JP2009063921A - Imaging apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To reduce influence by turn-back of an image signal so as to perform accurate autofocus (AF). <P>SOLUTION: The imaging apparatus includes an imaging device 102 having a first photoelectric conversion cell group 201 photoelectrically converting a subject image formed with luminous flux from an imaging optical system 101 and outputting a first signal used to form an image, and second photoelectric conversion cell groups 202 and 203 photoelectrically converting a plurality of images formed with divided luminous fluxes out of the luminous flux from the imaging optical system and outputting a second signal used to detect a phase difference. Control means 108 and 109 perform first focus control by a contrast detection system using the first signal and second focus control by a phase difference detection system using the second signal. The control means switches the first focus control and the second focus control in accordance with distribution of the frequency component of the image with specified frequency decided based on the arrangement of the second photoelectric conversion cell group as a standard. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an imaging apparatus such as a digital camera or a video camera, and in particular, contrast detection type focus control using an image generation signal from an image sensor and phase difference detection using a phase difference detection signal from the image sensor. The present invention relates to an imaging apparatus that performs system focus control.

  There are a contrast detection method (refer to Patent Document 1) and a phase difference detection method (refer to Patent Document 2) as an autofocus (AF) method of the imaging apparatus. Furthermore, a so-called hybrid system (see Patent Document 3) in which these contrast detection system and phase difference detection system are combined is also known.

  In contrast detection AF (hereinafter referred to as contrast AF), a high frequency component is extracted from an image signal obtained by using an image sensor, and a contrast evaluation value is generated from the high frequency component. Then, the focus lens position corresponding to the maximum contrast evaluation value among the plurality of contrast evaluation values obtained while moving the focus lens in the optical axis direction is determined as the focus position, and the focus lens is moved to this focus position. Let

  In the AF of the phase difference detection method (hereinafter referred to as phase difference AF), the defocus amount of the imaging lens is determined from the shift amount (phase difference) between the two images formed by the two light beams that have passed through different pupil regions of the imaging lens. Is calculated. And then. A focus position is obtained from the defocus amount, and the focus lens is moved to this focus position.

  Here, in the phase difference AF disclosed in Patent Document 2, the pupil division optical system is eccentrically arranged in a part of photoelectric conversion cells (phase difference detection pixels) constituting the image pickup device for image acquisition, and adjacent to each other. The phase difference is detected by changing the eccentric direction of the pupil division optical system of the phase difference detection pixel.

Further, in the hybrid AF disclosed in Patent Document 3 (hereinafter referred to as hybrid AF), the presence or absence of periodicity of the subject is determined during the execution of phase difference AF, and when there is periodicity, the AF is switched to contrast AF. .
JP 2000-258681 A JP 2000-156823 A JP 2006-301150 A

  In the image sensor disclosed in Patent Document 2, phase difference detection pixels are discretely arranged in a pixel group for image acquisition. When the phase difference is detected using the phase difference detection pixels arranged in this way, so-called folding of the image signal may occur due to the arrangement (interval). The aliasing of the image signal causes a phase difference detection error and may reduce the accuracy of the phase difference AF.

  The present invention provides an imaging apparatus capable of performing highly accurate AF with less influence of aliasing of an image signal.

  An imaging apparatus according to an aspect of the present invention includes a first photoelectric conversion cell group that photoelectrically converts a subject image formed by a light beam from an imaging optical system and outputs a first signal used for generating an image, and An imaging device having a second photoelectric conversion cell group that photoelectrically converts a plurality of images formed by split light beams out of light beams from an imaging optical system and outputs a second signal used for phase difference detection. Prepare. The control means performs first focus control by a contrast detection method using the first signal and second focus control by a phase difference detection method using the second signal. Then, the control means switches between the first focus control and the second focus control according to the distribution of the frequency components of the image based on the specific frequency determined based on the arrangement of the second photoelectric conversion cell group. Features.

  The control method according to one aspect of the present invention includes a first photoelectric conversion cell group that photoelectrically converts a subject image formed by a light beam from an imaging optical system and outputs a first signal used for generating an image. And a second photoelectric conversion cell group that photoelectrically converts a plurality of images formed by the divided light beams out of the light beams from the imaging optical system and outputs a second signal used for phase difference detection The present invention is applied to an imaging device provided with an element. The control method includes a step of performing a first focus control by a contrast detection method using a first signal, a step of performing a second focus control by a phase difference detection method using a second signal, And a step of switching between the first focus control and the second focus control according to the distribution of the frequency components of the image based on the specific frequency determined based on the arrangement of the photoelectric conversion cell groups.

  According to the present invention, the phase difference AF and the contrast AF are selected according to the distribution of the frequency components of the image based on the specific frequency determined based on the arrangement of the second photoelectric conversion cell group used for the phase difference detection. . For this reason, it is possible to perform high-precision AF at a high speed with little influence of aliasing of the image signal.

  Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 shows the configuration of a digital camera as an image pickup apparatus that is an embodiment of the present invention.

  The imaging lens (imaging optical system) 101 includes a focus lens, a variable power lens, a diaphragm, and the like (not shown). The light beam from the subject that has passed through the imaging lens 101 enters the imaging element 102. The image sensor 102 is a charge storage type image sensor, and is constituted by a CMOS sensor or a CCD sensor.

  The imaging element 102 includes a first photoelectric conversion cell group that photoelectrically converts a subject image formed by the light flux from the imaging lens 101 and outputs a signal for image generation (first signal: hereinafter referred to as an imaging signal). Have. Further, the image sensor 102 photoelectrically converts a plurality of images formed by the divided light beams out of the light beams from the imaging lens 101, and detects a phase difference detection signal (second signal: hereinafter referred to as a phase difference image signal). ) To output a second photoelectric conversion cell group. The second photoelectric conversion cell group is discretely arranged in the first photoelectric conversion cell group. Hereinafter, the first photoelectric conversion cell is referred to as an imaging pixel, and the second photoelectric conversion cell is referred to as a phase difference detection pixel. Specific arrangement and configuration of the imaging pixels and the phase difference detection pixels will be described later.

  The A / D conversion unit 103 converts an analog signal composed of an imaging signal and a phase difference image signal input from the imaging element 102 into a digital signal.

  Reference numeral 112 denotes a contrast focus detection unit, which extracts a high frequency component from an imaging signal and a phase difference image signal (hereinafter referred to as a mixed signal) converted into a digital signal by the A / D conversion unit 103, and extracts the high frequency component from the high frequency component. A contrast evaluation value signal indicating the contrast state of the subject image (image signal) is generated. The contrast focus detection unit 112 performs AF (first focus control: hereinafter referred to as contrast AF) of contrast detection method together with an AF method switching unit 109 described later.

  In this embodiment, a case where a mixed signal is input to the contrast focus detection unit 112 will be described. However, only an imaging signal extracted from the mixed signal may be input.

  The contrast focus detection unit 112 includes a high-pass filter (HPF) 104, an integration circuit 105, a peak hold circuit 106, and a contrast evaluation unit 114.

  The HPF 104 extracts a high frequency component from the mixed signal in the predetermined contrast evaluation area in the subject image. The integration circuit 105 integrates the high frequency component extracted by the HPF 104 to generate a contrast evaluation value signal (also referred to as an AF evaluation value signal). The peak hold circuit 106 detects and holds the peak value of the contrast evaluation value signal from the HPF 104. The contrast evaluation unit 114 reads the contrast evaluation value signal and the peak value output from the integration circuit 105 and the peak hold circuit 106, respectively. Then, the focusing direction of the imaging lens 101 (focus lens) is determined from the magnitude relationship of the contrast evaluation value with respect to the peak value. Then, a contrast AF control signal based on the determination result is output to the AF method switching unit 109. The specific operation of the contrast evaluation unit 114 will be described later.

  Reference numeral 115 denotes a phase difference focus detection unit, which calculates the defocus amount of the imaging lens 101 from the phase difference image signal extracted from the mixed signal, and determines the drive direction and drive amount of the focus lens up to the in-focus position. The phase difference focus detection unit 115 performs phase difference detection AF (second focus control: hereinafter referred to as phase difference AF) together with the AF method switching unit 109.

  The phase difference image signal used for the phase difference AF and the imaging signal used for the contrast AF have different frequency bands. In other words, the phase difference AF and the contrast AF are performed using signals in different frequency bands among signals obtained from the same image sensor 102.

  The phase difference focus detection unit 115 includes a phase difference image signal extraction unit 113, a defocus amount calculation unit 107, and a phase difference evaluation unit 111.

  The phase difference image signal is extracted from the phase difference image signal extraction unit 113 and the mixed signal, and is output to the defocus amount calculation unit 107. The defocus amount calculation unit 107 calculates the defocus amount from the phase difference image signal, and outputs the defocus amount to the subject determination unit 108 and the phase difference evaluation unit 111.

  The phase difference evaluation unit 111 calculates the drive direction and drive amount of the focus lens up to the in-focus position based on the defocus amount. Then, a phase difference AF control signal based on the calculation result is output to the AF method switching unit 109.

  An object determination unit 108 obtains a cutoff frequency (specific frequency) fp / 2 from a frequency (arrangement frequency) fp determined based on the arrangement of the phase difference detection pixels, and an object based on the cutoff frequency fp / 2. Calculate the frequency distribution of the image. Since the frequency fp is determined based on the arrangement of the phase difference detection pixels, it can be said that the cutoff frequency fp / 2 is also determined based on the arrangement of the phase difference detection pixels. Then, the subject determination unit 108 selects an AF method to be executed from among contrast AF and phase difference AF according to the frequency distribution.

  The frequency fp, the frequency distribution, and the AF method selection method corresponding to the frequency distribution will be described later.

  The subject determination unit 108 includes a frequency calculation unit 117, a frequency evaluation unit 118, and a scene evaluation unit 119. The frequency calculation unit 117 obtains a distribution of frequency components of the subject image with reference to the cutoff frequency fp / 2. Specifically, the imaging signal extracted from the mixed signal is divided into a low frequency component and a high frequency component with the cutoff frequency fp / 2 as a boundary, and the amount of the low frequency component and the high frequency component is calculated by integrating each of them. To do. Then, the integration result is output to the frequency evaluation unit 118. Note that the “frequency distribution of the subject image” can also be referred to as the frequency distribution of the imaging signal.

  The frequency evaluation unit 118 compares the amount of the low frequency component and the amount of the high frequency component of the subject image (that is, the subject) obtained by the integration. Note that a ratio between the low frequency component amount and the high frequency component amount may be obtained, and it may be determined whether or not the ratio is larger than a predetermined value. Furthermore, the frequency evaluation unit 118 determines whether or not contrast AF is possible for each of the low frequency component and the high frequency component. Then, the comparison result and the determination result of whether or not contrast AF is possible are output to the scene evaluation unit 119.

  The scene evaluation unit 119 selects an AF method to be executed from among the contrast AF and the phase difference AF based on the comparison and determination result from the subject determination unit 108 and the defocus amount from the defocus amount calculation unit 107. Then, an AF selection signal indicating the selection result is output to the AF method switching unit 109.

  In this embodiment, a case where a mixed signal is input to the subject determination unit 108 will be described. However, only an imaging signal extracted from the mixed signal may be input.

  The AF method switching unit 109 outputs the contrast AF control signal from the contrast evaluation unit 114 to the motor drive control unit 110 when the selection result of the subject determination unit 108 is contrast AF. When the selection result is phase difference AF, the control signal for phase difference AF from the phase difference evaluation unit 111 is output to the motor drive control unit 110.

  The subject determination unit 108 and the AF method switching unit 109 described above constitute a control unit.

  The motor drive control unit 110 outputs a drive signal to the motor 116 as an actuator that moves the focus lens in accordance with the contrast AF control signal or the phase difference AF control signal. The motor 116 operates to move the focus lens in accordance with the drive signal. When the motor 116 is a stepping motor, the drive signal from the motor drive control unit 110 is a pulse signal, and a pulse signal having the number of pulses corresponding to each control signal is given to the motor 116.

  Next, the arrangement of the imaging pixels (group) and the phase difference detection pixels (group) on the imaging element 102 and the frequency fp (cutoff frequency fp / 2) determined based on the arrangement will be described with reference to FIG. .

  FIG. 2 shows an enlarged part of the image sensor 102. Each white portion 201 surrounded by a rectangular frame indicates an imaging pixel. In addition, portions 202 and 203 surrounded by a rectangular frame and hatched on the inside thereof indicate phase difference detection pixels. The phase difference detection pixels form one pair with two pixels obliquely adjacent to each other, and the pairs are discretely arranged in the image sensor 102 (imaging pixel group). The pairs may be periodically arranged in the image sensor 102 (imaging pixel group).

  Here, the interval (pitch) between the imaging pixels 201 is d. Further, the interval (pitch) between the pair of phase difference detection pixels 202 and 203 and the pair of phase difference detection pixels 202 and 203 closest thereto is set to 10 × d. FIG. 2 shows a case where the number of phase difference detection pixels 202 and 203 is 2% of the total number of pixels on the image sensor 102. However, this ratio is only an example, and fewer or more phase difference detection pixels may be provided.

  When the phase difference detection pixels 202 and 203 are arranged at the ratio shown in FIG. 2, the frequency corresponding to the arrangement (pitch) of the plurality of imaging pixels 201 (imaging pixel group) is 1 / d (hereinafter referred to as fs). ). For this reason, the frequency band which satisfies the sampling theorem regarding the contrast evaluation value obtained using the imaging pixel 201 can be expressed as 0 to fs / 2.

  The frequency fp corresponding to the arrangement (pitch) of a plurality of pairs of phase difference detection pixels 202 and 203 (collectively referred to as a phase difference detection pixel group) is 1/10 × d (= fs / 10 = fp). It is. Therefore, the frequency band that satisfies the sampling theorem for the phase difference image signal obtained using the phase difference detection pixels 202 and 203 can be expressed as 0 to fs / 20 (= fp / 2).

  3 is a front view of the phase difference detection pixel 202, and the lower side of FIG. 3 is a cross-sectional view of the phase difference detection pixel 202 taken along line B-B 'in the front view. The phase difference detection pixel 202 does not have any of the R, G, and B color filter layers of the imaging pixel 201, and the microlens 301 is provided at the position closest to the light incident side. In the front view, the microlens 301 is not shown. Reference numeral 302 denotes a smoothing layer for forming a plane for forming the microlens 301.

  Reference numeral 303 a denotes a light shielding layer, which has a diaphragm opening eccentric in one direction with respect to the center O of the photoelectric conversion region 305. 304 is a smooth layer.

  4 is a front view of the phase difference detection pixel 203, and the lower side of FIG. 4 is a sectional view of the phase difference detection pixel 203 when cut along the line BB ′ in the front view. Yes. This phase difference detection pixel 203 also has no color filter layer, and a microlens 301 is provided at a position closest to the light incident side (on the smoothing layer 302). In the front view, the microlens 301 is not shown.

  Reference numeral 303b denotes a light shielding layer, which has an aperture opening that is decentered in the opposite direction to the light shielding layer 303a provided in the phase difference detection pixel 202 with respect to the center O of the photoelectric conversion region 305. That is, the light shielding layers 303 a and 303 b of the phase difference detection pixels 202 and 203 have diaphragm openings at symmetrical positions with the optical axis (O) of each microlens 301 interposed therebetween.

  According to such a configuration, when the imaging lens 101 is viewed from the phase difference detection pixel 202 and when viewed from the phase difference detection pixel 203, this is equivalent to the fact that the pupil of the imaging lens 101 is divided symmetrically. . That is, the microlens 301 and the light shielding layers 303a and 303b constitute a so-called pupil division optical system.

  Two more approximate images are formed on the phase difference detection pixels 202 and 203 adjacent to each other as the number of pixels of the image sensor 102 increases. In a state where the imaging lens 101 is in focus with respect to the subject, outputs (phase difference image signals) obtained from the phase difference detection pixels 202 and 203 coincide with each other.

  On the other hand, if the imaging lens 101 is out of focus, a phase difference occurs in the phase difference image signals obtained from the phase difference detection pixels 202 and 203. The direction of the phase difference is reversed between the front pin state and the rear pin state.

  5 and 6 show the relationship between the focus state and the phase difference. In these figures, the phase difference detection pixels 202 and 203 are shown as S1 and S2, respectively. Further, the imaging pixel 201 is omitted, and a plurality of pairs of phase difference detection pixels 202 and 203 are shown close to each other on the same cross section. In the drawing, the pair of phase difference detection pixels S1 and S2 is shown as being covered with the same microlens 301, but the phase difference detection pixels S1 and S2 may be covered with different microlenses. Good.

  Light from a specific point of the subject passes through the pupil for the phase difference detection pixel S1 and enters the phase difference detection pixel S1, and light flux L2 enters the phase difference detection pixel S2 through the pupil for the phase difference detection pixel S2. And divided. The two light beams L1 and L2 are collected at one point on the surface of the microlens 301 ° as shown in FIG. 5 when the imaging lens 101 is in focus at a specific point. The same image is formed on the phase difference detection pixels S1 and S2. Thereby, the phase difference image signal read from the phase difference detection pixel S1 and the phase difference image signal read from the phase difference detection pixel S2 are the same.

  On the other hand, when the imaging lens 101 is not focused on a specific point, the light beams L1 and L2 intersect at a position different from the surface of the microlens 301 as shown in FIG. The distance between the surface of the microlens 301 and the intersection of the two light beams L1 and L2, that is, the defocus amount is x. Further, it is assumed that the image shift amount (phase difference) on the phase difference detection pixels S1 and S2 generated in this state corresponds to n pixels.

When the pixel pitch is d, the distance between the centers of gravity of the two pupils is Daf, and the distance from the principal point of the imaging lens 101 to the focal position is u, the defocus amount x is
x = n × d × u / Daf (1)
Is required.

Since u is considered to be approximately equal to the focal length f of the imaging lens 101, the defocus amount x is
x = n × d × f / Daf (2)
You can also ask for it.

  FIG. 7 shows phase difference image signals read from each of the phase difference detection pixels S1 and S2. A shift amount (that is, an image shift amount) n × d is generated in these phase difference image signals. For this reason, the defocus amount x can be obtained from the equation (1) or (2) by obtaining the image shift amount n × d. If the focus lens in the imaging lens 101 is moved by an amount that keeps the defocus amount x within a predetermined range (preferably so as to be zero), a focused state by phase difference AF can be obtained.

  As described above, in this embodiment, the light beams L1 and L2 passing through different pupils among the light beams from the imaging lens 101 are divided by the pupil division optical system so that two images formed by the light beams L1 and L2 are received. By providing the phase difference detection pixels S1 and S2, phase difference AF is possible.

  As described above, the phase difference AF is performed by inputting the mixed signal output from the image sensor 102 to the phase difference image signal extraction unit 113 via the A / D conversion unit 103. The phase difference image signal extraction unit 113 extracts two phase difference image signals from the phase difference detection pixels S <b> 1 and S <b> 2 from the mixed signal and outputs them to the defocus amount calculation unit 107.

  The defocus amount calculation unit 107 performs a correlation operation between the two phase difference image signals, and calculates a deviation amount (phase difference) between the two phase difference image signals. Further, the defocus amount calculation unit 107 calculates a defocus amount based on the shift amount and outputs the defocus amount to the phase difference evaluation unit 111. The phase difference evaluation unit 111 generates a phase difference AF control signal based on the defocus amount and outputs it to the AF method switching unit 109.

  Next, a specific operation of the contrast evaluation unit 114 will be described. The contrast evaluation unit 114 moves the focus lens in a direction in which the contrast evaluation value from the integration circuit 105 is larger than the peak value held by the peak hold circuit 106, thereby determining the focus lens position where the contrast evaluation value is maximized. Explore.

  FIG. 8A shows an example of driving the focus lens according to the contrast evaluation value. In this figure, the horizontal axis indicates the position of the focus lens, and the vertical axis indicates the contrast evaluation value.

  For example, if the focus lens is moved from the infinity position to the close direction, the contrast evaluation value from the integration circuit 105 continues to be larger than the peak value held by the peak hold circuit 106, and the contrast evaluation value becomes maximum. It approaches the focal position P. The contrast evaluation unit 114 switches the contrast AF control signal so that the focus lens is moved in a direction in which the contrast evaluation value from the integration circuit 105 becomes larger than the peak value held by the peak hold circuit 106 in this way. Output to the unit 109.

  When the focus lens is further moved in the closest direction, the contrast evaluation value from the integration circuit 105 becomes smaller than the peak value held by the peak hold circuit 106. By reversing the magnitude relationship between the peak value and the contrast evaluation value, it is detected that the focus lens has passed the in-focus position. In this case, the contrast evaluation unit 114 sends the contrast AF control signal to the AF method switching unit 109 so as to move the focus lens in the opposite direction (infinity side) and return it to the in-focus position (or the vicinity thereof). Output to. Thereby, a focused state by contrast AF is obtained.

  When the in-focus state is obtained, a contrast AF control signal is output to the AF method switching unit 109 so as to stop the focus lens.

  Next, the subject determination unit 108 will be described with reference to FIG. Reference numeral 117 denotes the above-described frequency calculation unit, and reference numeral 118 denotes the above-described frequency evaluation unit.

  In the frequency calculation unit 117, 901 is a high-pass filter (HPF) having a cutoff frequency of fp / 2 (= fs / 20), and extracts a high-frequency component from the mixed signal from the A / D conversion unit 103.

  Reference numeral 903 denotes an integration circuit for integrating the high frequency component extracted by the HPF 901. The calculation result H of the integration circuit 903 is a value indicating the amount of the high frequency component that the subject (subject image or imaging signal) has.

  Reference numeral 902 denotes a low-pass filter (LPF) having a cutoff frequency of fp / 2, and extracts a low frequency component from the mixed signal from the A / D conversion unit 103.

  Reference numeral 904 denotes an integration circuit that integrates the low frequency components extracted by the LPF 902. The calculation result L of the integration circuit 904 is a value indicating the amount of the low frequency component of the subject (subject image or imaging signal).

  The frequency fp (or cut-off frequency fp / 2) determined based on the arrangement of the phase difference detection pixels is a fixed value given in advance by a microcomputer such as a CPU that constitutes the subject determination unit 108. However, the frequency fp (or the cut-off frequency fp / 2) may be a variable value.

  The low frequency band evaluation unit 905 compares the value of a value (hereinafter referred to as a corrected low frequency component amount) kL obtained by multiplying the low frequency component amount L by a correction coefficient k for weighting and the determination reference value A, thereby Evaluate the low frequency components of the subject.

  Here, the determination reference value A is a reference value for determining whether or not a correct contrast evaluation value is obtained by the low frequency component of the subject, that is, whether or not contrast AF is possible. That is, the low frequency band evaluation unit 905 determines whether or not contrast AF is possible based on the low frequency component of the subject.

  The determination reference value A is determined based on parameters such as sensitivity of the image sensor 102.

  The distribution evaluation unit 906 evaluates the distribution of frequency components of the subject by comparing the magnitude of the corrected low frequency component amount kL and the high frequency component amount H. In other words, a high frequency component amount H higher than the cutoff frequency fp / 2 is a low frequency component amount lower than the cutoff frequency fp / 2 (corrected low frequency component amount kL: hereinafter, simply referred to as a low frequency component amount kL). ) Or less than. In other words, it is determined whether the frequency component higher than the cutoff frequency is larger or smaller than the frequency component lower than the cutoff frequency. Thereby, it can be determined whether or not the subject has a low frequency component (a low frequency component amount kL larger than the high frequency component amount H) that can obtain a phase difference image signal (that is, a defocus amount) in the phase difference AF. .

  The high frequency band evaluation unit 907 evaluates the high frequency component of the subject by comparing the amount of the high frequency component amount H with the determination reference value A. That is, it is determined whether or not a correct contrast evaluation value can be obtained by the high frequency component of the subject, that is, whether or not contrast AF is possible.

  The scene evaluation unit 119 determines the subject based on the evaluation (determination) results of the low frequency band evaluation unit 905, the high frequency band evaluation unit 907, and the distribution evaluation unit 906 and the defocus amount obtained from the defocus amount calculation unit 107. Determine. Then, the AF method to be executed is determined based on the subject determination result.

  Note that k can be set arbitrarily, but in the following description, k = 1.

  Next, the subject determination performed by the scene evaluation unit 119 and the determination of the AF method corresponding thereto will be described with reference to FIGS. 11A to 11I and FIG.

  FIG. 11A shows conditions for classifying subjects by frequency distribution. The frequency distribution is classified into three categories a to c depending on the magnitude relationship among the judgment reference value A, the low frequency component amount kL, and the high frequency component amount H.

  In FIG. 11B to FIG. 11I, the frequency distribution of the subject image (imaging signal) is shown as an output (Power) obtained by integrating the horizontal axis with the frequency and the vertical axis with the integration circuits 903 and 904. Further, these drawings show frequency distributions in which no aliasing occurs when sampling is performed at the frequency fs.

  A low-frequency component amount L whose frequency is lower than fp / 2 is indicated by a region where horizontal hatching is performed, and a high-frequency component amount H whose frequency is higher than fp / 2 is indicated by a region where vertical hatching is performed. Moreover, the area | region which performed the diagonal hatching of FIG. 11C shows the criteria value A converted into an area. The boundary line of the determination reference value A region is indicated by a broken line in FIGS. 11B and 11D to 11I.

  In the case of being classified as a, the high frequency component amount H of the subject is larger than the judgment reference value A and the low frequency component amount kL. When it is classified as a, it can be determined that the imaging lens 101 is located in the vicinity of the in-focus position. The frequency distribution of the subject image (imaging signal) at this time is shown in FIGS. 11D and 11F. 11D shows a case where A <kL <H, and FIG. 11F shows a case where kL <A <H.

  In the case of b, the high frequency component amount H of the subject is smaller than the low frequency band component amount kL. Specifically, as shown in FIG. 11E, A <H <kL, or as shown in FIG. 11G, H <A <kL. In the case of H <kL <A shown in FIG. 11I, the high frequency component amount H is smaller than the low frequency band component amount kL, but both the high frequency component amount H and the low frequency band component amount kL are based on the determination reference value A. Since it is small, it is classified into c for the reason described later.

  Here, the classification of b is further subdivided as follows.

  b1) Since the subject itself has few high frequency components, even if the defocus amount is smaller than a predetermined value, the high frequency component amount H is calculated to be small.

  b2) Since the defocus amount is larger than the predetermined value, the high frequency component amount H is calculated to be small regardless of the subject.

  b3) When the high frequency component amount H is determined to be small, but the reliability of the phase difference and defocus amount obtained in the phase difference AF is low.

  When classified into b1 and b2, since the high frequency component amount H is smaller than the low frequency component amount L, it is considered that the influence of aliasing is small. For this reason, phase difference AF is possible.

  Further, the case of being classified as b3 corresponds to the case where the subject is difficult to detect the phase difference in the phase difference AF. Examples of such a subject include a subject having a periodic pattern pattern.

  FIG. 12 shows the shift of the phase difference image signal when the subject has a periodic pattern. With respect to the point P0 in the image on the phase difference detection pixel S1, the shift amounts are N, 3N, 5N, 7N, and 9N at points P1, P2, P3, P4, P5,. , ... are calculated in large numbers and cannot be determined as one value. That is, phase difference detection is difficult.

  The case where the subject is classified as c is a case where the subject has low illuminance or low contrast. Such a subject has a small power of high frequency component amount H and low frequency component amount kL. FIG. 11H shows a case where kL <H <A. FIG. 11I shows a case where H <kL <A. In these cases, it is difficult to obtain an accurate contrast evaluation value and phase difference at high speed. Therefore, in such a case, a non-focusing (unfocusable) process is performed.

  The flowchart in FIG. 10 shows AF processing (control method) performed by the subject determination unit 108 and the AF method switching unit 109. The processing is executed by a microcomputer such as a CPU constituting the subject determination unit 108 and the AF method switching unit 109 according to a computer program stored therein.

  In step 1001, the subject determination unit 108 performs scene determination at the current focus lens position. That is, the mixed signal from the A / D conversion unit 103 is divided into a low frequency component and a high frequency component, and the respective amounts H and kL and the determination reference value A are compared. It is determined which scene (subject) corresponds to.

  If it is classified as a, it is impossible to perform phase difference AF with higher accuracy using a phase difference detection pixel having an arrangement frequency of fp. Therefore, contrast AF is performed by a contrast detection method by minute driving of a focus lens described later. Do. Therefore, the subject determination unit 108 outputs a contrast AF selection signal to the AF method switching unit 109, and the process proceeds to step 1006.

  If it is classified as b, depending on the conditions, it is possible to perform phase difference AF, and the process proceeds to step 1002. Further, if it is classified as c, the process proceeds to step 1008 in order to perform the out-of-focus process.

  In step 1002, the subject determination unit 108 acquires the defocus amount from the defocus amount calculation unit 107. Then, the process proceeds to Step 1003.

  In step 1003, the subject determination unit 108 determines the reliability of the phase difference AF based on the defocus amount. When it is determined that the reliability of the phase difference AF is high, that is, when the image shift amount can be reduced to one value corresponding to b1 and b2 described above, the process proceeds to step 1004.

  On the other hand, when it is determined that the reliability of the phase difference AF is low, that is, when the image shift amount cannot be narrowed down to one value due to the periodicity of the subject as in b3 described above, the contrast AF is performed. The contrast AF selection signal is output to the AF method switching unit 109. Then, the process proceeds to Step 1006.

  In step 1004, the subject determination unit 108 selects an AF method based on the defocus amount. When the defocus amount is smaller than the predetermined value as in b1, a contrast AF selection signal is output to the AF method switching unit 109 in order to perform contrast AF. Then, the process proceeds to Step 1006. On the other hand, when the defocus amount is larger than the predetermined value as in b2, a phase difference AF selection signal is output to the AF method switching unit 109 in order to perform phase difference AF. Then, the process proceeds to Step 1005.

  In step 1005, the AF method switching unit 109 drives the motor 116 to move the focus lens to the in-focus position based on the defocus amount acquired in step 1002. Thereafter, the process returns to step 1001. Thereby, the subject determination unit 108 performs scene determination again.

  In step 1006, the AF mode switching unit 109 controls the movement of the focus lens by contrast AF.

  Here, when it is determined as a in step 1001 or when it is determined in step 1004 that the defocus amount is smaller than a predetermined value, AF with high accuracy cannot be performed by the phase difference AF. However, as shown in FIG. 8B, contrast AF by minute driving of the focus lens is possible.

  If it is determined as a in step 1001 and the process proceeds to step 1006, it is considered that the focus lens is already positioned near the focus position P. Therefore, in this case, in order to unify the driving direction of the focus lens by contrast AF, the focus lens is moved in a minute region starting from point A on the infinity side from the in-focus position, and the contrast evaluation value Monitor changes.

  If it is determined in step 1003 that the reliability of the phase difference AF is low and the process proceeds to step 1006, it cannot be determined where the focus lens is located with respect to the in-focus position P. For this reason, the focus lens is first driven to a minute amount, and the drive direction of the focus lens (the direction of the focus position) is determined based on the change in the contrast evaluation value at this time. Thereafter, the focus lens position where the contrast evaluation value is the maximum value is searched. In this case, the focus lens position where the contrast evaluation value becomes the maximum value may be searched while moving the focus lens in the entire region from the infinity end to the close end.

  Also, after performing phase difference AF in step 1005, if a is determined in step 1001, if it is determined in step 1003 that the reliability is low, it is determined in step 1004 that the defocus amount is smaller than a predetermined value. In this case, the same processing is performed in step 1006. Then, the process proceeds to step 1007, and after confirming the in-focus state, the AF process is terminated.

  In step 1008, the focus detection of the subject for which focus detection is difficult is interrupted by stopping the imaging lens 101 at the current position. The motor drive control unit 110 outputs a control signal for stopping the motor 116. After processing, focus detection is terminated.

  As described above, according to the present embodiment, in order to reduce the influence of aliasing due to the arrangement of the phase difference detection pixels, the subject is evaluated from the surface of the frequency component distribution and the AF method is selected. The accuracy of hybrid AF can be further increased. Moreover, by adopting an appropriate AF method from the contrast AF having a wide frequency band corresponding to the subject image and the phase difference AF having a narrow frequency band performed by discretely arranged phase difference detection pixels, the hybrid AF can be further improved. The speed can be increased.

  Each embodiment described above is only a representative example, and various modifications and changes can be made to each embodiment in carrying out the present invention.

1 is a block diagram illustrating a configuration of a digital camera (imaging device) that is an embodiment of the present invention. 1 is a diagram illustrating a configuration of an image sensor used in a digital camera of an embodiment. The front view and sectional drawing which show the structure of the phase difference detection pixel of the image pick-up element in an Example. The front view and sectional drawing which show the structure of another phase difference detection pixel of the image pick-up element in an Example. The figure which shows the concept of the phase difference detection (focused state) in an Example. The figure which shows the concept of the phase difference detection (out-of-focus state) in an Example. The figure which shows the image shift | offset | difference in an Example. The figure which shows the operation | movement of contrast AF in an Example. The figure which shows another operation | movement of contrast AF in an Example. FIG. 3 is a block diagram illustrating a configuration of a subject determination unit in the embodiment. The flowchart which shows the AF process in an Example. The figure which shows the scene determination in an Example. The figure which shows the frequency distribution of the to-be-photographed image in an Example. The figure which shows the frequency distribution of the to-be-photographed image in an Example. The figure which shows the frequency distribution of the to-be-photographed image in an Example. The figure which shows the frequency distribution of the to-be-photographed image in an Example. The figure which shows the frequency distribution of the to-be-photographed image in an Example. The figure which shows the frequency distribution of the to-be-photographed image in an Example. The figure which shows the frequency distribution of the to-be-photographed image in an Example. The figure which shows the frequency distribution of the to-be-photographed image in an Example. The figure which shows a mode that the image shift amount of the to-be-photographed object does not become settled.

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Image pick-up lens 102 Image pick-up element 104 High pass filter 105 Integration circuit 106 Peak hold circuit 107 Defocus amount calculation part 108 Subject determination part 109 AF system switching part 110 Motor drive control part 111 Phase difference evaluation part 112 Contrast focus detection part 113 Phase difference detection Signal extraction unit 114 Contrast evaluation circuit 115 Phase difference focus detection unit 116 Motor 117 Frequency calculation unit 118 Frequency evaluation unit 119 Scene evaluation unit 201 Imaging pixel 202, 203 Phase difference detection pixel 301 Micro lens 303 Light shielding layer 304 Insulating layer 901 High pass filter 902 Low-pass filter 903, 904 Integration circuit 905 Low frequency band evaluation unit 906 Distribution evaluation unit 907 High frequency band evaluation unit

Claims (5)

A first photoelectric conversion cell group that photoelectrically converts an object image formed by a light beam from the imaging optical system and outputs a first signal used to generate an image, and a light beam from the imaging optical system are divided. An image sensor having a second photoelectric conversion cell group that photoelectrically converts a plurality of images formed by the light flux and outputs a second signal used for phase difference detection;
Control means for performing first focus control by a contrast detection method using the first signal and second focus control by a phase difference detection method using the second signal;
The control means performs the first focus control and the second focus control according to a distribution of frequency components of the image based on a specific frequency determined based on an arrangement of the second photoelectric conversion cell group. An imaging apparatus characterized by switching.   The control means switches between the first focus control and the second focus control based on the distribution of the frequency components and the defocus amount of the imaging optical system obtained based on the phase difference. The imaging apparatus according to claim 1.   The control means performs the first focus control when a frequency component higher than the specific frequency among frequency components of the image is larger than a frequency component lower than the specific frequency. The imaging device according to claim 1.   The control means is a case where, among the frequency components of the image, the frequency component higher than the specific frequency is smaller than the frequency component lower than the specific frequency, and the defocus amount is smaller than a predetermined value. The imaging apparatus according to claim 2, wherein the first focus control is performed, and the second focus control is performed when the defocus amount is larger than the predetermined value. A first photoelectric conversion cell group that photoelectrically converts an object image formed by a light beam from the imaging optical system and outputs a first signal used to generate an image, and a light beam from the imaging optical system are divided. A method for controlling an image pickup apparatus having an image pickup element having a second photoelectric conversion cell group that photoelectrically converts a plurality of images formed by the luminous flux and outputs a second signal used for phase difference detection,
Performing first focus control by a contrast detection method using the first signal;
Performing second focus control by a phase difference detection method using the second signal;
Switching between the first focus control and the second focus control according to a distribution of frequency components of the image with reference to a specific frequency determined based on the arrangement of the second photoelectric conversion cell group. And a method of controlling the imaging apparatus.