JPH0584882B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to improving the focus detection ability for a periodic pattern object in a pupil division image shift detection type focus detection device.

[Conventional technology]

Conventionally, as one type of camera focus adjustment device, the exit pupil of the photographic lens is divided into two by a focus detection optical system, and the two subject images formed by the light flux passing through each pupil area are A method is known in which light is received by a photoelectric conversion element array (for example, a CCD sensor array), the focal state of the photographic lens is detected from the output, and the photographic lens is driven based on the detection result.

Photographing lens whose focus should be detected in Fig. 1
A field lens FLD is placed on the same optical axis as the LNS. Behind it, two secondary imaging lenses FCLA and FCLB are arranged symmetrically with respect to the optical axis. Furthermore, sensor arrays SAA and SAB are arranged behind it. Apertures DIA and DIB are provided near the secondary imaging lenses FCLA and FCLB. The field lens FLD separates the exit pupil of the photographing lens LNS into two
The image is almost formed on the pupil plane of the next imaging lenses FCLA and FCLB. As a result, the ray bundles incident on the secondary imaging lenses FCLA and FCLB are placed on the exit pupil plane of the photographing lens LNS, respectively.
They are emitted from areas of equal area that do not overlap with each other and correspond to the FCLB. The aerial image formed near the field lens FLD is sent to the sensor array SAA, by the secondary imaging lenses FCLA and FCLB.
When the image is re-imaged on the surface of SAB, sensor arrays SAA and SAB are
The two images above will change their positions. Therefore, the displacement (shift) of the relative positions of the two images on the sensor array
By detecting the amount, it is possible to know the focal state of the photographing lens LNS.

The focus state detection method described above may not work well depending on the subject conditions. The most typical example is a low-luminance subject. There is a practical limit to the photoelectric charge accumulation time of a photoelectric sensor, and if there is not enough light to generate a sufficient photocharge during that time, a signal cannot be formed. In such a case, even if the storage time is extended, the camera becomes extremely difficult to use, or the dark current increases and the effective S/N ratio does not improve. Therefore, a focus detection auxiliary light is often installed in the camera to compensate for the lack of light intensity at low brightness times. The upper optical system in Figure 1 is the auxiliary light projection system, and the light emitted by the LED light source illuminates the pattern chart CHT via the condenser lens CON, and the CHT pattern is projected onto the subject plane by the projection lens LEL. . Parallax occurs when the photographing lens and the light projecting lens are separated as shown in FIG. 1, but since light emitted from the photographing lens tends to cause ghosts, a separate light projecting system is usually provided outside the photographing system.

It is known that fill-in illumination is effective not only for low-brightness subjects but also for low-contrast subjects. When the contrast is low, there is no light-dark pattern to form the basis of calculations, and focus cannot be detected. In such a case, it is possible to forcibly give a brightness pattern to the subject by projecting a pattern onto the subject plane, and perform focus detection calculations using this as a foothold.

Figure 2 shows the sensor array SAA with the configuration shown in Figure 1.
An example of photoelectric conversion output of two images formed on SAB is shown. Let the output of SAA be A(i) and the output of SAB be B(i). Note that the number of pixels of the sensor is required to be at least five, and preferably several dozen or more. Image signal A
As a signal processing method for detecting the image shift amount PR from (i) and B(i), Japanese Patent Laid-Open No. 142306/1983 and Japanese Patent Laid-Open No. 59/1989
No. 107313, Japanese Unexamined Patent Publication No. 101513/1982, Japanese Patent Application No. 160824/1984, etc. are disclosed by the present applicant.

By determining the amount of image shift using the methods disclosed in these publications and adjusting the focus of the photographic lens based on this, it is possible to bring the photographic lens into focus.

In the method disclosed in the above publication, for example, for two image signals A(i), B(i), i=1.2, ..., N, V(m)= 〓 ⁱ max {A(i), B(i+k−m)}− 〓 ⁱ max{A(i+k), B(i−m)} (1) is calculated for the integer value m. The range of i to be summed is determined from the condition that each subscript i, i+k-m, i+k, i-m must fall within the closed interval [i, N]. k is an integer constant, usually k=
It is 1. Also, the range of m cannot be determined unconditionally because it depends on the purpose of detecting how large an amount of image shift is, but the correlation amount defined by equation (1), which usually changes m within -N/2mN/2, is an example. However, the following principle is exactly the same for other known correlation amounts.

As a correlation quantity equation other than equation (1), for example, 〓min{A(i), B(i+k−m)}−〓min{A
(i+k), B(i-m)} 〓|A(i)-B(i+k-m)|-〓|{A(i
+k), B(i-m) | 〓|A(i)-B(i+k-m)| ² −〓|{A(i
+k), B(i-m) | ² is adopted.

A typical result obtained by calculating the above equation (1) for each m is as shown in FIG. 3, where m where V(m) is inverted is the image shift amount expressed in pixel pitch units. This value usually does not take an integer. V (m ₀ ) and V
If there is a sign reversal between (m ₀ +1), the image shift amount M ₀ including fractions is M ₀ = m ₀ + |V(m ₀ )/{V(m ₀ +1) −V(m ₀ )}｜(2).

In addition to the above-mentioned conventional example, the means for dividing the exit pupil of the photographic lens may be such that a large number of units each having a microlens placed in front of a pair of photoelectric sensors may be arranged in a straight line, as disclosed in USP4185191. Not limited.

[Problem that the invention is trying to solve]

A focus detection device based on image shift amount detection as described above generally has a characteristic of malfunctioning for periodic subject patterns. This defect arises directly from the principle of detecting the misalignment of two images.

For example, suppose that the subject image is a repeating pattern of pitches P on the photoelectric sensor surface as shown in FIG. At this time, when trying to align the two images A(i) and B(i), even if A(i) is shifted in the direction of arrow α, B(i) is shifted in the direction of arrow β. It is also possible to match the two images, making the amount of image deviation unique (→
−) cannot be defined. Furthermore, in addition to the above-mentioned α and β, shift points at which the two images coincide exist for each pitch of periodicity.

Due to the above-mentioned circumstances, it is not possible to detect the amount of image shift of an object having a periodic pattern using the conventional method, and therefore the amount of defocus of the photographic lens cannot be calculated, and the in-focus state of the photographic lens cannot be determined. Periodic patterns are surprisingly common in man-made structures, such as window lattices, railings, blinds, checkered and striped clothing, arranged bookshelves, and many other objects that cannot be ignored by cameras.

[Means for solving problems]

The present invention has a means for calculating and determining the strength of periodicity of the brightness and darkness pattern of the subject surface using the output of a photoelectric conversion element array that receives the imaging light flux from the subject surface, and at least the strength of the periodicity is determined. When the predetermined standard is exceeded, the above-mentioned method is detected by detecting the in-focus state of the photographic lens using the output of the photoelectric conversion element array of the received imaging light beam when the auxiliary light emitting means emits light. It is a solution to a problem. By using the focus detection method of the present invention,
Focus detection is possible for all periodic subjects within the effective distance for auxiliary light emission. Furthermore, since the present invention has a means for determining the strength of periodicity, it is possible to emit auxiliary light only where necessary, which is also advantageous from the point of view of energy efficiency. Further, the periodicity determining means of the present invention is suitable for being constructed using digital calculations, and can be housed in a microprocessor that is standardly installed in automatic focus detection cameras. Further, the present invention is even more effective when used in combination with a photoelectric conversion element having an external light removal function. In such a combination, the periodic object pattern that is harmful or useless to focus detection is canceled out, and only the projection pattern of the auxiliary light remains as the image signal for focus detection, so that a focus detection device with no room for malfunction can be obtained.

〔Example〕

Next, examples of the present invention will be described. Fifth
The figure shows a pupil splitting type focus detection optical system using a secondary imaging separation optical system used in a focus detection device according to the present invention. The photographing lens of the camera and the like are omitted, and only the focus detection optical system is shown. In FIG. 5, a field mask having a distance measurement field aperture Z is placed near the field lens FLD, and a combination of the field mask and the field lens is placed near the planned imaging plane of a photographing lens (not shown). The image of field of view Z is separated and formed onto sensor array SAA by lens FCLA and onto sensor array SAB by lens FCLB. 2 sensor rows
On both sides of SAA and SAB, an AGC circuit for controlling the sensor storage time according to the amount of incident light and a readout circuit RD for sequentially outputting the data of each pixel to the outside are integrated on-chip. The operating principle of the sensor array is whether it has a CCD structure or a MOS
Any structure may be used as long as it detects and outputs the light intensity distribution on a line.

FIG. 6 is a circuit diagram showing an embodiment of a camera equipped with an automatic focusing device according to the present invention.

In the figure, PRS is a camera control device that includes a CPU (central processing unit), ROM,
It is a one-chip microcomputer with RAM and A/D conversion functions. PRS follows the camera sequence program stored in ROM.
It performs a series of camera operations such as automatic exposure control, automatic focus adjustment, and film winding and rewinding. For this purpose, PRS uses communication signals SO,
SI, SCLK, communication selection signal CLCM, CSDR,
The CDDR is used to communicate with peripheral circuits within the camera body and the control device within the lens to control the operation of each circuit and lens.

SO is the data signal output from PRS, SI is
The data signal input to PRS, SCLK is the signal SO,
This is the SI synchronous clock.

LCM is a lens communication buffer circuit that supplies power to the lens power supply terminal VL when the camera is in operation, and also receives the selection signal from PRS.
CLCM is at a high potential level (hereinafter abbreviated as "H",
When the low potential level is abbreviated as “L”,
It becomes a communication buffer between the camera and lens.

When PRS sets CLCM to “H” and sends predetermined data from SO in synchronization with SCLK, LCM communicates with SCLK and SO via the camera-lens communication contact.
The buffer signals LCK and DCL of each are output to the lens. At the same time, the buffer signal of the DLC signal from the lens is output to the SI, and the PRS inputs lens data from the SI in synchronization with SCLK.

The SDR is a drive circuit for the line sensor device SNS for focus detection, which is composed of a CCD and the like. The line sensor device SNS is constructed as shown in FIG. 5, and includes integrated light receiving element arrays SAA and SAB, a signal processing circuit, and the like. In addition, the drive circuit SDR is a signal
Selected when CSDR is “H”, SO, SI,
Controlled from PRS using SCLK. Signal CK
This is a clock for generating the CCD driving clocks Φ1 and Φ2, and the signal INTEND is a signal that informs the PRS that the accumulation operation has ended.

The output signal OS of the SNS is a time-series image signal synchronized with the clocks Φ1 and Φ2, and after being amplified by an amplifier circuit in the SDR, it is outputted to the PRS as an AOS. PRS inputs AOS from the analog input terminal,
In synchronization with CK, the internal A/D conversion function sequentially stores the data as a digital signal at a predetermined address in the RAM.

SAGC, which is also the output signal of SNS, is
AGC (Auto Gain Control)
It is the output of the sensor, is input to the SDR, and is sent to the SNS
Used for storage control.

SPC is a photometric sensor for exposure control that receives light from the subject through the photographic lens, and its output SSPC is input to the analog input terminal of PRS.
After A/D conversion, it is used for automatic exposure control according to a predetermined program. The AuT is an auxiliary light unit and communicates with the control device PRS via contacts.
SAL is an auxiliary light emission signal, and when SAL is at H level, the auxiliary light ALED is turned on.

DDR is a switch detection and display circuit,
Selected when signal CDDR is “H”, SO, SI,
Controlled from PRS using SCLK. That is, PRS
PRS can switch the display of the camera's display component DSP based on the data sent from the camera, and the on/off status of various camera operation components can be controlled via communication.
to be notified.

SW1 and SW2 are switches linked to a release button (not shown); when the release button is pressed in the first stage, SW1 is turned on, and subsequently, when the release button is pressed in the second stage, SW2 is turned on. PRS performs photometry and automatic focus adjustment when SW1 is on, and uses SW2 on as a trigger to control exposure and advance the film.

Furthermore, SW2 is a microcomputer.
Connected to the "interrupt input terminal" of PRS, SW1
An interrupt is generated even when a program is being executed when SW2 is on, and control can be immediately transferred to a predetermined interrupt program.

MTR1 is a motor for film feeding, MTR2 is a motor for mirror up/down and shutter spring charging, and each drive circuit MDR1, MDR
2 controls forward rotation and reverse rotation. Signal M1F input from PRS to MDR1 and MDR2,
M1R, M2F, and M2R are signals for motor control.

MG1 and MG2 are magnets for starting the leading and trailing shutter curtains, respectively, and are energized by signals SMG1 and SMG2 and amplification transistors TR1 and TR2, and shutter control is performed by PRS. In addition, switch detection and display circuit DDR, motor drive circuit
MDR1, MDR2, and shutter control are not directly related to the present invention, so detailed explanations will be omitted.

The signal DCL input to the in-lens control circuit LPRS in synchronization with LCK is data of a command from the camera to the lens FLNS, and the operation of the lens in response to the command is determined in advance. LPRS analyzes the command according to a predetermined procedure, and outputs the focus adjustment and aperture control operations, the operation status of each part of the lens from the output DLC (the driving status of the focusing optical system, the driving status of the diaphragm, etc.), and various parameters. (Open F number, focal length, coefficient of the amount of defocus versus amount of movement of the focusing optical system, etc.) are output.

In the example, an example of a zoom lens is shown,
When a focus adjustment command is sent from the camera,
The focus adjustment motor LTMR is driven by the signals LMF and LMR in accordance with the drive amount and direction sent at the same time, and the optical system is moved in the optical axis direction to perform focus adjustment. The amount of movement of the optical system is determined by the encoder circuit.
It is monitored by the ENC's pulse signal SENC and counted by the counter in the LPRS, and when the predetermined movement is completed, the LPRS itself sets the signals LMF and LMR to "L" and brakes the motor LMTR.

Therefore, once a focus adjustment command is sent from the camera, the camera control device PRS does not need to be involved in lens driving at all until the lens driving is completed. Furthermore, the configuration is such that the contents of the counter can be sent to the camera when a request is received from the camera.

When an aperture control command is sent from the camera, a stepping motor known as a stepping motor is used to drive the aperture according to the number of aperture stages sent at the same time.
Drives DMTR. Note that since the stepping motor can be controlled in an open manner, it does not require an encoder to monitor its operation.

Before explaining the operation shown in FIG. 6 above, the periodicity determination operation of the present invention will be explained.

First, a case will be described in which the periodicity of a subject pattern is determined from the result of a correlation calculation for focus detection.
When calculating the (differential form of) the cross-correlation between the two subject images A(i) and B(i) based on the above equation (1), in a normal subject pattern without strong periodicity, V as shown in Figure 3 is obtained. (m)
is obtained, and the zero-crossing point corresponding to the amount of photographic lens differential focus can be determined. However, in the case of a strong periodic pattern as shown in FIG. 4, when V(m) is calculated, it appears repeatedly as shown in FIG. 7a, and multiple zero crossing points appear. If the positive/negative reversal of V(m) occurs between a certain m ₁ and m ₁ +1, the magnitude of the absolute value of △V(m ₁ ) = V(m ₁ +1) − V(m ₁ ) (3) |△V(m ₁ )| represents the reliability of the zero-crossing point as a two-image matching point, and the fact that the values of |△V(m)| appear without much difference at different zero-crossing points is due to periodic objects. This is a characteristic of the result of correlation calculation. In other words, there are multiple zero-crossing points such as m=m ₁ , m ₂ , ... where the sign of V(m) is reversed, and |△V(m ₁ ) | | △V(m ₂ ) |... It is.

Note that multiple zero-crossing points may occur even in normal non-periodic objects, but as shown in Figure 7b, |△V(m ₁ )|≫|△V(m ₂ )| With respect to the zero-crossing point m ₁ corresponding to the two-image matching point, other zero-crossing points are clearly distinguished from each other in terms of reliability.

From the above facts, for example, as a periodicity determination method, if the zero crossing point showing the maximum |△V(m)| is m ₁ , then |△V(m)|>F×|△V(m ₁ ) | (4) The fact that there are L or more m that satisfy the following can be used as a condition for determining that the pattern is a periodic pattern.
Here, F is a positive constant of 0<F≦1, and L is a positive integer of 1 or more. Further, when using this determination condition, the reliability factor |ΔV(m)| of the zero-crossing point may be divided by the contrast C of the object for normalization.

Generally, C means C=MAX{A(i)}-MIN{A(i)| (5a) or C= _N 〓 ⁱ⁼² |A(i)-A(i-1)| (5b) It is. Here, MAX {A(i)} is the image signal A ₁ ,
A ₂ , ... the maximum value of A _N , MIN {A(i)} is the image signal A ₁ ,
A ₂ ,...means the minimum value of A _N. B(i) may be used instead of A(i), or C may be calculated using both image signals.

As another criterion, for example, a threshold may be set by simply counting the number of zero-crossing points that invert the sign of V(m), or a reliability factor of zero-crossing points |△V( The zero-crossing points m may be restored in order of magnitude of m)|, and a reliability threshold may be defined for the Lth zero-crossing point.

The periodicity of the object pattern is determined by the method disclosed above, and if it is determined that the periodicity of the object is large, the auxiliary light emitting means is emitted to project the pattern onto the surface of the object. The light intensity distribution of the auxiliary light must have a pattern. For example, by projecting the auxiliary light having the light intensity distribution as shown in Fig. 8b onto a periodic subject as shown in Fig. 8a, the brightness distribution of the subject is By changing the object to a non-periodic object as shown in c, it is possible to shift to a focus detectable state.

Next, the flow of the focus detection device according to the present invention shown in FIG. 6 will be explained.

When the power switch (not shown) is operated, power supply to the microcomputer PRS starts, and the PRS
starts executing the sequence program stored in ROM.

FIG. 9a is a flowchart showing the overall flow of the above program.

When execution of the program is started by the above operation, the state of switch SW ₁ , which is turned on at the first stroke of the release button, is detected in step (002), and when switch SW ₁ is off,
At step (003), all control flags set in the RAM in the PRS are cleared. The detection of this switch SW ₁ is performed by setting the signal CDDR to H from the computer PRS, selecting the circuit DDR, and transmitting the SO signal as a detection command for the switch SW ₁ to the DDR, thereby detecting the state of the switch SW ₁ .
This is done using DDR and the result is transmitted to PRS as an SI signal. Above step (002),
(003) is repeatedly executed until the switch SW ₁ is turned on or the power switch is turned off, and when SW ₁ is turned on, the process moves to step (004).

Step (004) means the "AE control" subroutine. This "AE control" subroutine performs a series of camera operation controls such as photometric calculation processing, exposure control, and post-exposure shutter charge and film winding.

Note that the "AE control" subroutine is not directly related to the present invention, so a detailed explanation will be omitted, but an outline of the function of this subroutine is as follows.

While SW ₁ is on, this "AE control" subroutine is executed, and each time the camera mode setting, photometry and exposure control calculations and display are performed. Switch SW ₂ is activated by the second stroke of the release button (not shown).
When turned on, the release operation is started by the interrupt processing function of the microcomputer PRS, and the aperture or shutter speed is controlled based on the exposure amount determined by the above exposure control calculation, and after the exposure is completed, the shutter is - One frame of film is photographed by performing charge and film feeding operations.

Now, when the "AE control" ends at step (004), the "AE control" subroutine at step (005) is executed.

FIG. 9b shows a flowchart of the "AF control" subroutine.

First, in step (102), the state of the flag PRMV is detected. PRMV is a flag related to lens control as described later, but as mentioned above, SW ₁
While off, all flags are cleared at step (003), so when the "AF control" subroutine at step (005) is called for the first time after SW ₁ is turned on, flag PRMV is also 0, so step (106) is cleared. ).

In step (106), the state of flag AUXJF is detected. AUXJF is a flag related to auxiliary light control, and as described above, it was cleared in step (003) and flag AUXJF is also φ, so the process moves to step (108).

Step (108) is the "image signal input" subroutine, and by executing this subroutine, the image signal A/D is sent from the sensor device SNS to a predetermined address on the RAM of the microcomputer PRS.
The converted signal is stored. In addition, in step (108), the level of the image signal and the contrast C defined in (5a), (5b), etc. are calculated at the same time as the input, and when it is determined that the brightness is low or the contrast is low (C is small), Set the flag LSIGFLG to 1.

In step (111), the state of the flag AUXMOD is detected. The flag AUXMOD is a flag indicating the auxiliary light mode. Control regarding the auxiliary light will be described later.

As mentioned above, all flags are cleared at (003) and the flag AUXMOD is also 0, so
Move to step (112). In step (112), the state of the flag LSIGFLG is detected.
LSIGFLG is "image signal input" in step (108)
This flag is set in the subroutine and is set to 1 when the subject brightness is low or the contrast of the image signal is low. Here, we will proceed with the explanation assuming that the subject has sufficient brightness and contrast (LSIGFLG is 0). flag
Since LSIGFLG is 0, the process moves to step (113), where the auxiliary light mode flag AUXMOD is cleared since the subject brightness and contrast are sufficient.

Next, in step (114), a "focus detection" subroutine is executed.

In this subroutine, the focal point of the photographing lens is determined from the image signal data stored in RAM using the formula (1).
Calculate and detect using equation (2), and if the focus is in focus, set the focus flag JF to 1, and if the focus cannot be detected due to low contrast even though the subject is illuminated with a fill-in light, Focus detection failure flag AFNG
is set to 1, and in either of the two states, a lens drive prohibition flag LMVDI for prohibiting lens drive is set to 1, and the process returns. Further, when the contrast is high and the image is not in focus, the amount of defocus is determined. In addition, at this time, the flag
LMVDI is kept at 0.

In the next step (115), periodicity is determined. For example, with respect to the zerox point for V(m) found using equation (1), use equations (3) and (4) to determine whether there are multiple |△V(m)| approximated at the zerox point. In other words, when the reliability of multiple zero-correlation points is determined and there are multiple points that are recognized to match the two images,
Set LSIGFLG to 1 and branch to step (131). If the subject is not periodic, proceed to step (116). The flag control related to the auxiliary light after step (131) is basically the same as that after step (121), which will be described later.

In the next step (116), a "display" subroutine is executed to indicate in-focus or inability to detect focus. This involves communicating predetermined data to the display circuit DDR and displaying it on the display device DSP, but since this operation is not directly related to the present invention, further explanation will be omitted.

Now, in the next step (117), the flag LMVDI is set.
Detect the state of. As mentioned earlier, when lens drive is not required, LMVDI is set to 1, so the flag is set in step (117).
If LMVDI is 1, the "AF control" subroutine is returned at step (118). LMVDI is 0
If so, the process moves to step (119) and executes the lens drive subroutine "lens drive" to drive the lens in accordance with the calculated defocus amount.

When the "lens drive" subroutine (119) is completed, the lens drive execution flag PRMV is set to 1 in step (120), and then the "AF control" subroutine is returned to step (130).

When the "AF control" subroutine returns, the process returns to step (002), and as long as switch SW ₁ is on, the AE control and AF control subroutines are repeated.

Now, suppose that "AF control" in step (005) is called again (for the second time) in the main flow of FIG.
PRMV status detection is performed.

If focusing or focus detection is not possible in the previous ``AF control'' routine, the flag PRMV has not been set to 1, and the above-described flow from step (106) onwards is executed again. If the lens was driven last time, PRMV is performed in step (120).
is set to 1, so step (103)
Move to.

In step (103), the current driving status of the lens is detected by communicating with the lens, and when the lens side informs that the specified driving instructed in step (119) has been completed, a flag is set in step (105).
Set PRMV to 0 and execute the flow from step (106) onwards. Note that this judgment is made by the encoder
Since a monitor signal SENC is sent from the ENC while the lens is being driven, this signal SENC is detected by the computer PRS. Also,
If the lens is informed that it is still being driven, proceed to step (104) and select "AF control".
Return subroutine.

Therefore, the "AF control" subroutine performs a new focus detection operation only when the lens is not being driven.
The lens will be controlled.

That is, in the normal mode, as long as switch SW ₁ is on, the AE and AF control subroutines are stopped back, and the AF control subroutine detects the amount of defocus based on the image signal, and the predetermined conditions for the image signal or calculated value are met. If not, a focus detection impossible display is displayed, and if focus is determined, an in-focus display is displayed. If the focus is out of focus and the amount of defocus is determined, the lens is driven by this amount of defocus to shift to the in-focus state. Become.

Next, operations related to the auxiliary light will be explained.

In the above AF control subroutine, if the subject brightness is low or the contrast is low,
In step (108), the flag LSIGFLG is set to 1 in step (115) in the case of a periodic subject, and the process moves to step (121) when the LSIGFLG state is detected in step (112).

In step (121), the state of the auxiliary light unit AUT device members is detected, and if the unit AUT is not attached, the process moves to step (113), where the same normal operations as described above are performed. If the unit is installed, proceed to step (122),
Set the auxiliary light mode flag AUXMOD to 1. Note that if the auxiliary light unit AUT is built into the camera body, step (121) may be omitted.

Next, in step (123), the state of the flag AUXUSE is detected. AUXUSE is a flag that is set to 1 when the "image signal input" subroutine of step (108) is executed (with the flag AUXMOD set to 1) when the auxiliary light is actually emitted. be. In the situation just described, the auxiliary light mode is entered for the first time, so the auxiliary light has not been emitted before then, and "AF control" returns at step (124). That is, in this case, the image signal data input in step (108) is discarded, and in the next "AF control" an image signal is input in the auxiliary light projection state, and this is used for focus detection.

Now, at step (122), you can use AUXMOD for the first time.
When the "AF control" subroutine is called again as described above after "AF control" returns with An image signal is input in the light emitting state, and the process moves to step (121) when the state of the auxiliary light mode flag AUXMOD is detected in step (111).

In addition, in the "image signal input" subroutine, the flag
Since AUXMOD is set to 1, when accumulating image signals, SAL is set to H level and the auxiliary light is set to H level.
Turn on the ALED.

Step (121), (122) then step (123)
Now, detect the state of the auxiliary light use flag AUXUSE. The "image signal input" subroutine has already been executed in step (108) with the fill light emitting state.
Since AUXUSE is set to 1, the program moves to step (114) and executes the "focus detection" subroutine. The rest is the same as normal AF control.

Also, when repeat periodicity is detected in step (115), steps (131), (132), and (134) are performed.
is performed in the same way as steps (121), (122), and (123) under low contrast detection described above, and returns "AF control".When "AF control" is called again, the above-mentioned auxiliary light In the same way as the mode, step (108), (111), (121), (122), (123)
)
Execute step (114) through. Therefore, even when it is determined that the repeating periodicity is high, the AF control operation is performed under the auxiliary light projection in the same way as when the contrast is low.

As mentioned above, when shooting under conditions of low brightness, low contrast, or periodic objects, and when the auxiliary light unit is attached, the camera enters auxiliary light mode and performs focus adjustment based on the detected image signal under the auxiliary light emission. However, when focusing is achieved with the auxiliary light emitted, the auxiliary light focus flag AUXJF is set to 1 in the "focus detection" subroutine of step (114), and in this case, in the "AF control" flow,
At step (106), the state of AUXJF is detected and the process moves to step (107), after which the "AF control" subroutine is returned. That is, when focusing is achieved in the auxiliary light projection state, the focus adjustment operation and lens drive will not be performed again until the switch _SW1 is turned off.

FIG. 10 is an explanatory diagram showing another example of periodicity determination. Object image signal A as shown in FIG.
When there is (i) or B(i), its maximum value Amax
Then, find the minimum value Amin and set the slice level SL in the middle.

SL=(Amax+Amin)/2 (6) If it is a periodic pattern, the magnitude relationship of the image signal A(i) with respect to the slice level changes periodically depending on i.

This phenomenon can be easily detected by focusing only on the sign of (A(i)-SL), and when the number of alternating positive and negative periodic reversals exceeds a predetermined number, the periodicity is detected. It is sufficient if it is determined that the subject pattern is strong. still,
The slice level may be set as the average value of the entire visual field of A(i) (〓NA(i))/N. Furthermore, since there is often a loose luminance distribution superimposed on the periodic pattern of the object, better discrimination can be achieved by changing the slice level as a local image signal average value along with i. That is, SL(i)= _j 〓 ^l=-J A(i+l)/(2J+1) (7). However, J is an appropriate positive integer, and is set so that the smoothing width (2J+1) is sufficiently larger than the pitch of the periodic pattern in question.

The above method only detects the severe fluctuation of A(i), but does not detect the periodicity of the fluctuation. The gist of the present invention includes periodic subjects.
Since it is also possible to issue periodic determination signals for a slightly larger set of image signals, this method is sufficiently practical, except that there are some cases in which auxiliary light is emitted for subjects that do not require light emission.

When determining periodicity directly from fluctuations in the image signal using equations (6), (7), etc., the determination can be performed before the focus detection calculation, which reduces the number of invalid calculations. That is, in the flowchart of FIG. 9, the image signal is calculated using equations (6), (7), etc. from the time of image signal input in step (108) to the detection of the LSIGFLG set state in step (112). Insert a subroutine to determine periodicity, and if the subject is periodic,
When it is determined, LSIGFLG is set to 1 and the flow branches from step (112) to step (121). At this time, the steps below step (115) and step (131) in FIG. 9b can be omitted.

Further, as another method for determining periodicity, a method using digital Fourier transform may be adopted. When the power spectrum of the digital Fourier transform of the image signal luminance distribution is calculated, a gentle power spectrum without very drastic fluctuations is obtained for an aperiodic subject, as shown in FIG. 11a. In the figure, the horizontal axis is the spatial frequency, _N
is the Nightquist frequency. By the way, the periodic pattern has a strong peak at a specific spatial frequency, as shown in FIG. Therefore, the ratio of the specific spectrum to the total power spectrum sum is determined, and if the ratio exceeds a certain value, it can be determined that the subject is a periodic subject. Specifically, when the power spectrum components of the image signal A(i), i=1, 2,...N are A~ ² (i), i=1, 2,...N, γ _i =A ~ ² (i) / _N 〓 ⁱ⁼¹ A ~ ² (i) (8) is obtained, and if there is an i that satisfies γ _i >R, it is considered a periodic subject. R is a predetermined constant of 0<R<1. Further, if the J-th values are added from the greater of A~ ² (i) and the periodicity is determined to be strong when the sum exceeds the threshold R, reliable operation can be achieved.

Note that the power spectrum referred to in this embodiment is the sum of squares of Fourier sine transform and cosine transform.

Also, equation (8) can be transformed as γ _i = A~2(i)/ _N 〓 ⁱ⁼¹ A(i) ² (9) using Parsvall's theorem, so the power spectrum A~2( There is no need to calculate i). In fast Fourier transform, the spectral components are found for all i at the same time, but the number of pixels N is not necessarily large and N = 2 _K.
(K integers) is not necessarily satisfied, so a high-speed conversion algorithm cannot be said to be advantageous. It is relatively easy to calculate the power spectrum for a small number of i, and it is also possible to roughly estimate i for which the spectrum should be calculated by calculation in real space.
Note that the position of the subroutine in the flowchart does not change whether direct arithmetic judgment is performed from the image signal or digital Fourier transform is performed.

In all of the above embodiments, the periodicity is determined using an image signal formed without the auxiliary light being emitted, and the patterned auxiliary light is emitted onto the subject determined to have strong periodicity, and then the pattern is emitted again. The assumption is that the sequence will form an image signal. However, the present invention presents a method for converting a highly periodic subject brightness distribution into an aperiodic one by projecting auxiliary light with a light intensity distribution pattern onto the subject surface, and a means for determining the effectiveness of the method. It is something to do. For example, the present invention is also effective in a sequence that has an image signal when the auxiliary light is emitted and an image signal when the auxiliary light is not emitted, and selects the image signal.

〔Effect of the invention〕

As described above, according to the focus detection device of the present invention, accurate automatic focus detection can be performed even for periodic objects that cannot be detected using conventional image shift detection type focus detection methods. Furthermore, since the present invention is equipped with a pattern determining means that specifically responds only to periodic patterns, it is possible to minimize the emission of auxiliary light, which is advantageous in terms of power consumption. Furthermore, since the periodicity determination means described above can be configured with simple software,
It can be operated on a microprocessor with a built-in autofocus camera, and can be easily realized, resulting in great effects.

[Brief explanation of the drawing]

Fig. 1 is a configuration diagram showing the optical system of the automatic focus detection device, Fig. 2 is an explanatory diagram showing the image signal in the sensor, Fig. 3 is an explanatory diagram showing the amount of correlation by the focus detection device, and Fig. 4 is a synchronization diagram. FIG. 5 is a configuration diagram showing an optical system of the focus detection device according to the present invention, FIG. 6 is a circuit diagram showing an embodiment of the focus detection device according to the present invention, and FIG. Figures a and b are explanatory diagrams for explaining the periodicity determination operation used in the focus detection device according to the present invention, and Figures 8a, b, and c are
9A and 9B are explanatory diagrams illustrating the image signal changes during projection of the auxiliary light, FIGS. 9A and 9B are explanatory diagrams illustrating the operation of the focus detection device shown in FIG. 6, and FIG. 10 is an explanatory diagram illustrating another example of periodicity determination. Explanatory diagrams, FIGS. 11a and 11b are explanatory diagrams showing another example of periodicity determination. AUT...Auxiliary light unit, PRS...Microcomputer.

Claims (1)

[Claims] 1. First and second photoelectric conversion element arrays that each receive an image from a subject, and detecting the focal state of the photographic lens based on the outputs of the first and second photoelectric conversion element arrays. In a focus detection device for a camera that uses
an arithmetic circuit that calculates and determines the strength of periodicity of the output signal of the photoelectric conversion element array based on the output of the photoelectric conversion element array, and drives an auxiliary light source when the arithmetic circuit determines that the periodicity is strong. An auxiliary light drive circuit is provided to perform pattern light projection, and when the periodicity of the output signal of the photoelectric conversion element array is strong, focus state detection is performed based on the output signal of the photoelectric conversion element array under pattern light projection. A camera focus detection device featuring: 2 The arithmetic circuit calculates the amount of correlation for each shift while shifting the relationship between each output signal of the first and second photoelectric conversion element arrays, and determines the period when the number showing the amount of correlation greater than a predetermined value is greater than or equal to the predetermined value. The camera focus detection device according to claim 1, which determines that the focus is strong.