JP2015025870A - Image-capturing device and control method therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image-capturing device capable of improving the effect of vibration-proof control of a camera-shake low-frequency component while preventing a swinging-back phenomenon after the occurrence of large shaking such as panning during image-capture.SOLUTION: First arithmetic means calculates a first shaking correction amount by using a high-pass filter 109. Second arithmetic means calculates a first offset component 111 of the output of first shaking detection means 103, and also calculates a second offset component 113 of the output of the first shaking detection means 103. A shaking subtraction amount is calculated on the basis of a signal having the first offset component 111 subtracted therefrom and a signal having the second offset component 113 subtracted therefrom. The second arithmetic means calculates a second shaking correction amount by integrating a signal having the shaking subtraction amount is subtracted from the output of the shaking detection means 103. The correction means corrects, before image-capture, the shaking of the image on the basis of the first shaking correction amount and corrects, during image-capture, the shaking of the image on the basis of the second shaking correction amount.

Description

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

  With today's cameras, important tasks such as determining exposure and focusing are fully automated. In addition, in a camera equipped with an image stabilization control device that prevents image blur due to camera shake or the like, there are almost no factors that cause a photographer to make a shooting mistake.

  In order to enable shooting without image shake even when camera shake occurs at the time of shutter release, camera shake due to camera shake is detected, and an image shake correction lens (hereinafter referred to as a “correction lens”) according to the detected value. ) Must be moved. At that time, it is a requirement to accurately detect the camera vibration and correct the change in the optical axis due to the shake. In principle, image blurring is suppressed by mounting a vibration detection unit that obtains detection results such as angular velocity and a drive control unit that displaces the correction lens based on the calculation processing result.

  However, it is generally known that an output signal of a sensor that detects camera vibration includes components other than a signal due to shake. The components other than the signal due to the shake include a direct current component such as a variation in the reference voltage due to individual differences among sensors. Moreover, this direct current component drifts with temperature change. When low-frequency components are included in this way, if image blur correction is performed using the output signal of the sensor as it is, the correction accuracy may be reduced. Therefore, in order to remove the offset component that is sensor noise, it is a common practice to insert an HPF (High Pass Filter) and remove the low frequency component from the sensor output signal to obtain a shake correction signal. Yes.

  In addition, when the imaging device is moved in one direction for a long time, such as panning or tilting, the output signal from the sensor contains many low-frequency components, so the low-frequency components are attenuated for image blur correction. It is necessary to let Therefore, a technique is known in which panning / tilting is determined based on angular velocity and angle data obtained by integrating the angular velocity, and panning control is performed to switch to correction characteristics suitable for panning or tilting.

  In this panning control, first, the cutoff frequency of the HPF or integral filter is shifted to the high frequency side. As a result, the shake correction does not respond to the low frequency range.

  The above-described panning / tilting control is disclosed in, for example, Japanese Patent Application Laid-Open No. H10-260260, and the panning state or tilting that can perform image blur correction at a high frequency while suppressing a response at a low frequency. This is known as state image blur correction control.

JP-A-5-323436 JP-A-10-010596

Although the offset component of the angular velocity sensor can be removed by the HPF, the configuration in which the low frequency component is attenuated through the HPF has the following problems.
First, the angle calculation filter will be described. The angle calculation filter used in the image stabilization system is a filter (right side of equation (1)) that combines an integrator (first term on the left side of equation (1)) and HPF (second term on the left side of equation (1)). Generally used. This filter has the same formula as a low-pass filter (LPF) having a time constant T multiplied by the time constant T.

  Here, on the left side of equation (1), T is the time constant of the high-pass filter (HPF), and on the right side of equation (2), T is the time constant of LPF.

  The angular velocity meter has a problem in that, since it includes a small amount of low-frequency noise components, it performs an anti-vibration operation different from the actual shake and conversely induces shake. For this reason, the angle calculation filter is not configured only by the integrator. Further, in order to prevent the calculation result of the angle signal from being saturated in the integrator calculation, the filter on the right side of Expression (1) is used as a vibration-proof angle calculation filter.

  Accordingly, the angle calculation filter includes an HPF, and when the HPF is connected to the preceding stage of the angle calculation filter, a second-order HPF is configured in the filter from the angular velocity meter to the angle calculation. For this reason, in the low frequency band (˜1 Hz) of camera shake, the phase is greatly advanced and the image stabilization effect is lowered.

  Further, even during panning, the filter characteristics including the secondary HPF are adversely affected. When a large shake such as panning occurs, a low-frequency component with a large amplitude is also attenuated, and for example, at the end of panning, a signal in the direction opposite to the panning direction (shake back) is generated. This signal then slowly converges to zero, but when image blur correction is performed based on this signal, the correction amount is calculated using a signal different from the actual image pickup device shake. If correction is performed based on the signal, the accuracy of camera shake correction may be reduced.

  Therefore, when the image stabilization filter includes the HPF, when the cutoff frequency of the HPF is set to be low, the image stabilization performance of the image blur correction for the low frequency component accompanying the body shake or the like is improved. However, after a large shake such as panning occurs, the magnitude of the swing back and the time until it converges to zero increase, and it is appropriate only when shooting with a solid hold so that the shake of the imaging device does not become large There is a problem that an effective anti-vibration effect cannot be obtained.

  In the technique proposed in Patent Document 2, HPF processing is not performed during imaging (exposure), and image stabilization control is performed with the offset during imaging fixed. This makes it possible to perform image stabilization control without providing an HPF during imaging. However, in the method of fixing the offset of the angular velocity immediately before imaging, if an error occurs in the angular velocity offset, the image stabilization control is performed in a state different from the actual camera shake. In other words, if there is an offset error, the image stabilization control is performed in a state where the angular velocity corresponding to the offset error is continuously added to the camera shake during imaging, and thus the shake correction moves in an unintended direction. Accordingly, when the angular velocity offset is fixed only during imaging, the correction effect may be deteriorated when an offset error occurs.

  Accordingly, an object of the present invention is to improve the anti-vibration control effect of the low-hand shake component while preventing a swing-back phenomenon after a large shake such as panning occurs.

  According to an aspect of the present invention, there is provided an imaging device, a shake detection unit that detects a shake of the imaging device, a first calculation unit that calculates a first shake correction amount that corrects an image shake before imaging, The first calculation includes: a second calculation unit that calculates a second shake correction amount that corrects an image shake during imaging; and a shake correction unit that corrects an image shake based on the first or second shake correction amount. The means calculates the first shake correction amount using a high-pass filter, the second calculation means calculates a first offset component of the output of the shake detection means, and the shake detection means. Second offset calculating means for calculating a second offset component of the output of the first offset component, and subtracting the first offset component from the output of the shake detecting means, and the second offset component from the second shake correction amount. Subtract Subtraction amount calculation means for calculating a shake subtraction amount based on the signal, and the second calculation means integrates a signal obtained by subtracting the shake subtraction amount from the output of the shake detection means, thereby integrating the second shake amount. A correction amount is calculated, and the correction unit corrects the image shake based on the first shake correction amount before imaging, and corrects the image shake based on the second shake correction amount during imaging. A featured imaging device is provided.

  ADVANTAGE OF THE INVENTION According to this invention, the anti-shake control effect of a low frequency component can also be improved, preventing the shake-back phenomenon after big shakes, such as panning, having occurred.

1 is a schematic diagram of an imaging apparatus in an embodiment. 1 is a diagram illustrating a configuration of an imaging device according to an embodiment. The figure explaining the calculation method of the angular velocity subtraction amount in the embodiment. The figure explaining the effect of anti-vibration control in an embodiment. 6A and 6B are diagrams for explaining shake correction during imaging and during other than imaging. The figure which shows the gain characteristic of the angle calculation filter in embodiment. The flowchart which shows the vibration proof control process in embodiment. 1 is a diagram illustrating a configuration of an imaging device according to an embodiment.

  DESCRIPTION OF EMBODIMENTS Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the drawings. In addition, this invention is not limited to the following embodiment, It shows only the specific example advantageous for implementation of this invention. Moreover, not all combinations of features described in the following embodiments are indispensable for solving the problems of the present invention.

<Embodiment 1>
FIG. 1 is a schematic diagram of an imaging apparatus 101 according to the first embodiment. FIG. 2 shows a configuration of an image pickup unit of the image pickup apparatus 101 and functional blocks of image blur correction processing executed by the CPU 105 which is a central processing unit.

  A release button 104 is provided on the main body of the camera 101, and a switch open / close signal generated by operating this button is sent to the CPU 105. A correction lens 108 and an image sensor 106 are positioned on the optical axis 102 of the imaging optical system. An angular velocity meter 103 serving as a shake detection unit detects a shake of the imaging device around arrows 103p (pitch) and 103y (yaw). The output of the angular velocity meter 103 is input to the CPU 105. The HPF 109 is a high-pass filter that cuts a DC component in the output of the angular velocity meter 103. This reduces the offset component added as detection noise to the angular velocity meter. The angular velocity after passing through the HPF is integrated by the angle 1 calculation unit 110 and converted into an angle signal.

  Also, panning / tilting is determined by the magnitude of the angular velocity, and at the time of panning / tilting, the cutoff frequency of the HPF is shifted to the high frequency side so that the shake correction does not respond to the low frequency. ing.

  The above is the conventional shake correction method, and is hereinafter referred to as “first angle calculation”. The angle 1 calculation unit 110 functions as a first calculation unit that calculates a first shake correction amount for correcting image shake with respect to the first angle calculation. In the present embodiment, the “second angle calculation” is performed in parallel with the “first angle calculation”. The angle 2 calculation unit 117 functions as second calculation means for calculating a second shake correction amount for correcting image shake with respect to the twelfth angle calculation. Then, before imaging, the image stabilization process for suppressing image blur is controlled based on the output of the first angle calculation, and the image stabilization process is controlled based on the output of the second angle calculation during imaging. Hereinafter, the second angle calculation will be described below.

  The output of the angular velocity meter 103 is also input to the subtractor 116, and the output value of the subtraction amount calculation unit 115 is subtracted from the angular velocity meter output. This subtraction result is input to the angle 2 calculation unit 117 and converted into an angle signal. Details of the calculation in the subtraction amount calculation unit 112 will be described later.

  The angle 1 calculated by the angle 1 calculation unit 110 and the angle 2 calculated by the angle 2 calculation unit 117 are input to the signal switching unit 118. In the signal switching unit 118, whether to perform the image stabilization control using the angle 1 or the image stabilization control using the angle 2 is selected depending on the state of the camera such as whether or not the imaging operation is started, and the selected angle is sensitive. Input to the degree adjustment unit 119. The sensitivity adjustment unit 119 amplifies the angle signal, which is the output of the signal switching unit 118, based on the zoom and focus position information 107 and the focal length and imaging magnification obtained from the information, and sets the angle shake correction target value. This is to correct a change in shake correction sensitivity on the camera image plane with respect to the shake correction stroke of the shake correction unit 108 due to a change in optical information such as lens focus and zoom.

  The angular shake correction target value obtained by the sensitivity adjustment unit 119 is output to the drive control unit 120, and the shake correction unit 108 is driven to perform image shake correction. In the example shown in FIG. 2, so-called optical image stabilization in which the correction lens is moved in a plane perpendicular to the optical axis based on the calculated correction amount is employed as the shake correction unit. The image blur correction method is not limited to optical image stabilization using a correction lens, and there is a method of performing image stabilization by moving the imaging apparatus in a plane perpendicular to the optical axis. In addition, there is a method based on electronic image stabilization that reduces the influence of shake by changing the cut-out position of each shooting frame output by the image sensor, and image shake correction can also be performed by combining a plurality of image stabilization methods.

  Next, a method for calculating the angular velocity subtraction amount will be described below. As for the output of the angular velocity meter 103, the offset 1 calculation unit 111 calculates an angular velocity offset component (first offset component) added as detection noise to the angular velocity meter output. For example, the output value of the angular velocity meter 103 when the shake of the imaging apparatus is very small, such as when the amplitude of the angular velocity after passing through the HPF is small, or when the amplitude of the angular acceleration obtained by differentiating the angular velocity is small, is acquired. Then, it is smoothly connected by an LPF whose cut-off frequency is set very small. An offset that is a DC component can be calculated by such a method. The offset 1 subtraction unit 112 subtracts the angular velocity offset calculated by the offset 1 calculation unit 111 from the output of the angular velocity meter 103. The angular velocity offset calculated by the offset 1 calculation unit 111 is input to the offset 2 calculation unit 113, and the angle offset component (second offset component) is calculated by the method of Expression (2).

Angle offset = angular velocity offset × T (2)
As described using Equation (1), the angle 2 calculation unit 117 calculates the angle signal by multiplying the LPF set to the time constant T by T. The offset of the signal calculated by the angle 2 calculator 115 can be obtained by multiplying the low frequency region where the gain characteristic of the LPF is 0 dB by T. The offset amount of the signal input to the angle 2 calculation unit 117 has already been calculated by the offset 1 calculation unit 111. Therefore, if the output of the offset 1 calculation unit 111 is multiplied by T, the offset component after passing through the angle 2 calculation unit 117 can be calculated.

  The angular velocity offset from the offset 1 calculation unit 111 calculated as described above is input to the offset subtraction unit 112. In the offset subtracting unit 112, the angular velocity offset is subtracted from the output of the angular velocity meter. The angle offset from the offset 2 calculation unit 113 is input to the offset subtraction unit 114. In the offset subtraction unit 114, the angular velocity offset is subtracted from the output value of the angle 2 calculation unit 117 (previous sampling value of the control period). The subtraction results in the offset subtraction units 112 and 114 are input to the subtraction amount calculation unit 115, respectively.

  Next, a method of calculating the angular velocity subtraction amount (shake subtraction amount) in the subtraction amount calculation unit 115 will be described. FIG. 3 shows time-series data in each control block (111, 112, 113, 114, 115, 116, 117) shown in FIG. In FIG. 3A, 301 is an angular velocity output from the angular velocity meter 103, and 302 is an angular velocity offset value calculated by the offset 1 calculation unit 111. When 301 is integrated and the angle is calculated, the signal becomes 308 in FIG. Reference numeral 303 in FIG. 3B denotes an angular velocity after being subtracted by the offset 1 subtraction unit 112 from the output of the angular velocity meter 103, and is a signal obtained by subtracting 302 from 301. Since the angle calculation filter used here has a characteristic of HPF + integration, when the angular velocity offset 302 is multiplied by the time constant T, the angle offset 307 is obtained as described using the equation (2).

  When a large angular velocity such as panning occurs, the output of the angle calculation filter deviates greatly from the offset center 307 and transitions to converge to the offset 307 over time. If the signal 308 moves far away from the offset after panning, it takes a long time for the signal 308 to return to the vicinity of the offset 307, so that shake correction cannot be performed. In particular, as the frequency band of the image stabilization control is expanded to the low frequency side, that is, as the cut-off frequency of the angle calculation filter is decreased, the time for which the shake correction is reduced becomes longer.

  Therefore, when the angular velocity is high due to operations such as panning, it is desirable to cut the angular velocity unnecessary for camera shake stabilization as much as possible and input it to the angle calculation filter. Therefore, if the output of the angle calculation filter can be controlled within a certain angle range centered on the angle offset, the vibration isolation performance immediately after panning is improved.

  Therefore, the angular velocity subtraction amount is calculated using the angle signal (previously sampled value of the control cycle) that is the shake correction target value, and the angular velocity subtraction amount is integrated after subtracting the angular velocity subtraction amount from the angular velocity. In this way, the shake correction target value can be calculated by limiting the vibration-proof movable range, and when a large shake such as panning or tilting occurs, the shake correction can be performed immediately.

The angular velocity subtraction amount is calculated using the following (1) to (3).
(1) Angular velocity 303 after offset removal obtained by subtracting the offset 302 from the angular velocity 301.
(2) An angle signal 309 after offset removal obtained by subtracting the angle offset 307 calculated by the offset 2 calculation unit 113 from the output of the angle 2 calculation unit 117 (previous sampling value of the control period).
(3) A gain calculation table shown in FIGS.

In FIG. 3G, the horizontal axis is the angle signal after offset removal, and the vertical axis is the gain α.
When the angle signal after offset removal is A1 or less, the gain α is zero.
When the angle signal after the offset removal exceeds A2, the gain α is 1. That is, the angular velocity subtraction amount is calculated when the signal after offset removal exceeds the predetermined threshold A2, and the angular velocity subtraction amount is increased as the difference from the predetermined threshold is larger.
When the angle signal after offset removal is positioned between A1 and A2, the gain α is a value obtained by linear interpolation between A1 and A2.
Also in FIG. 3H, the gain β is obtained by the same method. The horizontal axis is the angle signal after offset removal, and the vertical axis is the gain β.
When the angle signal after the offset removal is B1 or more, the gain α is zero.
When the angle signal after offset removal is B2 or less, the gain β is 1.
When the angle signal after offset removal is located between B1 and B2, the gain β is a value obtained by linear interpolation between B1 and B2.

  The angular velocity subtraction amount is calculated from the gain α and gain β obtained as described above and the angular velocity 303 after offset removal. The angular velocity subtraction amount is calculated from the equations (3) and (4), and the gain to be multiplied differs depending on the sign of the angular velocity after the offset subtraction. When the sign of the angular velocity that is the output of the offset subtracting unit 108 is positive, the angular velocity is multiplied by a gain α, and when the sign of the angular velocity is negative, the angular velocity is multiplied by a gain β.

Angular velocity is positive: Angular velocity subtraction amount = Angular velocity after offset subtraction x α (3)
Angular velocity is in the negative direction: Angular velocity subtraction amount = Angular velocity after offset subtraction × β (4)
Reference numeral 304 in FIG. 3C denotes a signal obtained by subtracting the angular velocity subtraction amount from the angular velocity 303 after offset subtraction.
305 in FIG. 3E is a signal obtained by subtracting the angular velocity subtraction amount from the angular velocity 301.
306 in FIG. 3F is a signal obtained by integrating 305 and calculating an angle.
As described above, the control described above can cut the angular velocity component due to panning, and stabilization of the image stabilization control can be accelerated immediately after panning, and appropriate shake correction can be performed.

  Here, even when an error from the actual offset occurs in the angular velocity offset calculation, a large angular velocity such as panning can be cut. Even if a slight error occurs in the subtraction amount due to the offset deviation, the angular velocity of the panning component can be removed. Even if there is an error, the time until the camera shake correction immediately after panning is stabilized can be significantly improved as compared with the case where the angular velocity subtraction process is not performed.

  FIG. 4 shows the control effect in this embodiment. When the shake angle of the imaging apparatus has a waveform 401 in FIG. 4A, it is preferable as the shake prevention that the target shake correction angle is also the same as 401. However, in reality, there is a limit to the correction range of the shake correction device, and there is a problem that the shake detection device includes an offset and the offset drifts due to temperature. It is difficult to correct the shake as in 401. . Here, if the angle offset generated by integrating the angular velocity offset of the angular velocity meter which is the shake detection device is 404 in FIG. 4B, it is desirable that the target shake correction angle is controlled around the offset 404. 2 When there is no angle subtraction process described with reference to FIG. 3, when panning / tilting is performed, it takes a long time to return to the offset center by far away from the offset center. (The longer the filter cutoff frequency characteristic is on the low frequency side, the longer it takes to return.) Unlike the actual camera shake angle 401, the target shake angle 402 until returning to the offset center has an appropriate anti-shake effect. I can't get it. Therefore, the target shake angle after the angular velocity subtraction process can obtain the image stabilization control amount at the center of the offset 404 as indicated by 403, and approaches the ideal shake target value 401 to improve the shake correction effect.

  In the above control, the angle can be calculated without using HPF. However, the angle calculated here includes an offset angle due to the influence of the output noise component of the angular velocity meter.

  Next, how to use the angle 2 including the offset for the image stabilization control will be described. FIG. 5 shows time-series data for explaining a shake correction method during imaging and other than during imaging. A waveform 501 is the angle 2 calculated by the angle 2 calculation unit 114, and a waveform 502 is the angle 1 calculated by the angle 1 calculation unit 110. Since no HPF is provided at angle 2, if a long time elapses after the power is turned on (for example, period 504), the angle 501 is calculated away from the 0 center due to the temperature drift effect of the offset component of the angular velocity meter. . FIG. 6 shows the gain characteristics of the angle calculation filter. Reference numeral 601 denotes a pure integral filter characteristic (frequency-gain characteristic), and reference numeral 602 denotes an integral + HPF filter characteristic used for angle calculation. 602 has a flat characteristic in the low frequency band, and this gain characteristic remains for the offset of the angular velocity. Therefore, during the period 504 in FIG. 5, the angle 501 is calculated away from the center as the angular velocity offset increases due to the temperature drift effect of the angular velocity meter. The angle 502 calculated by the angle 1 calculation unit 110 is an output of the angle 1 calculation unit 110 in FIG. 2 and is a signal calculated using the HPF, and thus has a characteristic 603 in which HPF is added to 602. . As can be seen from the fact that the gain decreases in the low frequency region, the offset component included in the angular velocity meter 603 can be removed, and the angle is calculated at the zero center. However, since the HPF is added as described above, the anti-vibration effect immediately after a large shake such as panning and tilting is weakened due to the swing back phenomenon.

  Therefore, it is possible to obtain an appropriate anti-vibration effect by performing anti-vibration control using the angle 501 rather than the angle 502. However, since the angle 501 has a filter characteristic such as 602 in FIG. 6, the gain does not attenuate like the filter 603 in the low frequency region, and becomes a flat gain characteristic. That is, the angle calculation is performed while including the offset corresponding to the angular velocity offset. Therefore, if the vibration control is always performed at 501, the offset of the angle 501 increases due to the temperature drift of the angular velocity offset, and when the time is up, the vibration control movable range is insufficient and the movable end becomes uncontrollable. Therefore, it is detected whether or not the image is being picked up, and the image stabilization control is performed using the signal 502 during the image capturing, and the image stabilization control is performed using the signal 501 during the EVF display or AF / AE operation in the preparation period before the image capturing. To do. As a result, the image stabilization effect is improved by the filter characteristic expanded to the low frequency region during the imaging period, and a certain amount of image stabilization effect can be secured even outside the imaging period. Accordingly, it is possible to improve the AF / AE accuracy and the ease of the photographer's framing operation while viewing the EVF display.

  Reference numeral 505 denotes an imaging start timing, and reference numeral 506 denotes an imaging end timing. If control is performed at the angle target position of the waveform 502 during the period from 505 to 506 during imaging, shakeback occurs due to the influence of the HPF immediately after panning, and image stabilization control different from actual camera shake is performed. Therefore, there is a possibility that the image stabilization control effect may be reduced, but the difference between 501 and 502 at the timing 505 is calculated as an offset, and the signal 503 obtained by subtracting the offset from the waveform 501 is used in the period 505 to 506. To do. When the imaging is completed at 506, a signal that returns to the waveform 502 at a constant speed is added to the waveform 503 until 503 and 502 match.

  As described above, during imaging, there is no backlash immediately after panning / tilting without HPF, and the anti-vibration control can be performed by the angle calculated by expanding the anti-vibration filter to the low frequency region, thereby improving the anti-vibration effect. be able to.

  With reference to the flowchart of FIG. 7, the image stabilization control process of this embodiment is demonstrated. This flow starts when the main power of the camera is turned on and is executed at a constant sampling cycle. First, in S701, the state of the image stabilization SW is detected. If it is ON, the process proceeds to S702, and if it is OFF, the process proceeds to S716. In step S702, the output of the angular velocity meter 103 is captured. In next step S703, it is determined whether or not shake correction is possible. If the shake correction is possible, the process proceeds to S704, and if the shake correction is not possible, the process proceeds to S716. In step S <b> 703, for example, it is determined that the state from when the power is supplied until the output of the angular velocity meter 103 is stabilized is not a state in which shake correction is possible. Further, after the output of the angular velocity meter 103 is stabilized, it is determined that shake correction is possible. As a result, it is possible to prevent deterioration in the image stabilization performance due to the shake correction performed when the output value immediately after the power supply is unstable.

  In step S704, the angular velocity offset 1 is calculated, and the first angular velocity after subtracting the offset 1 from the angular velocity is calculated. In S705, the first angular velocity is integrated, and angle 1 is calculated. Next, in S706, the angular velocity offset 2 is calculated, and the second angular velocity after subtracting the offset 2 from the angular velocity is calculated. In S707, the value at the previous sampling of the angle 2 is acquired. In S708, the angular velocity subtraction amount is calculated from the second angular velocity calculated in S706 and the previous value of angle 2 acquired in S707. Next, in step S709, the angular velocity subtraction amount is subtracted from the angular velocity (before offset subtraction), and in step S710, an angle 2 is calculated by integrating the signal obtained by subtracting the angular velocity subtraction amount from the angular velocity. Next, in S711, it is determined whether or not an image is being captured. If the image is not being captured, the target angle is set to 1 in S715, and the process proceeds to S713. If it is determined in S711 that an image is being captured, the angle 2 is set in S712, and the target angle is set by the method described with reference to FIG.

  In step S713, the shake correction target value is calculated by multiplying the target angle by the sensitivity based on the focal length and the shooting magnification obtained from the zoom and focus information 107. Next, in step S714, after the shake correction lens is driven based on the shake correction target value, the shake correction routine is terminated and the next sampling cycle is awaited. In S716, the driving of the shake correction lens is stopped, the shake correction routine is terminated, and the next sampling cycle is awaited.

  In the first embodiment described above, the angular velocity offset and the angular offset are respectively calculated. Next, the angular velocity subtraction amount is obtained from the signal obtained by subtracting the angular velocity offset from the angular velocity from the angular velocity meter and the signal obtained by subtracting the angular offset from the angular velocity offset. Then, the target deflection angle is obtained by integrating the signal obtained by subtracting the angular velocity subtraction amount from the angular velocity from the angular velocity meter. Thereby, even if an offset is included in the output of the angular velocity meter, it is possible to form an anti-vibration filter without providing an HPF, and the anti-vibration effect immediately after panning / tilting is improved. In addition, since the control band can be expanded to the low frequency side, the image stabilization control effect is improved.

<Embodiment 2>
FIG. 8 shows a configuration of the imaging unit of the imaging apparatus 101 according to the second embodiment and functional blocks of image blur correction processing executed by the CPU 105. The difference between FIG. 2 according to the first embodiment and FIG. 8 according to the second embodiment is as follows.
(1) FIG. 8 does not include the HPF 109 and the angle 1 calculation unit 110 of FIG.
(2) In FIG. 2, the output of the angle 1 calculation unit 110 is input to the signal switching unit 118. Instead, in FIG. 8, the signal obtained by subtracting the angular velocity offset from the output value of the angle 2 calculation unit by the offset subtraction unit 114 is input to the signal switching unit 118.

  In the second embodiment, shake correction is performed using the output of the offset subtraction unit 114 before shooting, and shake correction is performed using the output of the angle 2 calculation unit 117 during shooting. The angle 2 calculation unit 117 calculates an angle including the noise offset component of the angular velocity meter 103. However, the angle offset is obtained by the offset 2 calculation unit 113, and the angle offset due to the offset effect of the angular velocity meter 103 is removed to some extent by the offset subtraction unit 114. Therefore, since the output of the offset subtraction unit 114 is an angle signal after the offset is removed, the output is calculated in the vicinity of zero. However, it is difficult to accurately determine the angular velocity offset and the angle offset, and some errors are included. Further, when a filter is used for calculating the offset or when calculating the average value of the signal, a phase delay or the like is added to the actual offset. Due to the influence of the offset calculation error, the output signal of the offset subtraction unit 114 is not used as it is during imaging, and image blur correction is performed by the output of the angle 2 calculation unit. However, in the state before photographing, there is little trouble even if there is a shake remaining in the image stabilization control due to the effect of a phase delay or the like, and a certain amount of image blur can be suppressed. In the first embodiment, the shake correction amount before imaging is obtained using the HPF. However, since the HPF is not used in the present embodiment, the shakeback effect immediately after panning is eliminated, and the image stabilization performance immediately after panning can be improved.

  According to the second embodiment described above, the HPF 109 and the integral filter (angle calculation unit 110) of the first embodiment can be reduced. Therefore, it is possible to perform the image stabilization control with high accuracy while avoiding the increase in the scale of the image stabilization processing circuit and the image stabilization processing program.

In the present embodiment, the so-called optical image stabilization in which the correction lens is moved in a plane perpendicular to the optical axis has been described. However, it is not limited to optical image stabilization, and the following configuration may be used.
A configuration in which shake correction is performed by moving the image sensor in a plane perpendicular to the optical axis.
A configuration based on electronic image stabilization that reduces the influence of shake by changing the cut-out position of each shooting frame output by the image sensor.
-A configuration that performs shake correction by combining multiple image stabilization controls.

(Other embodiments)
The present invention can also be realized by executing the following processing. That is, software (program) that realizes the functions of the above-described embodiments is supplied to a system or apparatus via a network or various storage media, and a computer (or CPU, MPU, etc.) of the system or apparatus reads the program. It is a process to be executed. In this case, the program and the storage medium storing the program constitute the present invention.

Claims (6)

An imaging device,
Shake detection means for detecting shake of the imaging device;
First computing means for computing a first shake correction amount for correcting image shake before imaging;
Second computing means for computing a second shake correction amount for correcting image shake during imaging;
Shake correction means for correcting image shake based on the first or second shake correction amount;
With
The first calculation means calculates the first shake correction amount using a high-pass filter,
The second calculation means includes
First offset calculating means for calculating a first offset component of the output of the shake detecting means;
Second offset calculating means for calculating a second offset component of the output of the shake detecting means;
Including
Subtraction amount calculation means for calculating a shake subtraction amount based on a signal obtained by subtracting the first offset component from the output of the shake detection means and a signal obtained by subtracting the second offset component from the second shake correction amount. Prepared,
The second calculation means calculates the second shake correction amount by integrating a signal obtained by subtracting the shake subtraction amount from the output of the shake detection means,
The imaging apparatus corrects image blur based on the first shake correction amount before imaging and corrects image blur based on the second shake correction amount during imaging. An imaging device,
Shake detection means for detecting shake of the imaging device;
First computing means for computing a first shake correction amount for correcting image shake before imaging;
Second computing means for computing a second shake correction amount for correcting image shake during imaging;
Shake correction means for correcting image shake based on the first or second shake correction amount;
With
The second calculation means includes
First offset calculating means for calculating a first offset component of the output of the shake detecting means;
Second offset calculating means for calculating a second offset component of the output of the shake detecting means;
Including
Subtraction amount calculation means for calculating a shake subtraction amount based on a signal obtained by subtracting the first offset component from the output of the shake detection means and a signal obtained by subtracting the second offset component from the second shake correction amount. Prepared,
The second calculation means calculates the second shake correction amount by integrating a signal obtained by subtracting the shake subtraction amount from the output of the shake detection means,
The first calculation means sets a signal obtained by subtracting the second offset component from the second shake correction amount as the first shake correction amount,
The imaging apparatus corrects image blur based on the first shake correction amount before imaging and corrects image blur based on the second shake correction amount during imaging.   The subtraction amount calculation means calculates the shake subtraction amount when a signal obtained by subtracting the second offset component from the second shake correction amount exceeds a predetermined threshold, and a difference between the signal and the predetermined threshold is large. The imaging apparatus according to claim 1, wherein the shake subtraction amount is increased. A method for controlling an imaging apparatus,
A shake detection unit that detects a shake of the imaging apparatus;
A first calculation step in which a first calculation means calculates a first shake correction amount for correcting image shake before imaging;
A second calculation step in which a second calculation means calculates a second shake correction amount for correcting image shake during imaging;
A correcting step for correcting image blur based on the first or second shake correction amount;
Have
In the first calculation step, the first calculation means calculates the first shake correction amount using a high-pass filter,
The second calculation step includes:
Calculating a first offset component of the output of the shake detection means;
Calculating a second offset component of the output of the shake detection means;
Including
The subtraction amount calculation means calculates a shake subtraction amount based on a signal obtained by subtracting the first offset component from the output of the shake detection means and a signal obtained by subtracting the second offset component from the second shake correction amount. A subtraction amount calculating step;
In the second calculation step, the second calculation means calculates the second shake correction amount by integrating a signal obtained by subtracting the shake subtraction amount from the output of the shake detection means,
In the correction step, the correction unit corrects image shake based on the first shake correction amount before imaging, and corrects image shake based on the second shake correction amount during imaging. A method for controlling the imaging apparatus. A method for controlling an imaging apparatus,
A shake detection unit that detects a shake of the imaging apparatus;
A first calculation step in which a first calculation means calculates a first shake correction amount for correcting image shake before imaging;
A second calculation step in which a second calculation means calculates a second shake correction amount for correcting image shake during imaging;
A shake correction step in which the correction unit corrects the image shake based on the first or second shake correction amount;
Have
In the second calculation step, the second calculation means includes:
Calculating a first offset component of the output of the shake detection means;
Calculating a second offset component of the output of the shake detection means;
Including
The subtraction amount calculation means calculates a shake subtraction amount based on a signal obtained by subtracting the first offset component from the output of the shake detection means and a signal obtained by subtracting the second offset component from the second shake correction amount. A subtraction amount calculating step;
In the second calculation step, the second calculation means calculates the second shake correction amount by integrating a signal obtained by subtracting the shake subtraction amount from the output of the shake detection means,
In the first calculation step, the first calculation means sets a signal obtained by subtracting the second offset component from the second shake correction amount as the first shake correction amount,
In the correction step, the correction unit corrects image shake based on the first shake correction amount before imaging, and corrects image shake based on the second shake correction amount during imaging. A method for controlling the imaging apparatus.   The program for making a computer perform each step of the control method of Claim 4 or 5.

Images