This PDF file contains the front matter associated with SPIE Proceedings Volume 13200, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

With the exponential market increase in mini- and micro-drones, this technology became available for anybody around the world. Therefore, the possibility of using drones as weapons represents a danger in both the civil and military domains. In the last years, a lot of programs have been initiated to identify cost-effective measures for detection, classification, tracking and neutralization of this type of threat. In a prospective approach to mitigate these risks, the French-German Research Institute of Saint-Louis (ISL) is studying since the past ten years the technical responses that can be provided for the detection, identification, localization and tracking of small-sized aerial drones, in a variety of operational scenarios and contexts of use. In this paper, we will summary the main ISL results in this domain. Over the time, ISL participated in a lot of joint trials allowing to test acoustical and optical detection system in different operational scenarios and for a great variety of types of drones. The collected images form a database of almost 1 million of images of drones in flight at different scales, with different backgrounds and at different wavelengths (Vis, SWIR). This database is of great importance for the development of detection/localization/tracking algorithms for conventional or AI-based techniques.

Passive electro-optical systems, such as visible and infrared light cameras, and active systems, such as ladar or LiDAR, can acquire detailed two- and three-dimensional images of a scene. This paper presents a sensor fusion framework that combines passive and active electro-optical sensor data to reveal subtle patterns. It creates fused data structures that merge 3D coordinates and light intensities and customizes recent methods for classification of high-dimensional irregular data to segment the fused structures into different object classes. The framework can also be used to discriminate weak laser return pulses from noise by extracting point clusters in the higher-dimensional data under certain constraints. Methods for estimating and compensating for motion during acquisition of the data can be integrated in the framework to prevent misalignments. Experiments on IR images and 3D point clouds acquired by a ladar demonstrate scene segmentation, object recognition and motion estimation in various challenging conditions.

Airborne optoelectronic imaging systems often need long exposure times to capture high Signal-to-Noise Ratios (SNR) images. However, during flight, the movement, attitude change, and vibration of the system can cause image motion, which limits the exposure time. This creates a conflict between the exposure time and SNR. Conventional methods use the superposition of multiple frames with consecutive short exposure times to achieve high-quality images with short exposure times. However, sudden external perturbations can cause significant differences in the quality of images taken with different exposure times. This reduces the effectiveness of multi-frame superposition. In this paper, we propose a new method for improving the imaging quality of a photoelectric imaging system in complex environments. The method is called the multi-frame superposition method with adaptive weight assignment based on disturbance observer (DOB). It uses the disturbance estimation residuals of the DOB controller in the inertial stabilization loop of the airborne optoelectronic platform as a weight reference to assign weights to the continuous short-exposure images. This method effectively reduces the external disturbances in the multi-frame superposition process, resulting in improved performance and robustness of the system. The method establishes a mathematical link between the transient state of the control system and the imaging quality. Through simulation, we have proven that this approach can effectively improve the imaging SNR of the optoelectronic imaging system when compared with traditional methods.

Aerosol extinction is critical to EO sensor performance. Some information about extinction may be obtained from a visibility meter or from observing distant parts of the scene. However, this may be difficult under conditions of darkness, slant paths, clouds, smoke etc. Therefore, we investigate how a laser (for example in the form of a laser range finder) can be used to estimate atmospheric extinction at the laser wavelength. Translating this information to other wavelength regions may be done through established atmospheric models. The use of a laser to evaluate atmospheric attenuation offers an attractive way to achieve performance of various laser and passive electro-optical sensors. A laser weapon is for example very depending on the atmospheric propagation. The paper discusses and exemplifies lidar methods to derive extinction from measurements against targets in the scene or from atmospheric backscatter.

The rapid angular sweep of mirrors in a micro-mirror array (digital micromirror device - DMD) creates a dynamic diffraction grating and when coupled with a camera produces a form of streak camera, which is able to give temporal dispersion as well as spectral dispersion. Three pulsed lasers of different wavelengths (red, green and blue) were combined into a single fibre and then focused onto a target. Multiple collection heads were used to collect the reflected light and deliver it via separate fibres within an array to the DMD streak camera. Each fibre is imaged through the streak camera such that the temporal and spectral characteristics from each fibre channel are separate and distinguishable on the camera image. The lasers are fired via a delay triggered by the start of the micro mirror’s motion and the scale of the delay determines the distribution of intensity on the camera. Thus, the system is behaving as a multispectral and multichannel LIDAR, with temporal resolution at the level of 1nS, which is somewhat surprising considering that the DMD frame rate is limited to 10kHz.

This study introduces a revolutionary LiDAR sensor incorporating Optical Orthogonal Frequency-Division Multiple Access (OOFDMA) and a Risley prism for variable Field-of-View (FoV) functionality. This innovation provides a platform for adaptable scanning patterns and point cloud densities catering to diverse spatial scales and object distributions. To bolster the security of the sensor, the sensor integrates Advanced Encryption Standards (AES) and wavelength selection methods. In light of its potential to strengthen reliability and precision in aerial applications, this sensor holds promise for advancing Urban Air Mobility (UAM) technology. This sensor comprises three laser diodes with distinct wavelengths, which achieved exceptional accuracy, marking a significant leap forward in LiDAR technology.

Urban Air Mobility (UAM) presents a promising solution to urban traffic congestion, offering fast and efficient travel within cities. However, safety concerns persist, particularly regarding adverse weather conditions like Clear Air Turbulence (CAT), which can pose risks to UAM flights. Current weather detection instruments have limitations, especially in detecting water-free CAT. This study proposes using LiDAR technology with Optical Orthogonal Frequency Division Multiple Access (OOFDMA) and a Risley prism to detect CATs by analyzing the movement of urban dust affected by CAT-generated airflow. The LiDAR’s ability to rapidly rotate, send multiple laser wavelengths and accurately measure distances ensures precise detection of obstacles, including CATs. Computational Fluid Dynamics (CFD) modeling validates the LiDAR’s efficacy in distinguishing between CAT and non-CAT scenarios. By analyzing reflected waves from dust movements, LiDAR reliably identifies CATs, providing a robust solution for UAM safety in urban environments.

Results from the development of substrate illuminated planar Ge on Si Single Photon Avalanche Diodes (SPAD) imaging arrays will be presented operating at short wave infrared wavelengths. Simulations have been used to optimize the designs aiming to reduce dark count rates and increase the number of absorbed photons aiming for Pelter cooler operation whilst also minimizing cross talk. To date the highest performance of Ge on Si SPADs has been demonstrated at 125 K with 38% single photon detection efficiencies and a noise equivalent power of 8e-17 W/√Hz. Surface illuminated devices have demonstrated single photon detection efficiencies up to 38% for 1 μm thick Ge absorbers and the present work will present results from 2 μm and 3 μm thick Ge absorbers aiming to increase the absorption of incident photons. The paper will describe the compromises between absorbing more photons compared to dark count rates and jitter. Examples of single photon LiDAR applications at 1310 to 1550 nm will be presented and the performance from Ge on Si SPADs will be compared to InGaAs SPAD technology in terms of single photon detection efficiency, dark count rates, afterpulsing, jitter and operating temperatures. Afterpulsing measurements demonstrate significant reductions compared to InGaAs SPADs operated under nominally identical conditions by a factor of 5 to 10. The performance of the surface illuminated SPADs in linear mode as avalanche photodetectors will also be presented. Operation at 1550 nm wavelengths at room temperature has demonstrated responsivities at unity gain of 0.41 A/W, maximum avalanche gain of 101 and an excess noise factor of 3.1 at a gain of 20 for 50 μm diameter photodetectors.

This paper describes a comprehensive computational imaging field trial conducted in Meppen, Germany, aimed at assessing the performance of cutting-edge computational imaging systems (compressive hyper-spectral, visible/shortwave infrared single-pixel, wide-area infrared, neuromorphic, high-speed, photon counting cameras, and many more) by the members of NATO SET-RTG-310. The trial encompassed a diverse set of targets, including dismounts equipped with various two-handheld objects and adorned with a range of camouflage patterns, as well as fixed and rotary-wing Unmanned Aerial System (UAS) targets. These targets covered the entire spectrum of spatial, temporal, and spectral signatures, forming a comprehensive trade space for performance evaluation of each system. The trial, which serves as the foundation for subsequent data analysis, encompassed a multitude of scenarios designed to challenge the limits of computational imaging technologies. The diverse set of targets, each with its unique set of challenges, allows for the examination of system performance across various environmental and operational conditions.

In many type-II superlattice infrared detector architectures, performance at low temperatures is limited due to the dependence on minority hole conduction for operation, with holes having a high effective mass in the vertical direction. This inherently results in a decreased (especially as temperature is reduced) carrier diffusion length in the absorber, which can limit the detector quantum efficiency. The alternative pBpn architecture utilizes minority electrons for detection which have a higher mobility and enhanced collection. A general limitation of the pBpn design is that surface currents are often found to dominate the dark current density. This paper explores the effect of varying the absorber region pn junction growth parameters in an attempt to influence the surface current magnitude. An analysis of the surface vs. bulk contributions of the dark current is made as a function of absorber design, and potential sources of the surface current are presented. Ultimately, it is determined that the surface current magnitude is independent of the bulk absorber properties varied, implying that the surface properties, especially of the p-type absorber, must be altered to effectively mitigate the surface current.

We present a theoretical and computational model for simulating Single-Photon Avalanche Diode (SPAD) array images in the context of time-of-flight Lidar. We confirm the accuracy of out model by experimental verification. Further, we demonstrate that our simulated data can be used to train bespoke neural networks facilitating a number of enhanced imaging capabilities including, super-resolution, human-activity recognition, and long range SPAD Lidar.

Metasurface analogues of Electromagnetically Induced Transparency (EIT) can achieve high quality factor (Q factor) resonance transmission peaks, making them suitable for sensing applications including refractive index sensing and gas detection. We propose a displacement-driven symmetry-breaking EIT metasurface operating in the mid-wave infrared range, which consists of a tri-atom germanium nano brick array. By geometrically introducing a displacement to the central atom, an EIT transmission window can be generated. Ten EIT metasurfaces were designed with peak transmission wavelengths regularly distributed between 3.3 and 3.4 μm at intervals of 20 nm so as to target an important methane infrared absorption resonance. The transmission characteristics of the metasurfaces were studied using Rigorous Coupled-Wave Analysis (RCWA) simulations. The origin and contribution of the toroidal dipole to the EIT effect was explained through Finite-Difference Time-Domain (FDTD) simulation. Successful fabrication of precisely defined Ge nano bricks using electron beam lithography and reactive ion etching was achieved. Q-factors greater than 100 were experimentally obtained with resonant peak full-width-at-half-maxima of less than 30 nm.

Portable Thermal Infrared (TI) imaging is widely used by personnel negotiating complex environments in zero-light, but the field of view provided by conventional TI is akin to tunnel vision. A spherical or hemi-spherical field of view would provide an immersive situational awareness, but the scaling laws of optical aberrations mean that fast, f /1 optics, high space-bandwidth product, and wide field of view cannot be achieved simultaneously. We describe a scaleable multicamera architecture that leverages the availability of low-cost consumer-oriented miniature Longwave-Infrared (LWIR) cameras to yield imaging systems with arbitrarily large Fields of View (FOVs), high resolution and high radiometric sensitivity. Our concept transfers the burden of complexity from high-cost, high-performance optics to low-cost computation and cameras to enable high-performance imaging from person-portable systems. We describe a multi-camera LWIR system, which records 360◦ panoramic video with 1678 × 256 pixels. This system captures a 2π steradian field of view and can be readily adapted to full 4π spherical imaging. Furthermore, we employ integral imaging, and the changing parallax provided by perambulation to enable enhanced 3D imaging of people in highly cluttered environments.

Single-photon detectors record photon-detection events with high efficiency and picosecond-timing resolution allowing for lidar (ranging) with millimetre depth resolution and low light levels. However, only silicon SPADs offer high quantum efficiency without cryogenic cooling. InGaAs SPADs are limited to below a wavelength of 1800 nm with low efficiencies. We have developed an approach for high single-photon detection efficiencies in the mid-infrared (2-3.5um) using upconversion in a PPLN waveguide in combination with a silicon SPAD. Using a waveguide we confine the mode over extended lengths to increase conversion efficiency and reduce the required pump power. Using a precision-machining waveguide fabrication process we have fabricated waveguides that support modes in the mid-infrared with low loss. We show overall photon detection efficiency in excess of InGaAs SPADs. We use our new detection system for imaging and ranging applications.

The measurement of radiometric signatures in the Long Wave Infra-Red (LWIR) 8μm to 14μm, is an essential endeavour in many different contexts including: the design of threat warning systems; the measurement of own platform emission and stand-off temperature measurement of individuals for healthcare and bio-security. Recent advancement in uncooled micro-bolometer LWIR imagers afford a potential alternative lower cost approach to the measurement of LWIR radiometric signatures. Micro-bolometers have been predominantly designed for imaging applications, with considerable development devoted to the improvement of imaging quality. Good imaging performance relies on a response function which is linear with irradiance observed at the imaging aperture plane. It is therefore possible to determine the aperture plane irradiance using a-priori knowledge of this linear response function. During the development of a system designed to measure the temperature of individual people using micro-bolometer imagers, Thales UK observed that the response function of these devices was indeed linear. The parameters of this response function however, drifted considerably with time. Therefore to understand the response function at any given time, two data points of known irradiance are required in the scene. The conventional methodology for the measurement of radiometric signatures in the LWIR relies on the use of cooled imaging radiometer devices which can be very expensive with devices typically costing on the order of several hundred thousand dollars for one device, with many devices also requiring a supply of liquid nitrogen to function. These limitations make these devices ill-suited for portable low cost devices for radiometric measurements in the LWIR. This paper will present the approaches taken and results observed by Thales to effectively use micro-bolometer imagers to measure radiometric quantities for close-in (less than 5m) objects as well as outlining what further considerations need to be made for the radiometric signature of distant (hundreds of metres) objects at known range.

Emerging techniques for imaging around corners and through scattering media hold promise for visualizing hidden scenes in challenging situations. Synthetic Wavelength Holography (SWH) is one of these techniques, which appears to offer key advantages compared to alternative approaches. However, a critical analysis of the approach is currently lacking. Here we present our findings as to the capabilities of a nanosecond pulsed, short wavelength infrared SWH system we have built and compare these with theory. We also discuss practical aspects of deploying this technology.

This paper deals with the problem of tracking a specific target moving through a complex urban environment using co-operative airborne imaging sensors. We demonstrate algorithms to maintain a line of sight to the target where possible and adapts to occlusions, and other sensor limitations, as the vehicle moves through a complex urban environment. This work exploits an advanced simulation capability, built around Unreal Engine 5, that can generate realistic video data which is used to test and to stress tracking algorithms whilst providing ground truth data for true targets and for collateral entities.

Microscanning is an optical technique used principally to enhance the resolution of cameras using 2-D arrays of detectors. It provides small (typically half pixel pitch) movements of the image across the detector array so that a high-resolution image can be built up by interlacing a number of fields. Microscanning may also be used to provide information to allow non-uniformity correction of the image. Microscanning is frequently used in thermal cameras, where detector resolutions are limited, and where uniformity correction is critical because of the low contrast of the image. In this paper we look at the consequences of microscanning by lateral motions of a lens. It is found that in general the motion of the image caused by motion of the lens varies with position in the field of view, and is different in the radial and tangential directions. This difference is what we refer to here as microscan distortion. This paper describes how to minimise or completely eliminate distortions in the microscan pattern. Eliminating this distortion is particularly useful where the microscan is used to support scene-based uniformity correction.

The use of lasers on the battlefield is on the rise due to the recent development of ready-to-use laser dazzlers and laser directed energy weapons. Optical camera systems are particularly vulnerable to damage from laser radiation as the incident light is typically focused onto the sensor. Although there are several methods to protect against specific wavelengths, only a few can provide protection across a broad range of wavelengths. This work describes a technique in which the sensor is positioned behind the focal plane, allowing the use of an optical limiting filter on the freed focal plane. The incident radiation is masked by a coded aperture, and the blurred image is reconstructed using computational imaging. Tests show a significant improvement in resistance to laser damage, as well as satisfactory image reconstruction.

Laser scanning is one of the essential techniques in Remote Sensing, while galvanometer scanners (GSs) are one of the most utilized devices for such purposes. We report some of our main findings regarding galvoscanning analysis and optimization. The trade-off between GS parameters was considered, including duty cycle, scan frequency, and scan amplitude. Our first study considered the most common scanning inputs: triangular, sawtooth, and sinusoidal. We demonstrated that the former produces the smallest distortions [http://dx.doi.org/10.1364/AO.50.005735]. A modeling of the scanning process was completed, with regard to the theoretical duty cycle, following the analysis of the effective duty cycle as a function of scan frequency and amplitude. The optimal mosaicking of retrieved images was pointed out [http://dx.doi.org/10.1364/AO.54.005495]. Finally, in order to increase the effective duty cycle of the GS, custom-made input signals have been analyzed. We demonstrated that linear plus polygonal functions are the best for maximizing the duty cycle [http://dx.doi.org/10.1117/1.3497570], in contrast to the literature, where linear plus sinusoidal functions were considered best.

The Austrian research project iLIDS4SAM – Integrated LiDAR Sensors for Safe and Smart Automated Mobility – addressed the requirements and challenges of LiDAR systems for autonomous driving. In the course of the project a LiDAR sensor demonstrator aiming at these requirements was designed and developed to the level of an elegant breadboard. Integrating highly innovative technologies, this LiDAR sensor is the very first realization of a newly designed scanner concept and is, thus, in many aspects a proof of concept, but also a technology carrier for future LiDAR sensor developments. Benefitting from the high angular point density as well as the high frame rate offered by the new concept, it is not only attractive for automotive LiDAR systems but also for industrial LiDAR applications which require the acquisition of a high amount of range measurements within a narrow field of view in a very short time. The basic design of the LiDAR sensor as well as the intended requirements and specifications were already presented during Sensors + Imaging conference in September 2023. This contribution gives insight into the final phase of the research project, discusses the challenges arising during the realization of the LiDAR sensor and presents the achieved results. Based on point clouds acquired with the LiDAR sensor and artifacts identified therein, shortcomings of the scanner concept as well as the integrated components are discussed, and improvement measures suggested.

Hyperspectral imaging in short wave infrared wavelengths is typically limited to highly illuminated conditions, which is a limitation for security and defense applications. A hyperspectral imager capable of operating in dark conditions would have many uses cases in, for example drone imaging and imaging in front of vehicles for target recognition. In this work, we demonstrate an active shortwave infrared pushbroom imager for operation regardless of ambient lighting conditions. The imager is capable of capturing 1000 hyperspectral data lines per second by using a combination of pulsed broadly tuneable supercontinuum light source and 1D monochrome camera sensor. By using short pulses of 2.6ns, short sensor exposure times can be used and thus eliminating the effect of ambient light. We demonstrate imaging from 6 m distance. Further development and potential are discussed.

Triangle Orientation Discrimination (TOD) developed by TNO Human Factors and Minimum Temperature Difference Perceived (MTDP) developed by Fraunhofer IOSB are competitive measurement methods for the assessment of well and under sampled thermal imagers. Key differences between the two methods are the different targets, bars for MTDP and equilateral triangles for TOD, and the measurement methodology. MTDP bases on a threshold measurement whereas TOD uses a psychophysical approach. A former study compared MTDP and TOD for standard thermal imagers. It found that it is possible to transfer TOD to MTDP and vice versa using a sigmoidal transition function. If this is also valid when applying advanced signal processing remained open. To examine this, the study was now extended to boost filtered thermal imagers. Equipment in test was an under-sampled MWIR imager operating without and with five different boost filters, four different Laplace- and one Wiener-filter. Comparative MTDP and TOD measurements for these six configurations showed the validity of the derived sigmoidal transition function also for this type of signal processing. Conclusive, to date the comparison of MTDP and TOD gave no preference for one of the methods. They should result in comparable performance predictions, although it is generally not possible to rule out that there is advanced signal processing where one or both of the methods may fail.

Multispectral IR imaging sensors provide an additional dimension of data compared to conventional devices. Testing and validating these detectors require a spectrally controlled and reproducible test scene. IR scene simulators aim to generate a realistic IR scene using artificial IR emitters and optical components. To ensure both spatial and spectral accuracy of the scene, it is crucial to have precise knowledge of the spectral characteristics of these components. This work focuses on the investigation of low-temperature thermal emitter spectra. We utilize a grating spectrometer in combination with quantum IR detectors to determine the emission spectra of thermal IR sources. The entire system is calibrated to accurately record the spectra of arbitrary thermal IR emitters. In summary, this study provides the basis for a physically accurate IR scene simulation setup, with particular emphasis on spectral realism.

A mathematical model has been developed to assess the performance of different multi-sensor systems for maritime targeting applications. This model simulates the performance of two imaging cameras (visible band and infrared) together with an RF sensor. Time-dependent sensor data streams are generated, and salient target detection information is extracted using a data fusion architecture, the output of which gives closed-loop guidance commands for use within an engagement trajectory model. Gaining confidence in the model was required at two levels. Firstly, there was the need to verify that the model was correctly implemented and, secondly, the validation of the model predictions against the anticipated behaviour of a real system. To achieve the required level of confidence, a Test, Verification, and Validation (TV&V) framework was developed based around sensitivity analysis. The sensitivity analysis used single and multiple parameter variations in either defined steps or else with random values within a Monte-Carlo engagement. Multiple parameter variation was found to be particularly effective in determining issues within both the implemented model and the efficacy of the proposed system design. To visualise large volumes of data, various methods were examined, with the presentation in a time-varying polar plot form being considered the most effective.

Targets located on the Precision Impact Range Area (PIRA) of Edwards AFB are used to evaluate imaging systems’ sensitivity and spatial resolution to ensure they meet specified requirements. Spectral Sciences, Inc., is developing a field-ready electro-optical sensor calibration/test system for airborne instruments from the visible through longwave infrared. This spectral region is particularly challenging because of the contributions from both solar and thermal fluxes. The system is composed of spectral-spatial ground targets and atmospheric characterization instruments. The design challenges for a new ground target installation applicable over short to long ranges and a broad optical spectrum include: 1) development of an innovative spectral-spatial, high contrast, high uniformity, knife edge target for determination of the spatial characteristics of the imaging system under test, such as the Modulation Transfer Function (MTF) and Relative Edge Response (RER), Noise Equivalent Temperature Difference (NETD), linearity and more; 2) development and implementation of a suite of auxiliary instruments to quantify the atmospheric effects, such as line-of-sight (LOS) turbulence, surface temperatures, humidity, and visibility; 3) development of targets with stable, quantifiable spectral response that can be used for evaluation for the spectral characteristics of multi- or hyperspectral imaging systems; and 4) engineering the target set for simplified long-term maintenance and durability. In this paper we report on the development of a prototype 2m by 2m thermally controlled knife edge target. The target is composed of four 1m by 1m panels each of which has independent temperature control and face surface materials which can be exchanged with other panel faces to produce patterns or spectral features. The full prototype system can be rotated and tipped to maximize the surface area apparent to a sensor system under test. The paper includes initial field measurements of the target array using visible, MWIR and LWIR imaging systems.

The aim of the range performance model TRM4 is the evaluation of imager performances considering the entire electro-optical imaging chain, including the resulting image’s observation on a display. Usually, it is assumed that the image is viewed in a mostly optimized setting, where factors such as suboptimal display contrast are not a limiting factor to the overall system. However, this cannot always be guaranteed in actual operation. For example, glare caused by light shining on the display within a cockpit in an undesirable way, or even directly into the observer’s eye, reduces the ability of pilots to perform observation tasks. Therefore, we are aiming to elaborate on non-ideal observation settings and include the influence of glare detrimental to the observer’s contrast perception in the overall range calculation. In order to model the effect of glare, we implemented a contrast threshold model based on the visual perception model of Peter G.J. Barten. We introduced a number of additional parameters further describing the display’s luminance properties to our model. Glare is modelled by a veiling luminance. The effects of glare on the imager performance are discussed based on numerical results for different scenarios. These scenarios include two imaging systems, one operating in the thermal infrared and one in the visible spectral range, with two different displays each: one regular desktop monitor and one cockpit multifunction display.

The aim of the paper is to review the main achievements in the research of HgCdTe ternary alloy and InAs/InAsSb type-II superlattice material to indicate the Polish contribution to the development of medium and long wavelength infrared photodetectors. Research and development efforts in the WAT-VIGO joint laboratory have focused on the metal-organic chemical vapor deposition (from 2003) and molecular beam epitaxy (from 2015). At present stage of development, the photoconductive and photovoltaic HgCdTe detectors are gradually replaced with novel device designs based on III-V material system. T2SL devices complement the offer of MCT ones in applications where it is necessary to ensure, among others: higher resistance to difficult operating conditions and high uniformity of parameters of multi-element detectors.

The Microelectronics Research Group (MRG) at The University of Western Australia is a key partner of the Australian Research Council Centre of Excellence for Transformative Meta-Optical Systems. In this presentation, an overview of ongoing research will be given with an emphasis on the flagship research activities of MCT-based imaging arrays and Microelectromechanical Systems (MEMS). The MCT research and development utilise a vertically integrated capability from semiconductor material growth, through device modelling and design, to focal-plane-array fabrication and packaging. In support of the detector array capability, fully integrated MEMS technology can be used to further enhance the sensor device performance through the focal plane integration of tunable filters for spectral classification and infrared spectroscopy. The combination of high-performance detector designs and tunable spectral filters provides a major differentiator for military imaging systems, particularly for those operating in complex and degraded environments. This talk will highlight several research activities that are highly relevant to defence applications including metamaterial enhanced infrared detectors, and the fabrication of infra-red focal plane arrays on flexible substrates. For the MEMS technology, both wideband and narrowband tunable spectral filters will be discussed for multispectral imaging in the SWIR, MWIR and LWIR bands, and for hyperspectral imaging and spectroscopy. Considerations on future research activities and technology trends will be presented including opportunities for the rapid development of high-performance and spectrally adaptive low SWaP sensing systems for enhanced detection and discrimination of partially concealed or camouflaged targets in cluttered backgrounds.

In the evolving landscape of imaging technology, the development of next-generation CMOS Time Delay and Integration (TDI) detectors represents a significant leap forward in high-resolution imaging capabilities. This paper presents an advanced CMOS TDI detector, tailored for applications demanding high resolution, sensitivity, and speed. By integrating novel semiconductor materials and leveraging cutting-edge CMOS fabrication techniques, our detector exhibits superior performance characteristics compared to conventional CCD TDI systems. We detail the innovative architecture of our CMOS TDI sensor, which includes an enhanced pixel design for improved charge collection efficiency and a specialized readout circuitry to minimize noise, thereby achieving higher dynamic range and image quality. The experimental results demonstrate the detector's exceptional ability to capture high-resolution images under low-light conditions, making it an ideal solution for a wide range of applications, including satellite imaging, medical diagnostics, and high-speed industrial inspection. Additionally, we explore the implications of this technology for future imaging systems, highlighting its potential to drive advancements in various scientific and commercial fields. Through rigorous analysis and testing, this paper underscores the pivotal role of next-generation CMOS TDI detectors in pushing the boundaries of what is achievable in high-resolution imaging.

Time-of-flight imaging techniques are utilized for 3D-scene reconstructions and non-line-of-sight imaging capable of revealing persons or objects hidden behind obstacles. Such sensor systems usually demand for high-performance photodetectors with single-photon detection capabilities due to the high reflection losses of the emitted laser pulses during multiple, usually diffuse reflections from a relay wall and the actual scene of interest. To further compensate the high signal loss, a laser wavelength in the eye-safe Short-Wave Infrared (SWIR) spectral range, typically around 1550 nm, allows for higher laser intensities. In turn, SWIR-matching Single-Photon Avalanche Diodes (SPADs) need to be fabricated from adequate semiconductor materials, e.g., the InGaAs/InP material system. We report on the progress of the InGaAs/InP-SPAD fabrication using a customized process technology. The key technology is the planar process technology via zinc diffusion to produce spatially confined p-type regions. For the zinc-diffusion process, a novel method of selective epitaxial overgrowth was developed. Following a prior recess-etch step, the single-step Zn-diffusion process results in the intended double-well-shaped doping profile. Experimental data of thus fabricated InGaAs/InP SPADs show the expected dark-current, photo-current, and multiplication-gain characteristics in linear-mode operating as well as dark count rates in Geiger-mode operation around 200 – 220 K, which corresponds to a typical operating temperature for InGaAs/InP SPADs achievable by thermoelectric cooling.

As demands for optical materials with higher sensitivities, lower cost, and reduced size intensify across the electromagnetic spectrum, novel solutions beyond monolithic materials must be considered. With this motivation, we combine infrared colloidal nanocrystals (NCs) of different composition and diameter into binary superlattices, a type of ‘artificial solid’ where the NCs are analogous to ‘atoms.’ These binary superlattices provide the ability to tailor desired optoelectronic properties through choice of the constituent NCs, which are scalable following wet-synthetic routes and made from abundant materials, allowing for distributed infrared optoelectronics. We use so-called ‘t/lambda’-techniques and automated-determination of spatial extent of the superlattice to define the local volume in individual superlattice domains, and then compare these to photoluminescent yield from the same domains in a correlative microscopy. These results strongly suggest energy transfer between the excitonic emitters and plasmonic nanocrystals and allows for various tuning of geometry and energetics of the binary system for desired overall composite material properties.

The simultaneous detection of infrared signatures in separate spectral bands or sub-bands provides an advantage in many applications such as missile warning sensors. It allows an improved target identification and discrimination in front of structured natural or artificial background scenarios. AIM has been producing MCT-based dual color infrared SW/MW and MW/MW detectors grown by MBE on GaAs substrate for several years. These detectors allow a temporal and spatial coincident detection in two bands or sub-bands in the infrared spectral range. Since available space, allowed weight and power consumption are limited in many applications, an increase of the detector operation temperature is one way to cope with these constraints. It allows the usage of smaller and lighter coolers with less power consumption. Another way to enhance detector performance is the reduction of the pixel pitch. This allows for an increased spatial resolution or smaller detectors to improve the situational awareness and/or reduces the overall size and power consumption of the device. In this paper, we present investigations on AIM’s 320 x 256 pixels and 30 μm pitch infrared bispectral SW/MW and MW/MW detectors regarding elevated operating temperatures. The main limitation to operate this kind of infrared detectors at higher temperatures is the increasing dark current, resulting in increased temporal noise. We present measurements of the dark current and noise-related figures of merit like NETD. We investigate parameters like cooler power consumption to show the positive effects of higher operating temperatures regarding SWaP. Finally, we give an overview of the current activities regarding pixel pitch reduction and higher operation temperature of bispectral detectors at AIM.

Leonardo UK has developed a medium wave infrared detector module for use in a Directed Infra-Red Counter Measures (DIRCM) application. The sensor is based on Osprey S detector consisting of a 288 by 384 array of 20μm pitch, Mercury Cadmium Telluride (MCT) pixels. The MCT array is manufactured using Metal Organic Vapour Phase Epitaxy (MOVPE) grown on low-cost gallium arsenide (GaAs) substrates giving an affordable method of volume production. MOVPE enables precise band gap engineering of the MCT heterostructure to produce a focal plane array resilient to high flux events. The MCT operates at 110K allowing faster cooldown times and reduced power consumption. The MCT is hybridised to a custom Readout Integrated Circuit (ROIC). The sensor module incorporates bespoke interface electronics and a long life linear split cycle cooler. The resultant module is robust and has been qualified for a range of airborne applications. Whilst the sensors have been in serial production for over a decade, MOVPE manufacture of MCT arrays enable engineering improvements to be made at low risk to improve yield and Focal Plane Array (FPA) performance. This paper reports the improvement in the performance of the sensor through the engineering phase into production. The pixel operability has improved and is typically better than 99.9%. The module forms part of a DIRCM system. The system has been fully qualified with the sensor. Improvements in system performance over the previous generation of DIRCM sensors has been demonstrated.

The Type II superlattice (T2SL) technology is an industrial solution for manufacturing of infrared (IR) imaging sensors that cover the entire IR range, including the vital eSWIR band. These stable III/V materials are uniformly grown on large GaSb wafers (up to 6") and have demonstrated excellent manufacturing yield with maintained high performance and space environment robustness. These unique properties make T2SL the most suitable technology to design and manufacture large area eSWIR FPAs which can meet the stringent requirements of future European civilian and military space programs. In this paper, the performance of IRnova’s VGA format (640 x 512 pixels, 15 μm pitch) eSWIR T2SL detectors designed for industrial applications will be presented along with the development route towards large footprint (2048 x 512 pixels, 15 μm pitch) HD eSWIR FPAs with CTIA ROICs targeting space applications within the EU-funded STEP project (Silicon and T2SL EuroPean collaboration for a non-dependent supply chain for large eSWIR FPAs). From initial results, a cut-off wavelength of 2.5 μm, quantum efficiency > 80% and a dark current density lower than 0.1μA/cm2 @ 200 K have been demonstrated.

While agile multispectral imaging solutions presently exist, their size, weight and power (SWaP) specifications prevents deployment on small portable platforms such as drones. As much of the size and weight of existing solutions is attributed to the wavelength-selective optical subsystem, realizing low-SWaP hinges on miniaturization of this subsystem. The ultimate multispectral imaging implementation would integrate the wavelength-selective component at the imaging focal plane array. This paper presents a solution which aims to achieve such integration. Recent developments in Microelectromechanical Systems (MEMS) have realized a surface-micromachined optical tunable filter, operating in the shortwave infrared wavelength band (SWIR: 1 μm – 2.6 μm) for applications in miniature optical spectrometers. The tunable filter is a Fabry-Perot (FP) structure, composed of a fixed dielectric mirror on a silicon substrate, and a movable dielectric mirror suspended above. The separation (air gap) between these two mirrors defines the optical transmission centre-wavelength of this Fabry Perot structure. Consequently, electrostatic actuation of the top mirror towards the bottom mirror allows the gap, and thus the transmission centre-wavelength, to be controlled. This paper presents work towards integration of such a MEMS tunable filter technology directly on an infrared focal plan array. Realizing this integration relies on: (1) expanding the optical area of the MEMS Fabry Perot structure to cover a significant portion of the two-dimensional focal plan array, which is generally multi-millimetre in each of its two dimensions; and (2) devising a structure that will allow actuation of the MEMS filter with under 20 V.

Monolithic optical systems have the potential to provide compact solutions for traditional objectives. High machine and measuring accuracy, as well as the combination of several data sets, could allow complex optical components to be produced in series with low tolerances. Polyhedron structures with more than two arbitrarily curved optical and mechanical surfaces are manufactured with a high degree of precision, which could reduce the need for active, adjustment on system level. Findings and manufacturing results will be shown from an efficient series production.

Stray light, the unwanted light within an imaging system, can severely degrade the image quality of LWIR (Long-Wave Infrared) cameras, especially when it originates from external sources such as the sun or thermal emissions from internal components. While the Point Source Transmittance (PST) index measures an optical system's ability to mitigate stray light, it does not fully capture the extent of image degradation that results. In our study, we quantified image quality degradation by converting the distribution of stray light at the detector into digital grayscale levels. Based on this analysis, we conducted an optimal design that comprehensively accounts for both external and internal scattered light.

The design of continuous zoom systems has several challenges. One that is not often though about by the designer is that of setting the system once it has been assembled. The selection of the number and position of “setting points” (discrete positions in the zoom range) at which to perform the setting is one that has sometimes been left to educated guesswork. This paper will walk through some steps that help in assessing the viability of setting laws as well as proposing a way to optimise the number and position of setting points thereby minimising the cost of setting the lens.

The increasing demand for clearer dynamic images has elevated the requirements for adaptive optics. But deformable mirrors with mismatched disturbances limit the performance of optical systems. To address the needs of future optical systems, we propose a comprehensive control approach to improve the control performance of a deformable mirror, extend its lifespan, and enhance its disturbance rejection capabilities.

Advanced missile seeker technologies and missile propulsion systems with reduced electrooptical signatures constitute a serious threat to military platforms. Seekers using multiple spectral bands may be hard to jam using current countermeasure systems. Imaging seekers may have adaptive tracking algorithms to suppress the effect of current countermeasure strategies. Furthermore, countermeasures can only be used if a threat has been declared. Low-signature propellants in combination with the missile seen in head-on angle, where a bigger portion of the missile plume is shielded by the missile body, will make it extremely hard to detect the missile with current warning sensor systems. We report on the objectives and the status of the EDA CAT B project DEBELA (Detect Before Launch), which tries to address this threat, and which looked into potential technologies for future self-protection systems. The project focusses on within visual range threats and electrooptical sensors only. Candidate-technologies have been identified and tested in a field experiment on the premises of the Bundeswehr Technical Center (WTD 52) in Oberjettenberg, Germany.

The detection of a threat before launch offers the best options to avoid the threat. However, detecting a threat before it concretizes remains a difficult task because of the obvious risk of error. The detection of a laser emission or of optics that are pointed towards the platform gives a significant cue of a hostile behavior. In our work, we investigate if and how event (also called neuromorphic) cameras can be used to detect laser emission and combined with a laser source to detect retro reflections caused by pointed optics. We explore how the high temporal resolution of the event stream can be used to extract temporal information on the laser signal. The approach is demonstrated for a laser warning receiver use-case and a retro-reflection detection configuration using data that was collected during the DeBeLa trial in Germany and that reproduce operationally relevant scenarios.

A new approach for distinguishing neutral (e.g. walking) from threatening (e.g. aiming a handgun) poses, without training, is presented. There are various AI-based models that can classify human poses, but these oftentimes do not generalize to defence scenarios. The lack of data with threatening poses makes it hard to train new models. Our approach circumvents re-training and is a zero-shot, rule-based classification method for threatening poses. We combine a pretrained body part keypoint detection model with the neuro-symbolic framework Scallop. We compare the pretrained models MMPose and YOLOv8x-pose for keypoint detection. We use images from the YouTube Gun Detection Dataset containing persons holding a weapon and label them manually as having a ‘neutral’ or ‘aiming’ pose; the latter was further subdivided into ‘aiming a handgun’ and ‘aiming a rifle’. Scallop is used to define logic-based rules for classification, using the keypoints as input: e.g. the rule ‘aiming a handgun’ includes 'hands at shoulder height’ and 'hands far away from the body'. Recall and precision results for aiming are 0.75/0.81 and 0.83/0.73, for MMPose and YOLOv8x-pose, respectively. Average recall and average precision for ‘aiming a handgun’ and ‘aiming a rifle’ are 0.78/0.36 and 0.76/0.43, for MMPose and YOLOv8x-pose, respectively. Combining neuro-symbolic AI with pretrained pose estimation techniques shows promising results for detecting threatening human poses. Performance of neutral-versus-aiming classification is similar for both approaches, however, MMPose performs better for multi-class classification. In future research, we will focus on improving rules, identifying more poses, and using videos to obtain sequences of poses or activities.

This paper focuses on the specific case of anomaly detection from hyperspectral ground-based images in the context of hidden targets in the forest. Compared to nadir viewing, several limitations are identified making it difficult to use conventional anomaly detection methods. Since ground-based image simultaneously contains sky and background, very high contrasts are observed on the transition pixels between the sky and the treetops. Mixed pixels in transition have a very different spectrum from the sky spectrum and the vegetation spectrum. Consequently, detection methods based on distance from a model or local pixel density will interpret mixed pixels as anomalies. To overcome these limitations, this paper proposes a simple processing, to complement standard methods such as distance-to-model methods, for identifying mixed pixels and assigning them low detection score. The approach consists in isolating potentially mixed pixels and calculating a distance between these pixels and the closest point on the axis formed by neighboring classes. The benefit of the proposed method (named ADMC) is illustrated using hyperspectral data acquired during the DEBELA European project field campaign, where the anomalies are military targets concealed into the vegetation. In the illustrative image, the standard detection score for treetop pixels is higher than that of the targets. The proposed method drastically reduces the score of these mixed pixels, dividing the rate of false positives before detection by ten.

Abandoned/removed object detection is a critical task in video surveillance systems for ensuring public safety and security. In these type of systems, mostly static cameras are utilized to monitorize and observe the surrounding, hence background modeling based techniques are suitable for detection of objects that produce obvious changes in the image content. GMM(Gaussian Mixture Model) is one of the most endeavoured modeling technique for real-time surveillance applications. In this paper, we propose an edge-based approach for the detection of abandoned or removed objects under static background assumption. Traditional edge-based approaches rely on the amount of edge energy that suffers in cluttered areas. In order to solve this problem, we extend edge energy by use of edge orientation between current and background frame along foreground object's edge contours. This approach increases the robustness of abandoned/removed classification which is supported by detailed experiments.

This article presents a novel method for point target detection in remote sensing imagery, focusing on the development of an innovative approach to enhance detection accuracy and reduce false positives. The core contribution of this research lies in the refinement of traditional point target detection algorithms by introducing a targeted strategy to exclude edge artifacts and seamlessly integrating the methodology into a segmented matched filter framework. In this research we will expand the matched filter standard deviation filter (1) to the segmented matched filter.

This study introduces object detection within Electro-Optical (EO) systems utilizing Snapshot Compressive Imaging (SCI). Traditional EO systems often suffer from extensive processing time and high computational demands due to the sequential processes of capture, compression, reconstruction, and detection. To address these challenges, our research leverages SCI integrated with Artificial Intelligence (AI) to streamline the detection process directly from compressed optical measurements, thereby omitting conventional intermediate steps. This methodology significantly reduces time, storage, and computational overhead, while simultaneously enhancing accuracy by exploiting motion-encoded information inherent in the compressed data. By satisfying stringent Size, Weight, and Power (SWaP) requirements, our method holds promise for a variety of applications including autonomous driving, environmental monitoring, and public security. This paper represents not only a significant advancement in computational imaging, proposing a cost-effective, but also optimized solution suitable for a spectrum of EO systems.

This technical review presents a novel algorithm for target concentration estimation in hyperspectral imaging, comparing its performance with an existing method. Using the substitutive model of the matched filter for target detection, evaluation of both algorithms was based on target detection accuracy and false positive rates. Our findings reveal that while the new algorithm offers more accurate mean estimation of target concentration, the existing algorithm exhibits lower variance and superior detection capabilities. These insights highlight the trade-offs between mean accuracy, variance, and detection efficacy in hyperspectral target detection algorithms, advancing our understanding of their performance in practical applications.

Compressive Sensing (CS) imaging is a computational imaging technique that has the potential to reconstruct a high-resolution image from signals acquired by low-resolution detectors with one or a few elements. This paper presents a mobile dual-band CS system developed at Fraunhofer IOSB. The system has the capability of simultaneous image acquisition in two spectral bands between 0.35 μm and 2.9 μm. Furthermore, CS measurements in the visible (VIS) and Short-Wave Infrared (SWIR) spectral range at distances of up to 27 km are demonstrated.

Images from passive imaging polarimeters are often displayed in terms of their intensity (s₀), degree of linear polarization (DoLP), and angle of polarization (AOP). The AOP and DoLP together generally provide information about shape and surface quality of imaged objects; however, the information conveyed in these images are often difficult to interpret, especially to an analyst unfamiliar with polarization phenomenology. Furthermore, it can be challenging to train machine learning approaches with direct polarimetric information due to the limited availability of application-specific data. Augmenting s₀ images with polarimetric information in the form of contrast and shape enhancement thus allows for the data to remain in the domain of traditional EO-IR intensity imagery that is more familiar for both humans and trained networks. In this work we explore several strategies for effectively fusing polarimetric information into s₀ imagery to improve target shape and contrast for both human and algorithmic consumption, with a particular focus on remote sensing and target detection applications.

As we enter this new space age where the barrier to access space has never been lower, the technologies that enable various space-based missions are being reevaluated in light of evolving requirements and constraints. For example, event-based sensors show great promise for executing tracking functions with higher timing resolution and reduced power consumption and datalink demands, a great benefit for larger sensor network architectures that may be enabled by recent reductions in launch costs. Currently, the vast majority of event-based sensors on the market are designed to operate for visible wavelength applications using silicon-based photodetectors, however, operation in the infrared is essential for many space-based sensing applications. Evaluation of how the event-based read-out integrated circuit will interact with smaller bandgap photodetectors and how typical infrared photogenerated signal levels will propagate through the event-based sensor pixel unit cell will be necessary to extend the utility of event-based sensing into the mid- to long-wavelength infrared. To evaluate the functionality of the event-based sensor pixel unit cell, the circuit is implemented on a custom-designed printed circuit board using discrete devices selected to tailor the functionality to operate a mid-wave infrared photodetector. The measurements conducted provide understanding of merits such as photoresponse, latency, and general operation of the unit cell alongside possible limitations of the unit cell.

Remote sensing in the thermal infrared (TIR) band (3-14 μm) is an increasingly attractive tool for environmental sciences and security due to its chemical specificity and the large spectral superposition with the atmospheric transparency window. Currently, the technique is largely dominated by dispersive-type hyperspectral imagers, which usually require expensive, cumbersome and cooled quantum detectors setups to compensate for their low optical throughput. Here, we present a compact and low-cost hyperspectral camera based on the Fourier-transform approach. Its main element is a common-path birefringent interferometer made of TIR-transparent crystal calomel (Hg₂Cl₂). We characterize it with both coherent and incoherent sources in the TIR, determining high optical throughput, adjustable spectral resolution up to 4.5 cm^-1, interferometric contrast higher than 90% even for incoherent radiation and a robust and long-term interferometric stability. By coupling the interferometer to an uncooled microbolometer detector with 640x480 pixels, we demonstrate hyperspectral imaging in the 8-14 μm spectral range for transmission and emission remote measurements.

We introduce a radiation thermometer using a mid-wave infrared detector for observing ground target emissivity at room temperature. Key to accurate temperature measurement lies in detecting low signals reliably, requiring internal thermal equilibrium and the use of lock-in detection to amplify signals. Multiple temperature controls and an optimally placed optical chopper enable this precision. This presentation will also cover some of our works regarding Mid-IR ground reference targets.

In the field of microbolometer technology, there are two concepts for tuning the absorption to the desired spectral range. Absorbers based on a λ/4 resonator, which is the standard design for commercial uncooled IR imagers, and Plasmonic Metamaterial Absorbers (PMAs). This paper provides an overview of the optical results of both Fraunhofer IMS’s concepts and a comparison between the two technological approaches in terms of their targeted spectral ranges and bandwidths. We show that the quarter-wavelength resonator-based microbolometers provide a broadband absorption spectrum and the PMA-based microbolometers can have a narrow-band absorption spectrum with a bandwidth of about 0.5 μm. In addition, we show how λ/4 absorbers are primarily suited to thermography, while the PMAs can provide a promising platform for gas cameras.

In the maritime domain, fog represents a significant challenge, e.g. for the detection of castaways in search and rescue missions or obstacle avoidance/warning systems, particularly at night. In scattering environments, conventional active imaging techniques often struggle due to backscattering resulting in low contrasts or completely saturated pixels from backscattering at fog particles (water droplets). The effective suppression of atmospheric backscatter in foggy environments by using Gated-Viewing technology has been demonstrated in previous studies. However, another alternative of blocking backscattered light is using polarization if the scattering event preserves polarization, which is the case for water droplets (maritime fog). In the present study, we modified an existing, unpolarized light emitting Gated-Viewing instrument and added the option to emit and receive linear polarized light through installation of polarization filters. All images taken with this modification are acquired in Continuous Exposure (CE) mode, i.e. without gating. The objective was to investigate whether polarization can serve as an alternative to Gated-Viewing in the maritime domain where backscatter is expected to largely preserve polarization. In initial field tests the efficiency of the modified (polarized) instrument in suppressing retroreflections under clear weather conditions was evaluated. Further tests with artificial fog showed that the modified instrument could effectively suppress backscatter, achieving a contrast loss comparable to that observed in clear weather conditions.

The Microelectronics Research Group (MRG) at UWA has been developing its capabilities in the field of infrared materials and devices since 1989 and is the only HgCdTe research centre in Australia. In this paper, we report on the various HgCdTe based technologies being researched at UWA to enhance their capabilities for demanding applications such as heteroepitaxy-enabled low-cost, large array size, and high-performance HgCdTe IR FPAs, and ultra-high Quantum Efficiency (QE) HgCdTe detectors for squeezed-light applications.

In this study, we analyze the stability of Zernike coefficient computation using deep learning techniques and propose a new training method for deep learning model that can reliably output higher-order Zernike coefficients. Previous studies have shown that deep learning is a powerful tool for accurately deriving the Zernike coefficients of polygonal mirrors, but reliably extracting higher-order Zernike coefficients remains one of the challenges. To overcome these challenges, we present a new training method for the stability of deep learning model, enabling reliable high-order Zernike coefficient computation. The proposed deep learning model is designed based on the Network-in-Network concept, and a two-stage training process ensures that low-order and high-order Zernike coefficients are simultaneously reliably generated. Experimental results for performance evaluation show that the proposed deep learning model is effective in outputting stable and reliable higher-order Zernike coefficients, especially for polygonal mirrors.

Understanding how meta-surfaces generate far field intensity distributions is key for real-world applications such as remote sensing and for designing experimental setups. The electric field distribution generated by a meta-surface consisting of Germanium crosses deposited on a calcium fluoride substrate was simulated using the Finite Difference Time Domain (FDTD) method. The electric field was then analyzed to find the component which propagates to a detector at a distance of 100 mm from the meta-surface. To simulate far field propagation using ray optics, a ray diagram was made to approximate the square magnitude of the electric field. This was done using weighted Voronoi cells, where the density of rays corresponds to the magnitude squared of the electric field. The gradient of the wavefront was used to determine the propagation direction of each ray. The meta-surface was resonant at a wavelength of 4 μm and off resonant at 3.05 μm. At both wavelengths, the beam focused to a central spot at the detector plane. This was also confirmed by Fourier optics. When resonant, the transmission to the detector was reduced.