JP2020057892A - Imaging device - Google Patents

Abstract

To provide an imaging device capable of reducing an influence of periodic noise and suppressing the deterioration of focus detection accuracy while minimizing an increase in time required to read a signal in one row.SOLUTION: An imaging device performs a first reading operation V1-V7 to add and read signals of first and second photoelectric conversion units, a second reading operation V8-V13 to independently read the signals of the first and second photoelectric conversion units, reads by thinning out at a predetermined cycle in a vertical direction in the first reading operation, sequentially reads rows that are not read in a first scan at the predetermined cycle in the second reading operation, detects phase difference by using a signal read by the second reading operation, and in the second reading operation, performs control so that time difference between reading timing of a reference signal 1 of the first photoelectric conversion unit and reading timing of a reference signal 2 of the second photoelectric conversion unit has a predetermined relationship with the frequency of a noise source that occurs during a reading operation of an image signal.SELECTED DRAWING: Figure 7

Description

  The present invention relates to an imaging device and a control method thereof.

  2. Description of the Related Art There has been proposed an imaging apparatus that performs imaging using a solid-state imaging device such as a CCD image sensor or a CMOS image sensor. In such an imaging apparatus, various noises that cause deterioration of image quality occur in the process of converting an optical image into an electric signal by the imaging element. Typical noises generated inside the image sensor include reset noise of pixels and readout circuits, and random noise that fluctuates every time an image capturing operation is performed, such as dark current generated in a pixel region. In addition, in the vicinity of the image sensor, there are many elements that can be a noise source, such as an actuator for driving a lens. In particular, when a CMOS image sensor is used as an image sensor, when the physical distance between the image sensor and the noise source is short, periodic noise (electromagnetic waves) generated from the noise source spatially covers the output signal of the image sensor. Sometimes. This is attributable to a unique operation of the CMOS image sensor in which the charge stored in the photodiode is converted into a voltage for each pixel and a signal is read.

  Further, in recent years, there is an imaging apparatus configured so that a part or all of pixels of an imaging image sensor can be divided into a plurality of photodiodes to read stored charges. For example, when the image is divided into two in the horizontal direction, an image generated by only the left divided pixels (hereinafter, referred to as A image) and an image generated by only the right pixels (hereinafter, referred to as B image) are read out independently. To There is disclosed a technique for calculating distance information to a subject by detecting an image shift amount (phase difference) formed between the A image and the B image. Note that a signal for a captured image (hereinafter, referred to as an A + B image) can also be obtained by adding the accumulated charges of the two divided photodiodes and reading them out simultaneously.

  In the above-described imaging apparatus, it is necessary to read out a signal (A + B image) for a captured image and a signal (A image or B image) for detecting a phase difference in time series. At this time, when a noise source that periodically fluctuates the power supply voltage of the image sensor operates, a periodic signal (A image or B image) for detecting a phase difference (A image or B image) in addition to a signal for a captured image (A + B image) Noise is generated. If periodic noise occurs in the phase difference detection signal (A image or B image), a correct phase difference signal cannot be obtained, and the focus detection accuracy will be reduced. On the other hand, there has been proposed a technique for reducing the influence of periodic noise on both a captured image signal (A + B image) and a phase difference detection signal (A image or B image) (Patent Document 1). ).

JP 2016-111452 A

  In the related art disclosed in Patent Document 1, it is necessary to control the signal readout timing so that the noise amount becomes “zero” for both the A image and the A + B image. For this reason, depending on the frequency of the periodic noise, there is a problem that the time required to read a signal in one row becomes longer, resulting in a reduction in the frame rate.

  Therefore, an object of the present invention is to provide an imaging apparatus capable of minimizing an increase in the time required for reading a signal in one row, reducing the influence of periodic noise, and suppressing a decrease in focus detection accuracy. To provide.

In order to achieve the above object, an imaging device according to the present invention includes:
A plurality of pixels each having first and second photoelectric conversion units for photoelectrically converting two subject images formed by light fluxes that have passed through different divided regions of the exit pupil of the imaging optical system with respect to a unit pixel are provided. A first read operation for adding and reading the signals of the first and second photoelectric conversion units, and a second read operation of independently reading the signals of the first and second photoelectric conversion units. Signal reading means for performing a read operation of the first read operation, a first scan in which rows are thinned out at predetermined intervals in the vertical direction in the first read operation, and a read operation in the first scan in the second read operation. Scanning means for performing a second scan for sequentially reading out non-output rows at a predetermined cycle, and detecting a phase difference between the two subject images using a signal read by the second read operation And a time difference between the timing of reading the reference signal 1 of the first photoelectric conversion unit and the timing of reading the reference signal 2 of the second photoelectric conversion unit in the second reading operation. Timing control means for controlling the frequency of a noise source generated during a signal reading operation to have a predetermined relationship.

  According to the present invention, there is provided an imaging apparatus capable of minimizing an increase in time required for reading a signal of one row, reducing the influence of periodic noise, and suppressing a decrease in focus detection accuracy. Can be realized.

FIG. 1 is a block diagram illustrating a configuration of an imaging device according to an embodiment of the present invention. The figure which showed the pixel array of the image sensor Diagram showing the circuit configuration of the image sensor The figure which showed the detailed circuit structure of the unit pixel in the image sensor. Illustration of readout circuit Explanatory diagram of AD conversion circuit Schematic diagram of a specific readout row in signal readout of the image sensor Schematic diagram showing output waveforms of a readout circuit and an AD conversion circuit Explanatory drawing of a specific method for reducing the influence of periodic noise in the first embodiment Explanatory diagram of a specific method for reducing the influence of periodic noise in the second embodiment Explanatory diagram in the case where the noise amounts of the A image and the B image are both zero.

  FIG. 1 is a block diagram illustrating a configuration of an imaging device according to an embodiment of the present invention. In FIG. 1, reference numeral 100 denotes a first lens arranged at the tip of the photographing optical system. The aperture 101 adjusts the light amount at the time of shooting by adjusting the aperture diameter. Reference numeral 102 denotes a second lens, and reference numeral 103 denotes a third lens. The second lens and the third lens are driven by a focus actuator 116, which will be described later, and advance and retreat in the optical axis direction to adjust the focus of the photographing optical system. .

  Reference numeral 104 denotes a focal plane shutter that adjusts the exposure time when capturing a still image. Reference numeral 105 denotes an optical low-pass filter for reducing false colors and moire of a captured image. An image sensor 106 photoelectrically converts an optical image of a subject formed by the photographing optical system into an electric signal. The image sensor 106 is controlled by a CPU 109 described later.

  Reference numeral 107 denotes an image processing unit DSP (Digital Signal Processor). The DSP 107 is an image processing unit that performs processing such as correction and compression on image data captured by the image sensor 106. Further, it has a function of adding A image data and B image data, which will be described later, and a correlation operation function of performing a correlation operation using an image signal. 108 is a RAM. The RAM 108 functions as a signal holding unit that holds output data from the image sensor 106, a function of an image data storage unit that stores image data processed by the DSP 107, and a function of a work memory when the CPU 109 described below operates. Combined. In this embodiment, these functions are realized by using the RAM 108. However, any other memory can be applied as long as the access speed is sufficiently high and there is no problem in operation. It is. Further, in the present embodiment, the RAM 108 is disposed outside the DSP 107 to the CPU 109, but a configuration in which some or all of the functions are incorporated in the DSP 107 or the CPU 109 may be employed.

  Reference numeral 109 denotes a CPU that controls the operation of the imaging apparatus in a comprehensive manner. The CPU 109 executes a program for controlling each unit of the imaging device. Further, it also has a function of controlling a focus driving circuit 115 described later using the result of the correlation operation output from the DSP 107 to adjust the focus of the photographing optical system. A display unit 110 displays a captured still image, a moving image, a menu, and the like. Reference numeral 111 denotes an operation unit that sets a shooting command, a shooting condition, and the like for the CPU 109. Reference numeral 112 denotes a removable recording medium for recording still image data and moving image data. Reference numeral 113 denotes a ROM that stores a program to be loaded and executed by the CPU 109 to control the operation of each unit.

  Reference numeral 114 denotes a shutter drive circuit for controlling the focal plane shutter 104. Reference numeral 115 denotes a focus drive circuit which is a focus position changing unit that changes the focus position of the optical system. The focus drive circuit 115 controls the drive of the focus actuator 116 based on the output of the CPU 109, and moves the second lens 102 and the third lens 103 in the optical axis direction. The focus is adjusted by driving forward and backward. An aperture driving circuit 117 controls the driving of the aperture actuator 118 to control the aperture of the aperture 101.

  Next, the configuration of the image sensor 106 will be described with reference to FIGS. FIG. 2 shows a pixel array of the image sensor 106. In FIG. 2, reference numeral 200 denotes a unit pixel constituting the pixel array. 201 is a micro lens in a unit pixel. 202a and 202b are photodiodes (PD) as photoelectric conversion means for performing photoelectric conversion, and these two PDs constitute a unit pixel arranged under a single microlens. With this configuration, the PDs 202a and 202b have a pupil division configuration, and separate images having a phase difference from each other are incident. Here, the PD 202a constitutes the photoelectric conversion unit for the A image, and the PD 202b constitutes the photoelectric conversion unit for the B image.

  Next, a circuit configuration of the image sensor 106 will be described with reference to FIG. Reference numeral 300 denotes a pixel array in which the unit pixels 200 are arranged in a matrix direction, and a plurality of (m + 1) unit pixels in the horizontal direction and (n + 1) unit pixels in the vertical direction are arranged. Reference numeral 301 denotes a timing generator (TG), which supplies a control pulse to each circuit block at an appropriate timing in accordance with the settings programmed by the CPU 109. In this embodiment, the TG 301 has the TG 301 built in the image sensor 106, but the TG 301 may be arranged outside the image sensor 106.

  The pixel signal photoelectrically converted by each unit pixel 200 is output to a vertical output line 303 for each row by a drive signal from a vertical scanning circuit 302. A constant current source 304 forms a source follower circuit in combination with a pixel amplifier transistor 403 described later. A readout circuit 305 has a function of amplifying the output from the vertical output line 303 in each column. Reference numeral 306 denotes an AD converter (ADC), which converts the output of the read circuit 305 into a digital signal. The image signals converted by the ADC 306 are sequentially selected by the horizontal scanning circuit 307 and sent to the output unit 309 via the horizontal output line 308. Then, a digital signal is output from the output unit 309 to the outside of the image sensor.

  FIG. 4 is a diagram illustrating a detailed circuit configuration of the unit pixel 200. In FIG. 4, PD 202a and PD 202b are PDs that constitute a unit pixel under the same micro lens as described above. 400a is an A image photoelectric conversion unit transfer switch, and 400b is a B image photoelectric conversion unit transfer switch. By turning on the transfer switch, the photoelectrons accumulated in the photoelectric conversion unit can be transferred to the floating diffusion (FD) unit 401. Since the ON / OFF of this transfer switch can be controlled independently, the signal for the A image and the signal for the B image can be transferred independently. Reference numeral 402 denotes a reset switch for initializing the FD. The FD unit 401 is connected to an amplifier 403, and the electric charge transferred from the PD 202a and PD 202b to the FD unit 401 by the amplifier 403 is converted into a voltage value according to the amount of electric charge. Then, the pixel signal is output to the vertical output line 303 through the select switch 404.

  The reading circuit 305 will be described in detail with reference to FIG. The pixel signal output to the vertical output line 303 is input to the inverting input terminal of the operational amplifier 503 through the clamp capacitor 500. The reference voltage Vref is input to the non-inverting input terminal of the operational amplifier 503. The switch 502 is a switch for short-circuiting the feedback capacitor 501.

  The configuration of the ADC 306 and its driving will be described in detail with reference to FIG. Reference numeral 600 denotes a comparator to which the output of the readout circuit 305 and the ramp signal from the ramp signal generation circuit 601 are input. The output level of the ramp signal generation circuit 601 changes over time, and when the signal level becomes equal to the level of the signal output from the reading circuit 305, the output of the comparator 600 switches. The output of the comparator 600 is connected to a counter 602, and the counter 602 counts a comparison time in the comparator 600. The output of the counter 602 is connected to a latch 603, and the latch 603 holds the count result of the counter 602. The output of latch 603 is connected to horizontal output line 308.

  Next, a signal reading operation in the image sensor 106 of the present embodiment will be described.

  In the present embodiment, first, a signal (addition signal) of the A + B image is read while thinning out a predetermined number of rows in the vertical direction. Then, after scanning the entire surface in the vertical direction, the vertical scanning is returned to the upper row in the vertical direction again, and the vertical scanning is performed again by scanning only the rows that have not been read earlier. At the time of the first scan, the addition signal is read from the pixels of the target row, and at the time of the second scan, the phase difference detection signal is read from the A pixel or the B pixel of the pixels of the target row. Then, phase difference detection is performed using the read A image and B image. Since such a reading method is suitable in the moving image mode, the present embodiment will also be described assuming application in the moving image mode.

  FIG. 7 is a schematic diagram of a specific readout row in the signal readout of the present embodiment. The hatched portions on the upper and left sides in the figure indicate light-shielded optical black portions (OB portions).

  In FIG. 7, rows surrounded by a thick frame are rows to be read in the first scan, and portions that are not thick frames are rows to be thinned out at the time of reading. After the addition signal is read from the A pixel and the B pixel in the V1 row, the addition signal is read from the V2 row three rows later. Subsequently, the addition signal is read from the rows V3, V4, V5, V6, and V7 at the same thinning rate. The rows read by the first scan output the charge signals generated in the entire pixels disposed below the same microlens suitable for image generation, and are obtained from the added signals of the V1 to V7 rows read at the same thinning rate. , A signal (A + B image) for a captured image can be generated.

  After the first scan up to the V7th row, the process returns to the V8th row and the second scan is performed. Among the pixels included in the V8th row, the phase difference detection signal is read from the A pixel. After that, a phase difference detection signal is read from the B pixel among the pixels included in the V9th row, which is the next row after the V8th row. After that, the V4th row already read out in the first scan is skipped, and the A pixel in the V10th row and the B pixel in the next V11th row are read out. Similarly, the V5 line already read in the first scan is skipped, and the V12 line and the next V13 line are read. It is possible to detect a phase difference using signals (A image and B image) for phase difference detection obtained by the second scanning.

  In the present embodiment, signals for detecting a phase difference (images A and B) are read out separately from signals for captured images (images A and B). If there is periodic noise at the time of signal reading, not only the signal for the captured image (A + B image) but also the signal for phase difference detection (A image and B image) is affected, and the focus detection accuracy is reduced. There is a risk that it will.

  For a signal for a captured image (A + B image), periodic noise can be reduced by a method described in Patent Document 1, for example. The same method can be applied to the phase difference detection signals (A image and B image). However, depending on the frequency of the periodic noise, the time required to read a signal in one row becomes longer, and as a result, the frame rate may be reduced.

  Therefore, in the present embodiment, the influence of periodic noise is reduced by equalizing the amount of noise superimposed on the A image and the B image. If the noise amounts of the A image and the B image are the same, the "phase difference" is not recognized, so that the focus detection is not affected.

  FIG. 8 is a schematic diagram showing output waveforms of the circuits shown in FIGS. An analog signal (readout circuit output) output from the readout circuit 305 is compared with a ramp signal by a comparator 600 arranged for each column. At this time, the counter 602 arranged for each column is operating like the comparator 600, and the potential of the ramp signal and the counter value change while taking a one-to-one correspondence, so that the potential of the read circuit output (analog Signal) into a digital signal.

  The change in the ramp signal converts a change in voltage into a change in time, and converts the time into a digital value by counting the time in a certain cycle (clock). Then, when the read circuit output and the ramp signal cross, the output of the comparator 600 is inverted and the AD conversion is completed.

  At time t = T1, when the readout circuit output and the ramp signal intersect (become equal) in the N signal period of the A image, the output of the comparator 600 is inverted. In response to the polarity inversion of the comparator 600, the counter 602 stops the counting operation, and the latch 603 holds the count value corresponding to the N signal (ΔVa).

  Next, when the readout circuit output and the ramp signal intersect (become equal) in the S signal period at t = T2, the output of the comparator 600 is inverted. In response to the polarity inversion of the comparator 600, the counter 602 stops the counting operation, and the latch 603 holds the count value corresponding to the output signal. This output signal is a value obtained by subtracting the N signal (ΔVa) from the read circuit output when the polarity of the comparator 600 is inverted in the S signal period. Thus, an output signal A which is a difference between the N signal and the S signal can be obtained.

  Similarly, at t = T3, when the readout circuit output and the ramp signal intersect (become equal) in the N signal period of the B image, the output of the comparator 600 is inverted. In response to the polarity inversion of the comparator 600, the counter 602 stops the counting operation, and the latch 603 holds the count value corresponding to the N signal (ΔVb).

  Next, when the readout circuit output and the ramp signal intersect (become equal) in the S signal period at t = T4, the output of the comparator 600 is inverted. In response to the polarity inversion of the comparator 600, the counter 602 stops the counting operation, and the latch 603 holds the count value corresponding to the output signal. This output signal is a value obtained by subtracting the N signal (ΔVb) from the read circuit output when the polarity of the comparator 600 is inverted in the S signal period. As a result, an output signal B which is a difference between the N signal and the S signal can be obtained.

  Noise is mixed in the A image and the B image due to the influence of the periodic noise, and since the noise amount of the A image and the noise amount of the B image are different, a false phase difference is detected. That is, erroneous focus detection is performed as it is.

  Next, a method of reducing the influence of periodic noise in the present embodiment will be described. FIG. 9 is an explanatory diagram of a specific method for reducing the influence of periodic noise in the first embodiment. FIG. 9A is the same as FIG.

  In FIG. 9B, when the readout circuit output and the ramp signal intersect (equal to each other) at time t = T1 in the N signal period of the A image, the output of the comparator 600 is inverted. In response to the polarity inversion of the comparator 600, the counter 602 stops the counting operation, and the latch 603 holds the count value corresponding to the N signal (ΔVa).

  Next, when the readout circuit output and the ramp signal intersect (become equal) in the S signal period at t = T2, the output of the comparator 600 is inverted. In response to the polarity inversion of the comparator 600, the counter 602 stops the counting operation, and the latch 603 holds the count value corresponding to the output signal. This output signal is a value obtained by subtracting the N signal (ΔVa) from the read circuit output when the polarity of the comparator 600 is inverted in the S signal period. Thus, an output signal A which is a difference between the N signal and the S signal can be obtained.

  Here, the reset period is lengthened by Δt1, and then the B image is read. At t = T5, when the readout circuit output and the ramp signal intersect (become equal) in the N signal period of the B image, the output of the comparator 600 is inverted. In response to the polarity inversion of the comparator 600, the counter 602 stops the counting operation, and the latch 603 holds the count value corresponding to the N signal (ΔVb).

  Next, when the readout circuit output and the ramp signal intersect (become equal) in the S signal period at t = T6, the output of the comparator 600 is inverted. In response to the polarity inversion of the comparator 600, the counter 602 stops the counting operation, and the latch 603 holds the count value corresponding to the output signal. This output signal is a value obtained by subtracting the N signal (ΔVb) from the read circuit output when the polarity of the comparator 600 is inverted in the S signal period. As a result, an output signal B which is a difference between the N signal and the S signal can be obtained.

  Although noise is mixed in the A image and the B image due to the influence of the periodic noise, the phase difference is not detected because the noise amount of the A image is equal to the noise amount of the B image. This is because the reset period is extended by Δt1 to adjust the timing of T5. Specifically, this is because the time (T5−T1) is set to be equal to the period 1 / f (f: noise frequency) of the noise waveform.

  FIG. 10 is an explanatory diagram of a specific method for reducing the influence of periodic noise in the second embodiment. FIG. 10A is the same as FIG. In FIG. 10B, the timing of T5 is adjusted by waiting for the time from the end of the reset period before reading the B image to the start of AD conversion of the N signal by Δt1. As a result, since the noise amount of the A image and the noise amount of the B image become equal, no phase difference is detected.

  FIG. 11 is an explanatory diagram in the case where the noise amount of the A image and the noise amount of the B image are both set to “zero”. In FIG. 11B, since the time of (T7-T1) matches the period 1 / f of the noise waveform, the noise amount of the A image becomes zero. Similarly, since the time of (T9-T8) matches the period 1 / f of the noise waveform, the noise amount of the B image is also zero. FIG. 11A is the same as FIG. Comparing FIGS. 11A and 11B, it can be seen that the time required to read one row of signals is significantly increased. In comparison with this, it can be seen that in the case of FIGS. 9 and 10, an increase in the time required to read out signals in one row is suppressed.

  As described above, according to the present embodiment, while minimizing the increase in the time required to read out one row of signals, it is possible to reduce the influence of periodic noise and suppress a decrease in focus detection accuracy. It is possible to provide an imaging device that is enabled.

106 image sensor, 200 unit pixel, 201 micro lens,
202a photoelectric conversion unit, 202b photoelectric conversion unit, 305 readout circuit,
306 AD converter, 600 comparator, 601 ramp signal generation circuit,
602 counter

Claims (3)

A plurality of pixels each having a first and second photoelectric conversion unit for photoelectrically converting two subject images formed by light fluxes having passed through different divided regions of the exit pupil of the imaging optical system with respect to a unit pixel are provided. An arranged image sensor;
A signal for performing a first read operation for adding and reading the signals of the first and second photoelectric converters and a second read operation for independently reading the signals of the first and second photoelectric converters Reading means;
In the first readout operation, a first scan in which rows are thinned out at a predetermined cycle in the vertical direction is read out, and in the second readout operation, rows which are not read out in the first scan are sequentially read out at a predetermined cycle. Scanning means for performing two scans;
Phase difference detection means for detecting a phase difference between the two subject images using a signal read by the second reading operation;
In the second read operation, a time difference between the read timing of the reference signal 1 of the first photoelectric converter and the read timing of the reference signal 2 of the second photoelectric converter occurs during the read operation of the image signal. An image pickup apparatus comprising: a timing control unit that controls the frequency of the noise source to be in a predetermined relationship.   In the second read operation, the signal readout of the first and second photoelectric conversion units is adjusted by adjusting the length of a reset period of the photoelectric conversion unit that is read later in time, whereby the first readout is performed. The control is performed such that the time difference between the read timing of the reference signal 1 of the photoelectric conversion unit and the read timing of the reference signal 2 of the second photoelectric conversion unit is an integral multiple of the period of the noise source. 2. The imaging device according to 1.   In the second reading operation, of the signal reading of the first and second photoelectric conversion units, the time from the end of the reset period to the start of reading of the reference signal 2 of the photoelectric conversion unit that is read temporally later is defined as the time. By adjusting, the time difference between the readout timing of the reference signal 1 of the first photoelectric conversion unit and the readout timing of the reference signal 2 of the second photoelectric conversion unit becomes an integral multiple of the period of the noise source. The imaging device according to claim 1, wherein the control is performed.