The application of mosaic focal planes to the problem of surveillance has resulted in increased interest in sources of background clutter. An ideal staring sensor would be able to perfectly suppress spatial variations in the background scene. This can be accomplished; for example, by simple frame subtraction. However, any practical sensor will be limited in its suppression capability by platform motions. The result is that spatial variations are seen as temporal change. A variety of on and off focal plane background suppression algorithms have been proposed. In this paper causes of earth background clutter, particularly in the thermal infrared region are discussed. Specific examples of high clutter scenes such as: edges, anisotropic backgrounds, and inhomogeneous backgrounds are discussed both in terms of their effects on generic processors and in terms of the underlying scene physics. Examples are presented from available thermal band infrared data where possible. The scene physics based on differential insolation, material differences, heat transfer, altitude differences, as well as temperature differences is treated via the examples.

A technique is described for suppressing the lower spatial frequencies in the object using an optical system with a time-varying MTF. The technique is applicable to imaging radiometers operating in the infrared region. The required variations in the MTF are produced by causing the image to move in a programmed sequence of small amplitude circular gyrations on the face of the detector. The gyrations are performed at a frequency high compared to the system frame rate. A series of calculations are described which indicate that, at low spatial frequencies, average suppression ratios of at least 500 to 1 can be achieved with this technique.

A 32 channel charge coupled device (CCD) analog multiplexer used in a modular staring mosaic sensor design is described. The multiplexer Z-packaging technology used in the module allows a unique degree of freedom in the multiplexer design because of the negligible influence of chip size on sensor active areas. Integrated along with the multiplexer function are 32 adaptive bandpass filters, utilizing switched capacitors, and a complementary metal-oxide-semiconductor (CMOS) 1-of-64 decoder. Both features are described and test results are included. Future activities to enhance performance and to provide a compatible photovoltaic detector interface are also discussed.

The objective of the program described here is to arrive at the optimum characteristics for an agile step-stare IR sensor with a given performance. The resulting optimized design depends upon technology projections, cost models, and fundamental formulas relating the various sensor cnrocterastics to the sensor performance. Since the projections, models and formulas are approximate the resulting sensor design is also approximate. In addition to establishing approximate point designs and technology requirements, the purpose of this program is also to provide an objective basis for comparing competing concepts and assessing- the payoffs associated with various technology development programs.

The characteristics of a mosaic IR sensor used in the step-stare mode impose some surprising constraints on scheduling measurements. This paper assumes that the mission data requirements for a hypothetical measurement program have been translated into a prioritized list of measurements and a sample of scenes for each measurement. It then examines how the characteristics of a step-stare mosaic sensor impose requirements on the distribution of candidate sites. It also examines the problem of scheduling sites from the list of candidates for a set of measurements.

A sensor data processor is under development designed to handle pixelized frames of digitized data from mosaic-type "staring" or "step-staring" sensors as well as from scanning sensors. The intent of the Sensor Data Test Program is to provide a convenient and rapid method of accomplishing the following tasks: (1) evaluation of bulk filter, track collection filter, target classification and motion estimation algorithms with real data frames; (2) evaluation of these same algorithms with simulated data frames; (3) evaluation of the effect of sensor characteristics and pre-filter algorithms; and (4) operation of optimum algorithm sets with real data frames to obtain and classify targets and estimate target motion characteristics. The processor and some major operation tests are described in this paper and some conclusions are drawn from the test results. This is the first of several papers which will detail the Sensor Data Test Program, outline tests made with it, and describe significant conclusions drawn from the test results. In its final form, this program will be capable of iterative sensor data manipulation and program optimization as well as Monte Carlo type statistical sensor system evaluation.

Rockwell has developed under a U.S. Army contract an advanced IR Imaging Seeker System which is compatible with the Hellfire Missile System mission. A technical overview of this program and current status will be presented. The IR imaging seeker was tested during late 1979 and early 1980. This seeker utilizes a Rockwell-developed 1024-element InAsSb/silicon hybrid focal plane array (FPA) operating at 77°K and IR sensitive in the 3.4-4.0 micrometer wavelength region. A multimode tracker provides improved tracking capability for operation against targets in a high clutter background. The payoff for the technology being developed on this program is the ultimate realization of reduced IR seeker cost at equivalent or increased performance to present-day IR Imaging seeker designs.

Laminated mosaic modules, a z-technology are introduced as an alternative architectural configuration for mosaic sensor focal plane applications. The module features a 4096 element (8x512) PbS array with 64 thirty-two input signal processing chips. Modules can be stacked into subarrays without a composite array fill factor loss. Measurements and results confirm the performance status of this technology development approach. Module components are discussed first followed by the module assembly and the subarray (a four module stack).

An automatic method for track assembly is presented which can initiate new tracks, assign new measurements to the appropriate track, and eliminate false tracks due to clutter. Consecutive frames of thresholded infra-red images are combined and then screened using a three-by-three window. Successive levels of data processing reduce the noise still further, and in a cluttered environment (noise in 3% of the pixels) , it may be necessary to use three or more levels to eliminate false tracks due to noise. Curved paths and crossing tracks are handled without difficulty, and track-like noise can be eliminated by preprocessing. A special coding scheme has been implemented on an array processor to obtain a processing rate of 160,000 pixels per second, independent of the number of targets. This paper develops theoretical calculations for the probability of false tracks and presents verifying experimental results.

Presented here is a review of an inhouse image processing facility, which has been improved for NASA/GSFC, to reduce raw 8-bit data tapes from the 90-element pushbroom IR/CCD/MUX field test instrument to 90 x 512 pixel frames of thermal infrared (8-14 micrometers) hard copy imagery. The 512 pixel length (160° panoramic scan) frames are recorded in 16 seconds (maximum rate) on magnetic tape. These tapes are then taken to the image processing laboratory where second order temperature calibration and image enhancement techniques are used to generate "quick look" 8 x 8 dot matrix line printer hard copy images or CRT displays in under 30 minutes. Final enhanced image data tapes are then recorded on photographic film using a digital film writer which quantatize the data into 256 grey levels. High quality, 16" x 20" photographic enlargements of selected imagery are made from these negatives using standard photographic techniques. The first hard copy imagery was obtained July 15, 1980. Several 90 x 256 pixel frames were recorded and image processed. Hard copy is from the 8 x 8 dot matrix printer. Initially, only warm objects were recognizable in the unprocessed data; however, once the temperature range of the imagery was adjusted and calibrated against three blackbody temperature frames, considerable detail was resolved in human forms and facial features. We are currently optimizing the integration time, chopper/signal electronics phasing, and CCD bias and the application of the image processing programs described herein.

This paper describes the generation of synthetic, large array, digitized, two-dimensional in-band radiance scenes for below the horizon (BTH) and above the horizon (ATH) viewing. A generated BTH scene for the case of multilevel clouds over an ocean-terrain interface includes edge effects. Edge geometry is specified from a large FOV visible photograph. Mean values of in-band radiance are determined from a radiation transport code. Fine-scale structure is added from prescribed statistics. The phenomenology base includes sun scattering and thermal emission effects for clouds and ground. A typical ATH irradiance scene in-band for a low threshold detector contains the earth's limb contribution and, shining through the limb, a few high threshold catalog stars and high density, weak, randomly generated stars. One-dimensional iterative and analytic statistical models are presented and compared with data for, respectively, (1) a detected radiance trace and (2) the standard deviation of the radiance and for power spectral densities of particular scene types. The analytic PSD models are a superposition of Wiener spectra, each of whose asymptotic forms is a power law of frequency to the minus two power.

The present status of calculating non-homogeneous rocket plumes that emit, absorb and scatter radiation is reviewed. The "hot-through-cold" problem is also discussed.

A description is given of a practical approach for calculating emission from a source having cylindrical symmetry and containing both absorbing gases (not obeying Beer's law in general) and absorbing and scattering particles of arbitrary size distributions. Two levels of sophistication are described: (1) a two-flux solution which is obtained by direct path integration and (2) a six-flux solution which requires a prior solution of a reduced-size equation for an isotropic source term. A generalization of standard random gaseous band models is provided to include scattering by use of empirical forms.

Two approximations are developed for estimation of background clutter-induced noise in the Nth-differenced output of a single pixel of a drifting, jittering optical sensor. The viewing geometry and coordinate transformations are reviewed and an exact integral solution in the frequency domain is given. Its basis, an analysis done by Williams and Fried, is summarized in an appendix. Then the approximate forms are derived and their limitations discussed. The approximations are readily scalable and lend themselves to iterative, optimization-loop type performance calculations. Their applicability is limited to sensor LOS motion of a fraction of a pixel per frame time. Within that domain agreement with exact solutions is good.

Down-looking staring mosaic infrared measurements of an USAF F-15A aircraft at 50 kft were made as part of the May 1979 measurement series of the Balloon Altitude Mosaic Measurements Program (BAMM). Staring measurements were obtained from a balloon-borne platform floating between 70 and 88 kft and containing a 4X4 detector mosaic dual filter radiometer, a 4X4 detector mosaic Michelson interferometer, both boresighted with a real time TV scene-monitoring system. Traverses by the aircraft through the TV and infrared instruments' fields of view included flybys at 28 and 50 kft altitude, three of which were in after-burner. The entire spectral infrared signature of the aircraft in the 2.5 to 5.5 micron region was measured. Signals of aircraft transits through the infrared instruments exhibited correlation with visual recorded data. Aircraft infrared signals were further enhanced using standard second order differencing, bandpass filtering, and pixel-to-pixel subtraction techniques.

Focal plane staring mosaics for real-time source classification purposes must be used in such a way as to provide a compromise between the usefulness of the extracted discrimination features and the frame by frame classification rate. This paper surveys various mosaic layouts and scanning patterns with the aim of optimizing the feature acquisition rate. A radial mosaic layout is studied in detail.

The phenomenon of selective reflection of incident light by cholesteric type liquid crystals has been extensively investigated during the past decades and recently led to various applications. The most promising utilisations of liquid crystal sheets are in the field of thermograohy and infra-red technology. We have been involved in the conception of a new type of cholesteric mixtures which show nearly linear temperature-color dependence. The technological difficulties inherent to the realisation of several thin layers of microencapsulated liquid crystals layers on a soft polyester support are now solved. Sheets of large size which can be produced are expected to find applications in medicine, IR image convertion, industrial non-destructive testing.

Background data, description and operation of various refrigeration cycles, performance data, and development potential have been summarized for a number of refrigeration concepts adaptable to spaceborne operations. Highlights of these various concepts are briefly defined, some of the more significant operating characteristics are indicated, and basic integration limitations with the entire spacecraft are outlined. Performance data and development potential are summarized. These are all presented with consideration of the primary factors for selecting an appropriate cooling system for long duration space applications.

In regions of partial cloud cover, strategic aircraft will penetrate their attack corridors a finite distance before detection by an electro-optical sensor. Malick and Allen have quantified by geographic region the probability of an E-0 sensor establishing a track on an aircraft flying under partial cloud cover. This paper shows the importance of cloud distribution within the critical region and quantifies expected penetration distance. A statistical analysis has been developed which is based on cloud cover data from the USAF ETAC 3-DNEPH grid. Expected penetration distance before detection is presented for a variety of geographic regions, seasons, times of day and attack altitudes. Excellent detection of a raid is shown regardless of penetration altitude, and particularly cloudy attack routes are identified.

High altitude infrared background data in the middle infrared were obtained during a series of balloon flights made several years ago. One aspect of these measurements that was difficult to explain physically was the large short-term fluctuations in the background radiation observed occasionally in Alaska. During a series of balloon flights made in 1978, data were obtained which appears to explain the observed fluctuations.

Development of solid cryogen coolers for spacecraft instrument cooling has been in progress since 1965. To date six coolers of this type have provided orbital cooling for 7 to 12 months in a temperature range of 60 to 130 K. These systems have demonstrated the high reliability and excellent temperature stability expected. This paper summarizes some of these data and describes some of the more recent developments in solid cryogen systems.

The scattering of infrared radiation from ZnSe was measured for three laser wavelengths (0.6328, 3.39, and 10.6 μm) and three material thicknesses (6 mm, 19 mm, and 25 mm). The data from these experiments are shown and a scattering model for this material is presented.

Infrared surveillance systems are often required to discriminate targets powered by combustion devices, such as rocket or jet engines, from the highly cluttered backgrounds of the earth-atmosphere scene against which the target must be viewed. The use of a Fabry-Perot etalon to perform this discrimination function based on the spectral differences between the source and background has been studied both experimentally and analytically. The quantitative results of this study are the subject of this paper. The discrimination factor for hot water sources, based on high resolution spectra in the 2.7μm region is found analytically to be less than 5 and accompanied by a large attenuation of the source signal. Measurements in our laboratory shows that the experimentally observed discrimination factors are even smaller. On the other hand an HF spectrum studied analytically shows discrimination factors as high as 50 may be achieved with as much as 50% of the source energy transmitted by the etalon. These different results are a consequence of the fundamentally different nature of the H2O and HF spectra. Based on these results further studies of selected other plume species, such as CO and CO2 near the head of the 4.3 μm band are suggested. Spectral discrimination of plumes containing hot water should be based on techniques other than an etalon.

Applications of infrared sensing technology to meteorology are presented with the emphasis on meteorology in lieu of instrumentation. Included are the use of thermal imagery from satellites, atmospheric temperature sensing from satellites, and the use of data from other satellite and non-satellite sensors.

A digital test target system is described which may be used to evaluate digital imaging systems and line-scan and digital displays (soft or hard copy) in a fashion similar to the use of resolution targets for the evaluation of optical analogue image systems. The digital test target system consists of a basic pattern of checkerboards of different sizes, which is repeated with different commanded "brightness" and "contrast" levels; a procedure for reading the target; and a method of adjusting the reading scores to obtain an index of image quality. The target system presently consists of an eight-target set, arranged for evaluation of 512 pixel square matrix display having eight bit gray level capability. The system is applicable to other sized display matrices and gray level capability. An index is derived from the readings so that displays having similar image quality index values will have the same average image quality performance regardless of differences in matrix size and nominal gray level capability.

Helicopters comprise over two-thirds of the Coast Guard's total aircraft fleet and are the primary search and rescue vehicle. Present Coast Guard helicopters have, at best, a weather/navigation radar capable of detecting only relatively large targets. Both helicopter models can carry a high intensity searchlight but its narrow field of illumination limits the effective search area. Beginning in 1973 the Coast Guard examined a variety of technologies in an effort to improve night search capabilities. FLIR emerged as the standout performer, and a FLIR system is presently under development for installation on an HH-52A to prove the concept in operational conditions. The expected impact on the Coast Guard's Search and Rescue, Enforcement of Laws and Treaties, and Marine Environmental Protection missions is enormous. This paper examines design and performance tradeoffs for the Coast Guard FLIR development.

This paper will describe the design of two infrared optical pyrometers used to study dust explosions. One is a six-wavelength pyrometer that can measure both dust particle temperature and gas temperature by observing continuum and gas band emission in the 1 to 5 μm infrared region using PbSe detectors. The second pyrometer measures particle continuum emission in the 0.8 to 1.0 μm region of the near infrared using Si detectors. Temperature data are presented for both laboratory and full-scale mine explosions.

Experimental measurements of infra-red transmittance are reported, which show that the computer code LOWTRAN-4 predicts too high a transmittance in the region 4.3-5.5 micron, and that the discrepancy is dependent on the water vapor content. We have carried out experiments at various ranges up to 45 km in order to vary the precipitable water amount over a wide range, and from the results we have deduced that the anomalous transmittance results from an incorrect estimate of the continuous absorption, while the spectral component is substantially correct.

The design and manufacture of dielectric-film interference filters for cooled. FIR astronomy is described. The bands are 16.5-21.5 μm, 17.5-19.5 μm, 19.5-21.5 μm and 27 μm cut on. The films are PbTe/CdSe and the substrates are CdTe (some 12 mm thick), without absorption in the region: KRS-6 films are used for antireflection.