JPH11258492A - Focus detection apparatus, method, and computer-readable storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To shorten an operation time for detecting the focus of a digital camera. SOLUTION: In a focus detecting method, three or more output images are picked up through CCD 9 as right and left pupils are mutually shielded with a shading plate 5, and these output images are correlation-computed to obtain a plurality of phase differences, and an operation for detecting the defocused quantity of a focus lens 1a is carried out on the above obtained phase difference for executing focus control, and in the case where the focus is detected two or more times for confirming the focused lens, after its second detection, a search range of a plurality of above image signals is narrowed on the last detected defocused quantity to carry out a correlation computation.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a focus detection apparatus and method for an image pickup apparatus using an image pickup device such as a digital camera, and a computer-readable storage medium.

[0002]

2. Description of the Related Art Conventionally, a large number of phase difference detection type focus detection devices have been used as automatic focus devices used in single-lens reflex type silver halide cameras and the like. FIG. 14 is a cross-sectional view of a single-lens reflex camera having a conventional focus detection device of a phase difference detection method. A light beam 109a emitted from a photographing lens 100 is reflected by a light beam 109b reflected by a main mirror 102 made of a half mirror. It is divided into a transmitted light flux 109e. The reflected light flux 109b forms an image of the subject on the diffusion surface of the focus plate 103, and the photographer can use the eyepiece 105a,
It is configured to observe a subject image on the focus plate 103 via the 105b and the pentaprism 104.

On the other hand, the light beam 10 transmitted through the main mirror 102
9e is reflected by the sub-mirror 106 and guided to the focus detection device 107. The focus detection device 107
The focus state (defocus amount) of the photographing lens 100 with respect to the silver halide film 108 is detected by the light flux 109f from 00. If it is determined that the detected defocus amount is larger than the predetermined focusing width and the camera is out of focus, the focus adjusting lens of the photographing lens 100 is driven so that the detected defocus amount is canceled. Make adjustments.

Next, the principle of focus detection of a conventional focus detection device will be described with reference to FIGS. FIG. 15 (a)
Is a focused state, that is, a focused state, and a light beam 116 that has passed through two different pupils of the photographing lens 100.
a and 116b form an image on a primary image forming surface 114, and a subject image on the primary image forming surface is placed on a sensor surface on which two line sensors 113a and 113b are arranged by secondary image forming lenses 112a and 112b, respectively. Re-image. Here, the field lens 111 is disposed near the primary image forming surface 114 of the photographing lens 100, efficiently guides a light beam having a predetermined image height to the sensor surface, and prevents a decrease in the amount of light generated as the image height increases. I do. Generally, the two light beams 116a and 116b passing through different pupils of the photographing lens 100 are defined by apertures (not shown) disposed immediately before or immediately after the secondary imaging lenses 112a and 112b. No pupil-dividing member is provided.

The two images formed on the line sensors 113a and 113b are luminous fluxes that have passed through different pupils, and the relative positions of the images depend on the amount of extension of the lens.
As shown in FIG. 6, the focus, the front focus, and the rear focus are different.
FIGS. 15A and 16A show that the distance between two images formed on the line sensors 113a and 113b in the focused state is equal to the relative distance e0 between the two line sensors. It is always constant at the time of focusing.

FIGS. 15B and 16B show a state in which the front focus state is set by the defocus amount d1 and the distance e1 between the two images.
Is smaller than e0, and the difference δ1 between e0 and e1 increases as the defocus amount d1 increases. FIG.
FIGS. 16C and 16C show a state in which the back focus state is set by the defocus amount d2, and the interval e2 between the two images is larger than e0. When the defocus amount d2 increases, e2
The difference δ2 between と and e0 also increases.

As described above, the magnitude and direction of the defocus amount at that time can be determined from the distance between the two images. Therefore, the difference between the distance e between the two images in the current defocus state and the reference image distance e0 at the time of focusing, that is, the relative shift amount (phase difference) δ = e−e0 between the two images. Is calculated by correlating the output signals of the two line sensors 113a and 113b, the defocus amount and the direction of the optical system are obtained from the phase difference, and the focusing is controlled by controlling the focus lens.

[0008]

However, even if the focus lens is moved from the phase difference obtained in the first focus detection, the defocus amount is larger than a predetermined allowable defocus amount, and further focus detection is required. Performs a correlation operation to find the phase difference again, but the focus detection operation method at that time changes depending on the defocus state and other photographing conditions.

In spite of such a change in the condition, for example, when the defocus amount is smaller at the next focus detection than at the previous focus detection, the same correlation calculation as the previous one is performed. In such a case, extra calculations are required, which requires a long calculation time. Considering the time required for focus detection, the shorter the calculation time, the better, and there is a problem in using the same focus detection calculation method under any conditions.

Accordingly, an object of the present invention is to use a focus detection calculation method suitable for focus detection at the time of performing a focus detection calculation for obtaining a phase difference using a phase difference detection signal for focus detection. In other words, the operation time is to be shortened.

[0011]

According to the present invention, there is provided a focus detection apparatus comprising: an image pickup means for picking up an optical image of a subject formed on an image pickup surface and outputting an image signal; Optical system means for forming an image on a plane, moving means for moving the pupil region to a plurality of positions, and detecting a phase difference between a plurality of image signals obtained by imaging at the plurality of positions by a correlation operation. Calculating means for detecting the focus of the optical system means based on the phase difference; and control means for causing the calculating means to perform a plurality of calculations and selecting and providing one of a plurality of calculation methods for each calculation. Is provided.

According to the focus detection method of the present invention, an image pickup procedure for picking up an optical image of a subject formed on an image pickup surface and outputting an image signal, and a pupil in an optical system for forming the optical image on the image pickup surface A moving procedure for moving the area to a plurality of positions;
An arithmetic procedure for detecting a phase difference between a plurality of image signals obtained by imaging at the plurality of positions by a correlation operation, and performing focus detection of the optical system based on the phase difference;
A control procedure is provided in which a plurality of calculations are performed in the calculation procedure and one of a plurality of calculation methods is selected and provided for each calculation.

[0013] A storage medium according to the present invention includes: an image pickup process for picking up an optical image of a subject formed on an image pickup surface and outputting an image signal; and a pupil region in an optical system for forming the optical image on the image pickup surface. Is moved to a plurality of positions, and a phase difference between a plurality of image signals obtained by imaging at the plurality of positions is detected by a correlation operation, and focus detection of the optical system unit is performed based on the phase difference. Calculation processing to be performed;
A program is stored for executing a plurality of operations in the above-described arithmetic processing and performing a control process of selecting and providing one of a plurality of arithmetic methods for each time.

[0014]

FIG. 1 is a view showing an embodiment of a focus detecting device and a camera using the same according to the present invention, wherein 1b is a focusing lens group of a photographing lens, and 1a is a focusing lens group of the photographing lens. The lens group 2 other than the lens group 1b is a lens extension mechanism for extending the focus lens group 1b, and includes a motor for moving the lens. Reference numeral 3 denotes a focus detection aperture, 4 denotes a motor for inserting the focus detection aperture 3 into an optical path, and 5 denotes a light shield for shielding one of the two openings 3a and 3b in the focus detection aperture 3. Reference numeral 6 denotes a motor for moving the light shielding plate 5. Reference numeral 7 denotes an optical low-pass filter, reference numeral 8 denotes an infrared cut filter, and reference numeral 9 denotes a CCD as an image pickup device which photoelectrically converts an optical image formed on an image pickup surface into an electric signal. The imaging surface of the CCD 9 is provided with a color filter shown in FIG.

Reference numeral 10 denotes an amplifier for amplifying the output from the CCD 9, reference numeral 11 denotes an A / D converter for digitizing an analog signal amplified by the amplifier 10 with a predetermined gain, and reference numeral 12 denotes an A / D converter.
Is a digital signal processing unit that performs various digital signal processing of the A / D converted digital signal, 13 is a system control unit that controls the entire camera, and 14 is a drive control of the CCD 9 and a setting of an amplification factor of the amplifier 10. C for
A CD driver 15 is a lens control unit for controlling the extension of the focusing lens group 1b.

Reference numeral 16 denotes a buffer memory such as a DRAM used for temporarily storing digital signals, for example.
Reference numeral 17 denotes a card slot connected to a recording medium or a function card or the like and its control unit, 18 denotes an electronic viewfinder (EVF), 19 denotes a driver unit of the LCD, and 20 denotes a D / D for sending an analog signal to a driver unit 19. An A converter 21 holds an image to be displayed on the EVF 18 and outputs a digital signal to the D / A converter 20.
AM and 22 are external black and white liquid crystals (LCD) for displaying the settings of the camera, etc., 23 is an LCD driver for displaying the LCD, and 24 is an operation switch for detecting the setting of the shooting mode of the camera and the release operation. is there.

Next, a focus detection method and a focus detection device directly related to the present invention will be described with reference to FIG. Now, it is assumed that the power of the camera is turned on and the camera is in a photographable state. In order to perform pupil time division phase difference AF, the focus detection diaphragm 3 has two openings (hereinafter, referred to as CCs) having the same shape in the horizontal direction.
When viewed from the side of D9, the left opening has a left pupil 3a and the right opening has a right pupil 3b), and is inserted into the optical path of the photographing lens by a motor at the time of focus detection. By moving the light-shielding plate 5 to shield either the left pupil 3a or the right pupil 3b, an optical image composed of light beams passing through different pupil regions can be formed on the CCD 9.

In order to perform the pupil time division phase difference AF, the focus detection diaphragm 3 is inserted into the optical path in accordance with an instruction from the system control unit 13. FIG. 2 is a diagram showing a positional relationship between the focus detection diaphragm 3 and the light shielding plate 5. FIG. 2A shows a state at the time of photographing, in which the focus detection diaphragm 3 and the light shielding plate 5 are at positions retracted outside the optical path of the photographing lens. Reference numeral 25 denotes a pupil shape when the taking aperture of the taking lens is opened.

First, as shown in FIG. 2B, the right pupil 3b of the focus detection aperture is closed with a light shielding plate 5, the left pupil 3a of the photographing lens is opened, and an optical image composed of a light beam passing therethrough. Is imaged on the CCD 9 to capture an image. At this time, the image data for focus detection composed of the light beam that has passed through the left pupil 3a is referred to as a left image 1. Next, in order to obtain focus detection image data composed of light beams passing through different pupil regions, the motor 6 is driven to move the light shielding plate 5 as shown in FIG. 2C, and the right pupil 3b of the photographing optical system is moved. The optical image composed of the light beam that has been released and passed through the image is formed on the CCD 9 to capture the image. At this time, the image data for focus detection composed of the light beam that has passed through the right pupil 3b is defined as the right image 2.

Further, the light-shielding plate 5 is moved again as shown in FIG. 2B in order to take in focus-detection image data consisting of a light beam having passed through the same left pupil 3a as the left image 1, and at this time, the CCD 9 An image to be imaged is captured. At this time, the image data for focus detection composed of the light beam that has passed through the left pupil 3a is referred to as a left image 3. Similarly, right image 4, left image 5
Take in. Exposure when capturing image data for focus detection is performed by an electronic shutter, a mechanical shutter (not shown), gain adjustment of the amplifier 10, addition reading in the CCD 9 described later, and in some cases, auxiliary light is prepared. The adjustment is performed using the above.

By the way, the left image 1, right image 2, left image 3, right image 4, and left image 5, which are image data for focus detection, are taken in and focus detection is performed. In order to reduce the focus detection error due to relative movement between the camera and the subject (such as camera shake or movement of the subject), it is desirable that the capture interval of each image be as short as possible. Therefore, reading out the image data for focus detection requires a long time to read out the entire screen of the CCD 9 as in the case of reading out an image for photographing. Therefore, only a part of the image necessary for focus detection is usually read. Is read out faster than in the case of the main shooting.

The above reading method will be described below. FIG. 3 shows a schematic view of an interline CCD. 3
1 is a pixel, 32 is a vertical charge transfer element, 33 is a horizontal charge transfer element, and 34 is an output unit. The signal charge photoelectrically converted by each pixel 31 is sent to the vertical charge transfer element 32, and the four-phase drive pulses φV1, φV2, φV3 and φV
4 sequentially transfers in the direction of the horizontal charge transfer element 33. The horizontal charge transfer element 33 transfers the signal charges for one horizontal line transferred from the vertical charge transfer element 32 to the output unit 4 by the two-phase driving pulses φH1 and φH2, where the signal charges are converted into voltages and output.

FIG. 4 is a schematic view of an image pickup area of the CCD.
In the present embodiment, in order to speed up the read operation, only the read area necessary for focus detection is read at a normal speed,
Otherwise, high-speed sweep transfer is performed. In FIG. 4, reference numeral 41 denotes an area for reading at a normal speed used for focus detection, and reference numerals 42 and 43 denote high-speed sweep transfer areas in the first half and the second half, respectively.

FIG. 5 shows that the vertical charge transfer elements 32 of the CCD
A timing chart for one vertical synchronization period in the case of phase driving is shown. VD is a vertical synchronization signal, a vertical blanking period is indicated by a LOW potential, and HD is a horizontal synchronization signal, and a horizontal blanking period is indicated by a LOW potential. φV
1, φV2, φV3 and φV4 denote four-phase driving pulses of the vertical charge transfer element 32, and 51 and 52 denote readout pulses for transferring the signal charges photoelectrically converted by the pixels 31 to the horizontal charge transfer element 33. 53 and 54 of the four-phase drive pulses are 42 and 43 in FIG. 4, respectively.
A high-speed sweeping transfer pulse for transferring the signal charges read out to the vertical charge transfer elements 32 in the area of the area at a high speed. By sweeping out the area other than the necessary read area at high speed in this manner, the partial read operation can be speeded up.

When reading out the focus detection signal,
The signal charges of a plurality of lines are added using the horizontal charge transfer element 33, and the addition and reading of the plurality of lines can be performed. This is to compensate for the light quantity shortage that occurs at the time of focus detection because the pupils 3a and 3b of the focus detection aperture 3 are smaller than when the shooting aperture is opened at the time of actual shooting.
It is used to increase the sensitivity together with the gain of the amplifier 3.

Using the focus detection image data left image 1, right image 2, left image 3, right image 4, and left image 5 which have been read out at high speed in this way, a correlation operation between the images is performed. By detecting the phase difference between the images, focus detection is performed.However, unlike a monochrome sensor dedicated to focus detection such as that used in a single-lens reflex type silver halide camera, an image sensor for photographing is used. When the sensor also serves as a focus detection sensor, since a color filter is built in a single-plate color sensor that is an imaging element for photographing, an output signal from the CCD is used as a phase difference detection signal. In order to use it, signal processing must be performed.

As described above, the luminance signal generated by using the luminance signal processing circuit for photographing can be used as the signal for detecting the phase difference, but the luminance signal for photographing is processed by a predetermined matrix operation. Considering the time required for focus detection, the shorter the time for the arithmetic processing for generating the phase difference detection signal from the output signal of the image sensor, the better, but using a luminance signal processing circuit for photography. There is a problem that it takes too much time to calculate the phase difference detection signal.

A description will now be given of a calculation process for obtaining a phase difference detection signal for focus detection from an output signal of the CCD.
FIG. 6A is a diagram showing a color filter array formed on a CCD. Ye, Cy, M constituting color filter
g, G (yellow, cyan, magenta, green) 4
When the colors are added one pixel at a time, Ye + Cy + Mg + G =
Since 2B + 3G + 2R, 1 of adjacent 2 × 2 pixels
The average of the outputs of the blocks is used as a phase difference detection signal.

Actually, as described above, line addition reading can be performed at the time of reading from the CCD 9. Therefore, by performing two line addition reading, two vertical pixels are added in the CCD (FIG. 6B), and this analog signal is output. A / D converter 1
The digital signal is converted into a digital signal by 1 and the operation of two pixels in the horizontal direction is performed in the digital signal processing unit 12 (FIG. 6C).
A phase difference detection signal is obtained by averaging the outputs of one block of × 2 pixels.

In FIG. 6, the pitch of the phase difference detection signal is p with respect to the pixel pitch p of the CCD. In this embodiment, the vertical addition is performed in the CCD in order to improve the S / N ratio. However, the vertical addition may be performed without performing line addition, and the vertical addition may be performed by digital processing.

Next, a phase difference is obtained by performing a correlation operation using the phase difference detection signal obtained as described above. FIG.
8 and 9 are views for explaining the focus detection principle of the pupil time division phase difference AF according to the present invention. FIG. 7 shows a focused state, that is, a focused state. In the photographing state of FIG. 7C, the light flux 27 transmitted through the photographing lens 1 is CC.
Focus is on the light receiving surface of D9, and the defocus amount is zero.

FIG. 7A shows a state in which the focus detecting diaphragm 3 is inserted into the optical path of the photographing lens, the right pupil 3b is shielded by the light shielding plate 5, and the left pupil 3a is opened with the left pupil 3a opened. The luminous flux 27a forms an image at a position zero away from the optical axis 26 on the light receiving surface of the CCD 9. FIG. 7B shows a state in which the left pupil 3 a is shielded by the light shielding plate 5 and the right pupil 3 b is opened, and the light beam 27 a passing through the right pupil 3 b is separated from the optical axis 26 by zero on the light receiving surface of the CCD 9. Image at the position. As described above, in the focused state, the optical images composed of the light beams that have passed through the two different pupils 3a and 3b are both incident on the light receiving surface of the CCD 9 from the optical axis 26 at the same zero position. Is zero.

FIG. 8 shows a state in which the defocus amount is in the front focus state by d, and in the photographing state shown in FIG.
The light flux 27 passing through is focused before the light receiving surface of the CCD 9 by d. FIG. 8A shows a state in which the focus detection diaphragm 3 is inserted on the optical path of the photographing lens 1 and the left pupil 3a is opened.
Is formed at a position + δ / 2 away from the optical axis 26 on the light receiving surface of. FIG. 8B shows a state in which the right pupil 3b is opened, and the light beam 27a passing through the right pupil 3b forms an image on the light receiving surface of the CCD 9 at a position separated by -δ / 2 from the optical axis 26. Thus, in the front focus state, two different pupils 3a, 3b
The phase difference between the two images composed of the luminous flux passing through is (+ δ
/ 2)-(-δ / 2) = δ.

FIG. 9 shows a state in which the defocus amount is in the back focus state by d. In the photographing state shown in FIG.
The light flux 27 passing through is focused behind the light receiving surface of the CCD 9 by d. FIG. 9A shows a state in which the focus detection diaphragm 3 is inserted into the optical path of the photographing lens 1 and the left pupil 3a is opened.
Is formed at a position -δ / 2 away from the optical axis 26 on the light-receiving surface of the light-emitting device. 9B, with the right pupil 3b opened, the light beam 27a passing through the right pupil 3b forms an image on the light receiving surface of the CCD 9 at a position away from the optical axis 26 by + δ / 2. Thus, in the front focus state, two different pupils 3a, 3b
The phase difference between the two images composed of the luminous flux passing through is (−δ
/ 2)-(+ δ / 2) =-δ. As described above, 2 composed of the light beam that has passed through the left pupil 3a and the right pupil 3b
The focus state can be detected by calculating the phase difference between the phase difference detection signals of the two images.

Next, the algorithm for calculating the phase difference will be described. First, for simplicity of explanation, a case where the relative position between the subject and the photographing optical system has not moved will be described as an example. A phase difference between a phase difference detection signal (left image) composed of a light beam having passed through the left pupil region and a phase difference detection signal (right image) composed of a light beam having passed through the right pupil region is obtained by a correlation operation. Now, as shown in FIG. 10, the phase difference detection signal is composed of M (horizontal) × N (vertical) signals, and m × n signals are cut out of these signals to perform a correlation operation.

As shown in the following equation (1), the correlation of the left image: a is fixed, and the right image: b
Calculates the integral value C (τx, τy) of the product of a and b (τx, τy are integer variables) while sequentially shifting the phase (τx, τy) with respect to the left image, and calculates each phase (τx, τy). A value C (τx, τy) obtained for each τy) is defined as a correlation amount between the two images.

[0037]

(Equation 1)

The correlation amount C (τx, τy) takes the minimum value, and the phase (τx, τy) at this time is the phase difference (δx, δy) between the data a of the left image and the data b of the right image. Equivalent to. Further, the phase (τx,
τy) is an integer value (the signal pitch of the phase difference detection signal is not the same as the pixel pitch of the CCD as shown in FIG. 6). The phase difference may be calculated by interpolation using the value of.

If the relative position between the subject and the photographing optical system does not move, if there is no influence of noise or the like,
δy becomes zero in principle. In addition, to reduce the effects of subject pattern, contrast, etc., the effects of differences in shooting conditions, the effects of the color filter array built into the solid-state image sensor, and the effects of noise, the filter processing is performed on the phase difference detection signal. , A correlation operation may be performed.

Since the relationship between the phase difference, the image plane moving amount, and the defocus amount is determined by the optical system, the defocus amount is obtained from the phase difference, and the extension amount necessary for focusing is obtained. Is performed, and focus is achieved.
Up to this point, if there is no relative movement between the subject and the optical system,
The method of performing focus detection using two images composed of light beams that have passed through different pupil regions has been described.

Next, a description will be given of a phase difference calculation algorithm when there is a relative movement between the subject and the optical system. The phase difference can be obtained from two images (left image and right image) composed of light beams of different pupil regions. In the present embodiment, the left image and the right image are chronologically captured, and the subject If there is a relative movement of the optical system (such as camera shake or movement of a subject), a focus detection error is included due to the influence.

Therefore, as described above, the readout of the focus detection image data does not involve reading out the entire screen of the CCD and only a part of the image necessary for focus detection as in the case of reading out a photographing image. The reading is performed at a higher speed than in the case of the normal main photographing, and the image capturing interval is devised to be as short as possible. However, the operation of capturing the signals of the left image and the right image is temporally different, and if there is a relative movement between the camera and the subject during this time, it is inevitable that an error occurs in the phase difference detection.

Therefore, in the present embodiment, as described above, the left image 1, the right image 2, the left image 3, the right image 4, the left image 5, and the five Using a plurality of phase differences obtained from the captured images and between the images, detection errors due to relative movement between the camera and the subject (such as camera shake or subject movement) are corrected. When the camera and the subject move relatively, the effect appears in the phase difference (δx, δy) between the two images. Here, δy is the phase difference in the vertical direction of the image and is a direction perpendicular to the phase shift direction generated in accordance with the optical defocus amount, so that the phase shift component due to defocus is not included, and the relative position between the camera and the subject is not included. There is only a phase shift component due to movement. Therefore, the value of δy is used as it is as the correction amount of the vertical image movement due to the relative movement between the camera and the subject.

Δx is the phase difference in the horizontal direction of the image, which corresponds to the phase shift direction generated according to the optical defocus amount. Therefore, the value δx obtained from the two images including the relative movement of the camera and the subject includes: A phase shift component due to defocus and a phase shift component due to relative movement between the camera and the subject are included.

FIG. 11 shows the horizontal position (x direction) of the image due to the relative movement between the camera and the subject after correcting the image movement due to the relative movement between the camera and the subject in the vertical direction, and the phase difference between the images. The relationship is shown. Solid lines L1, R2,
L3, R4, and L5 are left images 1, right images 2,
A left image 3, a right image 4, and a left image 5, where t = t1, t
2, t3, t4, and t5 are capture times of the respective images, and the capture time intervals of the respective images are equal. Dashed line R
1 ', L2', R3 ', L4', R5 'are L
1, R2, L3, R4, and L5 indicate the position of an image composed of light fluxes that have passed through different pupils when the capturing is performed at the same time. A signal that cannot be captured in practice. It is.

M1, m2, m3, and m4 are the amounts of movement of the camera and the subject on the image due to the relative movement between the image capturing times, and δ is a phase shift component due to defocus of the optical system. This is the phase difference to be obtained as focus detection.

First, images L1 and R2 composed of different pupils,
The respective phase differences δ1, δ2, δ3, δ4 of R2 and L3, L3 and R4, and R4 and L5 are determined based on the left image.
As can be seen from FIG. 11, each phase difference obtained here includes a phase shift component due to defocus and a phase shift component due to relative movement between the camera and the subject.

FIG. 12 is a diagram for explaining a method of correcting a phase difference detection error. The horizontal axis represents the time axis t, and the vertical axis represents the position x of the subject image in the horizontal direction. L1, R2, L3, R
4. L5 is the position of the subject image captured in time series, and while capturing this signal, the movement of the subject image due to the relative motion between the camera and the subject is approximated by a quadratic function, so that the left image at times t2 and t4 is obtained. Subject position L2 ', L
Find the position of 4 '. And the phase difference δ between L2 ′ and R2.
The average value of 2 ′ and the phase difference δ4 ′ between L4 and R4 ′ is defined as the finally obtained phase difference δ.

(T1, x1), (t3, x3), (t
5, x5), the quadratic function y = At ² + Bt is calculated as t1 =
0, x1 = 0, and the finally obtained phase difference δ is δ
When obtained from the average value of 2 ′ and δ4 ′, since the capture interval of each image is now equal, the corrected phase difference δ is given by the following equation (2). δ = (δ1 + 3 × δ2 + 3 × δ3 + δ4) / 8 (2)

From the thus obtained phase difference δ corrected for the relative movement between the camera and the subject, the amount of extension necessary for focusing is determined, focus control of the optical system is performed, and after focusing, shooting is performed. .

In the case of a camera system that falls within a predetermined allowable defocus amount in the first focus detection and the focus control of the optical system, the second focus detection for confirmation is not performed (close-out focusing), and photographing is performed. You may enter. However, in order to confirm the focus, the focus detection is performed for the second and subsequent times. In this case, the correlation calculation is performed again to obtain the phase difference. However, if the defocus amount is smaller than the first time, if the same correlation calculation is performed as the second time, an extra calculation is performed. .

Therefore, in the first embodiment of the present invention,
In the second and subsequent focus detections, the correlation calculation is performed by narrowing the search range of the image signal at the time of detecting the phase difference based on the defocus amount which is the previous focus detection information.
In practice, the value of Tx, which is the search range of τx in the correlation operation expression (1), may be reduced to a value obtained up to the phase difference δ1 detected at the first time. As described above, at the time of large defocus, the time required for the second and subsequent correlation calculations can be shortened by increasing the search range and narrowing the search range when near the focus.

Here, an example has been described in which the search range at the (i + 1) -th time is reduced to a value obtained up to the phase difference δ1 detected at the first time, but when the (i + 1) -th focus detection is performed. Is smaller than the first phase difference, the search range may be changed from the first phase difference δ1 to a predetermined phase difference δ1 ′ (<δ1).

After focusing has been performed in this way, the focus detection diaphragm 3 is retracted from the optical path of the photographing lens, and photographing is performed. Then, the data of the CCD 9 is read for recording, subjected to image signal processing by the digital signal processing unit 12, subjected to processing such as image data compression if necessary, and recorded on a recording medium via the card slot 13. Note that this image data is subjected to video processing for finder display in the digital signal processing unit 12 and passed through the VRAM 21 to the EVF.
18 is displayed. This allows the photographer to check the subject.

Next, a second embodiment will be described. Second
In this embodiment, the number of steps of shifting the image signal in the correlation calculation for obtaining the phase difference is changed between the focus detection up to the previous time and the next focus detection. Hereinafter, only different points from the first embodiment will be described, and the description of the same portions as the first embodiment will be omitted.

When it is necessary to perform focus detection and focus control of the optical system twice or more in order to achieve focusing, in the case of a large defocus state, focus detection is first performed in a large defocus state, and finally focus is performed. Focus detection is performed in a state near the focus. For this reason, at first, the focus may be roughly detected, the lens may be moved to the vicinity of the in-focus state, and then the focus may be adjusted with high accuracy.

Therefore, in the first focus detection, the way of shifting τx is expressed by τx = −
Tx, -Tx + 2 ..., -4, -2, 0, 2, 4, ..., T
As x-2 and Tx, the step amount of shifting the image signal is doubled. When the previous phase difference δ becomes smaller than a predetermined value, the number of steps for shifting the image signal is set to one.
The correlation calculation is performed by changing the values so as to be shifted one by one. That is,
At first, focus is roughly adjusted by lowering the focus detection accuracy,
The focus detection accuracy is changed so that the detection is performed with high accuracy when the focus comes near the focus. As described above, the amount of calculation is reduced by the amount of the step amount, and the correlation calculation time is shortened. still,
The same may be performed for τy.

Next, a third embodiment will be described. In the third embodiment, the size of the area of the phase difference detection signal used for the correlation calculation is changed between the previous focus detection and the next focus detection. Hereinafter, only different points from the first embodiment will be described, and the description of the same portions as the first embodiment will be omitted.

Normally, the size of the region of the phase difference detection signal used for the correlation calculation is fixed to a predetermined size.
If the contrast of the subject is low and the phase difference cannot be detected, or if the detection accuracy is poor, if the area is widened, the edge of the subject enters the area, and the detection accuracy is improved. However, usually, if the area is too large, an object with a long distance and an object with a long distance in the area are likely to occur, and the detection accuracy is reduced (distant / far conflict). Therefore, the area is fixed to a predetermined size.

Then, looking at the focus detection results up to the previous time,
If the focus cannot be detected, or if focus is not achieved after performing the focus detection a predetermined number of times, the size of the region is increased by a predetermined value. This prevents useless repetition of focus detection in a state where the edge of the subject is not included in the area, and can reduce the number of times of focus detection, thereby shortening the calculation time. The area may be expanded not only in the horizontal direction but also in the vertical direction.

Next, in the fourth embodiment, it is selected whether or not to perform a shift operation in a direction substantially perpendicular to a phase shift direction generated according to the defocus amount of the optical system by a correlation operation. That is. Hereinafter, only different points from the first embodiment will be described. When the focus detection device according to the present embodiment captures an optical image composed of light beams having different pupil regions, the capture time differs between the left image and the right image.
For this reason, as described above, measures are taken when there is a relative movement between the subject and the optical system. However, when the camera shake is large, the phase shift direction generated according to the defocus amount of the optical system by the correlation calculation is taken. When the shift operation in the vertical direction, which is a substantially vertical direction, is performed, the phase differences δ1, δ2, δ
3, δ4 and δ5 vary, and the detection accuracy of the phase difference δ deteriorates.

Also, when the S / N of the phase difference detection signal is poor, for example, when the subject is dark and the light quantity is insufficient, and the gain of the output from the CCD 9 is greatly increased by the amplifier 10, the correlation calculation is performed. When a vertical shift operation, which is a direction substantially perpendicular to the phase shift direction generated according to the defocus amount of the system, is performed, the phase differences δ1, δ2, δ
3, δ4 and δ5 vary, and the detection accuracy of the phase difference δ deteriorates.

Therefore, under such photographing conditions, when performing the correlation operation, the vertical shift operation which is substantially perpendicular to the phase shift direction generated according to the defocus amount of the optical system is not performed. To do. Actually, it is determined that the S / N is poor from the value of the exposure determination for adjusting the light amount, and the vertical shift calculation is not performed. Further, respective phase differences δy1, δy2, δy3, δ in the vertical direction of images L1 and R2, R2 and L3, L3 and R4, and R4 and L5 composed of different pupils.
When the variations of y4 and δy5 are larger than a predetermined value, it is determined that the camera shake is large, and the shift operation in the vertical direction is not performed.

As described above, when the vertical shift operation, which is a direction substantially perpendicular to the phase shift direction generated according to the defocus amount of the optical system, is performed, the detection accuracy of the phase difference δ is rather deteriorated. By not performing the correlation calculation in that direction, it is possible to prevent a decrease in detection accuracy and shorten the calculation time.

Next, in the fifth embodiment, in the correlation calculation for obtaining the phase difference, there is provided a thinning means which can perform the correlation calculation by thinning out the data of the phase difference detection signal in the area used for the correlation calculation. In this case, whether or not to perform thinning is changed based on the focus detection result up to the previous time. Hereinafter, portions different from the first embodiment will be described.

As described in the first embodiment, when performing the correlation calculation for obtaining the phase difference, all the phase difference detections in the area used for the correlation calculation (m × n areas in FIG. 10) are performed. The data of the application signal is used as it is. At this time, the signal pitch of the phase difference detection signal is the same as the pixel pitch of the CCD as described above. However, in order to perform focusing, it is necessary to perform focus detection and focus control of the optical system two or more times. In the case of a large defocus state, focus detection is first performed in a large defocus state, and finally, near focus is performed. In this state, focus detection is performed. That is, the lens may be moved roughly to near the focus at first, and then the focus may be adjusted with high accuracy.

Therefore, at the time of the first focus detection, as shown in FIG. 13, the data in the horizontal direction corresponding to the phase shift direction generated according to the defocus amount of the optical system is m × n data used in the correlation calculation as shown in FIG. The data of the area is thinned out every other pixel, and the correlation operation is performed by setting the signal pitch of the phase difference detection signal used for the correlation operation to の of the pixel pitch of the CCD. The hatched portions in FIG. 13 are data used for the correlation calculation. Then, when the phase difference δ, which is the previous focus detection result, becomes smaller than a predetermined value, the correlation calculation is performed without performing the thinning.

That is, at first, the focus detection accuracy is lowered and the focus is roughly adjusted, and the focus detection accuracy is changed so that the detection is performed accurately when the focus comes near the focus. By thinning in this way, the number of data used for the correlation operation becomes m / 2 × n, and the time required for the correlation operation is shortened.
In the present embodiment, the thinning amount is set to 1/2, but the thinning amount may be changed according to the focus result. Also, thinning may be performed in the vertical direction.

Further, the sixth embodiment is different from the fifth embodiment in that whether or not to perform thinning is changed based on the number of times of detection instead of the result of focus detection up to the previous time. . As described in the first embodiment,
When performing the correlation operation for obtaining the phase difference, the data of all the phase difference detection signals in the region used for the correlation operation (m × n regions in FIG. 10) is used as it is. The signal pitch of the phase difference detection signal at this time is CC as described above.
D is the same as the pixel pitch.

However, when it is necessary to perform the focus detection and the focus control of the optical system twice or more in order to achieve focusing, in the case of a large defocus state, focus detection is first performed in the large defocus state, and finally the focus detection is performed. Focus detection is performed in a state near the focus. That is, the lens may be moved roughly to near the focus at first, and then the focus may be adjusted with high accuracy.

Therefore, in the first focus detection, as shown in FIG. 13, the data in the horizontal direction corresponding to the phase shift direction generated according to the defocus amount of the optical system has m × n regions used for the correlation operation. Thinning out the data every other pixel,
The signal pitch of the phase difference detection signal used for the correlation operation is set to CC.
The correlation operation is performed with half the pixel pitch of D. FIG.
The hatched portions indicate data used for the correlation calculation. In the second and subsequent focus detections, the correlation calculation is performed without thinning.

That is, at the time of the first focus detection in which the defocus amount may be large, the focus detection accuracy is lowered, and the focus detection after the first time is performed with high accuracy. By thinning out in this way, in the first focus detection,
The number of data used for the correlation operation is m / 2 × n, and the time required for the correlation operation is reduced.

The system using the functional blocks shown in FIG. 1 may be configured as hardware, or may be configured as a microcomputer system including a CPU, a memory, and the like. When configured in a microcomputer system,
The memory constitutes a storage medium according to the present invention. The storage medium stores a program for executing the processing described in each embodiment. The storage medium may be a semiconductor memory such as a ROM or a RAM, an optical disk, a magneto-optical disk, a magnetic medium, or the like, and may be used as a CD-ROM, a floppy disk, a magnetic tape, or a nonvolatile memory card. Good.

[0074]

As described above, the first aspect of the present invention is characterized in that a plurality of calculation methods for focus detection are selected and a plurality of calculations are performed. As a result, it is possible to select a calculation method suitable for focus detection at that time, and unnecessary calculation can be omitted as compared with a case where one calculation method is used in all cases, thereby shortening the calculation time. Can be measured.

The second invention is characterized in that a calculation method for the next focus detection is selected based on the focus detection information up to the previous time. This makes it possible to change the calculation method between the previous and next times based on the focus detection result, the number of times of focus detection, and the focus detection conditions that are the focus detection information up to the previous time, and to perform unnecessary calculations at the next time the focus detection is performed. Since this can be omitted, the calculation time can be shortened.

The third invention is characterized in that, in the correlation operation for obtaining the phase difference, the search range of the image signal at the time of detecting the phase difference is changed between the previous time and the next time. Thus, when the defocus amount is large, the search range for detecting the phase difference can be increased, and when the defocus amount is small, the search range can be reduced. By setting the next search range based on the previous defocus amount, unnecessary calculations can be omitted, and the calculation time can be shortened.

The fourth invention is characterized in that, in the correlation calculation for obtaining the phase difference, the step amount for shifting the image signal at the time of detecting the phase difference is changed between the previous time and the next time. Thus, when the focus detection accuracy may be low, the number of steps is increased, and when high accuracy is required for focus detection, the number of steps is reduced, so that the calculation corresponding to the required focus detection accuracy is performed. And the calculation time can be shortened.

The fifth invention is characterized in that, in the correlation calculation for obtaining the phase difference, the size of a plurality of image signal regions used for the correlation calculation is changed between the previous time and the next time. As a result, it is possible to change the region of the subject used for focus detection, to increase the focus detection accuracy, and to omit unnecessary calculations, thereby shortening the calculation time.

According to a sixth aspect, in the correlation operation for obtaining the phase difference, it is selected whether or not to perform a shift operation in a direction substantially perpendicular to a phase shift direction generated according to the defocus amount of the optical system. It is characterized by. This allows
By performing the shift operation in the direction perpendicular to the phase shift direction generated according to the defocus amount, when the focus detection accuracy is deteriorated, the shift operation in this direction can be omitted, and the focus detection accuracy can be increased. In addition, since unnecessary calculations can be omitted, the calculation time can be shortened.

According to a seventh aspect of the present invention, a correlation operation is performed by thinning data of a plurality of image signals in an area used for the correlation operation, and whether or not to perform the thinning based on the focus detection information up to the previous time is determined. It is characterized by changing. This makes it possible to change the method of thinning out the data used for the correlation calculation between the previous and next times according to the required focus detection accuracy, and it is possible to omit unnecessary calculations, thereby shortening the calculation time. Can be measured.

The eighth invention is characterized in that the focus detection information is a focus detection result. As a result, the defocus state can be determined from the previous focus detection result.
Accordingly, it is possible to change the method of thinning out the data used for the correlation operation, and it is possible to omit unnecessary calculations, thereby shortening the operation time.

In the ninth and tenth aspects, the focus detection information is the number of times of focus detection. This makes it possible to change the method of thinning out the data used for the correlation calculation in accordance with the number of times of focus detection. The thinning is performed in the first focus detection, in which the probability of a large defocus amount is high, and the thinning is performed as the number of times increases. By not doing so, unnecessary calculations can be omitted, and the calculation time can be shortened.

[Brief description of the drawings]

FIG. 1 is a configuration diagram showing an embodiment of a focus detection device and a camera using the same according to the present invention.

FIG. 2 is a configuration diagram illustrating a positional relationship between a focus detection diaphragm and a light shielding plate.

FIG. 3 is a configuration diagram of an interline CCD.

FIG. 4 is a configuration diagram of an imaging area of a CCD.

FIG. 5 is a timing chart for one vertical synchronization period when a vertical charge transfer element of a CCD is driven in four phases.

FIG. 6 is a configuration diagram for explaining arithmetic processing when a phase difference detection signal is obtained.

FIG. 7 is a diagram for explaining the principle of focus detection.

FIG. 8 is a diagram for explaining the principle of focus detection.

FIG. 9 is a diagram for explaining the principle of focus detection.

FIG. 10 is a configuration diagram for explaining how to take a correlation when a phase difference is obtained by a correlation operation.

FIG. 11 is a waveform diagram showing a relationship between a horizontal position of an image due to a relative movement between a camera and a subject and a phase difference between the images.

FIG. 12 is a characteristic diagram for explaining a method of correcting a phase difference detection error.

FIG. 13 is a configuration diagram for explaining how to thin out data used for correlation.

FIG. 14 is a configuration diagram of a conventional single-lens reflex camera having a phase difference detection type focus detection device.

FIG. 15 is a configuration diagram for explaining a conventional focus detection principle.

FIG. 16 is a waveform chart for explaining a conventional focus detection principle.

[Explanation of symbols]

1a Focusing lens group of photographing lens 1b Lens group other than focusing lens group 1b of photographing lens 2 Lens extension mechanism 3 Focus detection aperture 4 Motor 5 Shield plate 6 Motor 9 CCD 12 Digital signal processing unit 13 System control unit 14 CCD Driver 24 Operation switch

Claims (30)

[Claims] An imaging unit configured to capture an optical image of a subject formed on an imaging surface and output an image signal; an optical system configured to form the optical image on the imaging surface through a pupil region; Moving means for moving an area to a plurality of positions; detecting a phase difference between a plurality of image signals obtained by imaging at the plurality of positions by a correlation operation; and detecting a focus of the optical system means based on the phase difference. And a control means for causing the calculating means to perform a plurality of calculations and selecting and providing one of a plurality of calculation methods for each calculation. 2. The focus detection apparatus according to claim 1, wherein said control means selects a next calculation method based on focus detection information obtained up to the previous time. 3. The focus detection device according to claim 2, wherein in the calculation method, the search range of the image signal at the time of detecting the phase difference is changed between the previous time and the next time in the correlation calculation. 4. The focus detection apparatus according to claim 2, wherein the calculation method changes a step amount of shifting the image signal at the time of detecting the phase difference between the previous time and the next time in the correlation calculation. 5. The focus detection method according to claim 2, wherein in the calculation of the correlation, the size of the area of the plurality of image signals used in the calculation of the correlation is changed between the previous time and the next time. apparatus. 6. The calculation method according to claim 1, wherein, in the correlation calculation, whether to perform a shift calculation in a direction substantially perpendicular to a phase shift direction generated according to a defocus amount of the optical system means is selected. 3. The focus detection device according to claim 2, wherein: 7. A thinning-out unit for thinning out data of the plurality of image signals in an area used for the correlation calculation and performing a correlation calculation, wherein the thinning-out means is performed based on focus detection information obtained up to the previous time. 3. The focus detection device according to claim 2, wherein whether or not to perform is selected. 8. The focus detection device according to claim 7, wherein the focus detection information is the focus detection result. 9. The focus detection apparatus according to claim 7, wherein the focus detection information is the number of times of focus detection. 10. The method according to claim 7, wherein the focus detection result is a defocus amount, and the correlation calculation is performed by thinning out the first focus detection, and the correlation calculation is performed without thinning out when the focus comes close to focus. Focus detection device. 11. An imaging procedure for capturing an optical image of a subject formed on an imaging surface and outputting an image signal, and a pupil region in an optical system for forming the optical image on the imaging surface at a plurality of positions. A moving procedure for moving, a calculation procedure for detecting a phase difference between a plurality of image signals obtained by imaging at the plurality of positions by a correlation calculation, and performing focus detection of the optical system based on the phase difference; A control procedure for performing a plurality of calculations in the calculation procedure and selecting and providing one of the plurality of calculation methods for each calculation. 12. The focus detection method and focus detection device according to claim 11, wherein said control procedure selects a next calculation method based on focus detection information obtained up to the previous time. 13. The focus detection method according to claim 12, wherein in the calculation of the correlation, a search range of the image signal at the time of detecting the phase difference is changed between the previous time and the next time. 14. The focus detection method according to claim 12, wherein in the calculation of the correlation, a step amount of shifting the image signal at the time of detecting the phase difference is changed between the previous time and the next time. 15. The focus detection according to claim 12, wherein the calculation method changes the size of the area of the plurality of image signals used in the correlation calculation between the previous time and the next time in the correlation calculation. Method. 16. The calculation method according to claim 1, wherein in the correlation calculation, it is selected whether or not to perform a shift calculation in a direction substantially perpendicular to a phase shift direction generated according to a defocus amount of the optical system. 13. The focus detection method according to claim 12, wherein: 17. A thinning-out procedure for thinning out data of the plurality of image signals in an area used for the correlation calculation and performing a correlation calculation, wherein the thinning-out is performed based on focus detection information obtained up to the previous time. 13. The focus detection method according to claim 12, wherein whether or not to perform is selected. 18. The focus detection method according to claim 17, wherein the focus detection information is the focus detection result. 19. The focus detection method according to claim 17, wherein the focus detection information is the number of times of focus detection. 20. The method according to claim 17, wherein the focus detection result is a defocus amount, and the correlation calculation is performed by thinning out the first focus detection, and the correlation calculation is performed without thinning out when the focus comes near the focus. Focus detection method. 21. An image pickup process for picking up an optical image of a subject formed on an image pickup surface and outputting an image signal, and pupil regions in an optical system for forming the optical image on the image pickup surface at a plurality of positions. A moving process of moving, detecting a phase difference between a plurality of image signals obtained by imaging at the plurality of positions by a correlation calculation, and performing a focus detection of the optical system means based on the phase difference; A computer-readable storage medium storing a program for executing a plurality of operations in the operation process and performing a control process of selecting and providing one of a plurality of operation methods for each operation. 22. The computer-readable storage medium according to claim 21, wherein said control processing selects a next calculation method based on focus detection information obtained up to a previous time. 23. The computer-readable storage medium according to claim 22, wherein in the calculation method, the search range of the image signal at the time of detecting the phase difference is changed between the previous time and the next time in the correlation calculation. . 24. The computer-readable computer according to claim 22, wherein in the correlation method, a step amount of shifting the image signal at the time of detecting the phase difference is changed between a previous time and a next time in the correlation calculation. Storage medium. 25. The computer-readable medium according to claim 22, wherein in the calculation method, the size of the area of the plurality of image signals used in the correlation calculation is changed between the previous time and the next time in the correlation calculation. Possible storage medium. 26. The method according to claim 26, wherein in the correlation operation, whether to perform a shift operation in a direction substantially perpendicular to a phase shift direction generated according to a defocus amount of the optical system is selected. 23. The computer readable storage medium according to claim 22, wherein: 27. A thinning-out unit for thinning out data of the plurality of image signals in an area used for the correlation operation and performing a correlation operation, and based on the focus detection information obtained up to the previous time. 23. The computer-readable storage medium according to claim 22, wherein it is selected whether or not to perform. 28. The computer-readable storage medium according to claim 27, wherein the focus detection information is the focus detection result. 29. The computer-readable storage medium according to claim 27, wherein the focus detection information is the number of times of focus detection. 30. The focus detection method according to claim 27, wherein the focus detection result is a defocus amount, and the first focus detection is performed by thinning out the correlation calculation, and the correlation calculation is performed without thinning out when the focus comes close to the in-focus state. Computer readable storage medium.