomalies of both inanimate objects and animate creatures. With continued technology developments, the experimental techniques will likewise become more attuned to ever-smaller dimensions of features under study. In this special issue we learn that characterization and non-destructive testing of novel materials, ranging from the graphene to plastics, requires utilization of ingenious new imaging, analytical, simulation, and image processing techniques. We invite all who want to see the future actually taking place to join us at the next AITA meeting in Florence, Italy, in September 2019. Further information will be posted on the conference website as it becomes available (ronchi.isti.cnr.it).
The members of the IR community, the workers in this field, the researchers publishing in this field, and the editors who had the privilege of assembling this collection of papers would like to express our gratitude to Dr. Ronald Driggers, the Editor-in-Chief of Applied Optics for inviting us to compile the papers featured in this special feature, Advanced Infrared Technology and Applications 2017. An attempt was made to involve two additional editors from other continents in order to broaden the appeal of the collection and to increase the author participation. Unfortunately, both withdrew due to heavy workloads arising from previous commitments.
Furthermore, we wish to acknowledge the invaluable assistance of Ms. Nicole Williams-Jones, Senior Journal Coordinator for Applied Optics, who did a wonderful job assisting along the way with keeping peer review on track, sending papers out for review, and reminding the editors of pending tasks. We want to express our appreciation to Mr. Dan McDonold, OSA Editorial Director, for providing guidance in preparing the call for papers and for explaining the established editorial procedures at Applied Optics.
Additionally, we are grateful to all the reviewers who took time from their busy schedules to provide constructive feedback to the authors. We appreciate the efforts that the authors have made in preparing the manuscripts and responding to the suggestions of the editors and reviewers in making their work more understandable to the readership of Applied Optics. Likewise, we thank Ms. Michelle Scholl who read the introduction several times, providing critical feedback in clarity of expression and language usage. Finally, we are grateful to Ms. Cristina Kapler of the OSA editorial staff, whose mastery of language is unsurpassed, for finely tuning the final subtle improvements to the introduction! The guest editors who often recommend that the authors ask a native English speaker to look over their manuscript one more time are the first to follow their own advice!