 |
Description  |
|
|
FIELD OF THE INVENTION
The present invention relates to object imaging, detecting, or locating
apparatus or interferometer spectrometric apparatus responsive to
electromagnetic radiation (such as infrared or visible light).
BACKGROUND OF THE INVENTION
It is well known in the art that the resolution of apparatus responding to
electromagnetic radiation in the optical (or near optical) regions, can be
improved by increasing the aperture. Likewise, the signal-noise ratio of
the apparatus can also be improved by increasing the aperture.
Furthermore, if the apparatus is to detect targets against an irrelevant
background, then the ratio of target signal to background clutter may be
improved by improving the resolution.
However, an increase in the size of the aperture results in an increase in
the size and weight of the instrument with corresponding disadvantages;
particularly the difficulty of mounting and transporting the apparatus.
For example, if the apparatus is to be installed in a vehicle such as a
space satellite, a missile or an aircraft where space is limited and where
each kilogram of payload requires many kilograms of vehicle and
propulsion, then size and weight are critical attributes having a great
effect on the cost and practicality of the overall system. In some
particular applications, no vehicle may be available which can accommodate
the size of the aperture required for a particular mission. In such a
case, the practicality and cost of the overall mission may be driven by
the sensor requirement through development of a special vehicle or the use
of extraordinary deployment techniques. In earth-bound applications,
similar considerations apply. For example, in an astronomical application
it may be desirable to limit the size of the apparatus which is traversed.
It is also well known in the art that the signal to noise ratio of optical
instruments which must gather data on an image within a limited period of
time can be improved by increasing the number of radiation detectors used
to transduce the electromagnetic radiation to electrical signals. However,
an increase in the number of such detectors entails similar disadvantages.
For reasons such as are briefly set out above, therefore, it has long been
a goal in this art to obtain the ultimate performance from these
instruments. The prior art evidences the use of a technique referred to as
synthetic aperture which is believed to be useful in attaining high
performance from instruments of this sort. In this regard, see the Meinel
et al. article entitled "A Large Multiple Mirror Project", appearing in
Optical Engineering, Volume 11, Number 2 (March/April 1972), pages 33-37,
and the article by Meinel entitled "Aperture Synthesis Using Independent
Telescopes", appearing in Applied Optics, Volume 9, Number 11 (November
1970), pages 2501-2504. In the latter article, the author defines aperture
synthesis as "occurring when separate optical elements are combined with
phasing to form a common image field in which the resolution is greater
than that for a single element".
In 1890, A. A. Michelsen (Phil. Mag. (5), 30, 1) utilized two auxiliary
mirrors to augment the effective diameter of the 100 inch telescope at
Mount Wilson to increase its resolution for measuring the diameter of
celestial objects. The two mirrors were moved laterally until the contrast
of fringes in the pattern of interference of light from the two apertures
was extinguished. The separation of the mirrors was then a measure of the
diameter of the object.
J. S. Wilczynski, in U.S. Pat. No. 3,556,630, discloses a "Method and
Apparatus for Obtaining, by a Series of Samples, the Intensity
Distribution Across Sources of Incoherent Electromagnetic Waves to Produce
a Single Composite Picture". He discloses how a desired "large" aperture
area can be subdivided into a plurality of smaller apertures of equal
size. The image that would be obtained by the "large" aperture is then
derived using only a pair of the smaller apertures by physically
relocating the pair of small apertures to a number of different positions.
However, due to the necessity for mechanically relocating the plurality of
mirrors to a precision comparable to the wavelength of the radiation
sensed, which requires moving substantial masses and which takes time,
during which time the image must be fixed or substantially fixed,
applications for this technique are limited.
From the foregoing it should be apparent that there is still a need to
increase the ability to extract information from a given sized aperture or
from an aperture of a given configuration and from a given array of
detectors.
In recent work (see J. A. Jamieson, "Passive Infrared Sensors: Limitations
on Performance", Applied Optics, Volume 5, page 891, April 1976), I have
shown that a passive sensor gathers information in four dimensions of
time, wavelength, and two angles, that frequently, information gathered in
one dimension is not required to fulfill the mission of the sensor, but
that the information acquired in another dimension is inadequate. A
principal purpose of this invention is to allow information gathered in
wavelength to be utilized to enhance information gathered in angle. That
is, it is an object of the invention to use data measured by the apparatus
which characterize the spectral distribution of radiance of a radiating
scene to augment data measured on the scene in angle or location in the
object surface. Another object of the invention is to provide apparatus
which can be used adaptively for spectral resolution or spatial resolution
beyond that ordinarily available or any intermediate combination of
spectral and spatial resolution without the need to reconfigure the
apparatus.
Another principal object of the invention is to provide apparatus which can
yield spatial resolution beyond that ordinarily available from a given
size of aperture. Another principal object is to provide apparatus which
can measure an image of an object scene at a resolution beyond that
ordinarily available from a given number of detectors. Another object is
to provide apparatus in which the spatial response can be adjusted readily
without the need to reconfigure the apparatus (e.g., to emphasize high
spatial frequencies for edge sharpening or alternatively in the same
instrument at another time to treat all spatial frequencies equally for
best fidelity of response). Another object of the invention is to provide
apparatus in which failure or degradation in performance of some of the
detectors causes a minimum impact on the capability of the apparatus to
yield a complete image.
A particular advantage of the invention is that the entrance pupil is
incompletely filled so that other apparatus required for an overall
mission in which this apparatus is used can be colocated with the entrance
pupil or alternatively the entrance pupil can be distributed (e.g., on
either side of the nose of a missile or on the wings of an aircraft) so as
not to interfere with other functions of the overall system. Another
advantage is that the radiation detectors in the apparatus are not
required to fill the focal plane so that space remaining can be utilized
for other functions such as electrical connections, pathways to remove
heat, to minimize crosstalk, or for preamplifiers, charge-coupled devices,
and other auxiliary electronic apparatus, or for redundant detectors.
Another advantage of the invention is that it provides an improved
capability to discriminate a target object from a structured radiant
background. This advantage results initially from the improved spatial
resolution of the apparatus but is further enhanced if the target object
has a different spectral distribution of radiance than the structured
background.
A further advantage is that the invention reduces the need for accurate
a-priori knowledge of target and background signatures before a system
using a sensor is designed or committed to its mission. This advantage
derives from the adaptive nature of the spatial/spectral response of the
invention. This attribute allows the spectral response to be modified by
reprogramming the data processing without reconfiguring the apparatus
after the system is committed to its mission. If a mission should require
the sensor to perform well against several kinds of targets at different
times, the response can be successively altered an unlimited number of
times rather than adopting an inferior fixed compromise response.
A further advantage is that the system designer can select the maximum
dimensions of the entrance pupil independently of the collecting area of
the entrance pupil so that he may select resolution and sensitivity
independently to achieve a balanced, economical design.
SUMMARY OF THE INVENTION
The present invention meets these and other objects of the invention by
providing apparatus which first collects and modifies incident radiation
by optical and interferometric means and which subsequently transduces the
received radiation (e.g., infrared, visible light, or other forms of
radiation) to electrical signals and which further includes data storage
means and signal processor to operate on the signals provided by the
transducer to create data corresponding to an image with resolution
increased above that which is normally associated with the aperture. The
display apparatus, which can be a conventional display, can then be
employed to display an image from the data generated by the processor or
an alarm can be provided to be energized if the data corresponds to a
target of interest.
One particularly significant aspect of the invention is the manner in which
the radiation is sensed and collected. Electromagnetic energy is imaged,
by a split or distributed aperture of predetermined size, spacing and
orientation, onto a focal plane array of detectors. The path lengths from
several parts of the aperture to the focal plane array are variable in a
regular, periodic fashion. Interference effects produce a diffraction
pattern at the focal plane which is a function of the optical path length.
More particularly, the split aperture may include a plurality of
symmetrically disposed aperture elements which can be thought of as plural
pairs of aperture elements. The phasing of radiation received at the
detector array from pairs of elements is adjusted to produce different
diffraction patterns as a function of the phase adjustment. For example,
the path lengths between different elements in a pair can be
differentially varied. Since different path length variations produce
different phase adjustments for different spatial frequencies, the
variation of path lengths "encodes" different spatial frequencies allowing
them to be separately detected. In this fashion, the array response can be
analyzed with regard to the plural spatial frequencies of the incident
radiation. Selected responses are recorded in a data storage device.
Processing apparatus responds to the stored data and produces an image or
image-like signal.
The present invention, therefore, relates to an image-forming method and
apparatus, useful in the study of objects radiating, reflecting or
otherwise transmitting spatially incoherent electromagnetic waves. The
method and apparatus of the invention can be used or adapted in obtaining
images of spatially incoherent electromagnetic radiation in astronomy,
microscopy, remote sensing of earth resources and the like. The invention
is further valuable where the size of the instrument is limited but
resolution is an important characteristic. It would also be useful in
quasi-imaging applications, such as surveillance, to detect unauthorized
or hostile events. The invention would also be valuable in instruments
based on earth for astronomy or surveillance of space.
The inclusion of signal processing means in the apparatus of the invention
allows the invention to display an adaptive feature in which the same
apparatus can be used to achieve a continuously variable mix of spatial
and spectral resolution. More specifically, the signal processor can be
instructed to generate a display having very fine spatial resolution at
the expense of spectral resolution. On the other hand, the processor can
also be instructed to resolve spectrally at the expense of spatial
resolution. In one limit, the instrument can exhibit very fine spectral
resolution, comparable to a Fourier transform interferometer at less than
its potential spatial resolutions. At the other limit, the same apparatus
can achieve very fine spatial resolution but no spectral resolution.
In one specific embodiment of the invention, wherein spatial resolution is
enhanced at the expense of spectral response, the different array
responses are processed so as to produce an image whose spatial resolution
is greater than that available from the entrance aperture employing prior
art techniques.
By changing the processing, the spectral and spatial resolution of the
instrument can be varied to produce selected effects.
BRIEF DESCRIPTION OF THE DRAWINGS
The present invention will now be described in conjunction with the
attached drawings, in which
FIG. 1 is a schematic showing of the inventive apparatus, partly in block
form,
FIGS. 2A and 2C show Modulation Transfer Functions of various apertures,
FIG. 2B illustrates an exemplary split aperture of simple form,
FIGS. 2D-2I are schematic illustrations of a target, the response produced
by individual components thereof, and different illustrations of the
combined response,
FIGS. 3-5 show a preferred embodiment of the entrance pupil, optical paths
and detector array,
FIG. 6 is a functional block diagram of operation of the invention;
FIG. 7 illustrates an interpolation waveform for a preferred detector
array,
FIGS. 8 and 9 illustrate the responses to respectively two and three source
targets,
FIG. 10 shows the Modulation Transfer Function for this embodiment,
FIGS. 11-13 relate to a second preferred embodiment and illustrate,
respectively, the entrance pupil, the detector array and the Modulation
Transfer Function,
FIGS. 14, 15, 16A, 16B, 17A and 17B relate to a modification of the second
embodiment illustrating respectively the relation between target, entrance
pupil and detector array for subfield multiplexing, the frequency
multiplexed spectrum at the detector, a plan view of the entrance pupil, a
side view illustrating the relationship between entrance pupil, field of
view and phase retardation devices, a detailed cross-section of the
optical components of one leg and a side view thereof, and
FIG. 18 illustrates the processing steps.
DETAILED DESCRIPTION OF THE INVENTION
The present invention can employ various forms of electromagnetic radiation
for detection or imaging of objects. Two illustrative examples will be
disclosed herein employing visible or infrared devices. Those skilled in
the art will understand that the form of the optics using mirrors and
photoelectric detectors are specific to this example. Other forms of
optics such as lenses and other forms of detectors would be appropriate to
other regions of the electromagnetic spectrum and dimensions will depend
upon the wavelengths of radiations used.
In general, the invention comprises six principal components illustrated
schematically in FIG. 1. An optical train is provided for collecting and
focussing the radiation (mirrors 1, 2 and 4). Phase retardation is
implemented by relatively small, flat mirrors 3 translated back and forth
by drive mechanisms such as a piezoelectric drive, electromagnetic drives,
or lead screws forming a second component. The phase retardation device
provides for varying the optical path length in a predetermined regular
fashion. An array of radiation detectors such as photoconductive or
photovoltaic or thermal radiation detectors 5 is a third component. A
memory or data storage device 6, such as a charge-coupled device or an
electronic memory, to retain the array of signals generated by each of the
radiation detectors at the several positions of the phase retardation
devices is a fourth component. A signal processor 7, the fifth component,
is included to operate on the arrays of signals successively collected in
the memory. This component may comprise a computer, a mini-computer,
several microcomputers or a special purpose processor. A display 8 or
alternatively, or in addition, an alarm device 8, if the apparatus is used
in a warning or surveillance system is the sixth component. The display 8
may, for instance, be a conventional cathode ray display or a printer or
other convenient device. The processor may include a decision making
routine or threshold followed by an alarm 8 such as a light, bell, radio
transmitter, etc., which provides an alarm signal if the device is used in
a surveillance or warning system.
In some applications certain auxiliary equipment may be useful, including
optical baffles to minimize interference by stray light; a laser source
and a white light source to dynamically measure the phase differences from
the apertures of the entrance pupil to the detector array; a programmer
containing a fast clock to synchronize the motions of the phase
retardation devices, the cycling of the memory, and the cycling of the
signal processor and display; a calibration source as a reference when the
device is used radiometrically; and a telemetry link if part of the signal
processor and display are to be located remotely from the other parts of
the apparatus. These auxiliary devices are not shown in FIG. 1.
The optical section of the apparatus receives the radiation transmitted
both from the target to be detected as well as other radiating sources in
its field of view. The optical elements comprise a plurality of mirrors
for directing the received radiation to a planar transducer array 5. More
particularly, the optical elements can comprise a first series of mirrors
arranged in a substantially planar array, comprising the entrance pupil.
Some of these mirrors are illustrated as mirrors 1 in FIG. 1. The
radiation reflected from the several mirrors 1 is directed to a second
plurality of mirrors 2 which recollimate the radiation and from them to a
third plurality of flat or corner reflection mirrors 3. The mirrors 3 are
capable of rectilinear motion along a path directed at an angle to the
radiation reflected from the several mirrors 1 which path is parallel to
the optical axis 9. The mirrors 3 then reflect radiation to a third
plurality of mirrors 4 or a single eyepiece mirror 4 depending on the
application. The radiation reflected by the several mirrors 4 is then
imaged onto the detector array 5. Particularly significant to the
invention is the motion of the plurality of mirrors 3, for this motion
varies the optical path length in a predetermined manner for reasons which
will be explained hereinafter. The field of view of the device may be
limited by conventional baffles or the like, not illustrated.
Before describing two preferred embodiments in detail, the theoretical
basis of the invention will be explained.
THEORY OF THE INVENTION
When an optical instrument with a continuous entrance pupil images a scene
using electromagnetic radiation (for example, light) of wavelength
.lambda., then details which subtend an angle smaller than .lambda./D,
where D is the maximum dimension of the entrance pupil of the instrument,
will be blurred by diffraction. For example, if a telescope in a satellite
uses a concave mirror 1m in diameter to image part of the surface of the
earth with radiation of wavelength below 1 micron, then details subtending
angles less than 2.44 micro-radians will not be distinct. If the range
from the satellite to a point on the ground is 10,000 km. then details
smaller than about 24 m. will be blurred. If it is necessary to resolve
smaller details, the prior art teaches that the maximum dimension of the
light-gathering aperture must be increased. If a larger complete mirror is
used, the weight of the telescope will be increased. For most applications
there may be practical upper limits to either weight or size that can be
used.
It is well established (see, for example, M. Born and E. Wolf, Principles
of Optics, Pergamon Press, 1959, chapter 10) that the blurring effects of
diffraction for spatially incoherent light and object and image points
distant from all points of the optics can be expressed mathematically by a
convolution integral.
##EQU1##
where Q(x.sub.1, .nu.) is the intensity in the image at the point x.sub.1
and at optical frequency .nu. resulting from the intensity of an object
P(Xo, .nu.)
x.sub.1 is the typical image angle x.sub.1 = x.sub.1, y.sub.1
x.sub.0 is the typical object angle x.sub.0 = x.sub.0, y.sub.0
dx.sub.0 denotes the increment dx.sub.0 = dx.sub.0 dy.sub.0 and the
notation * denotes complex conjugate and the function K(x, .nu.) is
related to the shape of the telescope entrance pupil G(.xi.) by
##EQU2##
Denoting the Fourier Transform of both sides of equation (1) with respect
to x by the superscript
Q(m, .nu.) = P (m, .nu.) H (m, .nu.) (3)
where H is the transform of KK*
m is a spatial frequency m = m.sub.1, m.sub.2 reciprocal to x (i.e., if
x.sub.1 is measured in radians, m.sub.1 is measured in cycles per radian).
It follows from equation (2) (e.g., see Born and Wolf, page 525) that
##EQU3##
which allows the function H to be evaluated very easily as the
autocorrelation function of the entrance aperture for each wavelength
.lambda.= 1/.nu.. Employing this relation in prior art devices is
difficult since the transducer array does not separately respond to
different spatial frequencies.
The optical apparatus which is the subject of this invention is arranged so
that the light in the image can be separated into a number of spectral
intervals or bands. These intervals are nearly monochromatic; that is, the
difference between the largest and smallest wavelength is much less than
the mean wavelength. For each interval a relationship of the form of
equation (3) will hold.
The entrance pupil G(.xi.) is divided into two or more parts separated from
each other so that the largest distance between the boundaries of the
aperture is increased. This design offers a potentially great spatial
resolution for the area (and weight) of the aperture, but because a
portion of the aperture is missing, the function KK* is a complicated
multilobed function.
More specifically, a square aperture of length 2a on a side has the
transfer function H(m) whose cross-section is shown in FIG. 2A. Contrast
this with the cross-section of the transfer function shown in FIG. 2C
related to the split aperture shown in FIG. 2B, see J. Goodman, Fourier
Optics (McGraw Hill). The significance of the different transfer functions
will now be illustrated.
FIG. 2D shows an exemplary object or target with two radiation sources and
FIG. 2E is a cross-section of the target schematically showing the two
radiation sources and their relative amplitudes. With an entrance pupil of
the form of FIG. 2B the image will have the form of FIG. 2F. This image is
ambiguous in that several locations of the two targets might be inferred
from it, only one pair of which are correct (and if the number of targets
were unknown further incorrect inferences could plausibly be made). FIGS.
2G and 2H show the image contribution from the sources x and y,
respectively, and FIG. 2I illustrates the superposed result. Thus, the
image produced by the optics at each quasi-monochromatic band is blurred
by convolution with an ambiguous function which makes it difficult to
interpret. This difficulty is expressed in equation (3) by the fact that
the function H is zero over the middle part of its range, see FIG. 2C.
A part of the product PH making up the function Q is missing. Therefore, an
inverse Fourier Transformation of the function Q does not yield a
satisfactory image. If the optical transfer function H(m, .lambda..sub.1)
were known at .lambda. = .lambda..sub.1, for .vertline.m.vertline.
.ltoreq. 2a/F.lambda. and if the transform of the distorted image Q(m,
.lambda..sub.1) were observed for .vertline.m.vertline. .ltoreq.
2a/.lambda..sub.1 F then the transform of the true image P(m) could be
found from
P(m) = Q(m, .lambda..sub.1)/H(m, .lambda. .sub.1)
and the true image could be found from
##EQU4##
where B = 2a/.lambda..sub.1 F However,
H(m, .lambda..sub.1) < .epsilon. (an arbitrarily small quantity) for some
regions of m, e.g.,
##EQU5##
where Q.sub.o = PoHo, the value at m = o but the observations are only
known to finite accuracy because of noise and measurement accuracy
##EQU6##
Therefore, the estimate P(m) of P(m) is indeterminate in some regions of
m. If the value of Q is taken to be zero in these regions, then the
estimate p(x) of the image
##EQU7##
The transform of U(m) is proportional to sinc
##EQU8##
The convolution of this multilobed function with the true image p(x) is
subtracted from the true image as a distortion.
But H(m, .lambda..sub.1) = H(m .lambda..sub.1)
for example for an entrance pupil with two elements per dimension (the
pupil of FIG. 2B)
##EQU9##
where F is the focal length of the optics following the aperture .LAMBDA.
is the triangle function
##EQU10##
If the source radiates (or transmits, etc.) energy at several
quasi-monochromatic wavebands .lambda..sub.1, .lambda..sub.2, . . .
.lambda..sub.n then a new multispectral transfer function may be formed
H.sub.s (m, .lambda.) = A.sub.1 H(m, .lambda..sub.1) + A.sub.2 H(m
.lambda..sub.2) + . . . + A.sub.n H(m .lambda..sub.n)
where it is further assumed that the additional wavelengths available are
longer than the preferred wavelenth .lambda..sub.1
.lambda..sub.1 < .lambda..sub.2 < .lambda..sub.3 . . . < .lambda..sub.n
and the quantities A.sub.1, A.sub.2 . . . A.sub.n are constant gain factors
by which the components at the several wavebands are multiplied before
they are combined.
If the quasi-monochromatic wavebands available from the source cover an
adequate span of wavelengths at sufficiently small separation and adequate
intensity then the gain constants A.sub.i can be chosen so that H.sub.s
(m, .lambda.) is sufficiently great in the region .vertline.m.vertline. <
2a/.lambda..sub.1 F so that P(m) and subsequently p(x) can be estimated
properly. For example, for the aperture of FIG. 2B with area 4(a -
b).sup.2 to achieve resolution equivalent to a single continuous aperture
of area 4a.sup.2 at the shortest available wavelength .lambda..sub.1 is an
improvement in linear resolution
.gamma. = a/(a-b)
This requires that the centers of two halves of the small aperture be
displaced by a + b = (2.gamma. - 1)
The number of quasi-monochromatic wavebands required is
n .gtoreq. 1 - log(2.gamma. - 2)/log {(2.gamma. - 1)/2.gamma.}
where n must be interpreted as the integer next larger than the quantity on
the right and their proportions must be
.lambda..sub.2 /.lambda..sub.1 = 2.gamma./(2.gamma. - 1)
.lambda..sub.3 /.lambda..sub.1 = {2.gamma./(2.gamma. - 1)}.sup.2
.lambda..sub.n /.lambda..sub.1 = {2.gamma./(2.gamma. - 1)}.sup.n-1
(If the span of wavelengths available from the targets is less than
.lambda..sub.n - .lambda..sub.1 then the aperture must be divided into
more sections than two and these multiple sections should be arrayed with
their centers at constant ratios.)
For the .gamma.-enhanced, two-segment aperture, the gains may be arranged
to be in the ratios
##EQU11##
(This arrangement of gains will yield small minimum-side lobes. Since the
different gain ratios also apply to different bands of spatial
frequencies, advantage may be taken of their selection to provide spatial
pre-emphasis such as edge enhancement.)
These provisions correct the effects of the missing part of H. However, a
further difficulty arises because the distorted image data Q(x) are not
observed continuously but only at discrete intervals .DELTA.x
corresponding to discrete detectors. This can be overcome by using a
completely filled detector array. However, advantages accrue from using an
incompletely filled array.
In the case of the .gamma.-enhanced, two-element aperture, the spatial
bandwidth accommodated at the shortest wavelength .lambda..sub.1 is
4(a-b)/(F.lambda..sub.1) in each dimension so that a field X can be
reconstructed to arbitrary accuracy from F.lambda..sub.1, X/4(a - b)
samples although the samples shall be arrayed in two evenly spaced
sequences displaced by a fixed amount as described by Linden (in Proc.
IRE, p. 1219, July 1959). Thus, an array of {F.lambda..sub.1 X/4(a -
b)}.sup.2 detectors is required. It will, of course, be noted that the
corresponding single aperture would require (.gamma./2).sup.2 as many
detectors for the same field. Since each spatial bandpass is narrower than
the preceding one by a factor (2r-1)/2r, the same number of samples is
sufficient for each wavelength.
Since all the spatial frequencies within the bandwidth of the distorted
image at each wavelength can be reconstructed, therefore, the
multispectral estimate of the true image can also be reconstructed. A
sufficient method would be to perform the reconstruction specified by
Linden and perform the following operations on the resulting continuous
function.
The key to employing the relationships derived above to obtain a .gamma.
enhanced resolution is the ability to differentiate the responses to the
several wavelength bands, .lambda..sub.1, .lambda..sub.2, etc.
This is achieved by employing the phase retardation mirrors 3 which, in
effect, "code" the detector responses as a function of phase retardation.
DESCRIPTION OF A PREFERRED EMBODIMENT
A. REQUIREMENT
A sensor is required to resolve 100 feet at 3000 nmi over a field of
2.degree. square. The application will support a sensor weight
corresponding to an aperture area of 40 cm .times. 40 cm. The targets to
be sensed radiate strongly at wavelengths from 6 to 20 .mu.m.
B. DESCRIPTION
An aperture of 40 cm at 6 .mu.m can resolve only about 275 feet at 3000
nmi. Therefore, the available aperture is divided into four parts, each
comprising an identical mirror 20, each 20 cm .times. 20 cm spaced 90 cm
on centers as shown in FIG. 3. The total aperture area is thus 4(20).sup.2
= (40 cm).sup.2. Photons gathered by this entrance pupil are reflected to
four small off-axis parabaloids 21, FIG. 4, and from them as nearly
collimated beams to four flat mirrors 22. These flat mirrors are used as
differential phase retardation devices. They are translated as pairs in
push pull (e.g., by piezoelectric drives or the equivalent). The
"horizontal" pair are translated at one frequency; the vertipair at
another.
The photons reflected from these phase shifters are reflected by an
eyepiece 23 to a partially filled matrix array of detectors in focal plane
24. The detector array (shown in FIG. 5) can be considered as four doubly
periodic arrays superimposed. Each array has 1200 20 .mu.m detectors per
dimension spaced at 70 .mu.m. The total number to cover a field 2.degree.
.times. 2.degree. in a staring mode is (4800).sup.2. (The usual staring
sensor would require an aperture 100 cm .times. 100 cm and an array of
(6000).sup.2 detectors.)
Alternatively, this sensor could be used as a scanning sensor with a linear
array of 2400 detectors.
The photoelectric currents generated by the detectors are sampled 6k times
(k times per band) as the phase retardation mirrors are translated through
each cycle. These samples are digitized, (telemetered, if desired) and
processed by a digital computer, as shown in the functional block diagram
of FIG. 6. More specifically, the optics 20-23 provide signals to the
focal plane array 24. These signals are sampled and A/D converted at 25.
Telemetered, if desired to a remote location at 26. The digitized samples
are then processed as shown 27-29. The computer output is a display of the
entire 2.degree. .times. 2.degree. field to a spatial resolution of 6
.mu.radians (100 feet at 3000 nmi) or alternatively, a set of 3k displays
to a spatial resolution of 15 .mu.radians to 50 .mu.radians of the scene
at wavelengths from 6 to 20 .mu.m.
To illustrate the processing steps, assume only two spatial frequency bands
(1 .mu.m and 3 .mu.m) are of interest, at a single detector in the array
and a pair of apertures are of interest. Further, assume that the phase
retardation device is capable of a total travel of 30 .mu.m in a time
period of 1 sec. Assuming 1 .mu.m light from the apertures were in phase
at the detector prior to movement of the phase retardation device, then
the irradiance from each aperture would reinforce at the detector of
interest. After 1/2 .mu.m equivalent | | |
