 |
Claims  |
|
|
What is claimed is:
1. An automatic range determination system comprising:
image sensing means for sensing an image focused thereon;
optical means for focusing first and second images of an object the range
of which is to be determined onto said image sensing means, said first and
second images derived along different sightlines;
quantization means for quantizing the data generated for each image element
by said image sensing means;
storage means for receiving and storing the quantized image data
corresponding to said first and second images;
comparison means for comparing on an element by element basis the stored
quantized image data corresponding to said first image to the stored
quantized image data corresponding to said second image and providing an
output which is representative of the degree of similarity between the
stored quantized image data for corresponding elements of the first and
second images;
control means for controlling operations performed during range
determination sequences and for causing the comparator means to compare
during a range determination sequence the stored data with respect to M
successive elements of the first image in separate comparison sequences
with the stored data for a plurality of different sets of M successive
elements of said second image;
alignment storage means for storing the relative alignment between the
elements of the first image and the elements of the second image which
yields an optimum comparison between the stored data;
comparison optimization means for recognizing the alignment between the
elements of said first image and said second image which, for the
optimization criterion utilized, yields an optimum comparison between the
elements of the first image and the elements of the second image.
2. The automatic range determination system recited in claim 1 wherein said
optical means comprises:
first optical channel means for focusing said first image of said object
whose range is to be determined onto a first portion of said image sensing
means; and
second optical channel means for focusing said second image of said object
whose range is to be determined onto a second portion of said image
sensing means, said second optical channel means displaced from said first
optical channel means so that said images are derived along different
sightlines.
3. The system recited in claim 1 wherein said image sensing means comprises
an array of photocells.
4. The system recited in claim 3 wherein:
said photocells are arranged in rows;
said first image is focused on a first portion of a selected row of
photocells, and
said second image is focused on a second portion of said selected row of
photocells.
5. The system recited in claim 3 wherein:
said array is two-dimensional;
said photocells are arranged in substantially parallel rows, each row
including a plurality of uniformly spaced photocells and at least some of
said photocells of a first row are positioned in the interstices between
cells of the second row in the sense that if a first line is drawn through
the photocells of said first row and a second line is drawn perpendicular
to said first line and through a photocell in said first row, then the
second line will not intersect a photocell in the second row.
6. The system recited in claim 2 wherein said first optical channel means
and said second optical channel means are both folded in order to minimize
the overall size of said first and said second optical channel means.
7. The range determination system recited in claim 1 wherein said
comparison means provides for each comparison between the data stored for
an element of the first image and the data stored for an element of the
second image a first output indicating a match if the stored data being
compared are identical and a second output otherwise.
8. The system recited in claim 7 wherein said comparison optimization means
comprises:
match counter means coupled to said comparison means for counting the
number of matches between the elements of said first image and the
elements of said second image during each comparison sequence and;
match count comparator means for determining the largest match count.
9. The range determination system recited in claim 1 further comprising:
means for determining, from the alignment stored in said alignment storage
means at the end of the alignment determination sequence, the range of the
object the range of which is to be determined.
10. A method of automatically determining the range of an object comprising
the steps of:
deriving first and second images of said object along different sightlines;
electronically sensing and quantizing said first and second images to
produce a first quantized electronic version of said first image and a
second quantized electronic version of said second image;
storing said first and second quantized versions of said images;
electronically correlating said first and second electronic versions of
said images to determine the relative location of corresponding portions
of said first and second images, said correlation process comprising:
comparing M successive elements of said stored version of said first image
with a plurality of different sets of M successive elements of said stored
version of said second image to determine the degree of similarity between
said stored data for said M successive elements of said first image and
said stored data for each of said plurality of sets of M successive
elements of said second image;
determining from the results of the comparison process which set of M
successive elements of said second image is most similar to said M
successive elements of said first image in accordance with the comparison
criterion utilized;
determining from the location within the second image of said set of M
successive elements of said second image which is most similar to said M
successive elements of said first image, the range of the object the range
of which is to be determined.
11. An automatic range determination system comprising:
image sensing means comprising an array of photoresponsive cells for
sensing an image focused thereon;
optical means for focusing first and second images of an object, the range
of which is to be determined, onto said image sensing means, said first
and second images derived along different sightlines and each comprised of
a plurality of image elements;
quantization means for quantizing the data generated for each image element
by said image sensing means;
first and second shift registers for receiving the quantized image data
corresponding to said first image and said second image, respectively;
said first shift register containing M storage locations and said second
shift register containing N storage locations, wherein N>M;
comparator means for comparing the image data stored in said first shift
register with the image data stored in said second shift register on an
element by element basis and providing a first output indicating a match
when the stored data being compared are identical and a second output when
the stored data are not identical;
match counter means coupled to said comparison means for counting the
number of matches between the elements of said image stored in said first
shift register and the elements of said image stored in said second shift
register;
match count comparator means for determining the largest match count during
a range determination sequence;
alignment storage means for storing the alignment which yields said largest
comparison sequences, the data contained count;
control means for controlling an alignment determination sequence and for
causing the information in the M cells of said first shift register to be
compared with, in separate comparison sequences, the data contained in
each set of M successive storage locations in said second shift register
in order that the relative alignment of said first and said second images
which provides the highest match count may be determined; and
means for determining, from the alignment stored in said alignment storage
means at the end of the comparison sequence, the range of the object the
range of which is to be determined. |
|
 |
|
 |
Claims |
|
|
|
Description |
|
|
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention relates to the field of optical instrument focusing systems.
2. Prior Art
Prior art optical instrument focusing systems are of two kinds,
mechanically aided, but operator controlled, and fully automatic.
Mechanically aided, operator controlled, focusing systems are of many
varieties. These varieties include double image, split image and broken
image systems. In each of these systems, the operator must adjust the
focus of the optical system until the operator determines that the optimum
image has been achieved. In double image systems, the elimination of the
double image is the criterion which the operator must use. In split image
systems, the operator must align separate portions of the image to achieve
a unitary image. In broken image systems the operator must adjust the
focus to yield an image which to the operator's eye is the clearest image.
Automatic focusing systems have utilized moving lenticular screens disposed
between a photocell and a lens system at the location where the focus of
the image is positioned when the optical instrument is in focus. Such
systems make use of the fact that at the focus of an image projected by a
lens system the image is a point. When the focus of the image is at the
lenticular screen, the image passing through the lenticular screen suffers
a minimum of variation in image intensity. However, when the focus of the
image is displaced for the lenticular screen, the image passes through
portions of the screen having differing focal points and the image is
distorted. A photocell placed in the path of the image transmitted by the
lenticular screen provides an output signal which is dependent on the
position of the focus of the image relative to the lenticular screen. With
a motor driven lens focusing system, the motor is automatically stopped at
the lens position which is determined on the basis of the photocell output
as corresponding to the best focus. Unfortunately, such systems have a
disadvantage in that in order to establish that an image is in proper
focus the lens must be moved out of focus because the system determines
focus on a relative not an absolute basis. During focusing, the rate at
which the lens focus is adjusted must be limited to a low enough rate that
the optimum focus can be detected in time to stop the lens motion before
the lens passes out of focus or else the focusing motor must be reversed
to bring the image back into focus. Such systems have a further
disadvantage that when the lens is adjusted by the motor, it cannot be
determined a priority in which direction the motor should adjust the
focus. Therefore, on the average one-half of the time the motor will begin
by adjusting the lens focus in the wrong direction, in consequence, the
motor must be stopped and reversed with attendant disadvantages with
respect to vibration power consumption and focusing time.
SUMMARY OF THE INVENTION
The problems of prior art focusing systems for optical instruments such as
cameras are overcome by the present invention through the elimination of
mechanical adjustments during the determination of the range from the
optical instrument to the object on which the instrument is to be focused.
This is achieved by an all electronic range determination system. The
optical instrument is set to be in focus at the range of the object being
viewed as determined by the electronic range determination system. The
range is preferably determined through use of a double image system in
which a first image of the object whose range is to be determined is
focused on a first image sensor and a second image of the object which
includes at least a portion of the first image is focused on a second
image sensor. The location of the first image within the second image is
controlled by the distance from the range determination system to the
object. The first and second images are electronically correlated or
compared to determine where within the second image the first image
occurs. Once the location of the first image within the second image is
determined, the range to the object is known or can be determined from a
predetermined relationship between position of the first image in the
second image and the distance to the object. This range information is
then utilized in focusing the optical instrument. Such focus is preferably
achieved automatically, but may be achieved by operator adjustment.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a schematic diagram of an optical system which may be utilized to
generate raw ranging data in accordance with the invention.
FIG. 2 is a block diagram of a data reduction and optical instrument
control system which may be utilized to reduce the raw data to range
information and to adjust the optical instrument accordingly.
FIG. 3 is an illustration of a photosensor array which aids in resolving
one type of range ambiguity.
FIG. 4 illustrates a photosensor array for use in a split image embodiment.
FIG. 5 illustrates a photosensor array for use in a split image embodiment
having less stringent photosensor alignment requirements than those of the
array of FIG. 4.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
A preferred optical system for generating raw ranging data in accordance
with the invention is illustrated generally at 10 in FIG. 1. Optical
system 10 comprises first and second lenses 12 and 14 which are laterally
displaced from each other by a separation distance S in order to generate
separate subject (52) and range (54) images, respectively, of an object to
be ranged. In the preferred embodiment, lenses 12 and 14 are centered on a
common line 13 with their axes aligned perpendicular to common line 13 and
displaced from each other by the distance S. Because of the separation
between lenses 12 and 14, the subject and range images are generated along
differing sightlines 16 and 20, respectively. In this preferred
embodiment, the position of sightline 16 for the subject image is always
the same independent of the distance to the object being ranged and lies
along the axis of lens 12. The position of sightline 20 varies in
accordance with the range of the object to be ranged and is disposed at an
angle .phi. relative to common line 13. The angle .phi. varies in
accordance with the range R to the object being ranged. When the object
being ranged is at a range R.sub.1 from optical system 10 which is
essentially infinite, sightline 20 coincides with line 21 which is along
the axis of lens 14 and is thus parallel to sightline 16. With the object
to be ranged at the range R.sub.1 the angle .phi. is equal to the
.phi..sub.1 and is 90.degree.. When the object to be ranged is at some
range R.sub.2 from optical system 10 which is less than range R.sub.1,
sightline 20 coincides with line 22 for which the angle .phi. is equal to
.phi..sub.2 which is less than 90.degree.. When the object to be ranged is
at some range R.sub.3 which is less than range R.sub.2, sightline 20 is
along line 23 for which the angle .phi. is equal to .phi..sub.3 which is
less than .phi..sub.2.
In order to accurately focus an optical instrument using the range data
provided by the automatic ranging system in accordance with this
invention, (in order to assure that the image of the object will not be
blurred) the separation between lenses 12 and 14 must be selected such
that the angle .phi. of sightline 20 to the object being ranged is
detectably different from the angle .phi. the sightline 20 would assume
for an object at any range that would require a different focus setting of
the optical instrument.
The subject image 52 and range image 54 generated respectively by lenses 12
and 14 are focused on a photosensor array 30 comprised of a plurality of
photocells each of which detects the intensity of the light from the
object which strikes that photocell. Photosensor array 30 may preferably
comprise a single row of sensor cells in order to simplify the electronic
data reduction system. However, it will be understood that any sensor
array configuration may be used so long as it is compatible with the data
reduction system which is utilized.
It is preferred to use a folded optical system to focus the images on the
array 30. A folded optical system minimizes the optical system depth and
allows a single photosensor array which is shorter than the distance S to
be utilized to sense both the subject and the range images. The preferred
folded optical system employs two critical angle reflecting mirrors
associated with each lens. A first mirror 40 associated with lens 12
reflects the subject image from lens 12 essentially perpendicular to
sightline 16 and a second mirror 42 re-reflects the subject image
essentially parallel to sightline 16 and onto photosensor array 30.
Similarly, a first mirror 44 associated with lens 14 reflects the range
image from lens 14 off-angle from sightline 20 and a second mirror 46
re-reflects the range image approximately parallel to sightline 20 and
onto photosensor array 30. As illustrated in FIG. 1, after reflection from
mirrors 42 and 46 the light comprising each of the images travels in
substantially the same direction as that in which it traveled prior to
entering the lenses. However, if desired the mirrors 42 and 46 could be
re-oriented 90.degree. from the position shown in order to reflect the
images such that after reflection from mirrors 42 and 46 the light travels
in substantially the opposite direction to the direction in which it
travels prior to entering the lenses. In such an embodiment, photosensor
array 30 may be conveniently placed on or near the common line 13 through
the center of lenses 12 and 14.
It will be understood that although folded optical system 10 is preferred,
other optical systems may be utilized. Where a single large array 30 or
two separate arrays are utilized to sense subject image 52 and range image
54, the images from lenses 12 and 14 can be focused directly onto the
sensor array without the use of folded optics.
The use of a single sensor array is preferred for several reasons. First,
uniformity in the photoresponse of the individual sensor cells is more
easily obtained in a single array than in two separate arrays. Second,
misalignment of the sensing array relative to the optical system can cause
the subject and range images to be derived from different portions of the
object and the chances of such misalignment occurring are increased by the
use of separate sensor arrays. Third, utilization of a single array and
folded optics allows the overall dimensions of optical system 10 to be
minimized. The minimization of the overall dimensions of the optical
system is of particular benefit where it is desired to incorporate the
optical system within a photographic camera. With a folded optical system,
many popular size cameras can be adapted to utilize the inventive ranging
system without requiring a significant change in the dimensions of the
camera body.
The subject image from lens 12 is focused on a subject region 32 of
photosensor array 30. Region 32 contains enough sensing cells of
sufficiently small dimensions to provide a detailed representation of the
object to be ranged. There are preferably at least 20 photosensitive cells
within region 32.
The range image from lens 14 is focused on a range region 34 of image array
30. In this embodiment, the range region 34 is substantially longer than
the subject region 32 in order to simplify the electronic processing of
the data generated by optical system 10. However, it will be understood
that any subject image and range image sensor configuration may be used,
so long as it is compatible with the data reduction system which it
utilized. In the present embodiment range region 34 is preferably on the
order of three or more times as long as subject region 32. Regions 32 and
34 preferably each comprise a plurality of photosensor cells which are
identical in photoresponse, area and spacing.
The subject image portion 32 of photosensor array 30 is preferably
separated from the range image portion 34 of photosensor array 30 by an
unused portion 33 of the photosensor array. The separation between subject
image 32 and range imager 34 prevents overlapping of light from lens 14
with the subject image 52 and the overlapping of light from lens 12 with
the range image 54 either of which would result in a smeared image and
reduced resolution.
The position within region 34 of the portion of the range image which is
essentially a duplicate of the subject image is a function of the range to
the object being ranged. Where the object is at the range R.sub.1 which is
essentially infinite, the subject portion of the range image will appear
in a portion 41 or region 34 which is at one end of range region 34. Where
the object is at a minimum range R.sub.3, the subject portion of the range
image will appear in a portion 43 of region 34 which is at the opposite
end of range region 34 from portion 41. Where the object is at the
intermediate range R.sub.2 the subject portion of the range image will
appear in a portion 42 of region 34 which is intermediate portions 41 and
43.
Photosensor array 30 preferably comprises a CCD image sensing array or a
photodiode image sensing array and may comprise a single row of
photosensitive cells. In order to obtain sufficiently detailed electronic
image representations, it is preferred to have the individual photocells
of the array on the order of 0.0005 inch or less square and separated as
little as possible, preferably less than 0.0005 inches. Readout of the
image detected by photosensor array 30 can be achieved either serially
cell by cell or in parallel, in accordance with the characteristics of the
array. Serial readout takes more time than parallel readout, but allows
utilization of a less complicated data reduction system than is required
for parallel readout.
The amplitude of the response from a photocell is dependent upon the
intensity of the light illuminating that photocell and on the period of
time during which the photocell is allowed to integrate the incident
light. If the integration time is too short, the electronic output from
the photocells will be low. If integration times is longer, than the
output from the photocells will be larger. Thus, the period over which the
photocells are allowed to integrate the incident light can be used to
compensate for different light intensities to which imaging array 30 may
be subject when different objects are ranged.
The range to the object is determined by electronically correlating the
electronic version of the subject image 52 with the electronic version of
the range image 54 to determine the relative location of the corresponding
portions of the subject and range images. Any correlation system may be
used so long as it accomplishes this result.
In the preferred embodiment correlation is achieved by electronically
comparing on a cell by cell basis the subject image 52 (as sensed by
subject region 32) with the range image 54 (as sensed by range region 34).
That is, if the subject region 32 comprises M photosensor cells and the
range region 34 comprises N photosensor cells, the subject image is
compared first with the image on the 1st through Mth photocells of range
region 34, then with the second through M+1.sup.th photocells, then with
the 3rd through M+2.sup.th photocells and so forth. The position within
region 34 of the portion of the range image 54 which is most similar to
the subject image 52 (in subject region 32) is a measure of the range to
the object.
A preferred embodiment of a data reduction system for use in determining
the range of an object from the raw data provided by optical system 10 is
illustrated generally at 100 in FIG. 2.
Data reduction system 100 is designed to receive the information generated
by image array 30 serially cell by cell. Data reduction system 100
operates under the control of a central control 120. The collection and
readout of data by photosensor array 30 is controlled by a photosensor
control 108. Under control of photosensor control 108 photosensor array 30
integrates the light comprising the subject and range images. Data is then
read out of sensor array 30 serially photocell by photocell. It is
preferred to convert analog data generated by photosensor array 30 to a
digital form for processing. Consequently, a data quantizer 110 receives
the analog electronic response of the individual photocells of array 30
serially cell by cell and quantizes that information with respect to a
predetermined quantization standard. This quantization standard may be
binary or may provide gray scale information in accordance with the degree
of light intensity resolution desired for use in the image correlation
process.
The quantized data provided by quantizer 110 is supplied to a data director
112 which directs the data to the succeeding processing circuitry in
accordance with the information the data represent. Data director 112
discards the data from a portion 51 of the image which impinges on portion
31 of sensor array 30 because this data is not utilized for ranging
purposes. Data director 112 directs those data representative of the
subject image to a subject shift register 122. The subject data is
preferably simultaneously applied to an events counter 114 which
determines the number of photocells of the subject image portion 32 of
sensor 30 which had a response in excess of a predetermined level. The
count reached by event counter 114 is provided to a limits detector 116
which compares this count with upper and lower limits. Limit detector 116
provides an indication of whether the event count is below a minimum
limit, between minimum and maximum limits or exceeds a maximum limit. If
the event count is below the minimum limit, central control 120 causes the
data to be discarded and causes the photosensor control 108 to increase
the integration time of photosensor array 30 in order to increase the
number of photocells which provide a response in excess of the
predetermined level. If the event count exceeds the maximum limit, the
central control 120 causes the data to be discarded and causes the
photosensor control 108 to decrease the integration time of the
photosensor array to reduce the number of photocells which provide a
response in excess of the predetermined level. If the event count is
between the minimum and maximum limits, processing of the data proceeds.
The limits on the number of events are imposed because a quantized image
provides a maximum of information about the actual image when the
quantized values are not almost all at the maximum or almost all at the
minimum extremes. A quantization in which most of the values are at one
extreme contains a minimum of information and does not provide a reliable
basis for comparison of the subject image with the range image because the
comparison is prone to errors due to noise and slight variations in
photocell response. Thus, an electronic version of the image which
contains a maximum of information is preferred.
After the subject image data from region 32 has been provided to subject
shift register 122, data director 112 discards unused data 53 which are
generated by the photocells which lie in the "unused" portion 33 of imager
30 between subject region 32 and range region 34. Data director 112 then
directs the range image data from range region 34 to a range shift
register 124. Thereafter, data director 112 discards the unused data 55
which follow the data resulting from the range image 54.
Subject shift register 122 and range shift register 124 are preferably
recirculating shift registers which on each shift provide the data which
is shifted out of the register both at their output terminal and to their
own input so that the data shifted out is not lost from the shift
register.
Once the data of the subject image 52 is stored in subject shift register
122 and the data of the range image 54 is stored in range shift register
124, a comparison is performed by a comparator 128 under the control of
central control 120. This comparison is performed on a cell by cell basis
as the data from corresponding cells of the subject image 52 and range
image 54 are successively shifted into comparator 128. The number of cells
which provide matching data are counted by a match counter 130 during each
comparison of the subject image 52 and the portion of the range image
contained in the given set of M successive cells. The more similar the
image contained in the selected M cells of range image 54 is to the
subject image 52 on a cell by cell basis, the greater will be the match
count obtained by counter 130. If the match counter counts M matches
during a single image comparison, then the data from every cell of the
subject image 52 is identical to the data from corresponding cell of the
selected M cells of the range image 54 and the subject image is identical
to the selected portion of the range image.
Each time the subject image is compared with a set of M cells of the range
image, the match count 130 is compared with the previous maximum match
count by a match count comparator 132. If the match count associated with
the most recent comparison is greater than the previous maximum match
count, the maximum match count is replaced by the new, larger, maximum
count in a maximum count register 134 and the location of the M cells of
the range image which yielded this count is entered in a maximum count
location register 136 which stores the location of the M cells of the
range image which provided the current maximum count.
When the match count of match counter 130 is equal to the previous maximum
count stored in register 134, no change is made in the maximum count
stored in register 134, but a duplicate-match-count flip-flop 138 is set
to indicate that two different sets of M cells of the range image 54
provide equally good matches with subject image 52. Each time the count in
match counter 130 is larger than the previous count 134, the
duplicate-match-count flip flop 138 is reset. When the count of match
counter 130 is less than the match count stored in registers 134, no data
changing action is taken and the next set of M cells of range image 54 are
compared with the subject image 52.
Each time a comparison cycle is performed, subject shift register 122
shifts M cells of data so that at the beginning of each comparison cycle,
data from the first cell of the subject image 52 is provided to comparator
128 as the first data for comparison. Each time a comparison cycle is
performed, range shift register 124 shifts N+1 cells of data. Comparator
128 compares the first M cells of data provided by shift register 124 with
the M cells of the data provided by subject shift register 122. The final
N+1-M cells of data provided by shift register 124 are not utilized by
comparator 128. Because range shift register 124 shifts N+1 cells of data
each time a comparison cycle is performed, each successive comparison
cycle begins with a range cell which is one cell further from the first
range cell than the cell with which the previous comparison began. Thus,
during the first comparison, the first range cell (C.sub.R1) is compared
with the first subject cell (C.sub.S1), the second range cell C.sub.R2 is
compared with the second subject cell C.sub.S2 and so on. During the
second comparison cycle cell C.sub.R2 is compared with cell C.sub.S1, cell
C.sub.R3 is compared with cell C.sub.S2 and so on. Thus, during successive
comparison cycles the subject image is "aligned" one cell further from the
first range cell (C.sub.R1) and compared with the corresponding cells of
the range image. During the (N+1-m).sup.th comparison cycle cell
C.sub.R(N.sub.-M) is compared with cell C.sub.S1, cell
C.sub.R(N.sub.+2.sub.-M) 2-M) is compared with cell C.sub.S2, . . . and
cell C.sub.RN is compared with cell C.sub.SM. At the end of the
(N+1-m).sup.th comparison cycle, subject image 52 has been compared with
each portion of the range image which could correspond to it and since the
final M bits of the range image 54 have been compared with the M bits of
the subject image 52, the comparison process is complete.
Either continuously, as each new maximum count location is stored in
register 136 or at the end of the comparison cycle, the location yielding
the maximum count is provided to a Range Look Up Table 140 which converts
the maximum count location to range in feet, meters, or whatever other
unit may be desired.
If the duplicate match count flip flop 138 is set when the comparison
process is complete, then two or more different sets of M bits of the
range image 54 provided equally good matches to the subject image 52 and
accurate range data has not been obtained. Under these conditions, central
control 120 resets the counting and data registers and initiates a new
image acquisition cycle. This reset does not change the integration time
of the image sensor array 30. The image acquisition and comparison process
as described above are then repeated.
If the duplicate match count flip flop 138 is not set at the end of a
comparison cycle then central control 120 provides the range of the object
being ranged to an optical system focus servo 160 and a display 164
through a control gate 162. When actuated, gate 162 gates the range
supplied by range look-up table 140 to the servo 160 and display 164. In
response to the information with respect to the range of the subject which
is received from data reduction system 100, the optical system 160 adjusts
the optical instrument to focus at the specified range.
If it is desired to determine range information on the basis of the average
of several image acquisitions then match count comparator 132 may be a set
of registers which store the match count generated for each set of M cells
of the range image 54. The match counts for each set of M cells can be
totaled over the desired number of data acquisition cycles and the
location providing the maximum total count can be selected as indicating
the subject range.
Data reduction system 100 is most easily implemented if the quantization
standard applied by data quantizer 110 results in binary determinations
and utilizes a median response of a photocell as a threshold reference
against which to decide whether the information from the cell should be
quantized as a binary 0 (cell response less than the threshold) or a
binary 1 (cell response greater than the threshold). The median response
(R.sub.M) of a photocell is that response which is midway between the
response generated in the absence of illuminating light (R.sub.o) and the
response (R.sub.S) generated when the photocell response has saturated as
a result of illumination by light of a greater intensity than that to
which the photocell can respond in an intensity differentiating manner.
Thus, the binary quantization threshold is preferably equal to (R.sub.S
-r.sub.O)/2. When the response (R) is less than
##EQU1##
the response is quantized as a binary 0. Similarly when
##EQU2##
the response is quantized as a binary 1.
Where the data quantizer 110 is binary, event counter 114 will count binary
ones in the subject image data. Subject shift register 122 will be M bits
long and range shift register 122 will be N bits long.
It will be understood that a binary data quantization provides less
sensitivity in image comparison than does a gray scale data quantization
which yields multibit quantized representation of the response of each
cell. A gray scale quantization which utilizes quantization thresholds of
##EQU3##
will provide two bits of data with respect to the response R of each cell.
A value of 00 can be generated when the response
##EQU4##
a value of 01 can be generated when
##EQU5##
a value of 10 can be generated when
##EQU6##
and a value of 11 can be generated when
##EQU7##
Thus, the response values 00 and 01 are assigned to responses which would
be assigned value 0 in a binary system and values 10 and 11 are assigned
to responses which could be assigned value 1 in a binary system. Thus,
although a gray scale quantization system is more complicated than a
binary system, a gray scale quantization system will provide more
information than by a binary quantization system will. If two or more
different segments of the range image provide equal match counts, then
where gray scale quantization is utilized a further comparison can be
performed between the subject data and the data segments of range image
which provided equal maximum counts. In this further comparison, the
degree of difference between the non-matching data cells can be utilized
to receive the question of which segment of the range image really
provides the best match for the subject image.
Data reduction system 100 can be constructed as an integrated circuit in a
single semiconductor chip. Image sensing array 30 may if desired be
provided on the same semiconductor chip as data reduction system 100.
Thus, the overall size of an automatic ranging system in accordance with
the invention is essentially controlled by the spacing which must be
provided between the lens 12 and the lens 14 in order that the range image
would be viewed along sightlines having detectably different viewing
angles if the object being ranged were moved far enough that the optical
instrument would produce a blurred image if the focus were not adjusted.
In order to obtain maximum effectiveness, the automatic ranging system, in
accordance with the invention, must have an accuracy which is greater than
the depth of the optical instrument field of view. Otherwise, optical
instrument blurring could result even though the optical instrument focus
was set in accordance with the range data provided by the automatic
ranging system.
To maximize ranging accuracy, ranging optical system 10 should be designed
such that any image blurring on sensing array 30 which occurs either at
the minimum or maximum functioning range of the automatic ranging system
is less than the size of an individual photocell within the array 30. In
these circumstances, lenses having a large depth of field should be
utilized in optical system 10.
In order to obtain accurate comparison between the subject image 52 and the
range image 54, the image sensing array 30 must be accurately aligned with
respect to the optics 10 so that the portion of the subject image which
impinges on subject portion 32 of array 30 is identical with that portion
of the ranging image of the subject which impinges on the range portion 34
of array 30. Where the lenses 12 and 14 are displaced from each other
along a horizontal line. A slight rotation of the sensor array 30 relative
to optics 10 will cause the "subject image" and the "range image" which
impinge on the photosensing array 30 to be from vertically displaced
portions of the object which is being ranged. Such a misalignment may
reduce the accuracy of the range determination made by the automatic
ranging system.
In the embodiment of FIG. 1, if a vertical line were drawn on the object to
be ranged in such a position that it fell in a transition region between
two adjacent photocells of the subject image, then depending on the range
of the object, that same vertical line could appear at the transition
between two adjacent photocells in the range region 34 of imager 30, or it
could appear anywhere along the width of a photocell. Thus, for some
ranges to the object those features of the object which appear centered
within photocells of the subject image are centered on the transition
regions between photocells of the range image and vice versa. This
characteristic can lead to poor range determination where an image
contains sharp intensity fluctuations.
A photocell array for use in the embodiment of FIG. 1 which overcomes this
problem is illustrated generally at 300 in FIG. 3. A first line 310 of
photocells corresponds to the photocell array 30 in which equal sized
photocells are uniformly spaced along the array by a distance which is
equal to the width of the photocell. Thus, within a subject image region
of the line 310, subject image photocells 310 are evenly spaced and within
a range region 364 of the photosensor array the photocells 314 are evenly
spaced as a continuation of the pattern in the subject image portion 362.
Where a vertical line 350 on the object to be ranged lies in a transi | | |
