JP6448709B2 - Image shake correction apparatus, image shake correction method, imaging apparatus, and optical apparatus - Google Patents

Description

  The present invention relates to an image blur correction apparatus and an imaging apparatus that correct image blur due to camera shake and the like, and more particularly to a technique for smoothly correcting image blur during macro shooting.

  In a camera equipped with an image shake correction device, in order to enable shooting without image shake, camera shake due to camera shake or the like is detected, and an image shake correction lens (hereinafter referred to as a correction lens) according to the detected value. Drive. 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. Image shake is suppressed by a vibration detection unit (angular velocity meter or the like) that obtains a detection result such as an angular velocity and a drive control unit that drives a correction member (correction lens or the like) based on the calculation processing result.

  By the way, in the case of shooting at a close distance (shooting conditions with a high shooting magnification), there is a shake that cannot be detected only by the angular velocity meter. This is a so-called parallel shake applied in a direction parallel to or perpendicular to the optical axis of the camera, and image degradation due to the influence cannot be ignored. For example, under a condition where the subject is photographed close to the subject by macro photography up to about 20 cm, or under a condition where the focal length of the imaging optical system is very large (for example, 400 mm) with respect to a subject about 1 m away from the camera, It is necessary to detect and correct the parallel shake.

  Patent Document 1 discloses a technique for detecting a parallel shake by an accelerometer, obtaining a parallel shake from the second-order integration of the accelerometer, and driving a shake correction unit together with an output of an angular velocity meter provided separately. In this case, the output of the accelerometer is easily affected by environmental changes such as disturbance noise and temperature change, and the unstable factor is further expanded by the second order integration, so that it is difficult to correct parallel shake with high accuracy. Patent Document 2 discloses a technique for obtaining parallel shake as an angular shake when the center of rotation is at a location away from the camera. The correction value and angle using the rotational radius of the angular shake are obtained from the outputs of the angular velocity meter and the accelerometer, and shake correction is performed. By obtaining the center of rotation only in a frequency band that is not easily affected by disturbances, it is possible to reduce the influence on the correction due to the instability factor of the accelerometer.

JP 7-225405 A JP 2010-259592 A

  Generally, if the cut-off frequency of the filter used for image blur correction is set low and the frequency band is widened, low-frequency component shake correction for shakes of the photographer's body, etc. can be made and the performance is improved. . However, when the frequency band on the low frequency side of the filter is expanded, another problem may occur. That is, since the movable range of the correction member is finite, the position of the correction member reaches the limit of the movable range, and as a result, the performance may be deteriorated.

Further, the parallel shake correction using the radius of rotation disclosed in Patent Document 2 has a problem that it is difficult to accurately perform correction in a low frequency band. The rotation radius is calculated by specifying the rotation radius in a predetermined frequency band, and the frequency to be extracted is mainly set between 1 Hz and 10 Hz. Therefore, there may be a case where the rotation radius cannot be accurately obtained for a vibration of 1 Hz or less. In addition, when the actual turning radius at a shake of 1 Hz or less is smaller than the calculated turning radius, an extra shake correction different from the actual parallel shake can be performed in the correction at a low frequency band of 1 Hz or less. There is also sex. In the case of the above conditions, there is a concern that the image blur correction performance is deteriorated due to overcorrection accompanying the expansion of the frequency band on the low frequency side of the filter.
An object of the present invention is to improve the accuracy of image blur correction in an image blur correction apparatus.

In order to solve the above problems, an image shake correction apparatus according to the present invention uses an angular velocity detection signal output from a first shake detection unit and an acceleration detection signal output from a second shake detection unit. First calculation means for calculating a correction coefficient for translational shake correction, second calculation means for calculating a translational shake correction amount using the correction coefficient and the angular velocity detection signal, and the angular velocity detection signal A third calculation unit that uses the second calculation unit to calculate the angular shake correction amount , a filter that is used for the calculation of the second calculation unit and that selects a frequency band in which the parallel shake correction is performed, and the angular shake correction amount. , and a control means for controlling the shake correcting means in accordance with the parallel shake correction amount calculated by the second arithmetic means in a frequency band of the selected the parallel shake correction by said filter. The control unit controls the shake correction unit according to the angular shake correction amount instead of the parallel shake correction amount before imaging, and during the imaging period, when the angular shake correction amount and the imaging magnification are larger than a threshold value, before narrowing the low frequency band of distichum frequency band, wherein if the photographing magnification is equal to or less than the threshold value, before distichum frequency band low frequency band the calculated at determined frequency band by increasing the parallel shake correction The shake correction means is controlled according to the amount.

  According to the present invention, it is possible to improve the accuracy of image blur correction in an image blur correction apparatus.

1 is a top view of a camera equipped with an image shake correction apparatus according to the present invention. 1 is a side view of a camera equipped with an image shake correction apparatus according to the present invention. 1 is a block diagram of an image shake correction apparatus according to a first embodiment of the present invention. It is explanatory drawing of the rotation center of shake in this invention. It is a block diagram for demonstrating the image blurring correction apparatus which concerns on 1st Embodiment of this invention. It is a wave form diagram explaining shake correction amount calculation which concerns on 1st Embodiment of this invention. It is explanatory drawing of the filter gain characteristic which concerns on 1st Embodiment of this invention. 5 is a flowchart for explaining the operation of the image shake correction apparatus according to the first embodiment of the present invention. FIG. 6 is a block diagram of an image shake correction apparatus according to a second embodiment of the present invention. It is a block diagram of the image blur correction apparatus which concerns on 3rd Embodiment of this invention.

  Embodiments of the present invention will be described with reference to the accompanying drawings. The present invention includes an optical device such as an interchangeable lens and a lens barrel attached to a digital single lens reflex camera, a photographing device such as a digital video camera, a surveillance camera, and a Web camera, and a photographing device such as a mobile phone and a tablet terminal. Applicable to electronic equipment.

[First Embodiment]
1 and 2 are a plan view and a side view showing an image pickup apparatus equipped with an image shake correction apparatus according to the first embodiment of the present invention. The image shake correction apparatus mounted on the image pickup apparatus is adapted to shakes indicated by arrows 103p and 103y (hereinafter referred to as angular shakes) and shakes indicated by arrows 104p and 104y (hereinafter referred to as parallel shakes) with respect to the optical axis 102. Correct image blur.
The imaging apparatus 101 includes an operation switch (release switch) 105 using a release button, and a camera CPU (central processing unit) 106 executes a control program to perform various processes such as image blur correction. The image sensor 107 photoelectrically converts subject light imaged by the imaging optical system.
Angular velocity detection means (hereinafter referred to as angular velocity meters) 108p and 108y detect angular fluctuations around arrows 108pa and 108ya, respectively. The angular shake is a shake around an axis orthogonal to the optical axis of the imaging optical system. The first axis orthogonal direction orthogonal to the optical axis is the pitch direction p, and the first optical axis and the first axis are orthogonal to each other. The direction around the axis 2 is the yaw direction y. Further, acceleration detection means (hereinafter referred to as accelerometers) 109p and 109y detect parallel shakes indicated by arrows 109pa and 109ya, respectively. This parallel shake is a shake in a direction perpendicular to the optical axis of the imaging optical system. The arrow 109pa represents the vertical direction, and the arrow 109ya represents the horizontal direction.

The lens driving unit 110 drives the correction lens 111 freely in the directions of the arrows 110y and 110p in FIGS. 1 and 2 to perform image blur correction in consideration of both angular shake and parallel shake. The outputs of the angular velocity meters 108p and 108y and the accelerometers 109p and 109y are input to the camera CPU 106. The camera CPU 106 performs image blur correction by controlling the lens driving unit 110 based on these outputs.
In the present embodiment, so-called optical image stabilization, in which the correction lens 111 is moved in a plane perpendicular to the optical axis based on the calculated correction amount, is employed as the image blur correction unit. However, the correction method based on the correction amount is not limited to optical image stabilization. For example, a form that performs image blur correction by moving the image sensor in a plane perpendicular to the optical axis, or an electronic defense that reduces the effect of shake by changing the image cutout position of each shooting frame output by the image sensor. It may be in the form of shaking. Alternatively, the object of the present invention can also be achieved by performing correction with a combination thereof.

  FIG. 3 is a block diagram showing the image blur correction apparatus according to the first embodiment of the present invention. Components in the camera CPU 106 are shown as functional blocks. FIG. 3 shows only the configuration relating to shake (pitch direction: the direction of arrows 103p and 104p in FIG. 2) generated in the vertical direction of the camera. A similar configuration is also provided for shake (yaw direction: arrows 103y and 104y directions in FIG. 1) generated in the horizontal direction of the camera. Since these are basically the same configuration, only the configuration related to the pitch direction will be described below.

式（１）の左辺にてＴはＨＰＦの時定数を表し、式（１）の右辺にてＴはＬＰＦの時定数を表す。したがって、角度算出フィルタにはＨＰＦが含まれており、角度算出フィルタの前段に別のＨＰＦを接続すると、角速度計１０８ｐの出力から角度算出までのフィルタについては、２次のＨＰＦが構成されることになる。 First, a first calculation unit (see 301, 303, and 315) that calculates the first image shake correction amount based on the angular shake detection signal will be described.
The angular velocity meter 108p outputs an angular velocity signal to the camera CPU 106 as an angular shake detection signal. The angular velocity signal is input to the HPF integration filter 301, integrated after the DC (direct current) component is cut by an HPF (high pass filter), and converted into an angle signal.
The correction angle calculation filter includes an integrator (left side “1 / s” in equation (1)) and HPF (left side “Ts / (Ts + 1)” in equation (1) as shown on the left side in equation (1) below. ). This is the same as the expression obtained by multiplying the low-pass filter (LPF) with the time constant T by the time constant T, as shown on the right side of the following expression (1).

On the left side of equation (1), T represents the time constant of HPF, and on the right side of equation (1), T represents the time constant of LPF. Therefore, the angle calculation filter includes an HPF, and when another HPF is connected to the preceding stage of the angle calculation filter, a second-order HPF is configured for the filter from the output of the angular velocity meter 108p to the angle calculation. become.

  The output of the HPF integration filter 301 is input to the sensitivity adjustment unit 303. Position information 302 of the zoom lens and focus lens is input to the sensitivity adjustment unit 303. The position information 302 is acquired by a known detection method by a position detection unit (not shown) provided in the lens barrel. The sensitivity adjustment unit 303 amplifies the output of the HPF integration filter 301 based on the focal length and the imaging magnification obtained from the position information 302 of the zoom lens and the focus lens, and calculates a correction amount (hereinafter referred to as shake correction amount 1). To do. The shake correction amount 1 is input to the signal switching unit 315. Here, the imaging magnification means a ratio between the size of the subject image taken through the lens (the size of the image on the imaging surface) and the actual size of the subject.

Next, a second calculation unit (see 304 to 309 and 315) that calculates the second image shake correction amount based on the angular shake detection signal will be described.
The output of the angular velocity meter 108p is also input to the reference value subtraction unit 304. The reference value subtraction unit 304 calculates an offset component of the angular velocity meter 108p, subtracts the calculated offset component from the angular velocity, and outputs the result to the integration filter 305. The reference value subtraction unit 304 calculates an offset reference value of the output of the angular velocity meter 108p. This offset reference value is an angular velocity offset component added as detection noise to the output of the angular velocity meter 108p. For example, the output value of the angular velocity meter 108p when the angular velocity after passing through the HPF or the amplitude of the angular acceleration obtained by differentiating the angular velocity is smaller than a predetermined threshold is acquired. An angular velocity offset, which is a DC component, is calculated by a method of smoothly connecting it with an LPF whose cut-off frequency is set to be very small. The calculated offset reference value is held until the angular velocity meter 108p is turned off.

The output of the integration filter 305 is input to the signal switching unit 308 and the HPF 306. The HPF 306 cuts low frequency components and outputs them to the signal switching unit 308. The signal switching unit 308 acquires the shooting magnification from the shooting magnification calculation unit 307. The shooting magnification calculator 307 acquires the zoom lens and focus lens position information 302 and calculates the shooting magnification. The signal switching unit 308 compares the imaging magnification with the threshold, selects the output of the integration filter 305 or the output of the HPF 306 according to the size of the imaging magnification, and outputs the selected output to the sensitivity adjustment unit 309. The output of the HPF 306 selected when the imaging magnification exceeds the threshold corresponds to the first correction amount before changing the frequency band of the correction angle calculation filter. The output of the integration filter 305 selected when the photographing magnification is equal to or smaller than the threshold corresponds to the second correction amount after changing the frequency band.
When the HPF 306 is included in the angle calculation filter, the filter from the output of the angular velocity meter 108p to the angle calculation has a secondary HPF configuration as described above. For this reason, the phase is greatly advanced in the low waveband of camera shake (˜1 Hz), and the image blur correction effect is reduced.

In addition, in the panning or tilting operation, it is affected by the filter characteristics including the secondary HPF. When a large shake such as panning occurs, a low-frequency component having a large amplitude is attenuated, and, for example, a signal in a direction opposite to the panning direction is generated at the end of panning (so-called a swing-back phenomenon). This signal then slowly converges to zero. However, when image blur correction is performed based on this signal, the correction amount is calculated using a signal different from the actual camera shake. Therefore, the accuracy of camera shake correction may be reduced.
Therefore, a filter configuration that does not include the HPF 306 is preferable in terms of image blur correction, but in this case, the gain in the low frequency band becomes large. Accordingly, when the correction range is insufficient in the limited movable range of the image blur correction member, it may be difficult to perform appropriate image blur correction. Also, the image blur correction amount is larger when the photographing magnification is large than when the photographing magnification is small. For this reason, there is a possibility that the correction range is insufficient when the photographing magnification is large. Therefore, the signal switching unit 308 compares the shooting magnification with a predetermined threshold value set in advance. When the photographing magnification is equal to or smaller than the threshold value, the image blur correction effect is improved by calculating the blur correction amount using the output of the integration filter 305. When the photographing magnification is larger than the threshold value, an image blur correction amount is calculated using the output of the HPF 306, so that an appropriate image blur correction is performed without the correction member exceeding the movable range.

The following explains the parallel shake correction processing and the reason why the image shake correction amount increases when the photographing magnification is large.
The sensitivity adjustment unit 309 performs conversion processing into an image blur correction amount in consideration of angular shake and parallel shake. The output of the angular velocity meter 108p is input to an HPF phase adjustment filter (hereinafter referred to as an HPF phase adjustment unit) 310. The HPF phase adjustment unit 310 cuts the DC component superimposed on the output of the angular velocity meter 108p and adjusts the phase of the signal. The cutoff frequency here is matched with the cutoff frequency of the HPF of the HPF integration filter 311 described later, and is adjusted so that the frequency characteristics match. As for the output of the HPF phase adjustment unit 310, only the frequency component of a predetermined band is extracted by an angular velocity meter BPF (band pass filter) unit 312.

The accelerometer 109a outputs a parallel shake detection signal. The output of the accelerometer 109p is input to the HPF integration filter 311. After the HPF constituting the filter cuts the DC component, the integration filter converts it into a speed signal. The cutoff frequency of the HPF at this time is set in accordance with the HPF frequency characteristic of the HPF phase adjustment unit 310 as described above. As for the output of the HPF integration filter 311, only the frequency component in a predetermined band is extracted by the accelerometer BPF unit 313.
The outputs of the angular velocity meter BPF unit 312 and the accelerometer BPF unit 313 are input to the comparison unit 314, and a correction amount (correction coefficient) set for the sensitivity adjustment unit 309 is calculated. The correction amount calculation process in the comparison unit 314 will be described later.
The sensitivity adjustment unit 309 acquires the position information 302 of the zoom lens and the focus lens and the output of the signal switching unit 308. The sensitivity adjustment unit 309 amplifies the output of the signal switching unit 308 based on the focal length and the shooting magnification obtained from the position information 302 of the zoom lens and the focus lens and the correction coefficient from the comparison unit 314. The sensitivity adjustment unit 309 calculates a correction amount (hereinafter referred to as shake correction amount 2) and outputs the calculation amount to the signal switching unit 315.

Next, the correction coefficient calculation process output from the comparison unit 314 and the shake correction amount calculation process in the sensitivity adjustment unit 309 will be described.
FIG. 4 shows the angular shake 103p and the parallel shake 104p applied to the camera. In the photographic lens of the imaging apparatus 101, Y is the magnitude of the parallel shake 104p at the principal point position of the imaging optical system, and θ is the magnitude of the angular shake 103p. The relationship with the rotation radius L (402p) when the rotation center O (401p) is determined is expressed by the following equation.
[Formula 2]
Y = L × θ (2)
V = L × ω (3)
A = L × ωa (4)
ω is an angular velocity, ωa is an angular acceleration, V is a velocity, and A is an acceleration. The rotation radius L (402p) is a distance from the rotation center 401p to the accelerometer 109p.
In the equation (2), the value of the rotation radius L is calculated from the displacement Y obtained by second-order integration of the output of the accelerometer 109p and the angle θ obtained by first-order integration of the output of the angular velocity meter 108p. In Expression (3), the value of the rotation radius L is calculated from the velocity V obtained by first-order integration of the output of the accelerometer 109p and the angular velocity ω obtained from the output of the angular velocity meter 108p. In Expression (4), the value of the rotation radius L is calculated from the acceleration A obtained from the output of the accelerometer 109p and the angular acceleration ωa obtained by first-order differentiation of the output of the angular velocity meter 108p. In any method, the rotation radius L can be obtained.

From the parallel shake Y at the principal point position of the image pickup optical system, the shake angle θ of the image pickup optical system, the focal length f of the image pickup optical system, and the shooting magnification β, the shake δ generated on the image pickup surface is obtained by the following equation (5). It is done.
[Formula 3]
δ = (1 + β) × f × θ + β × Y (5)
F and β in the first term on the right side of Expression (5) are obtained from the positions of the zoom lens and focus lens of the imaging optical system, and the imaging magnification and focal length obtained from them. Further, the deflection angle θ is obtained from the integration result of the angular velocity meter 108p. Therefore, the angular shake correction can be performed as described with reference to FIG. Further, the second term on the right side of the equation (5) is obtained from the second-order integral value Y of the accelerometer 109p and the photographing magnification β obtained from the zoom and focus positions, so parallel shake correction is performed according to the information. It can be carried out.

However, in this embodiment, image blur correction is performed on a shake δ that is rewritten from Equation (5) as in Equation (6) below.
[Formula 4]
δ = (1 + β) × f × θ + β × L × θ = ((1 + β) × f + β × L) × θ (6)
That is, for the parallel shake, the product of the rotational radii L and θ is used without using the parallel shake displacement Y directly obtained from the accelerometer 109p. The shake δ is calculated from the rotation radius L obtained by the above formula (2), formula (3), or formula (4), the integration result (θ) of the output of the angular velocity meter 108p, the focal length f, and the imaging magnification β. Image blur correction.

FIG. 5 is a block diagram showing correction amount (correction coefficient) calculation processing in the comparison unit 314 shown in FIG.
The rotation radius calculation unit 501 of the comparison unit 314 acquires the outputs of the angular velocity meter BPF unit 312 and the accelerometer BPF unit 313, and calculates the rotation radius L using the following equation (7).
[Formula 5]
L = V / ω (7)
The rotation radius L may be calculated from the amplitude of the waveform sampled at a predetermined time (sampling time) interval. Furthermore, the update timing of the rotation radius L may be performed for each calculated moment, or a time-series averaging process or a process of cutting high-frequency components by LPF may be performed.

  The calculated rotation radius L is arithmetically processed using the upper limit value set by the limit processing unit 502. The limit processing unit 502 fixes the output value to the upper limit value when the output value of the turning radius calculation unit 501 is equal to or greater than the upper limit value, and outputs the output value as it is when the output value is less than the upper limit value. The output value of the limit processing unit 502 is processed by the correction signal rectification unit 503. The correction signal rectification unit 503 rectifies the output values of the limit processing unit 502, and performs signal processing so that a stepwise rapid change does not occur in the correction signal. For example, signal rectification is performed by cutting high-frequency components with an LPF. The cutoff frequency of the LPF is set to a low frequency of 0.5 Hz or less, for example. Alternatively, a calculation unit that calculates a moving average for a predetermined period is provided. The output of the correction signal rectification unit 503 is output to the sensitivity adjustment unit 309 as a signal indicating the final rotation radius used for parallel shake image blur correction.

  The image blur correction signal output from the sensitivity adjustment unit 309 is input to the signal switching unit 315 (see FIG. 3). The output of the sensitivity adjustment unit 303 and the output of the release SW (switch) 105 are simultaneously input to the signal switching unit 315. Depending on the state of the release SW 105, the output of the sensitivity adjustment unit 303 or the sensitivity adjustment unit The output 309 is selected and output to the drive unit 112. The image blur correction signal from the sensitivity adjustment unit 309 indicates a correction value without HPF processing when the imaging magnification is equal to or less than a predetermined threshold in the signal switching unit 308. When the HPF process is not performed, the calculated correction value includes an offset due to the influence of the output noise component of the angular velocity meter 108p. Hereinafter, how to use a correction value including an offset for image blur correction in the case of a configuration in which HPF processing is not performed will be described.

  FIG. 6 is a waveform diagram for explaining shake correction processing during imaging and during periods other than during imaging. A waveform 601 represents a temporal change in the shake correction amount 2 calculated by the sensitivity adjustment unit 309. A waveform 602 shows a temporal change of the shake correction amount 1 calculated by the sensitivity adjustment unit 303. Since no HPF is provided for the shake correction amount 2, as shown by a waveform 601, it gradually moves away from near zero at the time of power-on as time passes. For example, when the offset temperature drift of the angular velocity meter occurs in the period 604, the shake correction amount 2 away from the zero center due to the influence of the temperature drift is calculated over time.

FIG. 7 shows the frequency-gain characteristics of the angle calculation filter. A graph line 701 shows the characteristics of only the integral filter (corresponding to 1 / s of the above formula (1)). A graph line 702 indicates the integral and HPF filter characteristics used for angle calculation. The graph line 702 has a flat characteristic in the low frequency band, and a gain characteristic corresponding to the offset of the angular velocity remains in the angle output. Therefore, in the period 604 in FIG. 6, the shake correction amount 2 moves away from the zero center as the angular velocity offset increases due to the temperature drift effect of the angular velocity meter.
Since the shake correction amount 1 shown in the waveform 602 is calculated by adding the HPF 306 to the integration filter (integration and HPF), the characteristic shown by the graph line 703 is obtained by adding the HPF 306 characteristic to the characteristic of the graph line 702. In the graph line 703, as can be seen from the fact that the gain decreases in the low frequency region, the offset component included in the output of the angular velocity meter 108p can be removed, and the angle is calculated at the zero center. However, since the HPF 306 is used, the image blur correction effect immediately after a large shake such as panning or tilting is weakened due to the swing back phenomenon.

  Therefore, an appropriate correction effect can be obtained by performing the image blur correction using the shake correction amount 2 rather than the shake correction amount 1. However, in the case of the shake correction amount 2, since the filter characteristic indicated by the graph line 702 in FIG. 7 is obtained, the gain does not attenuate in the low frequency region and becomes a flat gain characteristic. That is, the shake correction amount 2 is calculated while including the offset component of the angular velocity. Therefore, if image blur correction is always performed based on the signal of the waveform 601, the offset of the shake correction amount 2 also increases due to the temperature drift of the angular velocity offset. When the movable range of the correction member becomes insufficient as time passes, there is a possibility that control is not possible at the movable end.

  Therefore, in this embodiment, it is detected whether or not imaging is being performed, and image blur correction is performed using a signal indicated by a waveform 603 in FIG. 6 during the imaging period. Also, during a period other than the imaging period, for example, during EVF display or AF (automatic focus adjustment) / AE (automatic exposure) operation in the preparation period before imaging, image blur correction is performed using the signal shown by the waveform 602 in FIG. Is done. Thereby, during the imaging period, the image blur correction effect is improved by the filter characteristic that is expanded to the low frequency region. In addition, a certain amount of image blur correction effect can be ensured even outside the imaging period. Therefore, the accuracy of the AF / AE operation is improved, and the ease of the photographer's framing operation is improved.

  In FIG. 6, the imaging period is a period from the imaging start time 605 to the imaging end time 606. If image blur correction is performed in accordance with the signal indicated by the waveform 602 during the imaging period, a shakeback phenomenon may occur due to the influence of the HPF 306, such as immediately after panning. In this case, when image blur correction different from actual camera shake is performed, the image blur correction effect is reduced. In the present embodiment, processing for calculating the difference between the waveform 601 and the waveform 602 as an offset at the timing of the imaging start time 605 is performed. A signal indicated by a waveform 603 obtained by subtracting an offset from the waveform 601 is used during the imaging period. When the imaging process ends at the imaging end time 606, a signal for returning to the waveform 602 at a constant speed is added to the waveform 603. The addition process is performed until the waveform 603 matches the waveform 602.

Through the above processing, image blur correction is executed during imaging according to the correction amount calculated with the filter configuration that does not use the HPF 306. Therefore, there is no shake back phenomenon immediately after panning or tilting, and the image blur correction effect is improved by expanding the filter characteristics to the low frequency region.
In addition, the correction coefficient (see equations (6) and (7)) related to the parallel shake correction is used only for the shake correction amount 2. This is because the movable range of the correction member is limited, and if the parallel shake correction is always performed, the movable range for shake correction is exceeded. The purpose of enhancing the image blur correction effect is during still image shooting. In addition to this, a state in which the focus operation is not fixed as in the AF operation in the state before photographing can frequently occur. Therefore, in order to prevent an erroneous calculation of the shooting magnification β during the AF operation or the like, an appropriate image blur correction effect is obtained only during the imaging period. That is, only in the shake correction amount 2 used during the imaging period, the correction coefficient (L) related to the parallel shake correction depends on the angle θ.

  The movable range of the correction member is limited, and when the imaging magnification β is large, the shake δ increases as shown in the equation (6), so the shake correction amount also increases. In the case of a filter configuration that does not use the HPF 306, the correction gain in the low frequency band is increased, so that the low frequency followability is increased. For this reason, there is a possibility that the shake correction amount immediately exceeds the movable range even in the control only during imaging. Further, in the parallel shake correction process using the rotation radius L, it is difficult to accurately perform the image shake correction in the low frequency band. The rotation radius L is calculated by specifying a certain frequency band. For example, the set frequency is mainly between 1 Hz and 10 Hz. Therefore, there may be a case where the rotation radius L is not accurately calculated for a vibration of 1 Hz or less. When the rotation radius relating to the shake of 1 Hz or less is smaller than the set rotation radius, image blur correction different from the actual parallel shake may be performed in the low frequency band of 1 Hz or less.

  Therefore, in this embodiment, when the imaging magnification is larger than the threshold value, image blur correction is performed within the movable range, and in order to prevent erroneous control in the low frequency band, image blur correction control with filter characteristics including the HPF 306 is performed. Switch. The signal switching unit 308 selects the output of the integration filter 305 or the output through the integration filter and the HPF 306 according to the magnitude of the imaging magnification and outputs the selected value to the sensitivity adjustment unit 309.

Next, image blur correction processing according to the present embodiment will be described with reference to the flowchart of FIG. This processing starts when the main power of the camera is turned on, and is executed by the camera CPU 106 at a constant sampling period.
When the image blur correction subroutine is started in S801, first, in S802, processing for taking in the outputs of the angular velocity meter 108 and the accelerometer 109 is performed. The next step S803 is processing for determining whether or not image blur correction is possible. If the image blur correction is possible, the process advances to step S804. If the image blur correction is not possible, the process advances to step S817. In the determination processing of S803, it is determined that image blur correction is not possible from the time when the power is turned on until the outputs of the angular velocity meter 108p and the accelerometer 109p are stabilized. When a certain time has passed and the outputs of the angular velocity meter 108 and the accelerometer 109 are stabilized, it is determined that image blur correction is possible. As a result, it is possible to prevent performance degradation due to image blur correction performed in an unstable output value immediately after power-on.

  In S804, the HPF integration filter 301 calculates an angle (hereinafter referred to as angle 1) from the output (angular velocity) of the angular velocity meter 108p. In step S805, the sensitivity adjustment unit 303 amplifies the angle 1 based on the focal length and the shooting magnification obtained from the position information 302 of the zoom lens and the focus lens, and calculates the shake correction amount 1. In S806, the reference value subtraction unit 304 and the integration filter 305 calculate an angle (hereinafter referred to as an angle 2). In step S807, the HPF 306 calculates the HPF-added angle 2 obtained by performing HPF processing on the angle 2.

In step S <b> 808, the comparison unit 314 calculates the rotation radius L based on the outputs from the angular velocity meter 108 and the accelerometer 109. In step S809, the shooting magnification calculation unit 307 acquires the zoom lens and focus lens position information 302 and calculates the shooting magnification β.
In step S810, the signal switching unit 308 determines whether or not the shooting magnification β is larger than a predetermined threshold value Thresh. If the shooting magnification β is larger than the threshold value Thresh, the process proceeds to S811. In S811, the signal switching unit 308 selects the angle with HPF 2 that is the output of the HPF 306, and the sensitivity adjustment unit 309 calculates the shake correction amount 2 from Equation (6). If the shooting magnification β is equal to or smaller than the threshold Thresh in S810, the process proceeds to S812. In S812, the signal switching unit 308 selects the angle 2 that is the output of the integration filter 305, and the sensitivity adjustment unit 309 calculates the shake correction amount 2 from Equation (6).

In S813, it is determined based on the signal from the release SW 105 whether or not imaging is being performed. If the result of determination is that imaging is in progress, processing proceeds to S814, where the signal switching unit 315 selects shake correction amount 2 and outputs it to the drive unit 112. In this case, image blur correction is performed using a signal indicated by a waveform 603 in FIG. If it is determined in S813 that the image is not being captured, the process proceeds to S815, where the signal switching unit 315 selects the shake correction amount 1 and outputs it to the driving unit 112. Image blur correction is performed using a signal indicated by a waveform 602 in FIG.
In step S816, the driving unit 112 drives the correction lens based on the image blur correction target value. In step S817, when the image blur correction is not possible in step S803, the driving unit 112 stops driving the correction lens.
The image blur correction subroutine is completed as described above, and the process waits until the next sampling period comes.

  In the present embodiment, image blur correction is performed by selecting whether the filter configuration includes the HPF 306 or the filter configuration not including the HPF 306 according to the magnitude of the imaging magnification β. As a result, it is possible to realize highly accurate image blur correction for angular shake and parallel shake within a limited movable range of the correction member.

[Second Embodiment]
Next, a second embodiment of the present invention will be described. In the second embodiment, the same components as those in the first embodiment are denoted by the same reference numerals as those used in the first embodiment, and detailed description thereof is omitted. Differences are mainly described. Omitting such description is the same in the embodiments described later.
FIG. 9 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 camera CPU 106.
Differences in configuration between FIG. 3 and FIG. 9 are as follows.
In FIG. 9, the HPF 306 and the signal switching unit 308 in FIG. 2 are deleted.
In FIG. 3, the output of the signal switching unit 308 is input to the sensitivity adjustment unit 309, but the output of the imaging magnification calculation unit 307 is input to the filter table 901 in FIG. 9. The output of the filter table 901 is input to the integration filter 305 together with the output of the reference value subtraction unit 304, and the output of the integration filter 305 is sent to the sensitivity adjustment unit 309.

  In the first embodiment, the HPF 306 is provided, the angle 2 with HPF or the angle 2 without HPF processing is selected according to the magnitude of the imaging magnification β, and the shake correction amount is calculated. In the second embodiment, no HPF is provided, and a filter table 901 that sets the cutoff frequency fc of the integration filter 305 according to the magnitude of the imaging magnification β is provided. By setting the cut-off frequency fc from the filter table 901 in the integral filter 305, a process for calculating the angle and calculating the image blur correction amount is performed.

  The filter table 901 is set to increase the cut-off frequency fc stepwise or continuously as the photographing magnification β increases. As a result, it is possible to prevent the correction member from immediately exceeding the movable range when the photographing magnification β is increased, and preventing an appropriate correction effect from being obtained due to the inability to perform correction. Further, as in the first embodiment, instead of determining whether or not the HPF is used with a predetermined photographing magnification as a boundary, in the present embodiment, it is gradually changed according to the size of the photographing magnification. A frequency band is set. Therefore, appropriate image blur correction can be performed according to the photographing magnification. Further, since it is not necessary to provide the HPF 306 and the signal switching unit 308, accurate image blur correction can be performed while avoiding an increase in the size of the processing circuit and processing program.

In FIG. 9, the HPF 306 is not provided, and the correction filter is configured only by the integration filter 305. Not only this but the filter structure which provided HPF306 like 1st Embodiment may also be the form which sets the cutoff frequency fc of HPF306 with the filter table 901. FIG. Also in this case, in the angle calculation and image shake correction amount calculation processing, it is possible to gradually change the frequency band according to the size of the shooting magnification and perform appropriate image shake correction according to the shooting magnification.
According to the present embodiment, image blur correction is performed by changing the cutoff frequency of the correction filter according to the size of the photographing magnification, thereby preventing angular shake and parallel shake within a limited movable range of the correction member. High-accuracy image blur correction can be realized.

[Third Embodiment]
Next, a third embodiment of the present invention will be described.
FIG. 10 shows a configuration of the imaging unit of the imaging apparatus 101 according to the third embodiment and functional blocks of image blur correction processing executed by the camera CPU 106. In the present embodiment, unlike the first and second embodiments, instead of the accelerometer 109, a motion vector is calculated from an image signal output from the image sensor 107 and image blur correction is performed.

  The image sensor 107 outputs an image signal by photoelectrically converting reflected light from the subject into an electrical signal. The image signal output from the image sensor 107 is input to the captured image capturing unit 1001 and converted into a digital signal for each set frame rate. The image information converted into the digital signal is input to the motion vector detection unit 1002. Here, the motion vector is calculated from the relative shift information between the images by comparing the pre-stored image of the previous frame with the current image, that is, two images that are temporally continuous. The image information to be extracted may be information on the entire image or a part of the image. Alternatively, the image may be divided into several areas, the motion information may be calculated by comparing the respective pieces of image information in the divided areas, and the optimum motion vector may be selected from the calculated motion vectors. For the application of the present embodiment, the motion vector calculation process is not limited.

The motion vector output from the motion vector detection unit 1002 is sent to the image plane angular velocity conversion unit 1003, and is equivalent to the angular velocity ω from the following equation (8) from the shake pixel amount σ (ie, shake pixel velocity) per frame rate and the cell pitch α. Converted to the value of.
[Formula 6]
ω = (σ × α) / ((1 + β) × f) (8)
The shake pixel velocity σ includes not only the angular shake amount but also other shake factors such as the parallel shake amount. Here, the image shift detected by the motion vector detection unit 1002 is considered to be based only on the angular shake amount. Let's hesitate to perform image blur correction.

The shake angular velocity from the image plane angular velocity conversion unit 1003 is input to a gain (a coefficient is denoted as Kp) unit 1004. After the shake angular velocity is multiplied by the gain coefficient Kp set by the gain table 1006, the integration filter 1005 calculates the filtered signal in accordance with the cutoff frequency fc set by the filter table 1007. The gain table 1006 and the filter table 1007 are input with the shooting magnification β from the shooting magnification calculator 307, respectively. The gain table 1006 increases the parallel shake correction effect by increasing the gain coefficient Kp as the value of the imaging magnification β increases. Note that the value of the gain coefficient Kp is set to a value smaller than 1, and when the photographing magnification β is small, it is set to a smaller value in order to prevent overcorrection. Further, the filter table 1007 increases the cutoff frequency fc of the integral filter 1005 as the value of the imaging magnification β increases. That is, when the imaging magnification β is large, the image blur correction effect in the low frequency band is weakened. However, since the gain coefficient Kp is set to a large value by the gain table 1006, the image blur correction effect particularly in the range of 1 to 10 Hz is greatly improved.
As in the case of the first embodiment, a filter configuration using the HPF 306 may be used, and the cut-off frequency fc of the HPF 306 may be set using the filter table 1007 to calculate the angle and calculate the image blur correction amount. . In this case, since the frequency band is gradually changed in accordance with the size of the shooting magnification β, appropriate image blur correction can be performed according to the shooting magnification.

  According to the present embodiment, image blur correction is performed by calculating a motion vector from a captured image signal output from the image sensor 107. By changing the cutoff frequency of the correction filter in accordance with the size of the photographing magnification, it is possible to realize high-accuracy image blur correction for angular shake and parallel shake within a limited movable range of the correction member.

101 Imaging device 106 CPU
107 Image sensor 108y, 108p Angular velocity meter 109y, 109p Accelerometer 110 Lens drive unit 111 Shake correction unit

Claims (8)

An image shake correction device,
First calculation means for calculating a correction coefficient for parallel shake correction using an angular velocity detection signal output from the first shake detection means and an acceleration detection signal output from the second shake detection means;
Second computing means for computing a translational shake correction amount using the correction coefficient and the angular velocity detection signal;
Third computing means for computing an angular shake correction amount using the angular velocity detection signal;
A filter that is used for calculation by the second calculation means and that selects a frequency band in which the parallel shake correction is performed;
With said rotational shake correction amount, and control means for controlling the shake correcting means in accordance with the parallel shake correction amount calculated by the second arithmetic means in a frequency band of the selected the parallel shake correction by said filter ,
The control unit controls the shake correction unit according to the angular shake correction amount instead of the parallel shake correction amount before imaging, and during the imaging period, when the angular shake correction amount and the imaging magnification are larger than a threshold value, before narrowing the low frequency band of distichum frequency band, wherein if the photographing magnification is equal to or less than the threshold value, before distichum frequency band low frequency band the calculated at determined frequency band by increasing the parallel shake correction An image blur correction apparatus that controls the blur correction unit according to an amount . The control means uses the filter including a high-pass filter when the photographing magnification is larger than the threshold, and uses the filter not including the high-pass filter when the photographing magnification is equal to or smaller than the threshold. The image blur correction apparatus according to claim 1. The control means sets the filter cutoff frequency large when the imaging magnification is greater than the threshold, and sets the filter cutoff frequency small when the imaging magnification is equal to or less than the threshold. The image blur correction apparatus according to claim 1. An image shake correction device,
First calculation means for calculating a correction coefficient for parallel shake correction using an angular velocity detection signal output from the first shake detection means and an acceleration detection signal output from the second shake detection means;
Second computing means for computing a translational shake correction amount using the correction coefficient and the angular velocity detection signal;
Third computing means for computing an angular shake correction amount using the angular velocity detection signal;
A filter for selecting the frequency band in which the translational shake correction is performed and integrating the angular velocity detection signal;
Control means for controlling the shake correction means according to the translational shake correction amount calculated by the second computing means in the translational shake correction frequency band selected by the filter. ,
The control unit controls the shake correction unit according to the angular shake correction amount instead of the parallel shake correction amount before imaging, and the filter when the angular shake correction amount and the imaging magnification are large during the imaging period. The shake correction unit is controlled according to the parallel shake correction amount calculated in the frequency band determined by making the cutoff frequency of the filter larger than the cutoff frequency of the filter when the imaging magnification is small. An image blur correction device. An image pickup apparatus having an image blur correction device according to any one of claims 1 to 4. Optical apparatus having an image blur correction apparatus according to any one of claims 1 to 4. An image shake correction method,
A first calculation step of calculating a correction coefficient for parallel shake correction using an angular velocity detection signal output from the first shake detection means and an acceleration detection signal output from the second shake detection means;
A selection step of selecting a frequency band in which the parallel shake correction is performed;
A second calculation step of calculating a translational shake correction amount using the correction coefficient and the angular velocity detection signal;
A third calculation step of calculating an angular shake correction amount using the angular velocity detection signal;
A control step of controlling a shake correction unit according to the amount of angular shake correction and the amount of parallel shake correction calculated in the frequency range of the parallel shake correction selected in the selection step;
In the control step, the shake correction unit is controlled according to the angular shake correction amount instead of the parallel shake correction amount before imaging, and during the imaging period, the angular shake correction amount and the imaging magnification are larger than a threshold value. , before distichum narrow low frequency band of the frequency band, the case the photographing magnification is equal to or less than the threshold value, before distichum frequency band of has been the parallel shake calculated frequency band determined by increasing the low frequency band An image shake correction method, wherein the shake correction means is controlled according to a correction amount . An image shake correction method,
A first calculation step of calculating a correction coefficient for parallel shake correction using an angular velocity detection signal output from the first shake detection means and an acceleration detection signal output from the second shake detection means;
A filter step of selecting a frequency band in which the translational shake correction is performed by a filter and integrating the angular velocity detection signal;
A second calculation step of calculating a translational shake correction amount using the correction coefficient and the angular velocity detection signal;
A third calculation step of calculating an angular shake correction amount using the angular velocity detection signal;
A control step of controlling shake correction means according to the translational shake correction amount calculated in the translational shake correction frequency band selected in the filtering step and the angular shake correction amount;
In the control step, the shake correction unit is controlled according to the angular shake correction amount, not the parallel shake correction amount before imaging, and the filter when the angular shake correction amount and the imaging magnification are large during the imaging period. The shake correction means is controlled according to the parallel shake correction amount calculated in the frequency band determined by making the cutoff frequency of the filter larger than the cutoff frequency of the filter when the photographing magnification is small. A characteristic image blur correction method.

Images