JP6592968B2 - Focus detection apparatus and imaging apparatus - Google Patents

Description

  The present invention relates to a focus detection apparatus and an imaging apparatus.

  A pair of light beams that have passed through the imaging optical system are photoelectrically converted by a pair of photoelectric conversion element arrays, and a deviation amount of a pair of signal sequences output from the pair of photoelectric conversion element arrays is detected by a correlation operation, and the detected deviation amount is obtained. When calculating the defocus amount of the imaging optical system based on the above, the shift amount is calculated for each of the plurality of partial areas obtained by dividing the focus adjustment area, and the most reliable partial area is selected from the divided partial areas. A technique is known (see Patent Document 1).

JP 2013-218082 A

  In the prior art, since the amount of deviation is calculated in each of the plurality of partial areas, there is a problem that the amount of calculation increases.

The focus detection apparatus according to the present invention includes a first photoelectric conversion unit that photoelectrically converts a first light beam that has passed through an optical system to generate charges, and a second light beam that is different from the first light beam that has passed through the optical system. A second photoelectric conversion unit that generates a charge by photoelectric conversion; a first signal based on the charge generated by the first photoelectric conversion unit; and a second signal based on the charge generated by the second photoelectric conversion unit. A second range that is narrower than the first range after performing a first correlation operation using a first frequency band in the first range of the first signal and the second signal in the imaging unit that outputs a signal The second correlation calculation is performed using a second frequency band higher than the first frequency band, and based on the first correlation calculation and the second correlation calculation, the image by the optical system and the imaging unit A calculating unit that calculates the amount of deviation.
An imaging apparatus according to the present invention includes the focus detection apparatus.

It is a figure explaining the principal part structure of a single-lens reflex camera. It is a figure explaining the principal part structure of a single-lens reflex camera. It is a block diagram which illustrates the circuit configuration of a single-lens reflex camera system. It is a schematic diagram explaining a focus detection optical system. FIG. 5A is a diagram illustrating an a-sequence signal, and FIG. 5B is a diagram illustrating a b-sequence signal. It is a figure which illustrates the relationship between the shift amount and the correlation amount. It is a figure which illustrates a pair of signal which consists of a series and b series. It is a figure which illustrates a pair of signal after a low-pass filter process. It is a figure which illustrates a pair of signal after a high pass filter process. It is a figure which illustrates the correlation amount in the signal after a different filter process. It is a flowchart explaining the flow of the camera process which MPU performs. It is a flowchart explaining the flow of the camera process which MPU performs in 2nd embodiment. It is a figure which illustrates the correlation amount in the signal after a different filter process. It is a figure explaining supply of a program.

Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings.
(First embodiment)
FIG. 1 and FIG. 2 are diagrams for explaining the main configuration of a single-lens reflex camera system equipped with a focus detection apparatus according to a first embodiment of the present invention. 1 and 2, a photographic lens 202 configured to be detachable from the camera body 201 is attached.

  Light from the subject is incident on the camera body 201 via the imaging optical system 210 of the photographing lens 202 and the diaphragm 211. The subject light incident on the camera body 201 is reflected toward the upper finder by a transflective quick return mirror (hereinafter referred to as a main mirror) 203 in the mirror-down state illustrated in FIG. An image is formed on 206. Part of the subject light transmitted through the main mirror 203 is reflected downward by the sub mirror 204 and is incident on a dedicated AF (autofocus) element 207 as a focus detection light beam. The dedicated AF element 207 photoelectrically converts incident light and outputs a signal. Since the signal output from the dedicated AF element 207 performs focus adjustment by the imaging optical system 210, focus detection processing for detecting the phase difference of the signal output from the dedicated AF element 207 (known pupil division AF processing) Used in

  The subject light imaged on the diffusing screen 206 is further incident on the pentaprism 208. The pentaprism 208 guides incident subject light to the eyepiece optical system 209. The photographer observes the subject image by the finder unit through the eyepiece optical system 209.

  After the release, the main mirror 203 rotates to the mirror-up position illustrated in FIG. 2, and the subject light is guided to the image sensor 102 via the mechanical shutter 205, and forms a subject image on the imaging surface. The image sensor 102 is configured by a CMOS image sensor or the like provided with a plurality of photoelectric conversion elements corresponding to pixels. The imaging element 102 captures a subject image formed on the imaging surface by photoelectrically converting incident light and outputting a signal.

  The lens driving mechanism 212 drives the focus adjustment lens constituting the imaging optical system 210 to advance and retreat in the optical axis direction (left and right direction in FIGS. 1 and 2). The driving direction and driving amount of the focus adjustment lens are instructed from the camera body 201 side. In order to simplify the drawing, the imaging optical system 210 is represented as a single lens.

  The mirror-down state in FIG. 1 is a shooting preparation state in which the photographer can observe the subject image via the eyepiece optical system 209. The single-lens reflex camera performs focus detection using the dedicated AF element 207 in the shooting preparation state.

  FIG. 3 is a block diagram illustrating a circuit configuration of the single-lens reflex camera system described above. The camera body 201 includes a timing circuit (TG) 101, an image sensor 102, an ASIC 103, a RAM 104, an MPU 106, a shutter drive mechanism 107, a mirror drive mechanism 108, an operation member 109, an AF control unit 110, and the like. The AE control unit 111 is provided, and a detachable recording medium 105 is mounted.

<Camera body>
A timing circuit (TG) 101 generates a predetermined timing signal and supplies the timing signal to each of the image sensor 102 and the ASIC 103. A signal (image signal) output from the image sensor 102 is input to the ASIC 103. The ASIC 103 performs predetermined image processing (including white balance adjustment) on the input image signal.

  The RAM 104 temporarily stores image data processed by the ASIC 103. The recording medium 105 includes a non-volatile memory that stores image data processed by the ASIC 103 as an image file.

  The MPU 106 is constituted by a microprocessor, inputs a signal output from each block of the camera body 201, performs a predetermined calculation, and outputs a control signal based on the calculation result to each block. The MPU 106 and the ASIC 103 transmit and receive each other's control signals by performing predetermined communication.

  The shutter drive mechanism 107 performs charge and drive control of the mechanical shutter 205 (FIGS. 1 and 2) in accordance with an instruction sent from the MPU 106. The mirror drive mechanism 108 controls the mirror-up drive and the mirror-down drive of the main mirror 203 (submirror 204) in accordance with an instruction sent from the MPU 106.

  The operation member 109 outputs a setting / switching signal corresponding to various settings and selection operations to the MPU 106. The operation member 109 includes a release button (not shown).

  The AF control unit 110 performs a known phase difference type focus detection calculation using a signal output from the dedicated AF element 207. By this calculation, the defocus amount by the imaging optical system 210 is obtained, and the movement amount of the focusing lens constituting the imaging optical system 210 is calculated according to the defocus amount. A signal indicating the movement amount of the focus adjustment lens is transmitted to the lens CPU 214 via the MPU 106.

  The AE control unit 111 calculates the luminance of the subject using a signal output from a photometric sensor (not shown). The AE control unit 111 further performs calculation in a predetermined exposure calculation mode using the set imaging sensitivity, the lens information acquired by the MPU 106, and the calculated luminance of the subject, and the aperture value AV and shutter at the time of shooting Determine the speed TV.

  The MPU 106 communicates with the lens CPU 214 of the photographing lens 202 attached to the camera body 201. Through communication between the camera body 201 and the photographic lens 202, lens information such as an aperture value and lens data is transmitted from the photographic lens 202 to the camera body 201, while lens control information such as the movement amount and driving instruction of the focus adjustment lens. Is transmitted from the camera body 201 to the photographing lens 202 side.

<Photographing lens>
The photographing lens 202 is provided with a lens CPU 214, a lens driving mechanism 212, and an aperture driving mechanism 213. The lens CPU 214 communicates with the MPU 106 on the camera body 201 side. The lens CPU 214 transmits lens information such as an aperture value and lens data to the camera body 201 side, and acquires lens control information such as a moving amount of the focus adjustment lens and a drive instruction from the camera body 201.

  The lens driving mechanism 212 moves the focus adjustment lens by a predetermined amount in a predetermined direction in response to an instruction sent from the lens CPU 214. The aperture driving mechanism 213 drives the aperture 211 (FIGS. 1 and 2) a predetermined number of stages so that the aperture value is in accordance with the instruction sent from the lens CPU 214.

<AF processing>
The focus detection calculation performed by the AF control unit 110 of the single-lens reflex camera system described above will be described. In the present embodiment, a shift amount of a pair of images (hereinafter referred to as a shift amount S) by a pair of light beams is detected by a pupil division type phase difference detection method.

<Description of deviation amount S>
FIG. 4 is a schematic diagram for explaining a focus detection optical system including the dedicated AF element 207. In FIG. 4, the focus detection optical system includes a field mask 350, a field lens 300, stop apertures 401 and 402, re-imaging lenses 501 and 502, and a pair of photoelectric conversion element arrays A and B. By this focus detection optical system, a pair of subject images is formed on the pair of photoelectric conversion element arrays A and B. That is, an image of the light beam that has passed through the aperture opening 401 provided behind the field lens 300 is formed on the photoelectric conversion element array A by the re-imaging lens 501. In addition, an image of the light beam that has passed through the aperture opening 402 provided behind the field lens 300 is formed on the photoelectric conversion element array B by the re-imaging lens 502.

  A plurality of photoelectric conversion elements are arranged in each of the photoelectric conversion element arrays A and B. The dedicated AF element 207 includes photoelectric conversion element arrays A and B. The photoelectric conversion element array A outputs signals from a plurality of photoelectric conversion elements (A1, A2,..., An), respectively. The photoelectric conversion element array B outputs a signal from each of the plurality of photoelectric conversion elements (B1, B2,..., Bn). A region 91 on the imaging optical system 210 is a back-projected image of the aperture opening 401. Similarly, a region 92 on the imaging optical system 210 is a back-projected image of the stop aperture 402.

  The light flux from the subject incident through the area 91 of the imaging optical system 210 forms an image on or before the planned focal plane 600 equivalent to the imaging surface of the image sensor 102, and then the aperture of the field mask 350, the field lens 300, An image is formed on the photoelectric conversion element array A through the aperture opening 401 and the re-imaging lens 501.

  Further, the light flux from the subject that has entered through the region 92 of the imaging optical system 210 forms an image on or before and after the planned focal plane 600, and then the aperture of the field mask 350, the field lens 300, the aperture opening 402, and the reconnection. An image is formed on the photoelectric conversion element array B through the image lens 502.

  A pair of subject images formed on the pair of photoelectric conversion element arrays A and B are separated from each other in a so-called front pin state in which the imaging optical system 210 forms a sharp image of the subject before the planned focal plane 600, and conversely In the so-called rear pin state, in which a sharp image of the subject is formed after the planned focal plane 600, they approach each other. When the imaging optical system 210 forms a sharp image of the subject at the planned focal plane 600, so-called in-focus state, the pair of subject images on the pair of photoelectric conversion element arrays A and B relatively match (the amount of deviation S is S). 0 state).

  The AF control unit 110 performs image shift detection calculation processing (correlation calculation processing, phase difference detection processing) on the a-sequence signal and the b-sequence signal photoelectrically converted by the pair of photoelectric conversion element arrays A and B, respectively. Thus, the deviation amount S is detected. The shift amount S is obtained using an a-sequence signal and a b-sequence signal.

<Correlation calculation>
As shown in FIG. 4, a plurality of photoelectric conversion elements are arranged in each of the photoelectric conversion element arrays A and B. The a-series signals (A1, A2,..., An) are composed of signals output from the plurality of photoelectric conversion elements (A1, A2,..., An), respectively. The b series signals (B1, B2,..., Bn) are composed of signals output from the plurality of photoelectric conversion elements (B1, B2,..., Bn), respectively. The signals (A1, A2,..., An) and the signals (B1, B2,..., Bn) constitute a pair of image signals for focus detection.

  The AF control unit 110 determines a range of signals used for image shift detection calculation processing in the a-sequence signals (A1, A2,..., An) and the b-sequence signals (B1, B2,..., Bn). For example, an image shift detection calculation is performed using substantially the entire range of the a-sequence signal and the b-sequence signal, or an image shift detection calculation is performed using a part of the range of the a-sequence signal and the b-sequence signal. can do.

FIG. 5A is a diagram illustrating an a-sequence signal (a1, a2,..., An) after a later-described filtering process is applied to the a-sequence signal (A1, A2,..., An), and FIG. FIG. 5B is a diagram exemplifying b-sequence signals (b1, b2,..., Bn) after filtering processing to be described later on the b-sequence signals (B1, B2,..., Bn). The AF control unit 110 relatively shifts the a-sequence signals (a1, a2,..., An) and the b-sequence signals (b1, b2,..., Bn) of the pair of image signals in a one-dimensional manner, The correlation calculation shown in the following equation (1) is performed.
C (k) = Σ | a (n + k) −b (n) | (1)
The Σ operation indicates a cumulative operation for n, and the calculation range is limited to a range in which image signals of a (n + k) and b (n) exist according to the shift amount k. The shift amount k is an integer, and is a relative image shift amount in units of the arrangement interval of the elements of the photoelectric conversion element arrays A and B.

  FIG. 6 is a diagram illustrating the relationship between the shift amount k (horizontal axis) and the correlation amount C (k). As shown in FIG. 6, the calculation result of the above equation (1) shows that the correlation amount C (k) is the smallest in the shift amount (k = kj = −3 in the example of FIG. 6) in which the correlation between the pair of image signals is high. (The higher the degree of correlation, the smaller the correlation amount).

The AF control unit 110 calculates a minimum value C (x) of the correlation amount based on the calculated correlation amount C (k). In the present embodiment, for example, the minimum value C (x) with respect to the continuous correlation amount and the shift amount that gives the minimum value C (x) by the three-point interpolation method shown in the following equations (2) to (5) x is calculated. Note that C (kj) in the equation is a value satisfying the conditions of C (k−1) ≧ C (k) and C (k) <C (k + 1) in FIG. 6 (kj = −3).
D = {C (kj−1) −C (kj + 1)} / 2 (2)
C (x) = C (kj) − | D | (3)
x = kj + D / SLOP (4)
SLOP = MAX {C (kj + 1) -C (kj), C (kj-1) -C (kj)} (5)

  In the present embodiment, the AF control unit 110 has two different pass frequency bands for a pair of a-sequence signals (A1, A2,..., An) and b-sequence signals (B1, B2,..., Bn). Each filter process is performed. The two filter processes include a first filter process for limiting the a-sequence signal (A1, A2,..., An) and the b-sequence signal (B1, B2,..., Bn) to the first frequency band, respectively. It is the 2nd filter processing which restricts to the 2nd frequency band higher than the 1st frequency band. In other words, the two filter processes are a filter process having a first pass frequency band and a filter process having a pass frequency band higher than the first pass frequency band. For example, a filter (low-pass filter) process that outputs a low frequency component of the input signal for each of the a-sequence signal (A1, A2,..., An) and the b-sequence signal (B1, B2,..., Bn). Then, a filter (high pass filter) process for outputting a high frequency component of the input signal is performed. The same applies to a second embodiment described later. The AF control unit 110 calculates a shift amount that increases the degree of correlation between the pair of signals (that is, reduces C (k) of the correlation amount) using the filtered signal.

FIG. 7 shows an a-sequence signal (A1, A2,..., An) that is an output of the photoelectric conversion element array A and a b-sequence signal (B1, B2,..., Bn) that is an output of the photoelectric conversion element array B. FIG. FIG. 7 is a graph in which the horizontal axis represents the position (element order) of each photoelectric conversion element in the photoelectric conversion element array, and the vertical axis represents the output value for each photoelectric conversion element. The AF control unit 110 sets a range 71 (calculation range) of a pair of signals that is surrounded by a broken line to be used for image shift detection calculation processing. The calculation range 71 is determined according to, for example, the size of a subject to be focused. The calculation range 72 will be described in the second embodiment.
The calculation range may be a partial range of the a-sequence signal and the b-sequence signal, or may be a partial range of the a-sequence signal and the b-sequence signal that have been subjected to calculation processing such as filter processing. May be. In other words, in the correlation calculation, it is also a shift direction in which the a-sequence signal and the b-sequence signal are relatively shifted.

  8 shows a pair of a-sequence signals and b-sequence signals (filter output value P1 () after low-pass filter processing is applied to the pair of a-sequence signals and b-sequence signals in the calculation range 71 of FIG. j) It is a figure which illustrates). The calculation range 72 will be described in Modification 2. 9 shows a pair of a-sequence signals and b-sequence signals (filter output value P2 () after high-pass filter processing is applied to the pair of a-sequence signals and b-sequence signals in the calculation range 71 of FIG. j) It is a figure which illustrates). The calculation range 72 will be described in the second embodiment.

  The AF control unit 110 performs a correlation operation on the filter output value P1 (j) and also performs a correlation operation on the filter output value P2 (j). FIG. 10 is a diagram exemplifying the relationship between the shift amount k (horizontal axis) and the correlation amount C (k) obtained by two correlation operations, and the upper stage is for the filter output value P1 (j) (FIG. 8). The lower part shows the correlation amount C (k) for the filter output value P2 (j) (FIG. 9).

  In general, a pair of a-sequence signals and b-sequence signals (filter output value P1 (j)) after low-pass filter processing are suitable for rough phase difference detection, but fine contrast tends to be crushed. In addition, the pair of a-sequence signals and b-sequence signals (filter output value P2 (j)) after the high-pass filter processing are suitable for detecting a phase difference with fine contrast, but may detect a phase difference in error. There is.

  Therefore, the AF control unit 110 first performs a correlation operation using the pair of a-sequence signals and the b-sequence signals (filter output value P1 (j)) after the low-pass filter processing, so that the correlation between the pair of signals is performed. The first shift amount kp1 is detected in which the degree is increased (that is, the correlation amount C (k) is decreased). The correlation calculation is the first correlation calculation.

  Next, the AF control unit 110 detects the first shift amount kp1 detected by the first correlation calculation using the pair of a-sequence signal and b-sequence signal (filter output value P1 (j)) after the low-pass filter processing. Using a pair of a-sequence signals and b-sequence signals (filter output value P2 (j)) after the high-pass filter processing in a narrow range Δd narrower than the shift range ΔW in the first correlation calculation. Perform the operation. The correlation calculation performed after the first correlation calculation is defined as the second correlation calculation. That is, two correlation amounts are calculated using a pair of signals in the low frequency band (filter output value P1 (j)) and a pair of signals in the high frequency band (filter output value P2 (j)). A shift amount that reduces both amounts is adopted.

  Based on the correlation amount C (k) calculated by the second correlation calculation, the AF control unit 110 calculates the second shift amount x that gives the minimum value of the correlation amount using the above-described three-point interpolation method. . Then, the second shift amount x is treated as the shift amount S.

The AF control unit 110 calculates the defocus amount df based on the shift amount S according to the following equation (6). The defocus amount df is a deviation of the current imaging plane with respect to the planned focal plane 600. The variable K in the equation is a conversion coefficient (K factor) for converting the shift amount S into the defocus amount df.
df = S × K (6)

<Description of flowchart>
FIG. 11 is a flowchart for explaining an example of camera processing executed by the MPU 106. For example, the MPU 106 repeatedly executes the process of FIG. 11 when the main switch is on. In step S101 of FIG. 11, the MPU 106 determines whether or not a release button (not shown) constituting the operation member 109 has been half-pressed. When the release button half-press operation signal is input, the MPU 106 makes an affirmative decision in step S101 and proceeds to step S102. When the half-press operation signal is not input, the MPU 106 makes a negative determination in step S101 and waits for the half-press operation.

  The process from step S102 to step S112 is mainly the AF process. In step S102, the MPU 106 sends an instruction to the AF control unit 110, starts the periodic operation of the dedicated AF element 207, and proceeds to step S103. Thereby, the dedicated AF element 207 repeats charge accumulation at a predetermined cycle. In step S103, the MPU 106 sends an instruction to the AF control unit 110, reads the accumulated signal from the dedicated AF element 207, and proceeds to step S105.

  In step S105, the MPU 106 sends an instruction to the AF control unit 110, and the low-pass filter processing described above for the a-sequence signal (A1, A2,..., An) and the b-sequence signal (B1, B2,..., Bn). The process proceeds to step S106.

  In step S106, the MPU 106 sends an instruction to the AF control unit 110 to perform the first correlation calculation process using the a-sequence signal and the b-sequence signal (filter output value P1 (j)) after the low-pass filter processing. Then, the first shift amount kp1 is calculated, and the process proceeds to step S108.

  In step S108, the MPU 106 sends an instruction to the AF control unit 110, and applies the above-described high-pass filter to the a-sequence signal (A1, A2,..., An) and the b-sequence signal (B1, B2,..., Bn). The process proceeds to step S109.

  In step S109, the MPU 106 sends an instruction to the AF control unit 110 to perform the second correlation calculation process using the a-sequence signal and the b-sequence signal (filter output value P2 (j)) after the high-pass filter processing. Then, the second shift amount x is calculated, and the process proceeds to step S110. In the second correlation calculation process, the vicinity of the first shift amount kp1 calculated by the first correlation calculation (S106) (including the first shift amount kp1 illustrated in FIG. 10 and the shift range in the first correlation calculation). In a narrow range Δd) narrower than ΔW, calculation is performed by relatively shifting the a-sequence signal and the b-sequence signal.

  In step S110, the MPU 106 sends an instruction to the AF control unit 110, calculates the defocus amount df based on the second shift amount x, and proceeds to step S111. In step S111, the MPU 106 determines whether or not the defocus amount df is equal to or less than a predetermined value. The MPU 106 determines the focus when the defocus amount df is equal to or smaller than the predetermined value, and proceeds to step S113. If the defocus amount df exceeds the predetermined value, the MPU 106 does not determine the focus and proceeds to step S112.

  In step S112, the MPU 106 instructs the photographing lens 202 to drive the lens, and the process returns to step S103. Specifically, lens control information for instructing the moving direction and moving amount of the focusing lens necessary for focusing is transmitted to the lens CPU 214.

  In step S113, the MPU 106 determines whether a release operation has been performed. When a full-press operation signal indicating a full-press operation of a release button (not shown) constituting the operation member 109 is input, the MPU 106 makes a positive determination in step S113 and proceeds to step S114. If the full-press operation signal is not input, the MPU 106 makes a negative determination in step S113 and returns to step S103.

  In step S114, the MPU 106 sends an instruction to the mirror drive mechanism 108 to drive the main mirror 203 up, and transmits an aperture drive instruction to the lens CPU 214, and proceeds to step S115. The aperture drive instruction is an instruction to drive the aperture value of the photographing lens 202 to the aperture value determined by the AE control unit 111.

  In step S115, the MPU 106 sends an instruction to the shutter drive mechanism 107, drives the mechanical shutter 205 to open for a predetermined time, performs a predetermined photographing process, and proceeds to step S116. As a result, when the aperture driving in the photographic lens 202 is completed, an imaging operation by the imaging device 102 is performed, and a recording image file is recorded on the recording medium 105 (FIG. 3).

  In step S116, the MPU 106 sends an instruction to the mirror drive mechanism 108 to drive the main mirror 203 down, and transmits an aperture drive instruction to the lens CPU 214, and the process in FIG. The aperture drive instruction is an instruction to return the aperture value of the photographic lens 202 to the full aperture value.

According to the first embodiment described above, the following operational effects can be obtained.
(1) The focus detection device receives one of a pair of light beams that have passed through the imaging optical system 210 and outputs an a-series signal, and receives the other of the pair of light beams and receives a b-series light beam. A dedicated AF element 207 having a photoelectric conversion element array B that outputs a signal, an AF control unit 110 that limits the a-sequence signal and the b-sequence signal to a low frequency band, and the a-sequence signal and the b-sequence signal. An AF control unit 110 that restricts each to a high frequency band, and an a-sequence signal and a b-sequence signal that are subjected to correlation calculation while relatively shifting an a-sequence signal and a b-sequence signal restricted to a low-frequency band. The AF control unit 110 that calculates the first shift amount kp1 corresponding to the extreme value of the first correlation amount and the correlation while relatively shifting the a-sequence signal and the b-sequence signal restricted to the high frequency band Performance When there are a plurality of extreme values of the second correlation amount and the AF control unit 110 that calculates the second shift amount x corresponding to the extreme value of the second correlation amount between the a-sequence signal and the b-sequence signal And an AF control unit 110 that detects a shift amount that substantially matches the first shift amount kp1 among the plurality of second shift amounts x as a shift amount S between the a-sequence signal and the b-sequence signal.
As a result, the advantage of using a pair of signals in the low frequency band that avoids erroneous detection of the phase difference and the advantage of using a pair of signals in the high frequency band that accurately detect the shift amount S based on the phase difference of the fine contrast. Thus, the defocus amount can be obtained with high accuracy. A camera equipped with this focus detection device improves the accuracy of focus adjustment. A camera equipped with this focus detection device can perform an accurate focus adjustment that maintains an in-focus image even when a photograph is greatly enlarged.

(2) The AF control unit 110 performs the first correlation calculation while relatively shifting the a-sequence signal and the b-sequence signal limited to the low frequency band in the shift range ΔW, and the AF control unit 110 Performs a second correlation calculation while relatively shifting the a-sequence signal and the b-sequence signal limited to the high frequency band in a narrow range Δd that is narrower than ΔW and includes the first shift amount kp1. Since the second correlation calculation range Δd is narrower than the first correlation calculation range ΔW, the amount of calculation can be reduced as compared with the case where the second correlation calculation range is not narrowed. In a camera equipped with this focus detection device, the time required for focus adjustment is shortened, and the time until the subject is focused is shortened.

(Second embodiment)
In the second embodiment, the AF control unit 110 varies the signal range (calculation range) of the pair of signals used for the image shift detection calculation process between the first correlation calculation and the second correlation calculation.

  The focus detection apparatus according to the second embodiment of the present invention is mounted on the single-lens reflex camera system illustrated in FIGS. 1 and 2 as in the case of the first embodiment. A block diagram (FIG. 3) showing the circuit configuration of the single-lens reflex camera system is also the same as that of the first embodiment.

  The AF control unit 110 sets, for example, a range 71 enclosed by a broken line in the pair of signals illustrated in FIG. 7 as a signal range (calculation range) used for the first image shift detection calculation process. The first calculation range 71 is determined according to the size of the subject to be focused.

  The AF control unit 110 performs low-pass filter processing on the pair of a-sequence signal and b-sequence signal in the first calculation range 71, and the a-sequence signal and the b-sequence signal illustrated in FIG. A filter output value P1 (j)) is obtained.

  Further, the AF control unit 110 sets, for example, a range 72 enclosed by a one-dot chain line in the pair of signals illustrated in FIG. 7 as a signal range (calculation range) used for the second image shift detection calculation process. The second calculation range 72 is set in the first calculation range 71 and is narrower than the first calculation range 71, for example, 1/3 of the calculation range 71.

  The AF control unit 110 further performs high-pass filtering on the pair of a-sequence signals and b-sequence signals in the second computation range 72, and in the computation range 72 surrounded by the one-dot chain line in FIG. And b-sequence signal (filter output value P2 (j)).

  The relationship between the shift amount k (horizontal axis) and the correlation amount C (k) obtained by the two correlation calculations is as illustrated in FIG. In general, a pair of a-sequence signals and b-sequence signals (filter output value P1 (j) (Fig. 8)) included in a wide calculation range 71 (Fig. 7) are less prone to erroneous detection of phase differences. When subjects with different phases, such as the subject and the far subject, are mixed in the calculation range 71, the minimum value in the correlation calculation is less likely to appear. In addition, a pair of a-sequence signal and b-sequence signal (filter output value P2 (j) (FIG. 9)) included in the narrow calculation range 72 (FIG. 7) makes it easy to separate and detect subjects having different phases. On the other hand, there is a possibility of erroneously detecting a phase difference that is similar by chance.

  Therefore, the AF control unit 110 performs correlation calculation (1) using a pair of a-sequence signals and b-sequence signals (filter output value P1 (j) (FIG. 8)) included in the wide calculation range 71 (FIG. 7). The first correlation calculation) is performed first to detect the first shift amount kp1 (FIG. 10) in which the degree of correlation between the pair of signals is high (that is, the correlation amount C (k) is reduced).

  Next, the AF control unit 110 includes the first shift amount kp1 detected by the first correlation calculation, and a narrow calculation range 72 (FIG. 7) in the narrow range Δd narrower than the shift range ΔW in the first correlation calculation. ) To perform a correlation operation (second correlation operation) using a pair of a-sequence signals and b-sequence signals (filter output value P2 (j) (FIG. 9)). That is, a pair of signals (filter output value P1 (j) (FIG. 8)) in a wide calculation range 71 and a pair of signals (filter output value P2 (j) (FIG. 9)) in a locally narrow calculation range 72 are obtained. The two correlation amounts are used to calculate, and a shift amount that reduces both the correlation amounts is employed.

  Based on the correlation amount C (k) calculated by the second correlation calculation, the AF control unit 110 calculates the second shift amount x that gives the minimum value of the correlation amount using the above-described three-point interpolation method. . Then, the second shift amount x is treated as the shift amount S. The process of converting the shift amount S into the defocus amount df is the same as in the first embodiment.

<Description of flowchart>
FIG. 12 is a flowchart for explaining an example of camera processing executed by the MPU 106. Processes similar to those in the flowchart (FIG. 11) of the first embodiment are denoted by the same step numbers and description thereof is omitted, and the processes in steps S104 and S107 added to the flowchart in FIG. 11 are mainly described. Explained. For example, the MPU 106 repeatedly executes the process of FIG. 12 when the main switch is on.

  In step S104 following step S103, the MPU 106 sets the first calculation range 71 (wide calculation range). Specifically, among the pair of signals illustrated in FIG. 7, the calculation range 71 corresponding to the size of the subject to be focused is instructed to the AF control unit 110, and the process proceeds to step S <b> 105.

  In step S105, the MPU 106 sends an instruction to the AF control unit 110, and the low-pass filter processing described above for the a-sequence signal (A1, A2,..., An) and the b-sequence signal (B1, B2,..., Bn). The process proceeds to step S106.

  In step S106, the MPU 106 sends an instruction to the AF control unit 110, and after the low-pass filter processing, the a-sequence signal and the b-sequence signal (filter output value P1 (j) (FIG. 8) included in the first calculation range 71. ) To perform the first correlation calculation process, calculate the first shift amount kp1, and proceed to step S107.

  In step S107, the MPU 106 sets the second calculation range 72 (narrow calculation range), and proceeds to step S108. The MPU 106 instructs the AF control unit 110 to set, for example, a calculation range 72 that is 1/3 of the calculation range 71 in the first calculation range 71. The MPU 106 selects the position of the calculation range 72 in the first calculation range 71 so that edge information corresponding to the closest subject can be obtained.

  In step S108, the MPU 106 sends an instruction to the AF control unit 110, and applies the above-described high-pass filter to the a-sequence signal (A1, A2,..., An) and the b-sequence signal (B1, B2,..., Bn). The process proceeds to step S109.

  In step S109, the MPU 106 sends an instruction to the AF control unit 110, and after the high-pass filter processing, the a-sequence signal and the b-sequence signal (filter output value P2 (j) (FIG. 9) included in the second calculation range 72. ) To perform the second correlation calculation process, calculate the second shift amount x, and proceed to step S110. In the second correlation calculation process, the vicinity of the first shift amount kp1 calculated by the first correlation calculation (S106) (including the first shift amount kp1 illustrated in FIG. 10 and the shift range in the first correlation calculation). In a narrow range Δd) narrower than ΔW, calculation is performed by relatively shifting the a-sequence signal and the b-sequence signal. The subsequent processing is the same as in the first embodiment.

According to the second embodiment described above, the following operational effects can be obtained.
(1) The focus detection device receives one of a pair of light beams that have passed through the imaging optical system 210 and outputs an a-series signal, and receives the other of the pair of light beams and receives a b-series light beam. The dedicated AF element 207 having the photoelectric conversion element array B that outputs a signal, and the a series signal and the b series signal included in the computation range 71 are subjected to correlation calculation while relatively shifting in the range ΔW, and a An AF control unit 110 that calculates the first shift amount kp1 corresponding to the extreme value of the first correlation amount between the series signal and the b-sequence signal, and a calculation range 72 that is within the calculation range 71 and is narrower than the calculation range 71. The a-sequence signal and the b-sequence signal are subjected to correlation calculation while being relatively shifted in a narrow range Δd that is narrower than the range ΔW and includes the first shift amount kp1. Of the second correlation amount with When there are a plurality of extreme values of the second correlation amount and the AF control unit 110 that calculates the second shift amount x corresponding to the shift, a shift that substantially matches the first shift amount kp1 among the plurality of second shift amounts x. And an AF control unit 110 that detects the amount as a deviation amount S between the a-sequence signal and the b-sequence signal.
Thereby, the correlation calculation for the a-sequence signal and the b-sequence signal included in the calculation range 72 narrower than the calculation range 71 may be performed only in the narrow range Δd in the vicinity of the first shift amount kp1. Compared with the case where the correlation calculation is performed over a wide ΔW, the calculation amount can be reduced. In a camera equipped with this focus detection device, the time required for focus adjustment is shortened, and the time until the subject is focused is shortened.
In addition, the advantage of using a pair of signals in a wide calculation range 71 that avoids erroneous detection of the phase difference and the deviation amount S can be accurately detected by separating subjects with different phases such as a near subject and a far subject. The defocus amount can be obtained with high accuracy in combination with the advantage of using a pair of signals in a narrow calculation range 72. A camera equipped with this focus detection device improves the accuracy of focus adjustment. A camera equipped with this focus detection device can perform an accurate focus adjustment that maintains an in-focus image even when a photograph is greatly enlarged.

(2) An AF control unit 110 that limits the a-sequence signal and the b-sequence signal to a low frequency band, and an AF control unit 110 that limits the a-sequence signal and the b-sequence signal to a high frequency band, respectively. The AF control unit 110 performs correlation calculation while relatively shifting the a-sequence signal and the b-sequence signal included in the calculation range 71 limited to the low frequency band, and the AF control unit 110 The correlation calculation is performed while relatively shifting the a-sequence signal and the b-sequence signal included in the calculation range 72 limited to the band. Accordingly, a pair of signals in an appropriate frequency band can be used in each of the correlation calculation performed in the wide calculation range 71 and the correlation calculation performed in the narrow calculation range 72.

The following modifications are also within the scope of the present invention, and one or a plurality of modifications can be combined with the above-described embodiment.
(Modification 1)
In the first embodiment, preliminary low-pass filter processing may be performed. In the preliminary low-pass filter processing, the cut-off frequency is set lower than that of the normally used low-pass filter processing. Thereby, the first modification performs three filter processes (normal low-pass filter process, high-pass filter process, and spare low-pass filter process) having different pass frequency bands.

  FIG. 13 is a diagram illustrating the relationship between the shift amount k (horizontal axis) and the correlation amount C (k) respectively obtained by two correlation operations. The correlation amount C (k) for the b-sequence signal (filter output value P1 (j) (FIG. 8)) is shown, and the lower row shows the a-sequence signal and b-sequence signal (filter output value P2 (filter output value P2 ( j) The correlation amount C (k) for (FIG. 9)). In FIG. 13, there are a plurality of extreme values in the correlation amount obtained by the correlation calculation performed on the filter output value P1 (j), and the first shift amounts kp1, kp2 correspond to the plurality of extreme values, respectively. There are several.

  In the first modification, when there are a plurality of extreme values in the correlation amount obtained by the correlation calculation performed on the filter output value P1 (j), the AF control unit 110 replaces the normal low-pass filter process. Perform preliminary low-pass filter processing. Instead of the preliminary low-pass filter processing having a cut-off frequency lower than that of the normal low-pass filter processing, only one extreme value is obtained by the correlation calculation performed on the a-sequence signal and the b-sequence signal after the filter processing. It is for doing so.

  If there is one extreme value obtained by the correlation calculation performed after the preliminary low-pass filter processing, the a-sequence signal after the high-pass filter processing is only in the narrow range Δd in the vicinity of the first shift amount kp1 corresponding to this extreme value. Since the correlation calculation may be performed on the b-series signals, the calculation amount can be reduced as compared with the case where the correlation calculation is performed a plurality of times in each of the plurality of narrow ranges Δd.

  According to the first modification, the AF control unit 110 lowers the low-frequency band of the preliminary low-pass filter process when there are a plurality of extreme values of the correlation amount obtained by the correlation calculation performed after the normal low-pass filter process. Since the frequency is lowered, the extreme value of the correlation amount can be reduced to one.

  Further, according to the first modification, the AF control unit 110 reduces the low frequency band to a lower frequency until the extreme value of the correlation amount obtained by the correlation calculation after the preliminary low-pass filter processing becomes one. Therefore, the extreme value of the correlation amount can be surely reduced to one.

(Modification 2)
In the second embodiment, the high-pass filter process in step S108 of FIG. 12 may be omitted. That is, image shift detection is performed using a pair of signals in the same frequency band, and the signal range (calculation range) used for the image shift detection calculation process is made different between the first correlation calculation and the second correlation calculation. .

  In the second modification, the MPU 106 omits the process of step S108 (FIG. 12) in order to perform image shift detection using a pair of signals in the same frequency band. In step S109 (FIG. 12), the MPU 106 sends an instruction to the AF control unit 110, and among the a-sequence signal and b-sequence signal (filter output value P1 (j) (FIG. 8)) after low-pass filter processing, The second correlation calculation process is performed using the signal in the second calculation range 72, the second shift amount x is calculated, and the process proceeds to step S110. In the second correlation calculation process, the vicinity of the first shift amount kp1 calculated by the first correlation calculation (S106) (including the first shift amount kp1 and narrower than the shift range ΔW in the first correlation calculation). In the narrow range Δd), the calculation is performed by relatively shifting the a-sequence signal and the b-sequence signal as in the case of the first embodiment.

(Modification 3)
In the second embodiment, the MPU 106 sets the second calculation range 72 to 1/3 of the first calculation range 71 (S107). However, the second calculation range 72 is the first calculation range 72. It is not limited to 1/3 of 71, and may be 1/2 or 1/5. Further, the MPU 106 makes the calculation range 72 when the detected deviation amount S is larger than the predetermined value (that is, the defocus amount is large) wider than the calculation range 72 when the detected deviation amount S is smaller than the predetermined value. Also good. By making the calculation range 72 variable in this way, the range of the calculation range 72 can be appropriately adjusted.

  Furthermore, the MPU 106 may set the position of the second calculation range 72 set in the first calculation range 71 at the left end of the calculation range 71 or may be set at the center of the calculation range 71. Alternatively, it may be set at the right end of the calculation range 71. When subjects with different phases, such as a near subject and a far subject, coexist in the first calculation range 71, it is difficult for a minimum value to appear in the correlation calculation for the calculation range 71. However, by changing the position of the second calculation range 72 that is locally performed, it is possible to avoid a state in which subjects having different phases are mixed in the second calculation range 72, so that a minimum value is obtained in the correlation calculation for the calculation range 72. Becomes easier to appear. Thereby, the second shift amount x that gives the minimum value of the correlation amount can be appropriately obtained.

(Modification 4)
In the first embodiment or the second embodiment, as illustrated in FIG. 13, there are a plurality of extreme values in the correlation amount obtained by the first correlation calculation, and each corresponds to the plurality of extreme values. Then, when there are a plurality of first shift amounts kp1, kp2, the second and subsequent correlation calculations may be performed in a narrow range Δd in the vicinity of the plurality of first shift amounts kp1, kp2. At this time, the AF control unit 110 selects, for example, a first shift amount (for example, kp1) corresponding to a smaller defocus amount from the plurality of first shift amounts kp1, kp2, and a narrow range in the vicinity of this kp1. At Δd, the second correlation calculation is performed first. When the second shift amount x1 calculated by the second correlation calculation substantially coincides with the first shift amount kp1, the AF control unit 110 sets the second shift amount x1 as the shift amount S and then performs another first shift. The second correlation calculation in the narrow range Δd near the quantity kp2 is omitted.

  When the second shift amount x1 calculated by the second correlation calculation does not match the first shift amount kp1, the AF control unit 110 performs the second correlation calculation in the narrow range Δd near the other first shift amount kp2. Do. Then, the AF control unit 110 sets the second shift amount x2 as the shift amount S when the second shift amount x2 calculated by the second correlation calculation substantially coincides with the first shift amount kp2.

  According to the fourth modification to the first embodiment, the AF control unit 110 includes a plurality of first shift amounts when there are a plurality of extreme values of the correlation amount obtained by the correlation calculation performed after the normal low-pass filter processing. Since correlation calculation is performed in each of a plurality of narrow ranges Δd including kp1 and kp2, shifts that substantially match between the first shift amount and the second shift amount for each of the plurality of first shift amounts kp1 and kp2. The amount can be detected appropriately.

  Further, according to the fourth modification to the first embodiment, the AF control unit 110 has a smaller first value when there are a plurality of extreme values of the correlation amount obtained by the correlation calculation performed after the normal low-pass filter processing. For the narrow range Δd including the shift amount (for example, kp1), the second correlation calculation is performed first, and when the calculated second shift amount x substantially coincides with the first shift amount kp1, the other first shift amount kp2 is calculated. Since the second correlation calculation for the included narrow range Δd is omitted, it is possible to quickly perform focus adjustment on a subject with a small first shift amount, that is, a small defocus amount. Further, by omitting the second correlation calculation which is unnecessary when the calculated second shift amount x substantially matches the first shift amount kp1, it is useful for reducing the calculation amount.

  Instead of selecting the first shift amount (for example, kp1) corresponding to the minimum defocus amount and performing the second correlation calculation first in the narrow range Δd in the vicinity of this kp1, it corresponds to the subject closest to the camera. A first shift amount (for example, kp1) may be selected, and the second correlation calculation may be performed first in a narrow range Δd near kp1.

  According to the fourth modification to the second embodiment, the AF control unit 110 responds to a subject closer to the camera when there are a plurality of extreme values of the correlation amount obtained by the correlation calculation performed after the normal low-pass filter processing. For the narrow range Δd including the first shift amount (for example, kp1) to be performed, the second correlation calculation is performed first, and when the calculated second shift amount x substantially matches the first shift amount kp1, it corresponds to another subject. Since the second correlation calculation for the narrow range Δd including the first shift amount kp2 to be performed is omitted, the focus adjustment can be quickly performed on the subject closest to the camera. Further, by omitting the second correlation calculation which is unnecessary when the calculated second shift amount x substantially matches the first shift amount kp1, it is useful for reducing the calculation amount.

(Modification 5)
In the above description, the correlation calculation (first correlation calculation) using (filter output value P1 (j)) is first performed on the pair of a-sequence signal and b-sequence signal after the low-pass filter processing, and the high-pass filter A correlation calculation (second correlation calculation) using the paired a-sequence signal and b-sequence signal (filter output value P2 (j)) after processing was performed later. Two correlation calculations may be performed simultaneously. Correlation calculation is performed in parallel, and two correlation amounts are obtained using a pair of low frequency band signals (filter output value P1 (j)) and a pair of high frequency band signals (filter output value P2 (j)). It may be calculated. The shift amount k may be calculated by detecting the extreme coincidence of the correlation amounts.

(Modification 6)
In the above description, a digital camera is exemplified as a single-lens reflex camera system. However, any single-lens reflex type can be used as long as it includes a focus detection device that performs phase difference type focus detection processing using a detection signal from the dedicated AF element 207. A silver salt camera may be used.

(Modification 7)
In the above description, a single-lens reflex digital camera is illustrated, but a mirrorless digital camera that does not include a quick return mirror may be used. For example, as disclosed in Japanese Patent Application Laid-Open No. 2007-282107, by using an imaging element that includes a focus detection pixel that receives a pair of light beams for focus detection in addition to the imaging pixel, a so-called imaging surface position is obtained. It may be a camera having a focus detection device that performs phase difference type focus detection processing.

(Modification 8)
In addition, a focus detection apparatus that includes an imaging element having imaging pixels and focus detection pixels and performs the imaging surface phase difference type focus detection processing may be mounted on a portable terminal. The portable terminal can be configured as an electronic device such as a high-function mobile phone (so-called smartphone), a tablet terminal, or a wearable terminal.

(Modification 9)
FIG. 14 is a diagram for explaining program supply. A program for executing the AF processing described with reference to FIGS. 11 and 12 may be supplied to the camera or the portable terminal 700 later. The program supply to the portable terminal 700 or the like can be transmitted from the personal computer 705 storing the program to the portable terminal 700 by infrared communication, short-range wireless communication, or wired connection, as illustrated in FIG. .

  The program may be supplied to the personal computer 705 by setting a storage medium 704 such as a CD-ROM storing the program in the personal computer 705, or to the personal computer 705 by a method via a communication line 701 such as a network. You may load. When passing through the communication line 701, the program is stored in the storage device 703 of the server 702 connected to the communication line.

  Further, the program can be directly transmitted to the portable terminal 700 via a wireless LAN access point (not shown) connected to the communication line 701. Further, a storage medium 704B such as a memory card storing the program may be set in the portable terminal 700. As described above, the program can be supplied as various types of computer program products such as provision via a storage medium or a communication line (signal).

  Although various embodiments and modifications have been described above, the present invention is not limited to these contents. Other embodiments conceivable within the scope of the technical idea of the present invention are also included in the scope of the present invention.

71, 72 ... Calculation range 102 ... Image sensor 106 ... MPU
DESCRIPTION OF SYMBOLS 107 ... Shutter drive mechanism 108 ... Mirror drive mechanism 109 ... Operation member 110 ... AF control part 201 ... Camera body 202 ... Shooting lens 203 ... Main mirror 207 ... Dedicated AF element 210 ... Imaging optical system 211 ... Diaphragm 212 ... Lens drive mechanism 214 ... Lens CPU
A, B: photoelectric conversion element array ΔW: first correlation calculation shift range Δd: second correlation calculation shift range

Claims (9)

Generating a first first photoelectric conversion unit that generates an electric charge by photoelectric conversion of the light beam, and the first light flux is different from the second charge by photoelectrically converting a light flux passing through the optical system passing through the optical system a second have a photoelectric conversion unit, the first signal based on the electric charge produced in the first photoelectric conversion unit, and an imaging unit for outputting a second signal based on the electric charge produced in the second photoelectric conversion unit to ,
In the first signal and the second first range signal, after the first correlation calculation was Tsu line using a first frequency band, a narrow second range than the first range, the first frequency A calculation for performing a second correlation calculation using a second frequency band higher than the band, and calculating a deviation amount between the image by the optical system and the imaging unit based on the first correlation calculation and the second correlation calculation And
A focus detection apparatus. The focus detection apparatus according to claim 1,
The calculation unit includes a first correlation amount of the first signal and the second signal based on the first correlation calculation, and a second correlation amount of the first signal and the second signal based on the second correlation calculation. A focus detection device that calculates the shift amount from The focus detection apparatus according to claim 2,
The calculation unit performs a first shift corresponding to an extreme value of the first correlation amount by the first correlation calculation while relatively shifting in a predetermined range of the first signal and the second signal included in the first range. A shift amount is calculated , and the first signal and the second signal included in the second range within the first range are narrower than the predetermined range and are relative to each other in a narrow range including the first shift amount. When a second shift amount corresponding to the extreme value of the second correlation amount is calculated by the second correlation calculation while shifting to, and there are a plurality of extreme values of the second correlation amount, a plurality of the second shifts are calculated. the shift amount substantially coincides with the first shift amount of the amount, the focus detection device is calculated as the amount of deviation. The focus detection apparatus according to claim 3 ,
The focus detection device, wherein the first range and the second range are ranges in the shift direction. The focus detection apparatus according to claim 3 or 4 ,
The calculation unit performs the first correlation calculation while relatively shifting the first signal and the second signal limited to the first frequency band, and the first unit limited to the second frequency band. A focus detection apparatus that performs the second correlation calculation while relatively shifting a signal and the second signal. In the focus detection apparatus according to any one of claims 3 to 5 ,
The calculation unit is a focus detection device that performs the first correlation calculation in a plurality of the narrow ranges each including a plurality of the first shift amounts when there are a plurality of extreme values of the first correlation amount. The focus detection apparatus according to claim 6 ,
The calculating unit, when the extreme value of the first correlation amount there are a plurality, performs the first correlation operation on the narrow range including said first shift amount corresponding to the closer of the subject, the second shift A focus detection apparatus that does not perform the second correlation calculation for the narrow range including the first shift amount corresponding to another subject when the amount substantially matches the first shift amount. In the focus detection apparatus according to any one of claims 1 to 7 ,
The focus detection device, wherein the calculation unit makes the second range when the shift amount is larger than a predetermined value wider than the second range when the calculated shift amount is smaller than the predetermined value. An imaging device comprising the focus detection device according to any one of claims 1 to 8 .

Images