The probability of detection (Pd) of targets in infrared and visually cluttered scenes is computed using the Fuzzy Logic Approach (FLA). The FLA is presented by the authors as a robust and high fidelity method for the computation and prediction of the Pd of targets. The Mamdani/Assilian, Sugeno and Neurofuzzy-based models have been investigated. A limited data set of visual imagery has been used to model the relationships between several input parameters; the contrast, camouflage condition, range, aspect, width, and experimental Pd. The fuzzy and neuro-fuzzy models gave predicted Pd values that had 0.98 correlation to the experimental Pd's. The results obtained indicate the robustness of the fuzzy-based modeling techniques and the applicability of the FLA to those types of problems having to do with the modeling of human object detection and perception in any spectral regime.

Recently the standard Night Vision search and detection thermal models have been challenged with the need to address scenarios which are quite different than those for which the models were originally intended. For example, there is a need to address the characteristics of the target and background in much more detail than was previously required. This paper will discuss and illustrate a selection of new concepts and formulations being considered by the Night Vision and Electronic Sensors Directorate for incorporation in improved search and detection thermal models. Included are new proposed formulations for the mean time for detection for both field-of-view and field-of-regard search and new concepts and formulas for thermal contrast signatures of targets and clutter characterization of backgrounds. Each new concept will be individually explained in detail with mathematical formulations and imagery examples. These formulations are then combined to illustrate how they can form new model metrics which can be used to predict both the static and dynamic probability of detection. The new candidate model formulations will be matched against available measured data to show the potential improvement in predictive capability offered.

Electro-optical and infrared systems are usually approximated to be Linear Shift Invariant (LSI) systems and are characterized by Modulation Transfer Functions (MTF). Each component in these systems has an MTF that describes the modulation throughput as a function of spatial frequency. Image compression is becoming a component in these systems, but unfortunately, cannot be described by the typical MTF. Image compression output signals can be very different than input signals in not only modulation, but frequency and phase. Therefore, the compression system is neither linear or shift invariant. We proposed an Information Transfer Function (ITF) that can be used in a manner similar to the MTF. This ITF describes the correlation between the input image and the output image as a function of spatial frequency, compression ratio, and image complexity. Like MTF, ITF can be used to determine image compression requirements to maintain system spatial resolution.

A computer code for a 1/f noise generator has been developed and is used to produce long sequences of temporal 1/f noise. This generator, which was developed in 1971, avoids the large memory requirements associated with other methods that have been used to produce extended records of 1/f noise. The power spectral density (PSD) of a sequence of such noise is calculated to verify its 1/f character. Noise that has a PSD of the form 1/f^(alpha ), where (alpha) is a non-integer exponent, may also be generated. As an example, the temporal 1/f noise current from MWIR HgCdTe photodetectors is simulated.

A model has been developed to compute the signal-to-noise ratio (SNR) for quasi-point sources imaged by infrared focal-plane-array sensors. In the far field such targets can be treated as point sources, while in the near field, where the target is much larger than the instantaneous field of view (IFOV), an extended target must be assumed. This paper describes the so-called Flat-Top Model which accounts for the effects on SNR of changes in the apparent size of the target relative to the sensor IFOV as a function of range; it also takes account of the pixel size and geometry. The model was developed to support work on the tracking of small missile targets in a naval environment.

A well designed performance modeling tool can assess trade- offs in sensor parameters for design; reasonably predict performance for varied targets, geometries, backgrounds, and environments; diagnose sensor system and signal processing problems, and provide a valued remote sensing educational tool. This paper describes the Infrared Performance Model (IRPM) which predicts IRST (Infrared Search and Track) detection SNRs for ranges of user specified operational scenarios, sensor design parameters, target and background models, and signal processing options. IRPM is a GUI driven software system which provides automated experimentation to graphically show the impacts of design or scenario parameter variations. Notable features include options for calculating radiometrically accurate target signatures from smoothed facet models, MTF calculation, system noise calculation from detector parameters, analytic Butterworth clutter models as well as clutter PSDs estimated from data, and sensor imperfections including aliasing, pattern noise, and jitter. The signal processing options include 2D and 3D matched filters and frame differencing detectors. Included is a discussion of the IRPM algorithms and sample results from the verification of IRPM for the AIRMS program.

A simulation is describe which determines in detail the response of visible through infrared focal plane devices, used in space-based electro-optical imaging sensors, to proton irradiation from the natural space environment, as a function of the choice of orbit around the earth. The quantitative description of the natural environment (protons) is based on the NASA code AP-8 and/or the CRRES database. Simplified shielding models and other models are then used to obtain the impinging proton environment, including the proton: flux, energy distributions, and trajectories, just above the surface of the focal plane in question. The interaction of a single proton with the focal plane is modeled based on the focal plane semiconductor material, the 3D pixel geometry, and the focal plane layout. This model is then exercised in a Monte-Carlo simulation to build up the aggregate focal plane response effects due to the variety of protons encountered under appropriate orbital and shielding conditions. The output of this simulation is statistically analyzed and used to construct a database to enable system planners and architects, algorithm developers, and sensor data users to mitigate the unwanted effects of this radiation.

Forward looking infrared (FLIR) detector arrays generally produce spatially undersampled images because the FLIR arrays can not be made dense enough to yield a sufficiently high spatial sampling frequency. Multi-frame techniques, such as microscanning, are an effective means of reducing aliasing and increasing resolution in images produced by staring imaging systems. These techniques involve interlacing a set of image frames that have been shifted with respect to each other during acquisition. The FLIR system is mounted on a moving platform, such as an aircraft, and the vibrations associated with the platform are used to generate the shifts. Since a fixed number of image frames is required, and the shifts are random, the acquired frames will not fall on a uniformly spaced grid. In this paper, we utilize gradient based shift registration to estimate the shifts between the acquired frames and then use a weighted nearest-neighbor approach for placing the frames onto a uniform grid to form a final high resolution image. We apply a Wiener filter to the high-resolution image in order to remove blurring due to the detector and optics of the imaging system. Simulation and experimental results are presented to verify the effectiveness of the proposed technique. The methods used here are significantly faster than alternate techniques, and are found to be especially suitable for real time applications.

Some imaging systems employ detector arrays which are not sufficiently dense so as to meet the Nyquist criteria during image acquisition. This is particularly true for many staring infrared images. Thus, the full resolution afforded by the optics is not being realized in such a system. This paper presents a technique for estimating a high resolution image, with reduced aliasing, from a sequence of undersampled rotated and translationally shifted frames. Such an image sequence can be obtained if an imager is mounted on a moving platform, such as an aircraft. Several approaches to this type of problem have been proposed in the literature. Here we extend some of this previous work. In particular, we define an observation model which incorporates knowledge of the optical system and detector array. The high resolution image estimate is formed by minimizing a regularized cost function which is based on the observation model. We consider both gradient descent and conjugate gradient optimization procedures to minimize the cost function. We show that the conjugate gradient optimization provides rapid convergence. Detailed experimental results are provided to illustrate the performance of the proposed algorithm using both visible and infrared images. Quantitative error analysis is provided and several images are shown for subjective evaluation.

Nonmechanical steering of passive imaging sensors is possible using liquid-crystal optical phase arrays as dynamic blazed-phase gratings. However, the resulting imagery is degraded due to diffraction effects which produce both multiple orders and dispersion. Restoration of this degraded imagery through post-processing techniques involves estimation of both the spectral radiance of the object within the field of regard of the sensor as well as deconvolution to remove blurring. A method of image restoration via Wiener filtering is presented for the case of an object scene composed of graybodies at various temperatures. Using simulated staring infrared sensor imagery, it is shown that image restoration is possible for small steering angles.

The advent of modern focal plane arrays has paved the way for the development of imaging spectrometers for use in remote sensing applications. This paper reviews four imaging spectrometer types: the dispersive spectrometer, the tunable Fabry-Perot filter, the Michelson Fourier Transform Spectrometer (FTS), and the spatially modulated FTS. Moreover, this paper identifies their strengths and weaknesses and considers their application to the LWIR region as defined by the Earth's atmospheric `windows' from 7.5 - 13.5 micrometers . Each instrument type places unique requirements on the performance of the focal plane arrays. The discussion centers on the dynamic range differences and dwell time difference for these instruments.

A variety of design constraints, imposed on the 2D hybrid staring infrared focal plane arrays used in emissive longwave imaging Fourier-transform spectrometers (FTSs), are developed and discussed. In particular, the origin of the large dynamic range and fast frame rate requirements for the focal plane are explained. The analysis is based on signal- to-noise ratio models for the interferogram (i.e., at the focal plane) and the associated spectrum. Specific focal plane design constraints associated with an optimized imaging FTS instrument design are illustrated, based on an aircraft-based remote sensing application.

Recent studies have demonstrated the potential for spectrally discriminating low contrast ground-based military targets in the thermal infrared for day/night reconnaissance, surveillance, and targeting applications. Although the underlying spectral features have been found to be very subtle in most cases, good detection performance is achievable due to the generally high band-to-band spectral correlation of terrestrial backgrounds. Recently, attempts have been made to develop imaging spectrometers of sufficient quality to preserve this high background spectral correlation and, in the process, provide robust target detection capabilities. One key issue which must be addressed in the sensor design is the impact of focal plane nonlinearity and nonuniformity on spectral correlation. In this paper, we present the details of a Monte-Carlo model which was developed to quantify this impact as a function of focal plane array characteristics for three sensing modalities: a dispersive spectrometer, a temporal Fourier transform spectrometer, and a spatial Fourier transform spectrometer. The results illustrate distinct differences in how these focal plane error sources propagate into the spectral domain and perturb the measured spectral statistics.

Line of sight (LOS) jitter produces temporal modulations of the signals which are detected in the focal plane of a temporally modulated imaging Fourier Transform Spectrometer. A theoretical treatment of LOS jitter effects is given, and is compared with the results of measurements with LIFTIRS (the Livermore Imaging Fourier Transform InfraRed Spectrometer). The identification, isolation, quantification and removal of jitter artifacts in hyperspectral imaging data by means of principal components analysis is discussed. The theoretical distribution of eigenvalues expected from principal components analysis is used to determine the level of significance of spatially coherent instrumental artifacts in general, including jitter as a representative example. It is concluded that an imaging FTIR spectrometer is much less seriously impacted by a given LOS jitter level than a non imaging FTIR spectrometer.

We investigated the feasibility of infrared imaging Spatial Spectral Interferometers (SSI) for spaceborne rapid- response, global-surveillance. The SSI is a sparse aperture interferometer that uses subaperture phase modulation techniques and multispectral sensors to increase angular resolution while reducing weight, volume, and cost. The subaperture phase modulation technique and multispectral sensors augment the spatial frequency sampling provided by the sparse aperture entrance pupil. We developed image modeling codes for SSI sensors incorporating multispectral sensors, change detection, and image deconvolution methods. Due to time and computer resource limitations, we did not simulate SSI systems with subaperture phase modulation schemes. Using change detection, the SSI need not form a whole image, but only updates a previously obtained image (from airborne or low Earth orbit sensors). Using deconvolution, the SSI need not be held to optical tolerances--pathlength compensators and laser metrology maintain the approximate subaperture piston and tip/tilt. But, the SSI self corrects by imaging a point source to measure the actual point spread function. We estimated optical and mechanical performance of the SSI sensor. We have shown that SSI sensors benefit significantly from change detection methods and the deconvolution algorithms greatly enhance the spatial resolution of SSI sensors. We have shown that space based interferometric imaging for rapid-response, global-surveillance is a feasible concept and that SSI sensors show promise of supporting a near real time, high-resolution, Earth-viewing, spaceborne imaging system.

The Infrared Sensor Performance Analysis Model (ISPAM) is a statistical tool used to evaluate detection and acquisition performances of IR sensors and seekers. ISPAM has become increasingly important in providing performance estimates for developmental infrared seekers. While models with varying degrees of freedom offer higher fidelity, the ISPAM model offers quick-look performances with relative ease particularly important in the early design phase of programs. Recent modifications to the model have been necessary to improve the fidelity in areas such as complex background clutter, signal processing techniques, and target discrimination. The model now more accurately represents the performance of state-of-the-art tactical IR sensors and seekers. In support of these modifications, the effectiveness of signal processing for target discrimination in a variety of clutter backgrounds was recently evaluated. An algorithm testbed utilizing a generic tracker and various spatial and temporal processing techniques was used in a Monte Carlo fashion to determine acquisition ranges based on each technique. Results of one of these spatial processes, an anti-median filter, were compared to field test data generated using this same processing technique by a current developmental seeker and tracker. The acquisition ranges from this field test were compared with that of the ISPAM model. This paper will address the general methodology of ISPAM's performance prediction, and the effectiveness of signal processing on target acquisition and tracking determined from the algorithm testbed.

Residual fixed pattern noise (RFPN) is a strong system performance driver for missile seekers using infrared focal plane arrays (IRFPA). RFPN is a result of variations in response from detector to detector in an IRFPA given a uniform input flux. This issue is further complicated by the need for reasonable correction methods for a fielded system where blackbody sources would not be available. One means of correcting a fielded system is to determine and store gains at the factory, and collect a one point offset correction in the field. For this method to be viable, gains must be stable through time. This paper presents test results for a multi-point piecewise linear correction technique using stored gains with offsets being collected just prior to correction. Data on gain stability over time and FPA temperature are presented. Results are also presented for a linear correction technique used at various FPA temperatures between approximately 77.9 to 88.1 degrees Kelvin.

Present laboratory test techniques for evaluating Forward Looking Infrared (FLIR) target acquisition sensors largely rely on simplistic infrared scenes such as four-bar targets against highly uniform backgrounds. One such test, Minimum Resolvable Temperature (MRT), is the primary laboratory test and evaluation (T&E) parameter for FLIRs. While these `simple' targets remove many `unwanted' variables for engineering analysis, they do not resemble the `real world'. Tactical FLIR sensors are being integrated into target acquisition subsystems (TAS) to provide information for purpose other than visual consumption, including automatic target detection, queuing, tracking, and automatic target recognizers. Ultimately, FLIR TAS operational performance must be demonstrated through live field testing. However, new acquisition strategies are driving toward performance specifications and increased modeling and simulation (and realism) into all levels of the testing processes. The time has come to look beyond MRT to assess the total operational performance of FLIR target acquisition subsystems in the laboratory. This paper describes the application of Dynamic Infrared Scene Projection (DIRSP) to project synthetic in- band infrared imagery (surrogate of the real-world) into the FLIR sensor entrance aperture. This paper concludes with a proposed utilization of DIRSP to support laboratory T&E of tactical FLIR target acquisition subsystems--beyond MRT.

The Modulation Transfer Function (MTF) is of fundamental importance in the testing of imaging systems as it is used to characterize the transfer of the spatial frequencies for the observed object. Different techniques based on the use of periodic targets made of unresolved lines or points have been proposed to assess this figure of merit for sampled imaging systems. The main potential problem in implementation of these methods is the fact that it is often difficult to insure a good balance in intensity between the individual lines or points belonging to the target. In a recent paper, we describe an analytical model allowing a first estimation of the importance of this problem. The purpose of this paper is to present this model and to apply it to the specific case of the 2D characterization of the MTF of an imaging system.

The design of most contemporary FURs is strongly driven by specific end user applications. Typical system architectures are unable to accommodate expansion to larger format focal planes, utilize alternative detector materials, or provide flexible sets of user accessible image and digital data manipulation features without significant investments in redesign. Amber provides an innovative solution to this dilemma by offering a modular infrared camera featuring flexible hardware and software architectures designed for stand-alone research and industrial instrumentation activity as well as in embedded gimbal FUR systems applications. Keywords: infrared, sensor, FLIR, seeker, camera.

The Surveillance Sciences Directorate of the Northrop Grumman Advanced Systems and Technology organization is developing a novel Multispectral IR camera known as Multiview. This prototype system is capable of simultaneously acquiring 4-color SWIR/MWIR 2D imagery that is both spatially and temporally registered utilizing a single 256² HgCdTe snapshot IR FPA capable of frame rates in excess of 240 Hz. The patented design offers an extremely compact package that contains the entire optomechanical assembly (lenses, interchangeable filters, and cold shield) in less than a 3 in³ volume. The unique imagery collected with this camera has the potential to significantly improve the effectiveness of clutter suppression algorithms, multi-color target detection and target-background discrimination for a wide variety of mission scenarios. This paper describes the key aspects of the Multiview prototype camera design and operation. Multiview's ability to dynamically manage flux imbalances between the four subbands is discussed. Radiometric performance predictions are presented along with laboratory validation of many of these performance metrics. Several examples of field collected imagery is shown including examples of transient rocket plume data measured at 240 Hz sample rate. The importance and utility of spatio-temporal multi-band imagery is also discussed.

In 1993, Inframetrics introduced the InfraCAM^TM to the commercial marketplace. It established the standard for hand-held, lightweight, low power, portable infrared sensors that is still unmatched in the industry. Under contract to the U.S. Army's Night Vision and Electronic Sensors Directorate, Inframetrics has integrated numerous detector technologies into a common system platform. This paper reports on a unique system for evaluating detector technologies using common electronics. The technologies covered are as follows: Platinum Silicide (PtSi) in the short wave infrared band (1.0 - 2.5 micrometers ), PtSi, Indium Antimonide and Mercury Cadmium Telluride in the mid wave infrared band (3.0 - 5.4 micrometers ), and HgCdTe and Quantum Well Infrared Photodetector in the long wave infrared band (8 - 12 micrometers ). The comparison criteria is primarily 3D noise (temporal and spatial) and Minimum Resolvable Temperature Difference. Additionally, non-uniformity correction circuitry, developed specifically to accommodate more sensitive, less stable detectors, is characterized.

The use of synthetic scene generation is gaining in popularity as a means of providing input stimuli to E-O system models. Such imagery can be used to complement any available real recorded data (where the latter is available) and to rapidly generate specific test scenarios including those at or even outside the operational envelope. The capabilities and scope of two of the many available synthetic scene generators are assessed.

For several years, Thomson-CSF Optronique has been using an internally developed scene and sensor model in the design of passive optronic systems. SEISM (Scene Electro-optical Image generator, and Sensor Model) is a powerful comprehensive tool that can easily produce physically accurate synthetic images including cluttered backgrounds. The scene is created using the image generation software EXPLORE^TM which undertakes 3D modeling, scene setting and non real-time rendering with texture mapping. The radiance reaching the sensor is computed by a multi-band (0.3 - 15 micrometers ) optronic scene model connected to a material database, and coupled with the atmospheric propagation code LOWTRAN7. Realistic complex backgrounds are created using either synthetic textures or real imagery (digitized aerial photographs or satellite images) mapped on Digital Terrain Model. These textures are converted into material or radiance textures in the waveband of the sensor. Special effects like concealing by clouds and realistic solar shadows (cloud and target shadows cast on the background) can be included in the scene. The generic sensor model simulates staring arrays, scanned and micro-scanned arrays. It faithfully reproduces the spectral, spatial and temporal effects introduced by a sensor. SEISM has been developed in order to meet the following needs: support to sensor design, visual evaluation of system performance, and algorithms validation as an alternative to field testing. This paper describes both scene and sensor model.

A Small Bispectral Infrared Detection (BIRD) push broom scanner for a small satellite mission is described, which is dedicated to the detection and analysis of high temperature events (HTE). Current operating and planned satellite sensors are not designed for high temperature event observation and therefore show some serious drawbacks such as saturation of the IR channels for target temperatures higher than 50 degree(s)C, low spatial resolution in case of daily coverage, low coverage of spatially high resolving systems, or not adequate IR channels. The BIRD instrumentation is a first attempt to overcome these disadvantages. For this purpose two infrared line scanners (3.4 - 4.2 micrometers and 8.5 - 9.3 micrometers ) will be combined with a Wide Angle Stereo Scanner in the visible. Because of the limited resources of a small satellite the design of all instruments is based on the usage of staring focal plane arrays. To observe HTE directly the covered sounding area should be as large as possible whereas at first glance the ground resolution of the sensor should be in order of some 10 m. These demands are in contradiction with the number of the infrared detector array elements currently available. For this reason methods of subpixel target detection and analysis have to be used. According to this concept a combination of the data from at least two radiometric high sensitive infrared sensor channels will be used to compensate the lack of high ground resolution. Adding to the infrared camera a suitable CCD-line scanner for a pre- classification with a higher ground resolution, an markedly improvement can be achieved.

Spatial resolution is the main characteristic of imaging system. It depends on various factors such as pixel area and pixel aperture, system noises, signal and background illumination levels, spectral range, and system MTF. To determine the resolution for given levels of signal and background illumination and spectral density of radiation, at first, it is necessary to calculate the number of photons which are transferred through multilayer coating of the imager and photogenerate carries giving a contribution into signal current or charge packet. Then full system noise consisting of input, photogeneration, thermogeneration, scanning, and output noises is determined. After that the system MTF including consideration nonlinear effects may be calculated and system resolution is determined. The resolution may be calculated for given wavelength or spectral range of input radiation. The main equations, algorithms for modeling and some results of simulation are described.

The importance of the thermal image system increases rapidly. Uncooled thermal imagers offer the advantages of room temperature operation, light weight, small size, low power requirements, simplified logistics and increased reliability. In this work, we propose a relatively simple procedure which can reasonably write specifications and design systems with higher speed. The first order design is coped with system performance models and PC based scientific spreadsheet, a mathematical software package constructs to provide a cost-effective, compliant solution to a design problem. An illustrative example is presented to show the details.

The noises of discretion, which accompany the quantization process of analogue signal, render essential influence on quality of the formed images in digital infrared remote sensing systems (DIRRSS). The uniform quantization question of analogue signals and their restoration are detail investigated and submitted in the scientific literature. However, as a result of casual change of movement parameters and orientation of an equipment carrier during the observation of various surfaces by DIRRSS, situations, when the value of an input signal quantization frequency is unknown or changes under the casual law with known density of probability, arise. In this case the input signal of system is subjected by the nonuniform quantization. For elimination of the image distortions during the restoration connected with nonuniform quantization it is necessary to decide an adaptive control problem by spatial quantization periods depending on casually varied parameters of observation. The influence of nonuniform quantization to quality of the image in DIRRSS is analyzed in the article. Expression of a spatial-frequency spectrum of 2D nonuniform quantization signal is received. The slanting scanning method and its combination with hexagonal raster are offered as effective methods of elimination of image nonuniform quantization. The schematic decision and work algorithm of the device for realization of a method of slanting scanning in combination with hexagonal sampling are offered.

Existing Infrared IRFPA models lack simplicity for setting up the detector's architecture/structure and lack continuity between IR detector material, IR detector processes, detector architecture, and detector operation. The models also lack the ability to reveal spatially and quantitatively the full impact of the detector's material, process and architecture on IRFPA performance. This paper will discuss the development of a new IRFPA computer model which is used to simulate existing and future IRFPA's with enhanced quantitative and visual information that allow the device engineer to access the impact of material quality, processing procedures and IR detector architecture on IRFPA performance in the SWIR-VLWIR region. This new model will be combined with statistical simulation to provide high IRFPA performance and lowest cost.