Geometric subdivision of the interferometer angular field allows essentially unrestricted total solid angle, leading to the superthroughput interferometer. In conjunction with a matrix detector array and specialized processing, this approach permits spectral imaging over large angular fields with virtually the same spectral resolution for all angular resolution elements. The use of correlation spectrometry in the output interferogram greatly simplifies the data processing aspect and permits the simplification of the output presentation to indicate the presence of key spectral signatures within the field with strong background suppression. Experimental data are presented indicating the quality of spectral imaging over the field. The change in optical path difference with angle in the field produces a corres-ponding shift in spectral calibration. This can be corrected by adjusting the reference scale or programming the spectral shift into the correlation functions. Higher frame rates and improved contrast can be obtained by limited scanning at selected offset positions in the wings of the interferograms, and trade-off between processing flexibility and false alarm rate is necessary. Since spectral correlation is done electronically, "filter" functions can be arbitrarily defined, allowing multiband sensing and simultaneous multiple channels, as well as the discrimination of spectral lines. Rapid remote reprogramming of functions is possible, consistent with processing complexity.
It is likely that future high resolution earth observation imaging systems will utilize self-scanned IR detectors. In an initial step toward this goal, an IR imaging system operating in the 10 to 12 micron spectral region has been developed. This system uses a 9-element Hg Cd Te/ CCD linear array operating in the photoconductive mode, nine pre-amplifiers and a silicon CCD multiplexer integrated into a focal plane assembly. Opto-mechanical techniques are used to scan the scene and images are produced in real time. The imaging performance of this system is described and measurements of noise, responsivity, D*, and detector sensitivity profiles are presented. The requirements for more advanced detector arrays for use in future NASA remote sensing missions are also discussed.
Thermal infrared Pushbroom scanners being developed for NASA's earth resources survey experiments in the middle to late 1980's offer high spectral, spatial and temporal resolution, and high reliability through design simplicity. This mode of operation does not require moving optics; has integral chopping and calibration, and consequently, the Pushbroom scanner is lighter, simpler and more compact than its electromechanical predecessors. This paper describes the development of a 90 linear element, 8-14 micrometer, photoconductive (PC) (Hg,Cd)Te, IR/CCD/MUX Pushbroom Field Test Instrument.
Recent advances in integrated focal plane array technology, wide-field-of-view optics, and lightweight structures have led to a review of the tradeoffs between wide-field imagers for earth. resources applications.
In this paper analytic and numerical methods are presented for the design of a staring mosaic sensor system. The calculations are simplified by using the properties of two-dimensional Fourier transforms. The methods permit investigation of the dependence of background suppression and signal-to-noise ratio on characteristics such as: non-isotropic background with two-dimensional spatial power spectral density, background drift rate, optical blur, detector footprint, electronic sensor noise, and target phasing and velocity. Representative examples are discussed along with sample numerical results.
The design and operation of an imaging Michelson interferometer-spectrometer used for near-infrared (0.8 μm to 2.5 μm) spectral imaging is reported. The system employs a rapid scan interferometer modified for stable low resolution (250 cm-1) performance and a 42 element PbS linear detector array. A microcomputer system is described which provides data acquisition, coadding, and Fourier transformation for near real-time presentation of the spectra of all 42 scene elements. The electronic and mechanical designs are discussed and telescope performance data presented.
Analytical expressions have been derived to simplify the design process for integrated focal planes. The focal plane parameters (detector size and CTD integration interval) are computed which peak system performance at a specific spatial frequency (fd). The spatial frequency is selected on the basis of target size and range and on an empirically derived relationship between recognition probability and cycles across the target. It is shown that a rectangular detector should have an angular subtense equal to 0.371/fd and that an ellptical detector should have a maximum dimension equal to 0.427/fd. In the case where a focal plane CTD integrator limits the electronics bandwidth, the optimum integration and sample interval is also 0.371/fd indicating that best performance is at one sample per detector dwell time. If a boxcar function is used in the scan direction for the display, and has a hold interval equal to the integration interval, the optimum integration and sample interval is 0.269/fd or 1.4 samples per dwell for the rectangular detector. The elliptical detector has slightly improved MRT and is slightly larger than the optimum rectangular detector. It has fabrication advantages for focal plane arrays using photovoltaic detectors. Signal aliasing is treated as a random noise source based on the fact that its impact on signal distortion is a function of the arbitrary phase relationship of the signal frequency and the sampling frequency. This model can be used to predict the MRT loss due to aliasing as a function of spatial frequency. At the design frequency (fd) a 5 percent loss in MRT is obtained for a rectangular detector and one sample per detector dwell time.
The potential for developing infrared imaging systems with both active and passive capabilities is examined. Active and passive systems are compared and several approaches to dual active/passive systems are outlined. Baseline systems are postulated and their performance is analyzed. Technology and utilization considerations are discussed.
The optical design considerations for a 10 micrometer scanning catadioptric optical system are discussed. The system scanning concept is an extension of previously developed can methods. Although systems of this form have traditionally been refractive, certain situations permit use of a reflective Cassegrain as the prime collecting optics. The Cassegrain with two piano convex germanium elements represent the simplest and fewest number of components with which to accomplish the task. The design and assembly philosophy, tradeoffs, and final optimized design are reviewed. The system was fabricated, and performed as predicted.
An important feature of the diamond-machining process is the ability to accomplish both high-precision machining and optical finishing on the same fixture and to like tolerances. Thus highly accurate mechanical relationships can be built into an optical component which, while increasing part and fixturing complexity, provide a much simpler overall system design It is this reduction in system complexity as it relates to system assembly and alignment, that provides the subject for this paper. Types of features that can be introduced, their typical accuracies and examples of their use will be presented.
An optical design was done for a system having a wide field of view with diffraction limited resolution over a 5° by 10° FOV. The length of the system including a baffle was set such that the earth does not illuminate the primary mirror at an angle of 1.66°. The system was analyzed using APART and, with the results of the stray light analysis, improvements were made allowing a focal plane irradiance of less than 10- 11 W/cm2 from the earth as a thermal source. Thermal self emission was also analyzed using APART allowing the temperature of the telescope walls to be specified.
Passive infrared imaging systems produce a signal in which the amplitude at a particular spot is related to the radiance of the corresponding spot in the scene viewed. The differences in signal levels in different areas of the image is usually interpreted in terms of radiance temperature differences in the scene viewed, and may be converted to true radiance temperatures if the scene includes an object whose radiance temperature is known. The radiance temperatures are usually converted to true temperatures by correcting for the emittance of objects in the scene. This would be correct in the absence of reflected ambient flux. However, for scenes at ambient temperatures, ambient flux is always present in significant amounts. Temperature errors due to reflected ambient flux are discussed from a theoretical standpoint, and a procedure for experimentally evaluating the ambient flux is suggested.
The position location accuracy of a point source in the fields of view of two telescopes with the same focal plane is considered. It is shown that for detectors sized to equal the distance to the first zero of the coherently added system, there is little, if any, difference in accuracy between a system that adds the flux coherently as opposed to a system that adds it incoherently.
A simulation model has been developed to analyze Infrared Imaging systems. This model synthesizes an optical point spread function and convolves it with target and detector configurations to produce a target signal. The model includes the effects on the signal of detector responsivity contours, varying target shapes, noise and various electronic filtering and signal processing methods. The model runs on CDC 6600/7600 computers and requires about 40 seconds processor time per run when exercising all options. Outputs are in the form of tables, printer plots and calcomp plots.
A visible-to-infrared image converter based on the Hughes 2-in. liquid-crystal light valve is under development at Hughes Research Laboratories. The system is designed to operate in the 2- to 5-μm and 8- to 12-μm spectral ranges. The device consists of an IR-visible transparent input substrate electrode; an IR-transparent, visible-sensitive photoconductor (CdS); and a thin layer of liquid crystal acting as the light modulator medium. The readout window is transparent in the IR range used. The required IR dynamic scenery, generated for convenience in a visible form (e.g., on a CRT), is projected onto the input window, thus activating the photoconductor. This spatial voltage pattern of the required image is transferred to the liquid crystal, thus converting the image to a birefringent spatial modulation. The polarized infrared readout beam projected through the light valve is modulated as it passes through the liquid crystal. The latter operates in the 90° twisted nematic mode. The required IR scenery is then formed as the polarization-modulated IR beam passes a 90° cross-polarizer. The device offers the advantages of (1) maximum flexibility in presenting versatile dynamic IR sceneries and (2) high resolution and contrast. Good performance was demonstrated by a device operating in the 2- to 5-μm range. The proposed approach strongly benefits from the technology that has been developed for the visible-to-visible liquid-crystal light valve over the last few years.
A method for evaluatina the contribution of background flux reflected from a sample is shown. An example is described showing that the reflected background contribution in a thermographic examination is about 2% and the effect of opening a window is about 0.5%.
A mosaic detector array which uses a charge coupled array to integrate and multiplex signals is significantly limited by charge saturation. Detection of warm targets by infrared sensing through the atmosphere tends to be limited more by clutter than by random fluctuations. For missions at television frame rates the important attributes of the detector array are good spatial resolution and uniformity of response. Responsive quantum efficiency is relatively unimportant. Operation at 3 to 4.2 μm offers advantages of resolution and larger rate of change of radiant signal with temperature.
A passive visible to infrared image transducer is described which allows practical simulation of high resolution dynamic infrared imagery for the first time. A theoretical performance analysis is presented which indicates an MTF of 50% at 5 cycles/mm is achievable with a time response of less than 20 msec and a required input power of less than 2 mw/cm2 per simulated degree ΔT. Experimental results on working devices are reported which match theoretical predictions. Successful simulation of high speed real-world infrared imagery is described. Suggestions for future work are presented which should further enhance resolution and dynamic range.
A comprehensive model which simulates the signature radiance from a discrete cloud of aerosols and vapors suspended in the atmosphere, for a completely general configuration of cloud, observer and sun, has been developed. The program of calculations has been computerized for validity over the range 0.25 to 28.5 µm, as it uses the atmospheric transmittance/radiance model LOWTRAN. Future extensions of the model are discussed.