The factors influencing the attainment of fine simulated temperature resolution in a dynamic infrared scene projection system are examined. It is shown that the input data word bit resolution, the simulated temperature range requirement, the shape of the input word transfer function and the presence of spatial noise all affect the temperature resolution. Expressions are derived for allowing the simulated temperature resolution of both emissive projectors and spatial light modulator projectors to be estimated. Projector spatial noise is also analyzed in order to allow the noise limit on temperature resolution to be determined in terms of array nonuniformity. The array uniformity goals to which infrared projector manufacturers need to strive after nonuniformity correction procedures have been applied are therefore determined.

There is a growing demand for high complexity infra-red scene projector devices with improved output accuracy and resolution. Sources of inaccuracies and limitations to resolution are discussed for both projector-alone and projector-in-simulation situations. Using experience of existing system design and performance, an attempt has been made to assign realistic percentage inaccuracy figures to the various relevant effects so that a perspective can be gained of their relative importance. Schemes for improving the accuracy and resolution performance are considered, both in terms of system techniques and array device design. A revised order of relative importance is presented to assess the residual accuracy perspective after application of practicable corrective measures. We conclude that the absolute accuracy to which a unit- under-test's collected input can be relied upon under all conditions is only 90 - 95%. However, it is essential to cater for a much finer output resolution than would be deciduate by the absolute accuracy figure, although it seems doubtful that the requirements of 12 bit resolution could be obtained except in very restricted simulation/scene conditions.

Kinetic energy weapon (KEW) programs under the Ballistic Missile Defense Office (BMDO) need high fidelity, fast framing infrared (IR) imaging seekers. As imaging sensors have matured to support BMDO, the complexity of functions assigned to the KEW weapon systems has amplified the necessity for robust hardware-in-the-loop (HWIL) simulation facilities to reduce program risk. Tactical weapon systems are also turning to imaging focal plane array (FPA) seekers. They too require more sophisticated HWIL testing. The IR projector, an integral component of a HWIL simulation, must reproduce the real world with enough fidelity that the unit-under-tests's (UUT) software will respond to the projected scenario of images as though it were viewing the real world. The MOSFET resistor array IR scene projector shows great promise in cryogenic vacuum chamber as well as room temperature testing. Under the wideband infrared scene projector (WISP) program, an enhanced version of the resistor array is currently under development. When the WISP system is delivered, the projector will consist of a 512 by 512 baseline array with the center 128 by 128 resistors having a higher output capability. For the development stage of the program, 512 by 512 prototype baseline and 128 by 128 prototype high dynamic range (HDR) arrays have been fabricated separately. Characterization measurements to include: spectral output, dynamic range capability, apparent temperature, rise time, fall time, cross talk, and current consumption have been accomplished on the prototype baseline and HDR arrays at the Kinetic Kill Vehicle Hardware-in-the-Loop Simulator (KHILS) Facility and the Guided Weapons Evaluation Facility (GWEF). Results from the measurements show the HDR array dynamic range is an order of magnitude greater than its predecessor. Other parameters such as droop, rise time, etc., either meet or are close to meeting system specifications. The final design of the arrays is currently in progress based on these results.

Arnold Engineering Development Center's (AEDC) Focal Plane Array Test Chamber (FPATC) is currently operational and has been involved in the testing and evaluation of focal plane arrays (FPAs) developed for interceptor applications. Testing included blackbody characterization, laser compatibility, and projection of AEDC test patterns and scenes. Efforts continue to be made to improve the capability of direct write scene generation (DWSG) to support the latest in FPA technology. A demonstration of closed-loop operation has been accomplished.

For the first time in more then 10 years, the same IR sensor has been tested in both the portable optical sensor tester (POST) space chamber and the Arnold Engineering Development Center (AEDC) 7V space chamber and the results compared. These chambers both have extensive calibration capabilities with direct tractability to the National Institute of Standards and Technology (NIST). This paper describes the tests performed, the excellent comparison results, and a brief description of the calibration approach used for each chamber.

BMDO must simulate the detection, acquisition, discrimination and tracking of anticipated targets and predict the effect of natural and man-made backgrounds and environmental phenomena on optical and radar sensor systems designed to perform these tasks. The SSGM is designed to integrate state-of-science knowledge, data bases and valid phenomenology models to simulate ballistic missile engagement scenarios for both passive and active sensors aboard surveillance system platforms and defensive interceptor missiles -- thereby serving as a traceable standard against which different BMDO concepts and designs can be evaluated. This paper concentrates on describing the current capabilities and planned development efforts for SSGM. The focus will be on the functionality of the SSGM Release 7.0 and the planned development effort for subsequent SSGM releases. We shall demonstrate the current SSGM capability (R7.0, January 1996) with sample multi-phenomenology output scenes and videos. New capabilities include realistic 6-DOF dynamics for targets, simulated target radar cross section correlated with IR information, and authorative target model data sets based on actual flight experiments.

The Arnold Engineering Development Center (AEDC) scene generation test capability (SGTC) program has completed the development of a laser based direct write scene generation (DWSG) facility that provides dynamic mission simulation testing for infrared (IR) focal plane arrays (FPAs) and their associated signal processing electronics. The AEDC DWSG focal plane array test capability (FPATC) includes lasers operating at 0.514, 1.06, 5.4, or 10.6 micrometer, and acousto-optic deflectors (AODs) which modulate the laser beam position and amplitude. Complex radio frequency (rf) electronics controls each AOD by providing multi-frequency inputs. These inputs produce a highly accurate and independent multi-beam deflection or 'rake,' that is swept across the FPA sensor under test. Each rf amplitude input to an AOD translates into an accurate and independent beam intensity in the rake. Issues such as scene fidelity, sensor frame rates, scenario length, and real-time laser beam position adjustments require rf control electronics that employs the use of advanced analog and digital signal processing techniques and designs. By implementing flexible system architectures in the electronics, the overall capability of the DWSG to adapt to emerging test requirement is greatly enhanced. Presented in this paper is an overview of the signal processing methodology and designs required to handle the DWSG requirement. Further, the paper summarizes the current status of recent AEDC technology efforts tasked with the implementation of real-time and closed-loop scene manipulation including sensor optical simulation using the DWSG. The paper describes a proof-of-principle (PoP) demonstration which used high speed digital signal processors inherent in the DWSG electronic design to compute the rotation, translation, optical transfer function convolution, and system calibration functions during scene projection.

The infrared (IR) sensor stimulator (IRSS) development program is a joint central test and evaluation investment program (CTEIP) effort involving the U.S. Air Force's electronic combat (EC) integrated test (ECIT) program (responsible for upgrading the Avionics Test and Integration Complex (ATIC), located at the Air Force Flight Test Center (AFFTC), Edwards AFB, Calif.) and the U.S. Navy's Air Combat Environment Test and Evaluation Facility (ACETEF) (located at the Naval Air Warfare Center -- Aircraft Division (NAWC-AD), Patuxent River, Maryland). The IRSS program involves an Installed Systems Test Facility (ISTF) technology upgrade to the ATIC and the ACETEF and will provide a high-fidelity dynamic IR stimulation capability to realistically test IR sensor systems while installed on host aerospace platforms. These IR sensor systems include: targeting and navigational forward looking IR (FLIR) systems, missile warning systems (MWS), IR search and track (IRST) systems, and IR missile seekers. The IRSS will consist of one or more scene generation systems, a digital injection system, a point source projection system, and an image projection system. Threats, targets and backgrounds will be dynamically presented to the IR system under test (SUT) with sufficient fidelity to reflect a realistic battlefield environment. The IRSS will enable the Air Force and Navy to meet IR developmental test and evaluation (DT&E) and operational test and evaluation (OT&E) requirements for multi-service customers.

This paper provides a description of an innovative 'Quick-look' technique (process) for the characterization and evaluation of the flight performance of Theater Ballistic Missile (TBM) and Strategic Target Systems. The methodology for processing real-time flight target telemetry data and integrating those data items into a visualization-based, Quick-Look Analysis Process (QLAP) is presented with highlights of the results obtained from a typical flight test. Emphasis is placed on providing visualization-based data products which 'jump start' the post-flight-test data evaluation process through the correlation of a robust first-principles target model developed with the USASSDC Optical Signatures Code (OSC 18.3) to determine (pre-mission) a priori target performance parameters and correlating these data to the actual flight target performance. Processed thermal, flight dynamics, orientation and angular rates telemetry data are treated as 'truth data inputs' for the OSC-18.3 simulation of the 'real-world, flight test behavior' of the test target system. Results from this process include a video tape with visualizations of temperature and radiance maps of the target system as 'viewed' from the perspective of each optical or radar corollary sensor participating in the flight test. Also included in this paper are examples of the efficacy of using this methodology for pre-mission optimization of the data collection plans for each corollary optical or radar sensor.

The reliability of computer models of both signatures and effects introduced as the signatures are measured is of critical importance in the development of seekers and other devices which use EO sensors. Typically the models consist of many independent modules or programs, each of which models many different physical processes and requires many different types of input data. Due to the models' complexity, their reliability is commonly not well established, and the domains within which the models are valid are commonly not understood. Efforts at establishing the reliability of the models frequently only address a very limited number of characteristics of the outputs and seldom characterize the reliability of the input parameters. Figures-of-merit usually address specific characteristics of certain outputs and often do not attempt to evaluate the performance of the model in the complete system. This paper discusses the problem of establishing the reliability of computer models of passive imaging sensors and the signatures they observe. General question about the process of comparing measurements with model results to determine how well the model works are examined, and methods for establishing reliability are presented. The physical models and types of input data required to describe the signatures and sensors are enumerated. Quantities used to describe the outputs of the different physical models are identified, and different types of figures-of-merit are identified to evaluate how well the computer model outputs compare with true/measured values. The Irma passive model is used as an example and several examples of validation completed for this code are presented.

Many of the standard considerations for modeling of target acquisition of stationary targets in a single field of view must be altered when considering target acquisition in the actual battlefield. Sensors are scanning, targets are moving, and the detection process is considerably more complex than the analysis of the highest detectable frequency detectable superimposed on the target size. In this paper, we discuss several of the issues involved. The use of a single detectable frequency to model the target acquisition process is not sufficient. We consider the replacement of this model with one consisting of an integral of the target spectrum, taking into account the spatial frequency MTF dependence of the imaging system.

Atmospheric effects at low elevation angles can complicate shipboard infrared search and tracking (SIRST) of distant low altitude targets, such as sea skimming cruise missiles. Here we focus upon the effects of ray refraction and atmospheric distortion. For constant-flux surface layer conditions we discuss target magnification and demagnification and atmospheric distortions. For sufficiently negative air-sea temperature differences (ASTD), the maximum intervision range (MIVR) of low altitude targets is reduced, but the target is significantly magnified compared to no-refraction predictions. Negative ASTD can give rise to an inferior mirage which we discuss with a model-data comparison. Positive ASTD extends a target MIVR, but the target image is severely demagnified, closer to the horizon, and more degraded by atmospheric turbulence. We discuss environments that are likely to violate constant-flux conditions and include an example of a superior mirage. Although horizontal inhomogeneity may well influence superior mirage formation, we show that inhomogeneity is not necessary to explain features such as numerous mirages or multiple (three) horizons.

In the process of creating synthetic scenes for use in simulations/visualizations, texture is used as a surrogate to 'high' spatial definition. For example, if one were to measure the location of every blade of grass and all of the characteristics of each blade of grass in a lawn, then in the process of composing a scene of the lawn, it would be expected that the result would appear 'real;' however, because this process is excruciatingly laborious, various techniques have been devised to place the required details in the scene through the use of texturing. Experience gained during the recent Smart Weapons Operability Enhancement Joint Test and Evaluation (SWOE JT&E) has shown the need for higher fidelity texturing algorithms and a better parameterization of those that are in use. In this study, four aspects of the problem have been analyzed: texture extraction, texture insertion, texture metrics, and texture creation algorithms. The results of extracting real texture from an image, measuring it with a variety of metrics, and generating similar texture with three different algorithms is presented. These same metrics can be used to define clutter and to make comparisons between 'real' and synthetic (or artificial) scenes in an objective manner.

This paper describes a methodology which has been successfully used to create high fidelity three-dimensional infrared (IR) signature models of terrain backgrounds for use in digital simulations by the U.S. Army Missile Command. Topics discussed include (1) derivation of database fidelity and resolution requirements based upon system parameters, (2) use of existing digital elevation maps (DEMs) (3) generation of digital elevation maps from stereo aerial and satellite imagery, and (4) classification of ground cover materials.

The next generation of ship defense missiles will need to engage stealthy, passive, sea-skimming missiles. Detection and guidance will occur against a background of sea surface and horizon which can present significant clutter problems for infrared seekers, particularly when targets are comparatively dim. We need a variety of sea clutter models: statistical image models for signal processing algorithm design, clutter occurrence models for systems effectiveness assessment, and constructive image models for synthesizing very large field-of-view (FOV) images with high spatial and temporal resolution. We have implemented and tested such a constructive model. First principle models of water waves and light transport provide a computationally intensive clutter model implemented as a raytracer. Our models include sea, sky, and solar radiance; reflectance; attenuating atmospheres; constructive solid geometry targets; target and water wave dynamics; and simple sensor image formation.

For generation and analysis of the multi-sensory image, we propose a new three dimensional (3D) modeling method considering the internal heat source. We represent the heat transfer process within the object as the equivalent thermal circuit. Therefore, without a complex computation, our modeling approach can obtain thermal features of the object. By using the faceted model, the proposed method can express the accurate visual features of the object. In simulation results, we have ascertained that the proposed method generates realistic and accurate thermal images of the object having the internal heat source.

We discuss the Ballistic Missile Defense Organization's current system for archiving measurement data from phenomenology missions. We propose a new paradigm for a distributed, interoperable 'virtual' data center that delivers access to measured data, descriptive catalogs, modeling and analysis resources to the user's desktop computer. We discuss requirements and architecture for the first version as well as future enhancements to such a system.

The principle concerning the substitution of the measuring resolution of night vision system in laboratory for the field detecting objective distance test is elaborated in this paper. A method concerning the substitution of the measuring resolution with the low illumiiiation and the high contrast of night vision system in laboratory for the field detecting objective distance test is given. But the method has some shortcomings. A new improving conceivability is proposed. Aphysical model for the conceivabffity and a performable scheme for it is given. Theoretical analysis shows that the new scheme will have a greatly improved creditability for the evaluation of detecting objective distance of it. Keywords: measurement in laboratory; night vision system; detecting objective distance.

This paper describes the algorithms Arete is developing for shipboard infrared search and track (SIRST) detection of low-observable targets, such as subsonic, sea-skimming cruise missiles. The key algorithm is Arete's Bayesian (probability) field tracker, which is a track-before-detect algorithm. The basic concept of this tracker is to update in successive time steps the probability of all possible target positions and velocities before thresholding. Sample results are presented for simulated low (approximately 6 dB) signal-to-noise ratio (SNR) targets injected into both simulated and real ocean horizon scenes. More conventional detection algorithms require greater SNR for each temporal update. Since the cruise missile signature decreases with increasing range between the sensor and the cruise missile, our Bayesian tracker can detect subsonic, low-observable cruise missiles at greater ranges. To mitigate false alarms the measurement likelihood is modified to account for non-Gaussian noise/clutter statistics (large intensity outliers). False alarm mitigation is demonstrated for injected signatures into real data.

Target acquisition is fairly straight forward when tracking a single bright target. A dim target in clutter on the other hand can present some challenges in target acquisition. The characteristics of detecting a target under typical conditions are fairly well understood. For example, one way to characterize signal detection is by constructing the receiver operating curve. By contrast, the characteristics of acquisition of a dim target in persistent background clutter can be very different from signal detection. For example, an acquisition can be a false acquisition even with the target in the field of view. A false acquisition is typically based on clutter points or false signals rather than the target. Given a target of interest in the field of view, the indicated false acquisition can be far from the location of the true target. The term acquisition used in this context is not the same as signal or target detection. In this sense, acquisition requires making the distinction between a detection or sequence of detections caused by a target versus that caused by false signals and persistent clutter. This paper presents an analysis of the differences between the characteristics of detection and acquisition. Analysis and understanding the characteristics of target acquisition are important for undertaking the specification, design or performance evaluation of a sensor system. To limit the scope to within the page limits, this paper concentrates on acquisition of a single, small, dim target with false signals and persistent background clutter (possibly moving) in the field of view.

The proliferation of infrared (IR) sensitive seekers in today's battlefield is a critical problem faced by airborne vehicles. A simple analytical target acquisition technique to determine the range at which an airborne vehicle may be susceptible to detection by an IR seeker sensor is presented. This approach can be applied to any IR sensor searching for various ground or air platforms. The detection range is determined through a series of basic radiometric calculations that take into account a target's radiance and the seeker's noise equivalent flux density (NEFD) and signal-to-noise ratio (SNR). The calculations allow for background and path radiances determined from LOWTRAN 7 program outputs for night and day conditions and a sky background. A step-by- step sample calculation is provided to illustrate this approach.

Present staring IR missile warning technology operates at thresholds that approach 1000 times noise equivalent irradiance (NEI) under stressing clutter conditions severely limiting the detection range. The detection thresholds must be lower for high probability of declaration, with a low false alarm rate, of low observable targets within high clutter environments. The DICE program includes three approaches for clutter rejection and improved signal to noise. One approach combines the clutter suppression capabilities of two- dimensional adaptive spatial filters with very accurate frame registration and adaptive temporal processing to meet DICE program requirements. A second approach consists of an adaptive neural maximum likelihood algorithm combined with a hyperbolic filter (HF) for clutter suppression. The HF is a 3- dimensional filter using a multi-dimensional fast Fourier transform which suppresses clutter based on relative motion to reduce the processing load of the adaptive neural algorithm. The maximum likelihood neural algorithm uses clutter models for comparison with sensor data to eliminate the IR clutter and declare targets. The third DICE approach combines dual color discrimination with non-linear morphological signal processing for single frame missile detection. The algorithm estimates the clutter background of one IR band by using a second IR band then enhances point like geometries for missile detection. The technique requirements include probability of declaring a threat of greater than 90%, a false alarm rate (FAR) of less than 1 per hour and declaration of threat in less than or equal to 2 seconds from the time the target first crosses the detection threshold. The goal is to maintain a detection threshold of 10 times present state-of-the-art noise equivalent irradiance (NEI) even in a highly cluttered background. Program results are available from two approaches as of March 1996.

Binocular stereo is recognized as a useful approach for extracting range information from images. This approach is passive and it can provide detailed depth information. In this paper we extend the area-based stereo approaches, to make them well suited for target ranging applications. In this domain it is required that the stereo approach should be able to handle relatively large depth of fields in outdoor environments. This is accomplished by using widebase lines in the placement of cameras. Also, in order to minimize the false alarms, multiple binocular cameras are utilized. Algorithmic development of this multiple widebaseline stereo approach is presented with the help of experiments use random dot stereograms. Preliminary evaluation with a series of outdoor images is also presented.

Reduction of the infrared signature of warships helps to increase their survivability. Two methods to reduce the infrared signature are discussed: the cooling of exhaust gases and the application of low emissivity paint. The infrared signature of a generic frigate has been calculated with and without signature reduction. The signature values are used to determine the lock-on range of four generic infrared seeker heads of anti ship missiles ranging from hot spot seekers to imaging seekers. From the reduction in lock-on range the reduction in hit probability is calculated. The results show that both exhaust gas cooling and low emissivity paint contribute to the survivability of a warship.

The detection of small targets in maritime infrared surveillance is hampered by the presence of clutter. Sea surface structure, reflection and emission changes related to incident angle variations and surface effects are standard features governing the clutter behavior. Also special effects as sun glint and horizon effects play an important role for clutter. In order to optimize the detection process, quantitative clutter estimates are of use for filter settings. We have recorded a large amount of infrared backgrounds in the last few years, during common NATO trials. A large amount of different meteorological conditions took place during the various experiments. A first set of these data have been analyzed to obtain statistical data that represent the infrared scene. We have derived vertical temperature profiles, vertical fluctuation profiles, horizontal correlation coefficients and temporal correlation functions. In this paper we present the first analysis of these data. We are in the process of obtaining a condensed database of information to regenerate clutter images from bulk meteo parameters, and clutter parameters. The clutter and meteo parameters have been used to simulate various infrared scenes. Examples of this simulation process are shown in the presentation. The simulated images are statistically similar to the original images that were used to derive the parameters. A description of the image- generation is presented. Future expansions of the model are discussed.

This paper concerns multispectral backgrounds and clutter as they hamper the detection process of small surface targets. The variability of target contrast is shown for a number of meteorological conditions in the various spectral bands, for which new types of sensors have become available. It is shown that simultaneous framing with pixel co-location is very essential, which requires automatically the application of staring 2-D arrays. Some additional sensor tricks are introduced such as gated viewing, dynamic range enhancement, and laser 3-D imaging. By proper selection of bands, optimum clutter reduction and detection probability can be achieved for targets such as swimmers, buoys and zodiacs.

Amplitude and phase fluctuations are an inherent characteristic of many types of propagation media including sound in the marine environment. Generally, fluctuations are regarded as a nuisance to be ignored, avoided, or eliminated. However, this paper shows that fluctuations can be effectively exploited to enhance the detection of targets that fluctuate less in amplitude than the clutter and background noise. In fact, a case can be made that the exploitation of fluctuations constitutes a third dimension for achieving gain to supplement the other two well know dimensions of frequency resolution gain and array aperture gain. Results from measurements at sea are presented to support the claim that a new dimension of gain is being accessed and to demonstrate that the additional gains can be substantial.

Recent data collections using an infrared hyperspectral measurement system have provided a significant measurement database of military vehicles in vegetated and desert backgrounds. This paper summarizes the results of a study performed to assess the detection performance potential of multispectral sensors using this database. Specific issues addressed include approaches to optimal band selection; robustness of band combinations with target, background, and environment diversity; and sensor noise requirements. All of these issues are vital to assessing the feasibility and utility of infrared multispectral sensors in operational scenarios.

The 8 - 12 micrometer polarization signatures of diffuse and specular aluminum geometrical darts were analyzed outdoors using a polarization sensitive thermal imager. Results of the degree and plane of polarization are presented for different thermal imager gain bands and weather conditions during a two week period. The 0 degree, 45 degree, 90 degree and 135 degree polarizer orientations were thermally calibrated and the S1 and S2 Stokes parameters shown as radiometric temperature differences. The effect on the polarization signatures of range is considered for these targets at 100 m and 370 m. A comparison of the degree of polarization to changes in the emission/reflection balance and to variations in the dart's complex refractive index is made.

The performance of many tactical sensor systems, especially inertial measurement units (IMUs), suffers when operating in the presence of spin. Sensor Applications has developed a magnetic field sensor, or spin counter, that is capable of switching once per revolution through the earth's magnetic field. The miniature sensor, which eventually will consist of a single ASIC chip, uses the properties of giant magnetoresistance ratio (GMR) materials. The sensor has been designed so that it is immune to local field strength variations. A spin sensor could aid or replace other sensing systems by providing a true measurement of the projectile's roll rate and by determining up from down, which is a very important factor for inertially navigated munitions. High-g shock tests and gyroscopic tests of the sensor have been performed by the U.S. Army Research Laboratory (ARL) to characterize its performance at conditions produced by artillery projectiles. Helmholtz coil testing of the sensor exposed to differing magnetic field conditions was also investigated by Sensor Applications. Initial test results look promising. This paper describes the development of the spin counter, summaries the test results, and discusses the many benefits of this sensor.

Background phenomenology databases and models are essential for the design and assessment of electro-optical sensing systems. The MWIR band has been proposed to satisfy a number of specific requirements in the DoD space based mission areas. However, the phenomenology database in the MWIR to support the design and performance evaluation is limited. Currently the high resolution infrared radiation sounder (HIRS/2) onboard NOAA 12, an operational polar orbiting environmental and weather satellite, offers continual global coverage of several bands in the MWIR. In particular, Channel 17 operates in the heart of the 4.23 micrometer carbon-dioxide band. Though with coarse resolution (approximately 20 km), the vast database offers a good baseline understanding of the MWIR phenomenology related to space based MWIR systems on (1) amplitude variation as function of latitude, season, and solar angle, (2) correlation to relevant MWIR features such as high-altitude clouds, stratospheric warming, aurora and other geomagnetic activities, (3) identification of potential low spatial frequency atmospheric features, and (4) comparison with future dedicated measurements. Statistical analysis on selected multiple orbits over all seasons and geographical regions was conducted. Global magnitude and variation in these bands were established. The overall spatial gradient on the 50 km scale was shown to be within sensor noise; this established the upper bound of spatial frequency in the heart-of-the-carbon-dioxide-band. Results also compared favorably with predictions from atmospheric background models such as the Synthetic High Altitude Radiance Code (SHARC-3).

Because most infrared (IR) observations of natural terrain are made at infrequent or widely spaced time intervals, information on the high time resolution variation in background radiance is lacking. This is in spite of the fact that meteorological and other influences can cause apparent radiance to change by amounts equivalent to a few degrees Kelvin in seconds or tens of seconds. Infrared imagery acquired during the recent Smart Weapons Operability Enhancement Joint Test and Evaluation (SWOE JT&E) allows the investigation of the nature of IR radiance fluctuations with high time resolution under a variety of diurnal and meteorological conditions. Specifically, during the Grayling I deployment, images were acquired at the rate of 6.25 frames per second, for 10 seconds, every 5 minutes during each of the 2-hr mission periods. There were 107 mission periods during the 41 days of the test. A portion of these images have been segmented according to homogeneous features (homogeneity determined independently) and histograms of apparent temperature within those features (8 in number) have been processed to arrive at several first order statistics (mean, variance, etc.) as a function of time. The nature of the fluctuations as a function of meteorological, atmospheric, and diurnal conditions is discussed.

New visual detection models developed under the Target Acquisition Model Improvement Program (TAMIP) require characterized color images as an input parameter. These images are used to make predictions of the probability of detection vs time and/or distance for a military target in a background of interest. Techniques to refine the methods of characterizing color images have been ongoing for several years within the Army. The subject of this paper is a method for specifying optimized reflectance standards for characterization of the images. The color of the reflectance standards are selected for compatibility with the outdoor environments used. The need for customization of color reflectance panels is also explained.

Synthetically generated radar imagery can provide a unique virtual environment which integrates the variability of sensor, scene, operational, and processing parameters. Since the advent of radar sensor technology, a significant level of effort has been expended on the development of electromagnetic modeling techniques to understand and predict the scattering and sensor phenomenology involved in synthetic-aperture radar (SAR) imaging systems. The inception of CAD/CAM tools in designing targets has allowed for the accurate prediction of geometrically and materially defined targets for use in signature prediction models. These three-dimensional target models are well suited for signature predictions which utilize deterministic approaches such as shooting/bouncing ray, finite-difference time-domain, and finite-element methods. In the area of clutter modeling, however, the temporal and inhomogeneous nature of foliage and terrain constituents require a hybrid formulation of deterministic and statistical approaches. In light of these modeling requirements, an effort to integrate sensor, target, and clutter simulations had led to the development of a SAR simulation system, advanced radar imaging emulation system (ARIES). This software provides the synthetic generation of high-fidelity SAR imagery for a wide range of user-defined platform, sensor, and scene parameters.

The scattering process of electromagnetic waves is dominated by the match between wavelength and the geometric dimensions of surface structures. With respect to the microwave radar bands millimeter-waves are better matched to small surface features of terrain. Therefore this frequency band is able to gain additional information on the terrain of interest. For high resolution imaging SAR is the favorite solution also for millimeter-wave frequencies. Compared to more classical radar bands millimeter-waves offer advantages in the SAR processing, because due to the higher primary resolution at a given antenna aperture sources of image distortions such as range migration or depth of focus can be neglected at these frequencies. Moreover the inherently short aperture time for a given resolution improves the relation to the time constant of flight instabilities and makes motion compensation a simple process. A coherent, polarimetric, high range resolution radar, operating at a nominal frequency of 94 GHz, has been installed onboard an aircraft to allow remote sensing measurements in a side looking synthetic aperture approach. The radar-raw-data were registered together with time code and inertial data of the aircraft and later on evaluated by an off-line SAR-processor. The resulting images then had to undergo an automatic recognition process to extract certain complex targets using a knowledge based production system. The paper describes the measurement system and discusses the evaluation procedures with emphasis on the applied SAR algorithm. Examples of radar images at 94 GHz are shown and samples of pattern recognition derived from the SAR images are shown.

Periscopes reflect radar energy both directly and by multiple reflection from the water's surface back to the radar unit. Radar reflections have also been observed by the water disturbance. A report is to be made on the development and check of the effectiveness of radar camouflage for periscopes. A special demonstration concept with which it is possible to measure the radar cross- section of the periscope, including the surrounding rough water's surface, under reproducible boundary conditions has been developed for this purpose. The set-up of such a measurement system and the related radar signature measurements are shown. The measurements were carried out with an uncamouflaged and a camouflaged periscope. The RCS values are compared with empirical values from sea surveillance radars against known radar cross- sections. The detection ranges against camouflaged periscopes as well as the camouflage concept for the periscope are described.

In developing RCS prediction codes a variety of physical effects such as the edge diffraction effect have to be considered with the consequence that the computer effort increases considerably. This fact limits the field of application of such codes, especially if the RCS data serve as input parameters for system simulators which very often need these data for a high number of observation angles and/or frequencies. Vice versa the issues of a system analysis can be used to estimate the relevance of physical effects under system viewpoints and to rank them according to their magnitude. This paper tries to evaluate the importance of RCS predictions containing an edge diffracted field for systems analysis. A double dihedral with a strong depolarizing behavior and a generic airplane design containing many arbitrarily oriented edges are used as test structures. Data of the scattered field are generated by the RCS computer code SIGMA with and without including edge diffraction effects. These data are submitted to the code DORA to determine radar range and radar detectibility and to a SAR simulator code to generate SAR imagery. In both cases special scenarios are assumed. The essential features of the computer codes in their current state are described, the results are presented and discussed under systems viewpoints.

To validate the computed RCS of a helicopter, its RCS was measured. Since no RCS measurement site exists in Belgium, we improvised one. The search for a measurement site, for a mobile radar and for a suitable calibration sphere were some of the problems we had to deal with. When starting the measurements we immediately realized that due to the high clutter level the calibration sphere and the helicopter had to be brought much closer to the radar, so that the far field conditions were not satisfied anymore. However, this seemed not to influence too much the results. On a rainy day, when the helicopter could not take off, we measured the RCS of a Volvo truck. The analysis of these measurements allowed us to observe if the presence of wet ground influenced the importance of the ground reflections. Finally we compared the measured RCS of the helicopter to the computed RCS, and while an acceptable agreement was found for the amplitudes of the peaks, we observed that the minima present in the computed RCS were not found back in the measured RCS. This is probably also due to the relatively high clutter level.

The Army Research Laboratory (ARL) has constructed a variation of the bruderhedral calibration and radar cross section (RCS) target model and measured its radar characteristics in the field. A computer version of the same model was generated, and later characterized in both elevation and azimuth for validation. Our goal is to develop a millimeter-wave (MMW) signature generation tool for guidance integrated fuzing (GIF) systems and applications. Before realizing this goal, one must develop a test-bed of tools and approaches upon which to build. ARL has identified approaches to developing generic analytical target-signature models based on some existing electromagnetic scattering codes. A high-frequency RCS and signature prediction software model was selected to perform the radar analysis and provide a mechanism, a synthetic aperture radar (SAR) model, for recognizing prominent scatterers off high-fidelity target models. This method will assist us in creating suitable far- to near-field 3-D transitional models at MMW frequencies. Two target model descriptions were used in the signature prediction model: a flat facet format and a curved surface format. This paper introduces these software models, and some optics and SAR considerations relating to the test wavelength and the size of the target. Also, the simulated azimuthal and elevation response patterns, along with some results from the SAR model, are presented.

A new era in signature simulation has arrived with the development of a family of instruments for the measurement of optical properties of materials. These new instruments, covering the spectral range from 0.2 to 200 microns plus millimeter waves, have been mated with workstations which control the measurement processes, and on which software for data analysis, signature simulation, and coatings design reside. Portability of the instruments allows measurements and analysis in the laboratory, or in the field. For the first time an analyst can affordably make rapid measurements of such quantities as bi-directional or directional reflectance/emittance of sample materials, analyze, plot, and archive the measurements, and then immediately use the new measurements in a signature simulation of targets, backgrounds and scenes. Alternatively, the analyst can design a unique coating and immediately subject the design to iterative analysis for target signature control or deception. This paper discusses the instruments and techniques for accomplishing the measurements and analysis.

Descriptions are given of major errors which appear in several published curves showing atmospheric attenuation at frequency intervals between 10 and 1000 GHz. Three cases are discussed, two for clear-air conditions and one for fog. In one example, the attenuation at 4 km elevation has been mislabeled as 9150 km (or 30,000 feet) for the 10 to 400 GHz range. This error has appeared in several journal articles, vendors' catalogs, short-course notes, and a recently-published book. In a second case the attenuation peak near 22.3 GHz (due to water vapor absorption) has been plotted at 20 GHz. The third case deals with an error pertaining to attenuation in fog for frequencies between 10 and 1000 GHz. Specific information and corrections are given for all three cases. The net result of these errors is that development of sensor and communications applications has been impeded because the errors usually make atmospheric losses appear to be greater than they really are.

In this paper, research studies aimed at the development of computational models, specifically for an important preattentive cue, texture, are discussed. The primary basis for the development of computational texture models is to match them with how a human perceives textured patterns. This matching between human and quantitative models is accomplished with the help of an extensive series of psychophysical experiments. Observers view textured patterns, which are generated with a precise mathematical model, and human responses are used to develop a ranking of distinctness of texture patterns. We illustrate the use of these computational and psychophysical experimental tools for the analysis and characterization of real-world high resolution imagery acquired in thermal and visible bands. These results are applicable to image clutter characterization, camouflage assessment, and automatic target detection and recognition tasks.

A wavelet transform method for analyzing the dynamic effects of atmospheric turbulence on infrared digital imagery is presented. Typical results are illustrated by applying the method to simulated turbulence induced distortions and to data from field experiments. Comparisons of this method to the more traditional short-time Fourier transform and long exposure (non-dynamic) atmospheric modulation transfer function (AMTF) methods are made to show the utility of the technique.

This paper discusses the dynamics of signature propagation associated with imaging and lidar operation through a realistic turbulent atmosphere in the vicinity of chemical plumes. Traditional phase screen models represent the propagation through artificial turbulence without dynamic compatibility. In contrast, three new models, AIRFLOS, CHAFFSIM, and SIMLIGHT have been developed to provide a first principles physics approach to modeling signature propagation through a dynamically consistent atmosphere. Solving the Navier Stokes equations the AIRFLOS model computes the turbulence and index of refraction over complex terrain and structures as a function of space and time, CHAFFSIM computes the chemical concentration motion, and SIMLIGHT models the electromagnetic energy propagation. Together these models allow simulation of imaging systems and lidar propagation through a realistic atmospheric environment. Included also is the ability to capture the dynamics of signature variability through environmental heating and cooling.

A simulation of human pattern recognition is outlined which classifies objects based on outputs of a computational vision model, called the Georgia Tech Vision (GTV) model. It is shown that the simulation is able to identify high- level features of military targets, and that identification of high-level features can be used as a tool for recognizing targets. The results suggest that the computational vision model will simplify the task of simulating target recognition by providing a 'front-end' that simulates the basic features that human observes use to recognize targets.

When an airborne imagery sensor moves from far to near, for a non-zoom imaging system, the image sequence from a scene will vary with different scales dramatically, state-of-the-art methods based on a single invariant feature or a simple feature, such as moments invariant, shape specific points, topological features and Fourier descriptor, etc., are rendered useless for representing and recognizing a multiscale object in this specific image sequence. Even the image gray-pyramid technique, which has great potential for pattern recognition by template matching with different resolutions, can not provide satisfactory performance due to not knowing exactly the resolution of real images, so there is an increasing need for improvement in multiscale object rendering and recognition. In this paper, we develop a class of algorithms for representation and recognition of a multiscale object in the specific image sequence taken from a sensor moving from far to near, which is called hierarchy features model (HFM) and sequential object recognition algorithm (SORA) based on this hierarchy features model, respectively, and intended to represent a size-changing object and recognize it. Experimental results with many real visual and infrared images and simulated images have shown that when a non-zoom imagery sensor moves from far to near, the HFM is suitable to represent a multiscale object, and the SORA available to recognize it.