Title: Focal point detection apparatus, focusing system, camera, and focal point detection method
Abstract: First to third liquid crystal filters are arranged at first to third different distances from a predeterminate image formation surface. Light flux passing through the liquid crystal filters is received by a photodiode and converted into an electric signal. A variable-density stripe pattern of a predetermined space frequency is formed on the first liquid crystal filter while the second and the third liquid crystal filter are in the state of entire transparency and a first signal is output from the photodiode. Similarly, a variable-density stripe pattern is formed on the second liquid crystal filter while the first and the third liquid crystal filter are in the state of entire transparency and a second signal is output. Similarly, a variable-density stripe pattern is formed on the third liquid crystal filter while the first and the second liquid crystal filter are in the state of entire transparency and a third signal is output. According to the first, second, and third signals, a focal point position is calculated.
Patent Number: 6,973,264 Issued on 12/06/2005 to Iwane
| Inventors:
|
Iwane; Toru (Chiyoda-ku, JP)
|
| Assignee:
|
Nikon Corporation (Tokyo, JP)
|
| Appl. No.:
|
490548 |
| Filed:
|
September 26, 2002 |
| PCT Filed:
|
September 26, 2002
|
| PCT NO:
|
PCT/JP02/09968
|
| 371 Date:
|
October 19, 2004
|
| 102(e) Date:
|
October 19, 2004
|
| PCT PUB.NO.:
|
WO03/029870 |
| PCT PUB. Date:
|
April 10, 2003 |
Foreign Application Priority Data
| Sep 27, 2001[JP] | 2001-296715 |
| Current U.S. Class: |
396/111; 396/121; 250/201.7 |
| Intern'l Class: |
G03B 013/36 |
| Field of Search: |
396/111,119,121
348/345-356
250/201.2,201.7
|
References Cited [Referenced By]
U.S. Patent Documents
| 4453818 | Jun., 1984 | Hayashi et al.
| |
| 4488799 | Dec., 1984 | Suzuki et al.
| |
| Foreign Patent Documents |
| A 59-0345/04 | Feb., 1984 | JP.
| |
| A 04-3330/10 | Nov., 1992 | JP.
| |
| A 07-0841/77 | Mar., 1995 | JP.
| |
Primary Examiner: Perkey; W. B.
Attorney, Agent or Firm: Oliff & Berridge PLC
Claims
1. A focal point detection device comprising:
a plurality of spatial modulation optical filters having adjustable transmission
characteristics which individually and selectively modulate an image-forming light
flux from an optical system, the plurality of spatial modulation optical filters
located at individually different positions on an optical axis;
a photoelectric conversion unit that sequentially receives the image-forming
light flux sequentially passed through the plurality of spatial modulation optical
filters and sequentially outputs signals, each of the signals corresponding to
an intensity level of received light for a different modulation state of the plurality
of spatial modulation optical filters; and
an arithmetic unit that calculates a focal point position of the optical system
based upon the signals output from the photoelectric conversion unit.
2. A focal point detection device, comprising:
a plurality of transmission-type liquid crystal filters disposed at different
positions on an optical axis, which an image-forming light flux from an optical
system sequentially enters;
a liquid crystal control unit that controls the individual transmission-type
liquid crystal filters so as to form a transmission characteristics pattern with
a predetermined spatial frequency at one of the plurality of transmission-type
liquid crystal filters with the remaining transmission-type liquid crystal filters
set in a state of full transmission;
a photoelectric conversion unit that receives a light flux having been transmitted
through the plurality of transmission-type liquid crystal filters and outputs a
signal corresponding to an intensity level of the received light; and
an arithmetic unit that calculates a focal point position of the optical system
based upon individual signals from the photoelectric conversion unit which are
obtained by sequentially forming the transmission characteristics pattern at the
plurality of transmission-type liquid crystal filters.
3. A focal point detection device, comprising:
a plurality of transmission-type liquid crystal filters disposed at different
positions on an optical axis, which an image-forming light flux from an optical
system sequentially enters;
a liquid crystal control unit that controls the individual transmission-type
liquid crystal filters so as to form an aperture pattern at one of the plurality
of transmission-type liquid crystal filters, to form a transmission characteristics
pattern with a predetermined spatial frequency at one of the remaining transmission-type
liquid crystal filters or at the aperture pattern and to set any transmission-type
liquid crystal filter at which neither the aperture pattern nor the transmission
characteristics pattern is formed in a state of full transmission;
a photoelectric conversion unit that receives a light flux having been transmitted
through the plurality of transmission-type liquid crystal filters and outputs a
signal corresponding to an intensity level of the received light; and
an arithmetic unit that calculates a focal point position of the optical system
based upon individual signals from the photoelectric conversion unit which are
obtained by sequentially forming the transmission characteristics pattern at the
plurality of transmission-type liquid crystal filters.
4. A focal point detection device according to claim 3, wherein:
the liquid crystal control unit implements control so as to form the aperture
pattern at a desired position within a display area of the transmission-type liquid
crystal filter.
5. A focal point detection device according to claim 2, wherein:
the transmission characteristics pattern is a variable-density stripe pattern
that repeats over a cycle corresponding to the predetermined spatial frequency.
6. A focal point detection device according to claim 5, wherein:
the arithmetic unit obtains filtering data defined by an even function at the
photoelectric conversion unit by forming a first variable-density stripe pattern
at the liquid crystal filter, obtains filtering data defined by an odd function
by forming a variable density pattern manifesting a 90° phase difference relative
to the first variable-density stripe pattern at the liquid crystal filter and calculates
the focal point position based upon the filtering data.
7. A focal point detection device according to claim 6, wherein:
the arithmetic unit calculates the focal point position by using an average value
of absolute values of the filtering data defined by the even function and the filtering
data defined by the odd function.
8. A focal point detection device according to claim 2, further comprising:
a frequency control unit that alters the spatial frequency.
9. A focal point detection device according to claim 8, wherein:
the frequency control unit reduces the spatial frequency under conditions in
which an extent of offset from the focal point position of the optical system is
significant and increases the spatial frequency under conditions in which the extent
of offset is small.
10. A focal point detection device according to claim 2, wherein:
the photoelectric conversion unit comprises a photodiode, a photomultiplier or
a CdS, having a light-receiving area large enough to receive the light flux from
the optical system in full.
11. A focal point detection device according to claim 2, wherein:
the plurality of liquid crystal filters are provided as an integrated optical
block having each liquid crystal filter set between transparent glass plates, a
first polarizing plate that converts an incident light flux to linearly polarized
light is disposed at a light flux entry surface of the optical block and a second
polarizing plate having a polarization angle equal to the polarization angle of
the first polarizing plate is disposed at a light flux exit surface of the optical block.
12. A focal point detection device according to claim 2, wherein:
three liquid crystal filters are provided.
13. A focus system comprising:
an optical system that forms a subject image;
a focal adjustment unit that adjusts a focal point position of the optical system;
a focal point detection device according to any of claims 1 through
12; and
a focus control unit that controls the focal adjustment unit to set the focal
point position of the optical system to a specific position based upon the focal
point position detected by the focal point detection device.
14. A camera comprising a focus system according to claim 13.
15. A focal point detection method comprising steps of:
modulating an image-forming light flux from an optical system at different positions
on an optical axis by using spatial modulation optical filters having transmission
characteristics with a predetermined spatial frequency;
sequentially receiving at a photoelectric conversion unit light fluxes obtained
by modulating the image-forming light flux at the different positions; and
calculating a focal point position of the optical system based upon signals sequentially
output by the photoelectric conversion unit.
16. A focal point detection method according to claim 15, wherein:
a pattern of the transmission characteristics is a variable-density stripe pattern
that repeats over a cycle corresponding to the predetermined spatial frequency.
17. A focal point detection method according to claim 16, wherein:
the focal point position is calculated based upon filtering data defined by an
even function which are obtained at the photoelectric conversion unit by forming
a first variable-density stripe pattern at the spatial modulation optical filter
and filtering data defined by an odd function which are obtained at the photoelectric
conversion unit by forming a variable density pattern manifesting a 90° phase
difference relative to a phase of the first variable-density stripe pattern at
the spatial modulation optical filter.
18. A focal point detection method according to claim 17, wherein:
the focal point position is calculated by using an average value of absolute
values of the filtering data defined by the even function and the filtering data
defined by the odd function.
Description
TECHNICAL FIELD
The present invention relates to a focal point detection device that detects
the focal point position of an optical system, a focus system achieved by using
the focal point detection device, a camera having the focus system and a focal
point detection method that may be adopted to detect the focal point of an optical system.
BACKGROUND ART
Autofocus systems used in cameras and the like in the known art adopt a
phase difference method, an active method or the like. In the phase difference
method, a virtual pupil is set at a position at which light exits the lens and
the focal point position is determined based upon parallax. The phase difference
method may be implemented in conjunction with a CCD array having numerous CCDs
each provided in correspondence to a specific parallax. Under normal circumstances,
a field lens is set at a position that is conjugate with a film surface or a plane
at which focus is assumed to be achieved, image-forming optical systems each corresponding
to one of the CCDs at the CCD array are disposed at a stage rearward relative to
the field lens and images are formed at the individual CCDs with the measurement
target light flux. Since the focal point is measured at a position conjugate with
the corresponding image-forming plane, the focus state at the image-forming position
can be ascertained substantially directly in the phase difference method.
In the active method, infrared light is usually irradiated on the subject, light
reflected off the subject is received at light-receiving elements such as SPDs
(silicon photodiodes) disposed to achieve a specific base length and a distance
is measured by adopting the principal of trigonometry. In other words, the distance
to the subject is determined through the active method instead of directly determining
the focus state. For this reason, an autofocus device adopting the active method
needs to drive out the objective lens to an extent corresponding to the distance
to the subject.
There is another autofocus method, i.e., the so-called contrast method, which
is adopted in electronic image-capturing apparatuses such as video cameras and
digital cameras. In this method, an image is captured with an image-capturing element
by driving a focusing lens and the focus state is ascertained based upon the output
from the image-capturing element. Under normal circumstances, it is assumed that
the focused state is achieved when the contrast information with regard to the
captured image indicates the maximum level. This method has an advantage in that
since the same image-capturing element is used to detect the focus state, the detection
can be executed free of any adverse effect of a mechanical error, an adjustment
error or the like.
When detecting the focal point by adopting the phase difference method, CCDs
need to be provided in correspondence to the detection areas and parallax, and
also it is necessary provide image-forming optical systems each in correspondence
to one of the CCDs. For this reason, it is technically difficult to distribute
AF areas over the entire subject image plane due to restrictions imposed by the
structure of the optical system, and only several AF areas are normally set. In
addition, once the system is manufactured by adopting a specific structure, the
system structure cannot be readily modified, and thus, the structure cannot be
adjusted in correspondence to varying operating conditions.
An autofocus system adopting the active method is a distance measuring device
as explained above and, for this reason, it cannot be utilized as an autofocus
system with a feedback loop that interlocks with the objective lens. In other words,
it can only be used as an open system, which is bound to be readily affected by
any apparatus assembly/adjustment error and whatever transpires in the system before
the autofocus operation is executed. Thus, it cannot be a strong, high precision
system that allows correction.
The contrast method, in which a peak in contrast is detected by moving the lens
while capturing an image, has been found less than ideal in terms of focusing speed.
Namely, the AF signal can only be updated at a video rate and the contrast peak
is only discerned after the lens passes through the focus point, and thus, there
is a problem in that a considerable length of time elapses before the lens is set
at the focus point.
DISCLOSURE OF THE INVENTION
An object of the present invention is to provide a focal point detection device
adopting a simple structure that is capable of detecting a focal point position
with ease, a focus system achieved by using the focal point detection device, a
camera having the focus system and a focal point detection method.
The focal point detection device according to the present invention comprises
a plurality of spatial modulation optical filters, having transmission characteristics
with a predetermined spatial frequency, which modulate an image-forming light flux
from an optical system at different positions on an optical axis, a photoelectric
conversion unit that receives light via the individual spatial modulation optical
filters and individually outputs signals in correspondence to varying intensities
of the light having been received and an arithmetic unit that calculates a focal
point position of the optical system based upon the signals output by the photoelectric
conversion unit.
The spatial modulation optical filters may each be constituted with a transmission-type
liquid crystal filter. In such a case, the individual transmission-type liquid
crystal filters should be controlled by a liquid crystal control unit so that a
transmission characteristics pattern with specific spatial frequency is formed
at one of the plurality of transmission-type liquid crystal filters while each
remaining transmission-type liquid crystal filters is set in a state of full transmission.
Namely, a plurality of sets of data should be obtained at the photoelectric conversion
unit and the focal point position should be calculated based upon the plurality
of sets of data.
An aperture pattern may be formed at one of the plurality of transmission-type
liquid crystal filters, a transmission characteristics pattern with a specific
spatial frequency may be formed at one of the other transmission-type liquid crystal
filters or at the aperture pattern and each transmission-type liquid crystal filter
having neither the aperture pattern nor the transmission characteristics pattern
formed therein may be set in a state of full transmission. In addition, the aperture
pattern may be formed at an arbitrary position in the display area of the transmission-type
liquid crystal filter.
The spatial frequency may be altered. In such a case, the spatial frequency should
be increased under conditions in which the extent of offset from the focal point
position of the optical system is significant and the spatial frequency should
be reduced under conditions in which the extent of offset is small.
It is desirable to constitute the photoelectric conversion unit with a photodiode,
a photomultiplier or a CdS, having a light-receiving area over which the entire
light flux from the optical system can be received in full.
The plurality of liquid crystal filters may be provided as an integrated optical
block having each liquid crystal filter set between transparent glass plates. In
this case, it is desirable to dispose a first polarizing plate that converts the
incident light flux to linearly polarized light at a light flux entry surface of
the optical block and to dispose a second polarizing plate achieving a polarization
angle equal to the polarization angle on the first polarizing plate at a light
flux exit surface of the optical block.
When the focal point detection device includes three liquid crystal filters,
first through third liquid crystal filters should be disposed over three different
distances, i.e., first through third distances, from an estimated image-forming
plane. In this case, a light flux transmitted through the liquid crystal filters
is received at the photoelectric conversion unit constituted of a photodiode or
the like and is converted to an electrical signal. The photodiode outputs a first
signal with a variable-density stripe pattern with a predetermined spatial frequency
formed at the first liquid crystal filter and the second and third liquid crystal
filters set in a state of full transmission. Likewise, the photodiode outputs a
second signal with a variable-density stripe pattern with a predetermined spatial
frequency formed at the second liquid crystal filter and the first and third liquid
crystal filters set in a state of full transmission, and the photodiode outputs
a third signal with a variable-density stripe pattern with the predetermined spatial
frequency formed at the third liquid crystal filter and the first and second liquid
crystal filters set in a state of full transmission. Then, the focal point position
is calculated by using the first through third signals.
The focus system according to the present invention comprises an optical system
that forms an image of a subject, a focal adjustment unit that adjusts a focal
point position of the optical system, the focal point detection device described
above and a focus control unit that controls the focal adjustment unit so as to
set the focal point position of the optical system to a specific position based
upon the focal point position detected by the focal point detection device. This
focus system may be mounted at a camera.
The focal point detection method according to the present invention comprises
steps for modulating an image-forming light flux from an optical system at different
positions on an optical axis by using spatial modulation optical filters having
transmission characteristics with a predetermined spatial frequency, sequentially
receiving at a photoelectric conversion unit individual light fluxes obtained by
modulating the image-forming light flux at the different positions and calculating
a focal point position of the optical system based upon signals sequentially output
by the photoelectric conversion unit.
The transmission characteristics pattern mentioned above, i.e., the modulation
pattern, is a variable-density stripe pattern that repeats in cycles defined by
the predetermined spatial frequency. Such a variable-density stripe pattern may
have a transmission factor corresponding to the Fourier cos wave or a transmission
factor corresponding to the Fourier sin wave. A light flux passing through the
variable-density stripe pattern with such a transmission factor undergoes Fourier
transformation to enable an analysis of the spatial frequency component with a
predetermined spatial wavelength at a specific position of the corresponding filter.
In other words, since the spatial frequency component is attenuated in correspondence
to the extent of image offset from the focal point position, the image offset quantity
can be calculated based upon a plurality of signal intensity levels corresponding
to the light fluxes transmitted through the plurality of spatial modulation optical
filters such as a plurality of liquid crystal filters.
In addition, a first variable density stripe pattern may be formed at the liquid
crystal filters or the like to enable the photoelectric conversion unit to obtain
filtering data defined by an even function, a second variable density pattern manifesting
a 90° phase difference relative to the phase of the first variable-density
stripe pattern may be formed at the liquid crystal filters or the like to obtain
filtering data defined by an odd function and the focal point position may be calculated
based upon these data. More specifically, the focal point position can be calculated
by using an average of the absolute values of the even function filtering data
and the odd function filtering data. Thus, with the data obtained by using the
filters manifesting a 90° phase difference from each other, an accurate focal
point detection can be achieved even when the focal point detection target is a
dynamic object.
According to the present invention, the position at which the light intensity
level peaks, detected by the photoelectric conversion unit, matches the focal point
position of the optical system, and thus, the focal point position of the optical
system can be ascertained with ease by calculating a signal peak position based
upon the individual signals resulting from modulating the image-forming light flux
at the different positions on the optical axis. Furthermore, the device structure
can be simplified by using the photoelectric conversion unit to detect the intensity
levels of the light fluxes resulting from modulating the image-forming light flux
at the spatial modulation optical filters or the transmission-type liquid crystal filters.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a block diagram of the focal point detection device achieved in a
first embodiment of the present invention;
FIG. 2 shows in detail the detection unit 1;
FIG. 3 presents another example of the detection unit 1;
FIG. 4 shows the filter 2A at which a variable-density stripe filter
pattern is formed;
FIG. 5 illustrates an instance of defocusing;
FIG. 6 shows the relationship between the defocus quantity x and I;
FIG. 7 presents measurement data P 1 through P 3;
FIG. 8 shows light intensity levels I corresponding to varying wavelengths R;
FIG. 9 shows filter patterns, with FIG. 9(
a) showing longitudinal stripes
and FIG. 9(
b) showing lateral stripes;
FIG. 10 shows the detection unit 100 having two filters 2A and 2B;
FIG. 11 shows the filter pattern display at the filter 2C in a second embodiment;
FIG. 12 illustrates a method adopted to control the filters 2A to 2C,
with FIG. 12(
a) showing the first procedure, FIG. 12(
b) showing the
second procedure and FIG. 12(
c) showing the third procedure;
FIG. 13 illustrates the second method adopted to control the filters 2A
to 2C with FIG. 13(
a) showing the first procedure, FIG. 13(
b)
showing the second procedure and FIG. 13(
c) showing the third procedure;
FIGS. 14(
a) and 14(
b) show the arrangements of the window
forming areas 202
a to 202
e;
FIG. 15 is a block diagram of the focal point detection device achieved in a
third embodiment of the present invention; and
FIG. 16 is a block diagram of a structure adopted to drive one of the filters,
i.e., the filter 2A, with the actuator 300.
BEST MODE FOR CARRYING OUT INVENTION
The following is an explanation of the embodiments of the present invention,
given in reference to the drawings.
First Embodiment
(Device Structure)
FIG. 1 is a schematic block diagram of the structure of an autofocus system
that includes the focal point detection device achieved in the first embodiment
of the present invention. Subject light exiting a photographic lens 8 is
detected at a detection unit 1. The detection unit 1 includes a filter
block 2 and a photodiode 3 constituting a photoelectric conversion
element. The photoelectric conversion element may instead be constituted with a
CdS, for instance. While the detection unit 1 is shown rearward relative
to the photographic lens 8 in FIG. 1, the detection unit 1 is actually
disposed at a position that is substantially conjugate with the image-capturing
plane or an estimated focus-matching plane. While the subject light having passed
through the lens 8 is guided to silver halide film, a CCD image-capturing
element or the like (not shown), it is split in the optical path and thus some
of the subject light enters the detection unit 1 as well.
The filter block 2 includes three filters 2A, 2B and 2C
which are sequentially disposed along the optical axis so that the filter 2A
is set closest to the photodiode 3, the filter 2B is set next to
the filter 2A and the filter 2C is set furthest from the photodiode
3. As detailed later, the filters 2A to 2C are each constituted
of a liquid crystal filter, and the states of the liquid crystals at the filters
2A to 2C are controlled by filter control units 4A to 4C
respectively, independently of one another. An output from the photodiode 3
is first amplified at an amplifier 5, the amplified output is converted
to a digital signal at an A/D converter 6 and the digital signal is then
input to a control device 7. An arithmetic unit 7
a of the
control device 7 calculates a focal point detection state based upon the
signal input thereto. The control device 7 controls the filter control units
4A to 4C and also controls a lens drive device 9 to enable
an autofocus operation.
FIG. 2 shows the structure adopted in the detection unit 1. In FIG. 2,
reference numerals 11A and 11B each indicate a polarizing plate and
reference numeral 14 indicates a glass substrate. The glass substrates 14
are positioned over a predetermined pitch d. The filters 2A to 2C
are each set between a pair of glass substrates 14. It is to be noted that
the focal point detection is executed by receiving the subject light flux in its
entirety in the embodiment. Accordingly, the light-receiving surface at the photodiode
3 ranges over an area large enough to fully contain the entire subject image
at the image plane. However, if the focal point detection is executed by using
only a portion of the subject light flux, the light-receiving surface only needs
to range over an area large enough to cover that portion.
As shown in FIG. 2, the filters 2A to 2C are each constituted of
a pair of transparent conductive films 13 and a liquid crystal 12
set between the transparent conductive films. The four glass substrates 14
and the three sets of filters 2A to 2C are enclosed between the polarizing
plates 11A and 11B. Since the glass substrates 14 are disposed
with the pitch d as described earlier, the filters 2A to 2C, too,
are set over the interval d. It is to be noted that while the common polarizing
plates 11A and 11B are shared by the three layers of liquid crystals
12 in FIG. 2, three independent transmission-type liquid crystal panels
15 each having a pair of polarizing plates may be stacked to constitute
the filters 2A to 2C, as shown in FIG. 3, instead.
The subject light flux having entered the detection unit 1 from the left
side in FIG. 2 is polarized at the polarizing plate 11A to become linearly
polarized light. The polarization angle of this light is altered as it passes through
the liquid crystals 12 of the individual filters 2A to 2C.
As a result, the intensity level of the light exiting the polarizing plate 11B
is determined by the angle formed with the transmission axis of the polarizing
light 11B and the direction along which the linearly polarized light oscillates.
For instance, when δ represents the angle formed by the transmission axis
and the oscillating direction, the intensity level of the transmitted light is
in proportion to (cos δ)
2. In other words, by controlling the
liquid crystal filters 2A to 2C with the filter control units 4A
to 4C so as to adjust the extent of change in the polarization angle occurring
as the light flux passes through the filters 2A to 2C, the intensity
of the transmitted light can be altered.
For instance, when δ=0(deg) all the incoming light is transmitted, whereas
when δ=90 (deg), the incoming light is completely blocked by the filter.
Such control may be implemented individually for the filters 2A to 2C
and various transmission patterns may be formed at the individual filters 2A
to 2C. For instance, a variable-density stripe pattern such as that shown
in FIG. 4 is observed from the side on which the photodiode 3 is present,
at the filter 2B having formed therein a pattern in which transmission and
non-transmission alternate in stripes, since the intensity level of the transmitted
light changes in a stripe pattern. In the following explanation, this situation
is described as "a variable-density stripe filter pattern such as that shown in
FIG. 4 is formed at the filter 2B".
In the embodiment, a variable density filter pattern having the spatial frequency
shown in FIG. 4 is individually formed at the filters 2A to 2C. Disposing
such filters 2A to 2C within the light flux is equivalent to spatially
modulating the light flux entering the filters. A modulated light flux is converted
to an electric current corresponding to the level of the intensity of the light flux.
(Defocusing)
First, the defocusing extent, which is equivalent to the focal point offset
quantity is explained. Let us now consider the situation shown in FIG. 5, in which
a focal point 20 of the lens 8 is at a position x to the right of
the filter 2B. The extent of blurring in the subject image at the position
x can be expressed as the diameter r of a plane along which a circular cone 21
with its vertex set at the focal point 20 is cut through by the filter 2B.
With θ representing the angle formed by the optical axis and the generating
line of the circular cone, the diameter r is expressed as in (1) below.
Since tan θ in the system represents the product of the reciprocal of
the F value (=focal length/aperture) of the lens and ½, r presenting the blurring
extent can be expressed as in (2) below.
r in this expression can be regarded as an attenuation wavelength at the primary
filter. Accordingly, the attenuation f of a subject image manifesting at a spatial
wavelength R due to the blurring effect can be expressed as in (3) below.
I in the following expression (4) assumes the absolute value of f and represents
the attenuation attributable to the defocusing quantity x of the spatial frequency
component at the wavelength R.
As explained later, the value of I corresponds to the intensity level of the
light
detected at the photodiode 3, and in the explanation of the embodiment,
the intensity level of the detected light, too, is expressed as I.
(Focal Point Detection Method)
Expression (4) indicates that the attenuation of the spatial frequency
component at the wavelength R is defined by the defocusing quantity x. This implies
that once the spatial frequency component at the wavelength R is ascertained at
a plurality of positions, the defocusing quantity x can be estimated. A curve L1
in FIG. 6 indicates the change in the value of I relative to the defocusing quantity
x. When x=0, i.e., when the filter 2B is set at the focal point position,
I=1, and the value of I becomes smaller as the absolute value of x increases. The
spatial frequency component at the wavelength R at a specific position can be ascertained
through a Fourier transformation of the light flux executed at the position.
The variable-density stripe filter pattern shown in FIG. 4 achieves a horizontal
symmetry relative to the origin point set at the center of the image plane and
has a transmission factor corresponding to the Fourier cos wave. It is to be noted
that the wavelength R is equal to the distance from a given non-transmission area
to the next non-transmission area. When the filter 2B having such a variable-density
stripe filter pattern is inserted in the light flux, the light flux undergoes Fourier
cos transformation as it is transmitted through the filter 2B. Namely, the
intensity of the light flux advancing toward the photodiode 3 can be expressed
as in (5) below.
Io in the expression above is a coefficient indicating the transmission factor
of the liquid crystal filter inclusive of the optical system. In addition, cos
θ
f represents the Fourier cos transformation.
In expression (5) θ represents the spatial frequency corresponding to the
spatial wavelength R and I
θ represents the spatial frequency component
at the wavelength R. It is to be noted that I
θresulting from a
Fourier sin transformation can be obtained by forming a variable-density stripe
filter pattern with a transmission factor corresponding to the Fourier sin wave
at the filter 2B. The focal point position may be calculated by adopting
either transformation.
When the light flux having been modulated at the filter 2B as described
above is received at the photodiode 3, the output of the photodiode 3
indicates the intensity level achieved by executing a Fourier transformation in
hardware. If the wavelength R of the variable-density stripe filter pattern is
altered, the Fourier transformation is executed in correspondence to the adjusted
wavelength R. For this reason, when the outputs measured in correspondence to varying
positions x are plotted on the x-I plane in FIG. 6, the data form a curve having
a profile identical to that of the curve L 1 in FIG. 6. In other words,
the curve peaks at x=0.
As shown in FIG. 2, the detection unit 1 includes the three filters 2A,
2B and 2C disposed over the pitch d. Thus, three sets of data can
be obtained for a single subject image by forming the variable-density stripe filter
pattern at the filter 2A and forming a full transmission pattern at the
remaining filters, by forming the variable-density stripe filter pattern at the
filter 2B and forming a full transmission pattern at the remaining filters
and by forming the variable-density stripe filter pattern at the filter 2C
and forming a full transmission pattern at the remaining filters.
It is assumed that the filter block 2 is disposed so as to set the filter
2B at a position conjugate with the image-capturing plane of the image-capturing
device in the embodiment. Accordingly, the light intensity level Yb measured by
forming the variable-density stripe filter pattern at the filter 2B is expressed
as in (6) below. When the variable-density stripe filter pattern is formed at the
filter 2A, the light intensity level Ya expressed as in (7) below is measured,
whereas the light intensity level Yc expressed as in (8) below is measured when
the variable-density stripe filter pattern is formed at the filter 2C. It
is to be noted that a indicates the intensity of light measured when the light
flux is not attenuated at all. In FIG. 7, the individual sets of measurement data
P1 to P3 are plotted on the light intensity curve Y.
By solving these equations (6) to (8), the position x of the filter 2B
and the light intensity level a when there is no attenuation, which are unknowns,
can be calculated. The position x is indicated with a positive value when the position
x is located further toward the lens 8 relative to the focal point 20,
as shown in FIG. 5. In the data example presented in FIG. 7, the focal point position
of the lens 8 is set rearward relative to the filter 2B over the
distance x from the filter 2B. In this case, by driving the lens 8
with the lens drive device 9 and moving the focal point position forward
by the distance x, a focused state is achieved.
Providing three equations to ascertain two unknowns x and a, as in the
example described above manifests redundancy. Accordingly, the attenuation constant
r (see expression (3)) representing the blurring extent mentioned earlier, too,
can be calculated based upon the measurement data obtained during the focal point
detection. Since this constant r may not always accurately indicate the extent
of a specific type of blurring depending upon the structure of the lens 8
or the position of the subject, it is desirable to execute the calculation by taking
into consideration the actual structural and positional details.
(Filter Patterns)
Expression (4) presented earlier indicates that the intensity I is dependent
on the spatial wavelength R at the individual filters 2A to 2C. Curves
L11 to L13 in FIG. 8 indicate light intensity levels I corresponding
to varying spatial wavelength R. The wavelength R2 corresponding to the
curve L12 is expressed as R2=R1/2 with R1 representing
the wavelength corresponding to the curve L11, whereas the wavelength R3
corresponding to the curve L13 is expressed as R3=R1/4. Namely,
as the wavelength R of the stripe pattern becomes smaller, the curve peaks more steeply.
For this reason, when the focal point position is close to the filters 2A
to 2C, the accuracy with which the focal point position is calculated by
using the curve L13 indicating a more drastic change in the vicinity of
the focal point position is higher. The data obtained when the focal point position
is distanced from the filters 2A to 2C, on the other hand, indicate
values toward the nadir of the curves L11 to L13, and for this reason,
the focal point position can be calculated with greater use by using the curve
11 which manifests a relatively large change and achieves a larger output.
In other words, when the defocusing quantity x is large, the wavelength R of the
variable-density stripe filter pattern should be increased, whereas when the defocusing
quantity x is small, a highly accurate focal point position detection can be achieved
by reducing the wavelength R. For instance, when the autofocus system is mounted
at a camera, R should be set to a large value when starting the AF (autofocus)
operation and R should then be reduced as the focus state approaches a matched
state. In addition, if the output value that has been obtained is smaller than
a reference value indicating the minimum value needed to enable calculation, data
should be reacquired by setting the wavelength R to a greater value. The reference
value may be, for instance, k·Ic obtained by multiplying the DC component
Ic of the light to be detailed later by a constant ratio k.
The sets of data P1 to P3 in FIG. 7 are sequentially obtained as
the variable-density stripe filter pattern is formed at the filters 2A,
2B and 2C in this order. Since the light intensity levels I to be
compared indicate levels of intensity of light corresponding to a single image,
a subject should ideally remain still while the data are being taken. However,
it is likely that the subject will move within the image plane, i.e., the phase
shifts, in reality. Under such circumstances, the phase shift can be handled by
taking the absolute values of both the real part (cos transformation) and the imaginary
part (sin transformation) of the Fourier transformation instead of using only either
part of the Fourier transformation.
More specifically, an absolute value Y as expressed in (9) below should be taken
by using data Ycos obtained when a filter pattern defined with an even function
(cos) is displayed at the center of the image plane and data Ysin obtained when
a filter pattern defined with an odd function (cos) is displayed, relative to the
center of the image plane.
While only the real part needs to be targeted under normal circumstances, the
focal point position should be calculated to based upon the absolute value Y by
displaying the stripes defined with the even function and the stripes defined with
the odd function to achieve a more precise measurement. By adopting this method,
an accurate focal point detection can be executed even when a pronounced vibration occurs.
(Actual Operation)
When detecting the light intensity level with the photodiode 3 by forming
a filter pattern at the filters 2A to 2C, it is necessary to ascertain
the contrast of the liquid crystals 12 at the filters 2A to 2C
and also to obtain calibration data with regard to the photodiode 3, since
a dark current flows at the photodiode 3 and thus, even when the liquid
crystals 12 are in a transmitting state or a light blocking state, they
do not allow 100% of the light to be transmitted or they do not block 100% of the light.
First, the liquid crystals 12 at all the filters 2A to 2C
are set in a transmitting state and an output Ic from the photodiode 3 in
this state is recorded. Next, an output Id from the photodiode 3 is recorded
by setting one of the filters 2A to 2C in a full light blocking state.
At this time, the remaining two filters are sustained in the transmitting state.
While Id is measured by setting only one of the filters 2A to 2C
in the light blocking state by assuming that the light blocking states at the filters
2A to 2C are all identical to one another, three outputs Id may be
individually measured each by setting one of the filters in the full light blocking
state instead. By executing the arithmetic operation with the three outputs Id,
an even higher degree of accuracy is achieved in the focal point position calculation.
When the output from the photodiode 3 is Ir, the corresponding effective
light intensity level I is calculated by subtracting the output Id attributable
to the dark current from the detection value Ir as expressed in (10) below.
The effective value Icc of the full open output achieved by setting all the filters
2A to 2C in the transmitting state is expressed as in (11), which
constitutes a Fourier output (DC component) of degree zero.
It is to be noted that Ic represents the output from the photodiode 3 achieved
by setting the liquid crystals 12 is all the filters 2A to 2C
in the transmitting state.
By using Icc as reference data, a decision can be made as to whether or not the
value obtained through the measurement is reliable.
As explained earlier, when a filter having formed thereat the variable-density
stripe filter pattern with a transmission factor corresponding to the cos wave
such as that shown in FIG. 4, is inserted in a light flux, the light flux undergoes
the Fourier cos transformation. By adjusting the spatial wavelength R of the variable-density
stripe filter pattern, the Fourier transformation corresponding to the varying
spatial wavelengths R are output from the photodiode 3. The output Icos
is expressed as in (12) below by taking into consideration the output Id attributable
to the dark current and the like.
In addition, by displaying the variable-density stripe filter pattern in FIG.
4 with its phase offset by π/2, a Fourier sin transformation expressed as
in (13) is effected.
In the expression given above, sin θ
f represents the Fourier
sin transformation.
Ycos and Ysin in expression (9) explained earlier are obtained by eliminating
the offset component (the second term) from Icos and Isin respectively.
If the filters 2A to 2C are constituted with general-purpose liquid
crystal panels 15, as shown in FIG. 3, the level of quantization becomes
a factor to be taken into consideration when displaying a variable-density stripe
filter pattern with an extremely small wavelength. In such a case, a rectangular
wave with the harmonic component the nature of which is well known, should be used
instead of a single cos wave (or sin wave). FIG. 9 shows such filter patterns,
with FIG. 9(
a) showing longitudinal stripes and FIG. 9(
b) showing
lateral stripes. In FIG. 9(
a), a fully transmitting longitudinal stripe
area 20A and a fully light blocking longitudinal stripe area 20B
are set alternately and repeatedly along the horizontal direction. In FIG. 9(
b),
a fully transmitting lateral stripe area 21A and a fully light blocking
lateral stripe area 21B are set alternately and repeatedly along the vertical direction.
Although not ideal, a rectangular wave can be used in practical applications
virtually without problems, unlike a single sin wave or cos wave. In addition,
taking into consideration the difficulty of displaying gradations with a liquid
crystal, the rectangular wave which enables a simple display can be handled with
greater ease than a sin wave or cos wave.
It is to be noted that if the subject manifests almost no change in contrast
along
the lateral direction, the output value becomes extremely small and thus the arithmetic
operation cannot be executed if a modulation is executed by using the longitudinal
stripe pattern such as that shown in FIG. 4 or in FIG. 9(
a). Accordingly,
a lateral stripe pattern such as that shown in FIG. 9(
b) is formed to calculate
the focal point position in such a case. If, on the other hand, the subject does
not manifest a substantial change in contrast along the longitudinal direction,
the output value obtained by using a lateral stripe pattern will be extremely small
and accordingly, the focal point position is calculated by forming a longitudinal
stripe pattern. In addition, a stripe pattern with diagonal stripes instead of
longitudinal stripes or lateral stripes may be used. In this case, it is possible
to calculate the focal point with ease when the subject manifests little change
in contrast along either the longitudinal direction or the lateral direction.
(Example of Variation)
While three filters 2A to 2C each having a liquid crystal 12
constitute the filter block 2 in the first embodiment described above, the
number of filters does not need to be three. Namely, a greater number of filters
may be used as long as the Fourier transformation output values from all the filters
can be detected within a range in which the image can still be regarded as a still
image when a large number of frames are displayed at the liquid crystals 12.
A greater number of filters will lead to an improvement in the accuracy of the
arithmetic operation. However, at present, the length of time for the on/off operation
at the liquid crystals 12 limits the number of filters that can be provided.
As expressions (6) through (8) indicate, in the equations of the intensity levels
Y, there are only two unknowns, x and a. For this reason, the minimum requirements
with regard to the number of filters is two, and the structure of the device can
be simplified by using a smaller number of filters. At a detection unit 100
shown in FIG. 10, a filter block 101 includes two filters 2A and
2B. The detection unit 100 in FIG. 10 differs from the detection
unit 1 in FIG. 2 only in the number of filters it includes, and structural
details of the filters 2A and 2B are identical to those of the filters
in FIG. 2. In this device, two sets of data defined as in expressions (6) and (7)
explained earlier are obtained through measurement, and thus, x and a can be determined
by solving the equations (6) and (7). However, if Ya=Yb, the filters 2A
and 2B are assumed to be at positions on the two sides of the focal point,
over distances equal to each other from the focal point position and the focal
point is considered to be at the middle position between the filters 2A
and 2B, i.e., x+d/2.
Second Embodiment
While the focal point detection is executed by using the entire light flux
from the lens 8 in the first embodiment, a focal point detection may instead
be executed by using a light flux in a given specific area of the image plane.
FIG. 11 shows a filter pattern displayed at the filter 2C. A rectangular
window 201 is formed in the filter pattern, and a variable-density stripe
pattern similar to that shown in FIG. 4 is formed in the area corresponding to
the window 201. The area outside the window 201 is in a full light
blocking state. It is to be noted that the shape of the window 201 does
not need to be rectangular. If the filter 2C is a liquid crystal filter,
the window 201 may be formed at any position, in any size and in any shape,
as long as the requirements regarding the quantization to be achieved with the
pixels are taken into consideration.
While the full-open output Ic is detected as the calibration data by setting
all the filters 2A to 2C in a state of full transmission in the first
embodiment, the filter having formed therein the window 201 is set in a
state of full transmission only over the area corresponding to the window 201
in the second embodiment. Thus, Ic is provided as a local variable which changes
in correspondence to the form adopted for the window 201. It is to be noted
that the filters other than the filter 2C with the window 201, i.e.,
the filters 2A and 2B, are set in a state of full transmission as
in the first embodiment.
Filter control methods that may be adopted during the focal point detection,
i.e., during the Fourier transformation, include the following.
In a first method, the window 201 is formed at a specific filter, e.g.,
the filter 2C, among the filters 2A to 2C. The data can be
obtained by first forming the window 201 at the filter 2C and setting
the entire area inside the window 201 in a transmitting state, as shown
in FIG. 12(
a). The filter 2B is set in a state of full transmission,
and a variable-density stripe filter pattern similar to that shown FIG. 4 is formed
at the filter 2A. The output from the photodiode 3 is obtained as
first data in this state.
Next, the filter 2A is set in a state of full transmission and the variable-density
stripe filter pattern is formed at the filter 2B, as shown in FIG. 12(
b).
The state of the filter 2C remains unchanged from that shown in FIG. 12(
a).
Second data are obtained in the state shown in FIG. 12(
b). Lastly, the filters
2A and 2B are both set in a state of full transmission, as shown
in FIG. 12(
c). A variable-density stripe filter pattern similar to that
formed at the filters 2A and 2B as shown in FIGS. 12(
a) and
12(
b) is displayed in the area within the window 20l formed
at the filter 2C. Third data are obtained in the state shown in FIG. 12(
c).
Then, the focal point position is calculated through a procedure similar to that
adopted in the first embodiment by using the three sets of data. It is to be noted
that the calculation procedure is not explained here.
In a second method, a window 201 having a variable-density stripe filter
pattern is sequentially formed at the filters 2A to 2C as shown in
FIG. 13. First, the window 201 is formed at the filter 2A and the
variable-density stripe filter pattern is displayed inside the window 201,
as shown in FIG. 13(
a). At this time, the other filters 2B and 2C
are in a state of full transmission. First data are obtained in the state shown
in FIG. 13(
a).
Next, as shown in FIG. 13(
b), the window 201 with the variable-density
stripe filter pattern displayed therein is formed at the filter 2B and the
filters 2A and 2C are set in a state of full transmission to obtain
second data. Lastly, as shown in FIG. 13(
c), the window 201 with
the variable-density stripe filter pattern displayed therein is formed at the filter
2C and the filters 2A and 2B set in a state of full transmission
to obtain third data. Then, the focal point position is calculated through a procedure
similar to that adopted in the first embodiment by using the three sets of data.
In the first method, all the data are obtained with the window 201 formed
at the same filter 2C. Thus, there is an advantage in that the area of the
light flux cut off with the window 201 remains unchanged at all times. This
is a crucial advantage from the viewpoint of assuring the required level of accuracy
in a device that detects the light fluxes resulting from the Fourier transformation
executed at the individual filters 2A to 2C and calculates the focal
point position by comparing the detection results. It is to be noted that the window
201 used to cut off the light flux should be formed at the filter 2C,
which is the closest to the lens 8 when adopting this method.
The second method, in which light flux areas cut off with the window 201
to obtain the individual sets of data vary, requires correction. For instance,
the position of the window 201 may need to be corrected based upon the distance
from the optical axis to the edge of the window 201. For further accuracy,
the correction needs to be executed by taking into consideration the position of
the focal point relative to the filters, and for this reason, if the focal point
position of the lens 8 is altered when executing a focusing operation, the
correction must be re-executed accordingly. These factors make it extremely difficult
to achieve a highly accurate correction. It is to be noted that while the variable-density
stripe filter pattern is formed inside the window 201 in the first and second
methods described above, a filter having a window 201 alone formed therein
may be provided and, in this case, three filters at which the variable-density
stripe filter pattern is formed may be provided in addition to the filter with
the window 201.
There are cameras known in the related art that execute a focusing operation
by selectively using one of a plurality of focus areas at which a focus-match can
be achieved formed on the image plane. In this embodiment, too, a similar focusing
operation can be executed by selectively using a window 201 formed at one
of a plurality of window forming areas. For instance, as shown in FIG. 14(
a),
five window forming areas 202
a, 202
b, 202
c,
202
d and 202
e may be set on the image plane at the
filter 2C with a window 201 formed at one of the five window forming
areas. FIG. 14(
b) shows the window 201 formed at the area 202
c.
It is to be noted that the number of window forming areas does not need to be five,
and that the shape of the window 201 does not need to be rectangular.
(Photometering)
The intensity level of the light detected by setting the window 201 in
a state of full transmission corresponds to the term of degree zero in the Fourier
transformation, i.e., the DC component, which is defined by the window 201.
Thus, by adjusting the shaped or the position of the window 201 in a state
of full transmission as appropriate, the intensity level of light at any area on
the screen can be measured. This is equivalent to the multiple split photometering
adopted in cameras and the like. In this application, too, the window 201
is formed at the filter 2C. During a photometering operation, the window
201 is set in a transmitting state, and the other filters 2A and
2B are set in a state of full transmission.
As described above, a focal point detection can be executed by forming the variable-density
stripe filter pattern over the area where the window 201 is formed and a
photometering operation can be executed by setting the window 201 in a state
of full transmission in the embodiment. In photometering systems in the related
art, the absolute value of the output from the photodiode constituting the light-receiving
element is directly used, and the accuracy is bound to be lowered by the adverse
effect of the dark current. In contrast, a dark state can be created by setting
the filters 2A to 2C an a full light blocking state and thus, the
effective light intensity level can be measured exclusively by ascertaining the
differences between the full transmitting state and the full light blocking state.
As a result, a highly accurate photometering operation can be executed.
It is to be noted that while areas at which the window 201 can be formed
are set in advance as shown in FIG. 14(
a) in the example explained above,
the filters 2A to 2C are liquid crystal filters and thus, the window
201 can be formed at any position on the display screen. A photometering
operation can be executed at any desired position on the screen by forming the
window 201 at the desired position.
Third Embodiment
FIG. 15 shows the focal point detection device achieved in the third embodiment
of the present invention in a block diagram similar to that presented in FIG. 1.
In the focal point detection device in FIG. 15, a photomultiplier 210, instead
of the photodiode 3 in FIG. 1, is used to constitute the light-receiving
element. Reference numeral 211 indicates a power supply for the photomultiplier
210. As weak subject light passes through the filter block 2 and
enters a photoelectric surface of the photomultiplier 210, photo electrons
are emitted from photoelectric surface. The photo electrons are amplified through
a secondary electron emission and ultimately, a signal having been amplified by
a factor of 500,000 to 1 million is output from the photomultiplier 210.
The signal output from the photomultiplier 210 is amplified at the amplifier
5 and is then input to the A/D converter. Other structural features are
completely identical to those in the device shown in FIG. 1. Since details of the
structure adopted in the filter block 2, the focal point detection operation
and the lik
