BACKGROUND OF THE INVENTION
The following invention relates to a system for measuring the total
velocity of a target, and more particularly relates to a passive electrooptical
system for measuring two components of the total velocity: one parallel to the
system lineofsight and one perpendicular to the system lineofsight.
Most speed detection systems require a transmitter to transmit energy
towards a moving target that is reflected back to a receiver. Laser ranging systems
measure the time of transmission and the return of the energy in order to calculate
the range to the target and its speed. Radar ranging systems, such as the radar
guns used by law enforcement agencies for traffic control, use the principle of
Doppler frequency shift to calculate target speed. One problem with radar as a
traffic control device is that target acquisition and measurement are ambiguous.
Generally it can not be determined which target out of a multitude of possible
targets is responsible for generating any particular speed indication. Another
problem is that radar can be detected by receivers tuned to the proper frequency.
Yet another problem with Doppler radars is that they cannot measure range. Available
laser ranging systems can measure range, but are detectable by receivers (laser
detectors), are more expensive than radar systems, and are more difficult to aim
than radar systems.
U.S. Patent No. 5,586,063 to Hardin et al., which is assigned to
the assignee of this application and is incorporated herein by reference, is directed
to a passive optical speed and distance measuring system (the '063 system). Specifically
the '063 system includes a pair of camera lenses positioned along a common baseline
a predetermined distance apart and controlled by an operator to capture images
of a target at different times. The camera lenses are focused on light sensitive
pixel arrays that capture target images at offset positions in the line scans
of the pixel arrays. A video signal processor with a computer determines the location
of the offset positions and calculates the range to the target by solving the
trigonometry of the triangle formed by the two camera lenses and the target. Once
the range to the target is known at two different times the speed of the target
is calculated.
The '063 system measures only one component of velocity, the component
in the direction of its lineofsight. This component, in the general case, is
less than the velocity of a target such as a moving object.
What is needed then is a system that is capable of measuring the
total velocity of a target.
SUMMARY OF THE PRESENT INVENTION
The system of the present invention measures the total displacement
of a target that is shown graphically as displacement vector δ_{R}.
The total velocity is then obtained by dividing by the time interval over which
the displacement occurred. By comparison, the '063 system can measure only one
component of this velocity (the component along the '063 system LOS). Further,
the system of the present invention is able to resolve the displacement vector
into two components; one component x that is parallel to the system LOS and one
component y that is perpendicular to the system LOS. The corresponding x and y
velocity components are obtained by dividing by the time interval over which the
displacement occurred.
Further, the system of the present invention may be used to track
an object or to control a sensor platform to keep a system lineofsight (LOS)
pointed at a moving object because this system is capable of deriving the angle
between the LOS of the system and the velocity vector of the moving target.
The present invention is directed to a passive electrooptical range
and total velocity measuring system having first and second cameras positioned
along a common baseline. A first camera control system activates the first camera
at a first instance to capture a target image of a target at location T_{1}
and at a second instance to capture a target image of the target at location T_{2}.
A second camera control system activates the second camera at the first instance
to capture a target image of the target at location T_{1} and at the second
instance to capture a target image of the target at location T_{2}. A
range computer calculates ranges from the first camera, the second camera, and
a baseline midpoint to a target at location T_{1} and location T_{2}.
An angles computer calculates target displacement. A velocity computer calculates
total target velocity, track velocity, and crosstrack velocity, where track velocity
and crosstrack velocity are components of the total velocity.
A separate preferred embodiment of the present invention is directed
to a method for measuring the range and total velocity of a target using a passive
electrooptical system. Specifically, the method includes the steps of activating
first and second image acquisition devices at a first instance to capture a target
image of a target at location T_{1} and at a second instance to capture
a target image of the target at location T_{2}. The next steps are calculating
steps in which the ranges from said first and second image acquisition devices
to said target at locations T_{1} and T_{2}, the target displacement
δ_{R}, the total target velocity, the track velocity x, and the
crosstrack velocity are calculated.
The foregoing and other objectives, features, and advantages of the
invention will be more readily understood upon consideration of the following detailed
description of the invention, taken in conjunction with the accompanying drawings.
BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS
 FIG. 1 is a simplified block schematic diagram of the system of the invention.
 FIG. 2 is a simplified flow chart diagram of a preferred embodiment of the
present invention.
 FIG. 3 is a schematic illustration of the electrooptical relationships of
the system used for generating a range measurement.
 FIG. 4 is a schematic illustration of the electrooptical relationships of
the system used for generating a velocity measurement.
 FIG. 5 is a schematic illustration of a simplified hypothetical example of
the correlation process.
 FIG. 6 is a null curve diagram illustrating an exemplary relationship between
the shift in pixels (xaxis) and the sum of the absolute differences (yaxis).
 FIG. 7 is a simplified schematic illustration depicting the angular relationships
between camera A and the target T at times t_{1} and t_{2}.
 FIG. 8 is a simplified schematic illustration depicting the angular relationships
between camera B and the target T at times t_{1} and t_{2}.
 FIG. 9 is a schematic illustration depicting the angular relationships used
for generating velocity vector components and approximations.
 FIG. 10 is a simplified schematic illustration depicting the angular relationships
used for generating velocity vector components and approximations.
 FIG. 11 is a simplified block schematic diagram of the system of the invention.
 FIG. 12 is a simplified schematic illustration of a twocamera system of the
present invention.
 FIG. 13 is a simplified schematic illustration of a fourcamera system of the
present invention.
 FIG. 14 is a simplified schematic illustration of a threecamera system of
the present invention.
 FIG. 15 is a depiction of the video scan lines orientation of the fourcamera
system of FIG. 13.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
Referring to FIG. 1, the present invention includes a video camera
subsystem and video display 10 connected to a control and computational subsystem
12. The camera subsystem 10 provides camera video from cameras A and B 14, 16
to the control and computational subsystem 12. The control subsystem supplies
alphanumeric video to the video display subsystem 10. Cameras A and B 14, 16 may
be any type of electrooptical imaging sensors with a focal length f. Each imaging
sensor can be, for example, a chargecoupled device (CCD), a chargeinjection
device (CID), a metaloxidesemiconductor (MOS) phototransistor array or various
types of infrared imaging sensors, one example of which is a Platinum Silicide
(PtSi) detector array. Control and computational subsystem 12 may be any type of
computer. For example, the computational subsystem 12 may be that shown in FIG.
11, a general purpose computer with special software, or an alternate computer
specifically designed to accomplish the functions described herein.
More specifically, as shown in FIG. 2, each of the cameras 14, 16
in the camera subsystem 10, when instructed by the control subsystem 12, take a
video image or linear scan of moving target T at a first instance t_{1}
and at a second instance t_{2} (for a total of four recorded images) 100a100d.
The target is at location T_{1} at the first instance t_{1} and
at location T_{2} at the second instance. The camera subsystem 10 then
passes the camera video to the computational subsystem 12 that makes the calculations
necessary to determine the range of the target T at time instance t_{1}
102a and the range R_{2}
of the target T at time instance t_{2} 102b.
As will be discussed below in detail, the ranges R_{1} and R_{2}
to target T at both time instances t_{1} and t_{2} are obtained
by correlating the images obtained from both cameras at that time. The image from
camera A at time t_{1} is then correlated with the image from camera A
at time t_{2} 104. From the correlation result, the angles &thetas;_{1A}
 &thetas;_{2A} and &thetas;_{1B} &thetas;_{2B}
can be calculated
106. Using R_{1}, R_{2}, and the angle &thetas;_{1A}
 &thetas;_{2A},
the target displacement between times t_{1} and t_{2} as seen by
camera A 108 can be calculated. Using R_{1}, R_{2} and the angle
&thetas;_{1B}  &thetas;_{2B}, the target displacement between times
t_{1} and t_{2} as seen by camera B can be calculated 110. The
two displacements are then averaged to obtain the target displacement between
times t_{1} and t_{2} 112. Then, the total target velocity v is
calculated using the target displacement and the measured time interval (t_{2}
 t_{1}) 114. Using the target displacement and the difference R_{1}

R_{2}, the components of the total target velocity parallel (x) and perpendicular
(y) to the lineofsight can be computed 116. Finally, from the knowledge of the
velocity components parallel and perpendicular to the lineofsight, the angle
between the total target velocity vector and the lineofsight can be computed
118.
It should be noted that knowledge of the total target displacement
δ_{R} and the time instance interval (t_{2}  t_{1})
enables computation of the velocity of the target as well as the components x and
y of the displacement vector δ_{R}. It should also be noted that
the order of computations shown in FIG. 2 is meant to be exemplary and may be
varied without changing the scope of the invention.
Turning first to the exemplary computation of range R, FIG. 3 shows
an optical schematic diagram the placement of cameras A and B 14, 16 used in the
method for measuring of the range R or distance from the center of a baseline
17 to the target T. The method for measuring range R, the first step in the method
of the present invention, is substantially the same method as that used in the
'063 system. Calculating R would be done twice in the method of the present invention:
once for calculating R_{1} (the distance from the baseline midpoint 17
to the target at location T_{1}) and once for calculating R_{2}
(the distance from the baseline midpoint 17 to the target at location T_{2}).
R_{1} and R_{2} will be used as approximations for R_{1A},
R_{1B}, R_{2A}, and R_{2B} as set forth below.
Both the '063 system and the present invention, as shown in FIG.
3, include a camera A 14 positioned at a first position 18 and a camera B 16 positioned
at a second position 20 on a baseline 17. In these positions, the cameras are
separated by a distance of bl and have linesofsight LOS that are parallel and
in the same plane. Range R, as measured by this method, is defined as the distance
from the midpoint 22 of the baseline 17 to the exemplary target T. LOS is the lineofsight
of the twosensor system. LOS_{A} and LOS_{B} are the linesofsight
for cameras A and B 14, 16, respectively. LOS intersects baseline 17 at its midpoint
22, is in the same plane as the cameras' linesofsight, and is perpendicular
to baseline 17. The angle shown as &thetas;_{1A} is the angle between LOS_{A}
and the target T and the angle shown as &thetas;_{1B} is the angle between
LOS_{B} and the target T. Using the image information supplied by the video
camera subsystem 10, the control and computational subsystem 12 first determines
the angle of interest (&thetas;_{1B}  &thetas;_{1A}) by electronically
correlating the images from the focal planes of cameras A and B 14, 16 to measure
the linear displacement d_{1B}  d_{1A}. The magnitude of d_{1B}
 d_{1A} can be measured by correlating the A and B camera images obtained
at time t_{1}. d_{1B}  d_{1A} is measured at the focal
plane which is behind the baseline by a distance f, the focal length.
Image correlation is possible in the present invention because the
system geometry (as shown in FIGS. 3 and 4) is such that a portion of the image
from camera A 14 will contain information very similar to that contained in a
portion of the image from camera B 16 when both images are acquired at the same
time. This common information occurs in a different location in the camera A image
when compared to its location in the camera B image due to the separation of the
two cameras by the baseline distance bl.
The correlation process is discussed in U.S. Patent No. 5,586,063
to Hardin et al., which is assigned to the assignee of this application and is
incorporated herein by reference. However, FIGs. 3 and 4 may be used to illustrate
this process. FIG. 5 illustrates the correlation of two linear images, one from
Camera A, the other from Camera B. For simplicity, a hypothetical video line of
12 pixels is shown. (In practice, cameras with video linelengths of hundreds of
pixels are used.) In addition, for simplicity of illustration, a single 3 pixelwide
image of unit (I) intensity is shown, with a uniform background of zero intensity.
In practice, any pixel can have any value within the dynamic range of the camera.
The pixel values for each of the two video lines are mapped in computer memory.
In this case, the Camera A line is used as the reference. The map for the Camera
B line is then matched with the A line map at different offsets from zero pixels
to some maximum value dictated by other system parameters. (Zero pixels offset
corresponds to a range of infinity.) This unidirectional process is sufficient
since the relative position of any target in the FOV of one camera with respect
to the other is known. At each offset position the absolute difference is computed
for each adjacent pixelpair that exists (the pixels in the overlap region). The
differences are then summed. It should be noted that there are a number of other
mathematical procedures that could be used to correlate the lines that would achieve
similar results. One advantage of the procedure described is that no multiplication
(or division) operations are required. (Addition and subtraction are computationally
less intensive.) FIG. 6 is a plot of the sum of absolute differences (yaxis) versus
the offset for this example. Note that the function has a minimum at the point
of best correlation. This is referred to as the "global null," "global" differentiating
it from other shallower nulls that can result in practice. The offset value corresponding
to the global null is shown in FIG. 6 as d_{1B}  d_{1A}. This
quantity is also shown in FIG. 3.
In order to measure the total displacement of the target (in order
to compute the total velocity) at least one more correlation is required. The additional
correlation is performed in a similar manner to that described above, but is a
temporal correlation. It uses images from the same camera (Camera A), obtained
at two different times (t1 and t2). One difference is that the relative positions
of the target image at the two different times is not known to the System. This
requires that the correlation be bidirectional. Bidirectional correlation is
achieved by first using the t1 image map as the reference and shifting the t2 image
map, then swapping the image maps and repeating the process.
Once image correlation has been completed, the angle (&thetas;_{1B}
 &thetas;_{1A}) can be found from the equation:
&thetas;_{1B}  &thetas;_{1A}
= arctan [(d_{1B}  d_{1A})/f]
. Using this information, range R is calculated by the equation:
R = bl/[2 tan 1/2(&thetas;_{1B}  &thetas;_{1A})]
. Alternatively, the computational subsystem 12 can find range R by solving the
proportionality equation:
(d_{1B}  d_{1A})/f = (bl/2)/R
. The method for finding R is set forth in more complete terms in U.S. Patent No.
5,586,063, however, alternative methods for computing range may be used.
FIG. 4 is an optical schematic diagram of the placement of cameras
A and B 14, 16 as well as the angles and distances used in the method for measuring
of the velocity v, the second step in the method of the present invention. To
make the necessary calculations to find the velocity v, first the target displacement
(δ_{R}) between the target location (T_{1}) at a first instance
(t_{1}) and the target location(T_{2}) at a second instance (t_{2})
must be determined. Once δ_{R} is determined, the velocity (v) is
computed as:
v = δ_{R}/(t_{2}  t_{1})
. It should be noted that the '063 system can compute only the ranges R_{1}
and R_{2}
which, when differenced (to form R_{2}  R_{1}),
constitute only one component of the total displacement δ_{R}.
To find an accurate δ_{R}, both triangle A (defined
by camera A lens 14 at position 18 on the baseline 17, the target location T_{1}
at the first instance t_{1}, and the target location T_{2} at the
second instance t_{2}) and triangle B (defined by camera B lens 16 at position
20 on the baseline 17, the target location T_{1} at the first instance
t_{1}, and the target location T_{2} at the second instance t_{2})
should be solved. By solving triangle A to find δ_{RA}, an approximate
of δ_{R} is found. Solving for δ_{RB} and averaging
it with δ_{RA} (
δ_{R} = (δ_{RA} + δ_{RB})/2
) greatly reduces error in using a single calculation. It should be noted that
images of the target acquired by cameras A and B at times t_{1} and t_{2}
may have already been acquired and stored for use in range computations of the
'063 system.
Fig. 7 shows an enhanced view of triangle A (defined by camera A
lens 14 at position 18 on the baseline 17, the target location T_{1} at
the first instance t_{1}, and the target location T_{2} at the
second instance t_{2}). Specifically, the angle &thetas;_{1A} 
&thetas;_{2A} is the angular difference between target locations T_{1}
and T_{2}, as measured by camera A. The images are acquired by camera A
at times t_{1} and t_{2}, as set forth above, and are then correlated
to obtain the angle &thetas;_{1A}  &thetas;_{2A}. The next step
is to use R_{1} and R_{2} as approximations for R_{1A}
and R_{2A} respectively. R_{1} and R_{2} can be calculated
using the equations set forth generally above and in detail in U.S. Patent No.
5,586,063, incorporated herein by reference. Using these calculations, triangle
A can be solved for the displacement δ_{RA}, using the law of cosines:
δ_{RA} = [R_{1} ^{2} + R_{2} ^{2}
 2R_{1}R_{2} cos (&thetas;_{1A}  &thetas;_{2A})]^{1/2}
.
δ_{RA} is slightly different than the desired δ_{R}
(of
FIG. 4) because R_{1} and R_{2} are distances from the midpoint
22 of the baseline to target locations T_{1} and T_{2}, whereas
R_{1A} and R_{2A} are distances from camera A to target locations
T_{1} and T_{2}. Using the built in symmetry of the system, this
error can be greatly reduced by solving triangle B (defined by camera B lens 16
at position 20 on the baseline, the target location T_{1} at the first
instance t_{1}, and the target location T_{2} at the second instance
t_{2}) of FIG. 8 for δ_{RB} and averaging the two results.
δ_{RB} may be found using calculations similar to those set forth
above for triangle A. Specifically, triangle B can be solved for the displacement
δ_{RB}, using the law of cosines:
δ_{RB} = [R_{1} ^{2} + R_{2} ^{2}
 2R_{1}R_{2}cos(&thetas;_{1B}  &thetas;_{2B})]^{1/2}
.
It should be noted that the solution of triangle B does not require
a correlation operation (as did the solution of triangle A) to determine the angle
&thetas;_{1B}
 &thetas;_{2B}. The reason for this can be seen by
referring to FIG. 4 where it can be seen that the triangles A, C, T_{1}
and
B, C, T_{2} both contain the same angle &phis; (from the law that opposite
angles are equal). C is the point of intersection between R_{1B}, the range
from camera B to the target at the first instance, and R_{2A}, the range
from camera A to the target at the second instance.) Thus, since three of the
four difference angles shown are known, the fourth can be computed using the law
that the sum of the interior angles of a triangle is always equal to 180 degrees.
Correlation using the images from camera B 16 may be performed for the optional
purpose of verifying optical alignment.
As set forth above, once δ_{R} is determined, the velocity
v of target T is computed as:
v = δ_{R}/(t_{2}  t_{1})
. The time base 12a and sync generator 12b (FIG. 11) would provide the elements
necessary to compute t_{1} and t_{2}.
The next step of the present invention is to compute the parallel
component x of the displacement vector δ_{R} and the perpendicular
component y of the displacement vector δ_{R}. Component x of the
displacement vector is parallel to the LOS in the plane defined by the LOS and
the baseline 17. Component y of the displacement vector is perpendicular to the
LOS in the plane defined by the LOS and the baseline 17. The velocity vector components
are determined by dividing the displacement vector component values by the time
interval over which the displacement occurred.
As shown in FIGS. 9 and 10, the x component parallel to the LOS is
defined as the difference of the two range measurements R_{1} (the distance
between the baseline midpoint 22 and the target T_{1} at first instance
t_{1}) and R_{2} (the distance between the baseline midpoint 22
and the target T_{2} at second instance t_{2}). The difference
between the two range measurements can be approximately defined by the equation:
x = R_{2}  R_{1}
. This is an approximation, since the actual difference of the two range measurements
is defined by the equation:
R_{2}cos&thetas;T_{2}  R_{1}cos&thetas;T_{1}
.
R_{1}cos&thetas;T_{1}
is the distance on the LOS from the baseline midpoint 22 to point 40, the perpendicular
distance from T_{1} to the LOS.
R_{2}cos&thetas;T_{2}
is the distance on the LOS from the baseline midpoint 22 to point 42, the perpendicular
distance from T_{2} to the LOS. However, &thetas;T_{2}
(the angle
between LOS and R_{2}) and &thetas;T_{1} (the angle between LOS
and R_{1}) cannot be determined. The
x = R_{2}  R_{1}
approximation will produce accurate results when &thetas;T_{1} and &thetas;T_{2}
are both small.
The y component of the velocity vector, also known as a "crosstrack"
velocity component, is then solved using the relationship set forth in FIG. 10.
Using δ_{R} (as computed above) as the hypotenuse and x (as computed
above) as one leg of the relationship triangle of FIG. 10, the triangle shown in
FIG. 10 can be solved for the perpendicular displacement component y using Pythagorean
theorem:
y = [(δ_{R})^{2}  x^{2}]^{1/2}
. The angle between the velocity vector and the LOS can then be calculated by
the following equation:
&thetas;_{LOS} = arctan (y/x)
. Knowledge of the angle &thetas;_{LOS} is of value in applications where
it is desirable to move the system lineofsight to track the target or simply
to keep the target in the field of view.
FIG. 11 shows an exemplary functional block diagram of one possible
implementation of the velocity measuring system of the present invention. Camera
or sensor A 14 and camera or sensor B 16 are electronic imaging cameras substantially
controlled by the system controller 12. The time base 12a and sync generator 12b
are used to synchronize the cameras. Further, the time base 12a provides the time
interval measurement capability that allows calculation of t_{1} and t_{2}.
The time between image acquisitions may be determined by keeping count of the
number of camera images that have been scanned between image acquisitions.
The digitizers 50a, 50b convert the analog camera outputs to a digital
format, enabling the camera images (or portions thereof) to be stored in conventional
computertype memory 52.
The image correlator 54 correlates the images supplied by camera
A 14 and camera B 16. The correlation process is used to determine the angular
difference between cameras when sighting an object or target T at the same time
("correlation") or at two different times ("crosscorrelation").
The range computer 56 then determines the range R to the target T
by triangulation using the measured angular difference acquired by the cameras
at the same time.
The angles computer 58 uses both the range and angle measurements
to compute the components of displacement of the target T parallel and perpendicular
to the system LOS.
The velocity computer 60 uses the measured displacement components
and knowledge of the time between measurements (t_{2}t_{1}) to
compute velocity v the components x and y total velocity.
The system controller 12 sequences and manages measurement and computation.
The image correlator 54, range computer 56, angles computer 58, velocity computer
60, and system controller 12 can be implemented as hardwired electronic circuits,
or these functions can be performed by a generalpurposed digital computer with
special software.
Although the invention has been described with reference to detection
systems for detecting the range and total velocity of a general moving target it
should be understood that the invention described herein has much broader application,
and in fact may be used to detect the range to a stationary object, the total
velocity of any moving object and/or relative motion between moving or stationary
objects. For example, the invention may be incorporated into a range and velocity
detection system for moving vehicles. Another example is that the invention may
be incorporated in a robotics manufacturing or monitoring system for monitoring
or operating upon objects moving along an assembly line. Still another important
application is a ranging device used in conjunction with a weapons system for acquiring
and tracking a target. Yet another application is a spotting system used to detect
camouflaged objects that may be in motion against a static background. Other possible
uses and applications will be apparent to those skilled in the art.
The foregoing invention can also be adapted to measure velocity in
threedimensional space. To do this a twodimensional camera configuration, such
as that shown in FIG. 12, is adapted to either the configuration shown in FIG.
13 or FIG. 14. The embodiment shown in FIG. 13 uses four cameras, A, B, C, and
D centered around a central LOS (extending outward form the page). The baseline
b1_{1} defined between cameras A and B is perpendicular to baseline b1_{2}
defined between cameras C and D, although b1_{1} and b1_{2} need
not be the same length. FIG. 15 shows the video scan lines orientation for this
system_{a} in which cameras A and B operate as one subsystem and cameras
C and D operate as a second subsystem that is a duplicate of the camera A and B
subsystem, except for its orientation. The velocity vectors produced by the two
subsystems are summed (vector summation) to yield the total target velocity in
three dimensions. FIG. 14 shows an alternate configuration that can measure velocity
in threedimensions, but uses only three cameras A, B, and C. It should be noted
that the FOV is smaller than that of the four camera system of FIG. 13 and the
calculations to determine the velocity are more complex.
The terms and expressions which have been employed in the foregoing
specification are used therein as terms of description and not of limitation, and
there is no intention, in the use of such terms and expressions, of excluding
equivalents of the features shown and described or portions thereof, it being
recognized that the scope of the invention is defined and limited only by the claims
which follow.