JP6659126B2 - Image blur correction device, image blur correction method, imaging device, and program - Google Patents

Description

  The present invention relates to an image shake correction device, an image shake correction method, an image pickup device, and a program for correcting an image shake due to a camera shake or the like.

  Patent Literature 1 determines a correction gain for correcting a parallel shake amount based on a ratio of a subject to an entire image, and obtains a combined value of an angle shake amount and a value obtained by correcting the parallel shake amount by the correction gain. Based on this, an image stabilization control device that drives a shake correction unit has been proposed.

JP 2011-615721 A

  In Patent Literature 1, the correction gain is set to be larger as the ratio of the subject to the entire image is larger. Therefore, the correction gain is reduced during macro shooting in which the focus range of the subject is narrow, and the parallel shake amount may remain. There is. In addition, since the conventional parallel shake amount information is obtained based on the information of the gyro and the accelerometer, the detection accuracy of the parallel shake amount is low especially for low-frequency vibration, and the high-precision parallel shake amount is obtained. Correction could not be made.

  An object of the present invention is to provide an image blur correction device, an image blur correction method, an imaging device, and a program that can perform highly accurate parallel shake correction.

An image shake correction apparatus according to the present invention is an image shake correction apparatus that corrects a shake of an image captured by the imaging device based on information on an angular shake amount due to rotation of the imaging device and a parallel shake amount due to parallel movement of the imaging device. A first acquisition unit for acquiring information on a first parallel shake amount based on information on an angular velocity and an acceleration of the imaging device; and selecting a motion vector of a subject based on information on a motion vector in the image. Subject vector selecting means, second acquiring means for acquiring information on a second parallel shake amount based on the motion vector of the subject, and focusing range detecting means for detecting a focusing range of the subject in the image. Multiplying means for multiplying the second parallel shake amount by a correction coefficient, adding the first parallel shake amount, and the second parallel shake amount multiplied by the correction coefficient by the multiplying means. Anda setting means for setting a correction amount for correcting the shake of the image by said multiplying means, said when the ratio of the focusing range of the focusing range detecting unit detects is less than the threshold The correction coefficient is set to be larger than the correction coefficient when the ratio of the focusing range is equal to or more than a threshold value.

  According to the present invention, it is possible to provide an image shake correction device, an image shake correction method, an imaging device, and a program that can perform high-precision parallel shake correction.

It is a top view and a side view of the imaging device according to the present embodiment. FIG. 2 is a control block diagram of the image blur correction device shown in FIG. 1. FIG. 4 is a diagram for explaining parallel shake correction in the embodiment. FIG. 4 is a diagram for explaining vector detection in a screen. 3 is a graph illustrating a relationship between the multiplier illustrated in FIG. 2 and a vector reliability determination unit. 3 is a flowchart for explaining the operation of the image blur correction device shown in FIG.

  2. Description of the Related Art Conventionally, an image blur correction device that moves a part of an image blur correction lens or an image sensor in a direction perpendicular to an optical axis based on an angular velocity detected by an gyro is known. In shooting at a close distance (high-magnification shooting conditions such as macro shooting), the image degradation due to so-called parallel shake, which cannot be detected by the gyro alone and is applied in a direction parallel or perpendicular to the optical axis of the camera, is ignored. Can not. For example, a condition such as macro photography in which the subject is photographed as close as about 20 cm, or the subject is positioned at about 1 m, but the focal length of the photographing optical system is very large (for example, 400 mm), so that aggressive. It is necessary to detect a parallel shake and perform correction.

  There are two types of vibrations applied to the imaging device, namely, angular vibration around the center of rotation due to rotation and parallel vibration due to parallel movement of the imaging device. The deterioration of the image due to the angular shake becomes larger as the subject distance and the focal length of the imaging device are longer. The image deterioration due to the parallel shake is related to the subject distance and the focal length, and the larger the image magnification (the closer the subject distance and the longer the focal length), the larger the deterioration amount.

  Under general shooting conditions (for example, a subject distance of 1 m), the effect of image deterioration due to parallel shake can be almost ignored. However, in close-up shooting (for example, a subject distance of 10 cm), image deterioration due to parallel shake is large due to a high shooting magnification. I can't ignore it. For this reason, parallel shake is detected (for example, in an optical image stabilization system, detection is performed using an accelerometer, and in an electronic image stabilization system, image shift is detected), and image shake is corrected according to the result. Subjects at various shooting distances are mixed in the screen. If image blur correction is performed according to the shooting distance of the subject, image deterioration due to parallel shake can be prevented for the subject, but sufficient shake correction cannot be performed for subjects at other shooting distances such as the background. It causes deterioration.

  FIG. 1A is a plan view illustrating an imaging device 101 according to the present embodiment, and FIG. 1B is a side view thereof. The imaging apparatus 101 includes an image shake correction device that performs shake correction for shakes indicated by arrows 103p and 103y (hereinafter, “angle shake”) and shakes indicated by arrows 104p and 104y (hereinafter, “parallel shake”). . The image shake correction device corrects a shake of an image captured by the imaging device based on the amount of angular shake and the amount of parallel shake.

  Reference numeral 105 denotes a release button (S1); 106, a photometry unit; 107, an image sensor; and 150, a camera CPU.

  The release button 105 transmits a SW1 signal to the CPU 150 in response to a half-press by the user, and transmits a SW2 signal to the CPU 150 in response to a full-press by the user. The release button 105 functions as an instruction unit for instructing shooting preparation and shooting. In this embodiment, half-pressing the release button 105 corresponds to an instruction for shooting preparation, and fully pressing the release button 105 corresponds to an instruction for shooting.

  The photometry unit (luminance value information acquisition unit) 106 measures the amount of light (luminance) that has passed through the imaging optical system using the output from the imaging element 107 and transmits the luminance value information to the CPU 150.

  The imaging element 107 is configured by a photoelectric conversion element such as a CCD sensor or a CMOS sensor, and captures an object by photoelectrically converting an optical image formed by the imaging optical system. An analog electric signal output from the image sensor 107 is converted into a digital signal by an A / D converter (not shown), and the digital signal is input to an image processing unit (not shown). The image processing means performs predetermined processing such as white balance and γ processing on the digital signal, and transmits the result to the CPU 150. The image processing means includes a motion vector detection unit 180 described later.

  The CPU 150 is a control unit that controls the operation of each unit of the imaging apparatus 101, and includes a microcomputer. The CPU 150 starts shooting preparation operations such as AE (auto focus), image blur correction, AWB (auto white balance) processing, EF (flash pre-emission) processing, and face detection in response to the SW1 signal. In addition, the CPU 150 drives a shutter driving circuit (not shown) in accordance with the SW2 signal, guides a photographing light beam to the image sensor 107, and performs exposure (photographing). In addition, the CPU 150 controls a series of photographing processing operations from reading a signal from the image sensor 107 to writing image data to the image recording unit according to the SW2 signal.

  Reference numerals 108p and 108y denote angular velocimeters (angular velocity detecting means) for detecting angular shake around the arrows 108pa and 108ya, respectively. Reference numerals 109p and 109y denote accelerometers (acceleration detecting means) for detecting parallel shakes indicated by arrows 109pa and 109ya, respectively. p represents pitch and y represents yaw.

  Reference numeral 112 denotes a lens driving unit that freely drives the image shake correction lens 111 in the directions of arrows 110p and 110y, and performs image shake correction based on both angular shake and parallel shake. Outputs of the gyros 108p and 108y and the accelerometers 109p and 109y are input to the CPU 150. Driving by the lens driving unit 112 is controlled by the CPU 150.

  The image blur correction lens 111 is an example of an image blur correction unit, and forms an imaging optical system. The image blur correction lens 111 is moved by the lens driving unit 112 in a direction orthogonal to the optical axis 102 to correct the image blur by decentering the optical axis 102. The “perpendicular direction” is sufficient if there is a component perpendicular to the optical axis 102, and may be moved obliquely to the optical axis 102. The position of the image blur correction lens 111 is detected by a position detection unit 113 described later, and a detection signal of the position detection unit 113 is converted by an A / D converter 114 described later and input to the CPU 150.

  The imaging optical system forms an optical image of the subject (subject image) on the imaging surface of the imaging element 107, and includes a zoom lens (magnification lens), an aperture, a focus lens, and the like. The zoom lens (variable lens) is moved in the optical axis direction to change the focal length. The stop is arranged at the exit pupil position of the imaging optical system, and is driven by a driving unit (not shown) to adjust the amount of light incident on the imaging element 107. The focus lens is moved in the optical axis direction to perform focus adjustment.

  In the present embodiment, optical image stabilization in which the image shake correction lens 111 is moved in a plane perpendicular to the optical axis based on the calculated correction amount is used as the image shake correction unit. However, the correction method is to perform image stabilization by moving the image sensor in a plane perpendicular to the optical axis, and to reduce the effect of shake by changing the cutout position of each shooting frame output by the image sensor Electronic vibration isolation, a combination thereof, or the like.

  FIG. 2 is a control block diagram of the image blur correction device according to the present embodiment. The image blur correction device shown in the figure includes a photometer 106, an angular velocity meter 108p, an accelerometer 109p, a CPU 150, a motion vector detector 180, a lens driver 112, an image blur correction lens 111, a position detector 113, an A / D converter. It has a portion 114. FIG. 2 shows only a configuration of a shake (pitch direction: directions of arrows 103p and 104p in FIG. 1) that occurs in the vertical direction of the imaging device 101. However, a similar configuration is also provided for a shake (yaw direction: directions of arrows 103y and 104y in FIG. 1) that occurs in the horizontal direction of the imaging apparatus 101. Since these have basically the same configuration, only the configuration in the pitch direction will be described below.

  The CPU 150 includes an HPF integration filter 151, 154, 158, a sensitivity adjustment unit 152, a zoom / focus information acquisition unit 153, a gain adjustment unit 155, an HPF phase adjustment unit 156, an angular velocity BPF unit 157, an acceleration BPF unit 159, and a comparison unit 160. Having. Further, the CPU 106 includes a parallel shake adding unit 161, a calculating unit 162, a subject vector selecting unit 163, a vector calculating unit 164, an integrator 165, a multiplier 166, a subject detecting unit 167, a focusing range detecting unit 168, a vector reliability determination. A portion 169 is further provided.

  In the image shake correction, the angular velocity signal from the gyro 108p is input to the HPF integration filter 151 of the CPU 150, integrated after the DC component is cut off by the HPF (high-pass filter or high-pass filter), and converted into an angle signal. Is converted. Since the camera shake frequency band is between 1 Hz and 10 Hz, the HPF has, for example, a first-order HPF characteristic that is sufficiently distant from the camera shake frequency band, for example, cuts a frequency component of 0.1 Hz or less.

  The output of the HPF integration filter 151 and the information acquired by the zoom / focus information acquisition unit 153 are input to the sensitivity adjustment unit 152. The sensitivity adjustment unit 152 amplifies the output of the HPF integration filter 151 based on the information obtained from the zoom / focus information acquisition unit 153, and generates an angle shake correction target value (angle shake correction amount). Accordingly, it is possible to correct a change in the shake correction sensitivity, which is a ratio of a shake amount on the image plane to a movement amount of the image shake correction lens 111 due to a change in optical information such as focus or zoom of the imaging optical system. The CPU 150 outputs the corrected angular shake correction target value to the lens driving unit 112, and corrects the image shake by driving the image shake correction lens 111.

  FIG. 3 is a diagram illustrating the angular shake 103p and the parallel shake 104p applied to the imaging apparatus 101. The relationship between the parallel shake Y (104p), the angular shake θ (103p), and the rotation radius L when the rotation center O is determined at the principal point position of the imaging optical system of the camera 101 is as follows (1) (2) (3) ) Expression. Note that the rotation radius L is a distance from the rotation center O to the accelerometer 109p.

Y = Lθ (1)
V = Lω (2)
A = Lωa (3)
The equation (1) is the radius of gyration L when the displacement Y is obtained by integrating the output of the accelerometer 109p by the second order and the angle θ is obtained by integrating the gyro 108p by the first order. Equation (2) is the radius of gyration L when the velocity V is obtained by integrating the output of the accelerometer 109p by the first order, and the angular velocity ω is obtained from the output of the gyro 108p. Equation (3) is the radius of gyration L when the acceleration A is obtained from the output of the accelerometer 109p and the angular acceleration ωa is obtained by first-order differentiation of the output of the gyro 108p. The turning radius L can be obtained by any of these methods.

  On the other hand, the shake δ generated on the imaging surface 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 photographing magnification β is obtained by Expression (4).

δ = (1 + β) fθ + βY (4)
Here, the focal length f can be obtained from the information acquired by the zoom / focus information acquisition unit 153 that acquires information on zoom / focus of the imaging optical system. The imaging magnification β represents the magnification of the size of the image of the subject formed on the image sensor 107, which corresponds to the size of the subject on the screen, and is also acquired by the zoom / focus information acquisition unit 153. Can be determined by information. The shake angle θ can be obtained from the integration result of the gyro 108p. Therefore, it is possible to perform the angular shake correction from these pieces of information. That is, the HPF integration filter 151 outputs θ, the sensitivity adjustment unit 152 multiplies θ by (1 + β) f, and outputs (1 + β) fθ of the first term on the right side to the combining operation unit 154.

  Since the second term on the right side is obtained from the second-order integral value Y of the accelerometer 109p and the imaging magnification β, parallel shake correction can be performed according to the information.

  In the present embodiment, a value obtained by adding the first parallel shake amount and the second parallel shake amount is set as the parallel shake amount of the image. Patent Document 1 uses only the first parallel shake amount, and conventionally does not consider the second parallel shake amount, so that parallel shake correction cannot be performed with high accuracy. In the present embodiment, highly accurate parallel shake correction is realized by considering the second parallel shake amount.

  The first parallel shake amount is a parallel shake amount obtained based on information on the angular velocity of the imaging device detected by the gyro 108p and the acceleration of the imaging device detected by the accelerometer 109p, and includes the components 154 to 160. It is obtained by the first obtaining means. The second parallel shake amount is a parallel shake amount obtained based on the motion vector of the subject, and is obtained by the second obtaining unit including the components 163 to 168.

  It is difficult for the angular velocity meter 108p and the accelerometer 109p to accurately measure the low frequency region of the vibration applied to the imaging device 101. On the other hand, the motion vector detection unit 180 can more accurately measure the low frequency region. However, since the measurement result of the motion vector detection unit 180 is easily affected by an error due to the ratio of the subject in the screen and the luminance, in the present embodiment, Y = Y · Kv is set, and the correction coefficient ( (Correction gain) Kv is adjusted.

  In obtaining the first parallel shake amount, the angular velocity signal from the gyro 108p is input to the HPF integration filter 154, where the DC component is cut off by the HPF, integrated, and converted into an angle signal. The output of the HPF integration filter 154 is input to a gain adjustment unit (gain adjustment filter) 155. The gain adjuster 155 and the HPF integration filter 154 adjust gain and phase characteristics in a frequency band in which parallel shake correction is to be performed.

  The output of the gyro 108p is also input to the HPF phase adjuster (phase adjustment filter) 156, where the DC component superimposed on the output of the gyro 108p is cut and the phase of the signal is adjusted. The cutoff frequency here is matched with the cutoff frequency of the HPF of the HPF integration filter 158 described later, and is set so that the frequency characteristics match. From the output of the HPF phase adjustment unit 156, only the frequency components in a predetermined band are extracted by an angular velocity BPF unit (band-pass filter or band-pass filter) 157, which is a band-pass unit.

  The output of the accelerometer 109p is input to the HPF integration filter 158, where the DC component is cut off by the HPF, integrated, and converted into a speed signal. At this time, the cutoff frequency of the HPF is set in accordance with the frequency characteristic of the HPF of the HPF phase adjustment unit 156, as described above. From the output of the HPF integration filter 158, only a frequency component in a predetermined band is extracted by an acceleration BPF unit (band-pass filter or band-pass filter) 159 which is a band-pass unit.

  The outputs of the angular velocity BPF unit 157 and the acceleration BPF unit 159 are input to the comparison unit 160, and a correction amount (correction coefficient) for correcting the output of the gain adjustment unit 155 is calculated. The comparison unit 160 obtains L = V / ω from equation (2). From the equations (1) and (4), δ = (1 + β) fθ + βLθ. Therefore, the gain adjustment unit 155 creates the first parallel shake amount by multiplying θ output from the HPF integration filter 154 by β obtained from the zoom / focus information obtaining unit and L obtained from the comparison unit 160. .

  The motion vector detecting section (motion vector detecting means) 180 detects a motion in an image based on a luminance signal included in the image signal obtained from the image sensor 107 and a luminance signal included in the video signal one field before stored in the image memory. Find the vector. The motion vector can be detected using, for example, a block matching method. In the block matching method, an image signal is divided into a plurality of appropriately sized block areas, and a difference from a certain range of pixels in a previous field (or frame) is calculated in block units, and a sum of absolute values of the differences is calculated. Is searched for a block in the field (or frame) before is minimum. The relative displacement between the screens indicates the motion vector of the block. The motion vector obtained by the motion vector detection unit 180 is input to a subject vector selection unit (subject vector selection unit) 163.

  The subject vector selection unit 163 selects and outputs a motion vector of the subject from the motion vectors in the screen detected by the motion vector detection unit 180 based on the subject detected by the subject detection unit 167.

  FIG. 4A is a diagram for explaining a method of detecting a motion vector in a screen. As shown in FIG. 4A, when the focusing range of the subject O is narrow, the second parallel shake correction difference between the subject O and the background becomes conspicuous. In the screen, n blocks 10 for detecting a vector are arranged in the horizontal direction of the screen and m in the vertical direction of the screen. Reference numeral 11 denotes an AF frame indicating a focus detection range. Since the motion vector detection unit 180 cannot distinguish the subject or the background vector from the detected vector, the motion vector detection unit 180 distinguishes the subject and the background vector by comparing the angular velocity and the frequency distribution of the vector using the angular velocity information.

  FIG. 4B is a histogram showing the frequency distribution of the angular velocity and the vector, where the horizontal axis represents the angular velocity and the vertical axis represents the frequency. Reference numeral 12 denotes a subject vector, 13 denotes a background vector, and 14 denotes an angular velocity. Ideally, if the angular velocity of the subject O and the angular velocity of the imaging device 101 match, the subject O appears to stop when viewed from the imaging device 101, so the angular velocity is 0 dps. Since the background moves at the angular velocity of the imaging apparatus 101, for example, if panning is performed at 10 dps, the angular velocity of the background becomes 10 dps. Thus, the vector near the angular velocity 14 is the background vector 13, and the vector apart from the angular velocity 14 is the subject vector 12. Since the user matches the AF frame 11 with the subject O, the vectors detected by the blocks 10 around the AF frame 11 are weighted, and the accuracy of extracting the subject and background vectors is increased. The focus range detection unit (synthesis range detection means) 168 can detect the frequency distribution range near the subject vector 12 shown in FIG. 4B as the focus range.

  The subject detection unit 167 may distinguish the subject from the background vector using the angular velocity information and the information of the AF frame 11, but the subject detection unit 167 uses the subject frame information displayed using the subject identification information such as face recognition to determine the frame. May be used. The vector of the subject distinguished by the subject vector selection unit 163 is output to the vector calculation unit 164.

  The focus range detection unit 168 detects a focus range of the subject O in the screen. Information on the ratio of the subject to the entire angle of view (the ratio of the subject in the image) is obtained from the focusing range detection unit 168. For example, the ratio of the subject to the entire angle of view may be calculated by dividing the subject vector 12 in the screen by all the vectors using the block 10 and the subject detection unit 167, or may be calculated from the zoom position and the subject distance information. The calculation may be performed using the obtained photographing magnification. In FIG. 4A, the focus range detection unit 168 can recognize the range of the subject O inside the AF frame 11 as the focus range.

  The vector calculation unit (vector calculation unit) 164 converts the input vector into an angular velocity using the focal length and the frame rate information, and outputs the converted vector to the integrator (integration unit) 165. The integrator 165 converts the angular velocity into angle information by integrating it. A multiplier (multiplication unit) 166 multiplies the image magnification β acquired from the zoom / focus information acquisition unit by a correction coefficient (correction gain) Kv as a vector feedback gain, and outputs the result to the parallel shake addition unit 161. As the correction coefficient Kv, a value that can be varied from 0 to 1 is selected based on the outputs of the focusing range detection unit 168 and the vector reliability determination unit 169.

  The multiplier 158 sets the correction coefficient Kv when the ratio of the focus range of the subject in the image detected by the focus range detection unit 168 is smaller than the threshold to be larger than the correction coefficient Kv when the ratio is equal to or larger than the threshold. are doing. In the present embodiment, the correction coefficient Kv when the ratio of the in-focus range of the subject is equal to or larger than the threshold is set to 0 (zero), and the correction coefficient Kv when the ratio is smaller than the threshold is set to 1. , But is not limited to this.

  In Patent Document 1, the balance between image blur correction of a subject and image blur correction of a background during moving image shooting is set by setting a correction gain larger as the ratio of the subject to the entire image increases. However, when the proportion of the subject in the entire image is small, the correction gain is small, and the correction of the parallel shake amount is not effective, and there is a concern that the influence of the parallel shake remains in macro shooting where the focusing range of the subject is narrow. According to the present embodiment, appropriate parallel shake correction can be performed even when the focusing range of a subject is narrow under close-up shooting conditions.

  The vector reliability determination section 169 determines whether or not the vector output by the motion vector detection section 180 is a correctly detected vector. The vector reliability determining unit 169 determines whether the luminance detected by the photometric unit 106 is less than a threshold. It is difficult to detect a vector with an image signal in which a feature point of an object is difficult to understand such as a low contrast. Therefore, when the vector reliability determining section 169 outputs an error value from the motion vector detecting section 180, the multiplier 158 sets a correction gain Kv smaller than 1 to prevent the image blur correction from being overcorrected. The multiplier 158 sets the correction coefficient Kv when the luminance value of the image is less than the threshold value smaller than the correction coefficient Kv when the luminance value is equal to or more than the threshold value.

  FIG. 5 is a graph showing a relationship between the multiplier 158 and the vector reliability determination unit 169. The horizontal axis represents the vector reliability, which is the output of the vector reliability determination unit 169, and the vertical axis represents the correction coefficient Kv. The horizontal axis may be replaced with a luminance value.

  The multiplier 158 determines a correction coefficient based on the detection results of the motion vector detection unit 180 and the photometry unit 106, and multiplies the output of the integrator 165. When the vector determination value output for each field is less than the threshold value T1, the multiplier 158 sets a correction gain of about 0.3. For example, this is the case of a shooting scene in which the detection accuracy of the motion vector detection unit 180 is reduced in a night shooting scene or the like. Next, when the vector determination value is equal to or more than the threshold value T1 and less than T2, the multiplier 158 sets a correction coefficient of about 0.8 times. For example, this is a case of a shooting scene (for example, ISO sensitivity of 1200 or more) that raises the ISO sensitivity of the imaging device 101 and makes noise noticeable, which affects vector detection. When the vector determination value is equal to or larger than the threshold value T2, it is determined that the vector detection has been correctly performed, and the parallel shake correction is positively performed. Therefore, the multiplier 158 sets a correction gain of 1 time.

  The parallel shake adding unit 161 sets a correction amount for correcting the image shake by adding the output (first parallel shake amount) of the gain adjustment unit 155 and the output of the multiplication value 166 (second parallel shake amount). Setting means (addition means). The calculation unit 162 adds the output (angle shake amount) from the sensitivity adjustment unit 152 and the output (synthesized parallel shake amount) of the parallel shake addition unit 161 to set a correction amount for correcting image shake. , An arithmetic unit for subtracting the output of the A / D converter 114 (current correction value).

  FIG. 6 is a diagram for explaining an image blur correction method (processing from detection of the subject vector 12 to calculation of the image blur correction amount, which is calculated at a constant interrupt cycle after the power of the camera is turned on) by the image blur correction device (CPU 150). It is a flowchart of FIG. “S” represents a step (process). The flowchart shown in FIG. 6 can be embodied as a program for causing a computer to execute each step. Such a program can be stored in a storage unit such as a non-transitory computer-readable storage medium. Note that the image blur correction method of the present embodiment is applicable to both moving image shooting and still image shooting.

  In step S600, the CPU 150 determines whether image blur correction is enabled (whether an unillustrated image stabilization switch is turned on and the SW1 signal is generated). If the image blur correction is enabled, the process advances to step S601. If there is, the process ends. In step S601, the CPU 150 acquires information detected by the gyros 108p and 108y and information detected by the accelerometers 109p and 109y. In S602, the CPU 150 calculates the amount of angular shake (the amount of angular shake correction) using the output values of the gyros 108p and 108y, focal length information, and the like.

  In S603, the CPU 150 arranges a vector detection block in the screen by the template matching method, and detects a motion vector in the screen based on a difference from one frame before. In S604, the CPU 150 distinguishes between the subject and the background vector using the frequency distribution of the vector and the angular velocity. In step S605, the CPU 150 converts the detected vector into angular information by converting the detected vector into angular velocity and integrating the detected vector. In S606, the CPU 150 makes an error determination on the detected vector. If the ratio of the focus range and the luminance value are equal to or greater than the threshold value for all the vectors, the process proceeds to S608, and if any of them is less than the threshold value, the process proceeds to S607.

  In S607, since the vector has been correctly detected, the CPU 150 sets a correction coefficient Kv of about one time. In S608, the CPU 150 sets a correction coefficient Kv smaller than 1 in order to prevent overcorrection.

  In S609, the CPU 150 determines whether or not the ratio of the focus range of the subject in the image is less than the threshold. If it is equal to or greater than the threshold, the process proceeds to S610, and if it is less than the threshold, the process proceeds to S611. In S610, since the focusing range is narrow, the difference in parallel shake correction from the background is conspicuous. Therefore, the CPU 150 calculates the translational shake correction amount by adding the angle-converted vector to the output of the gain adjustment unit 155. In step S <b> 611, the CPU 150 drives the image shake correction lens 111 as a final shake correction amount by adding the parallel shake correction taking the vector into consideration and the angular shake correction. On the other hand, if the value is equal to or larger than the threshold, the in-focus range of the subject is wide, so that the translational shake correction difference from the background is inconspicuous. Therefore, the CPU 150 does not perform the vector feedback, adds the parallel shake correction amount and the angular shake correction amount, and drives the image shake correction lens 111 as a final shake correction amount. According to the present embodiment, when the in-focus range of the subject is less than the threshold, the accuracy of the parallel shake correction can be improved by adding the vector to the parallel shake correction amount.

  Although the embodiments of the present invention have been described above, the present invention can be variously modified and changed within the scope of the gist. The present invention supplies a program for realizing one or more functions of the above-described embodiments to a system or an apparatus via a network or a storage medium, and one or more processors in a computer of the system or the apparatus read and execute the program. This process can be realized. Further, it can also be realized by a circuit (for example, an ASIC) that realizes one or more functions.

  The present invention can be mounted not only on an image blur correction device of a digital single-lens reflex camera or a digital compact camera but also on a photographing device such as a surveillance camera, a web camera, and a mobile phone.

151 ... HPF integration filter, 152 ... Sensitivity adjuster, 161 ... Parallel shake adder (adder), 162 ... Calculator (calculator), 168 ... Synthesis range detector (synthesis range detector), 180 ... Motion vector Detection unit (motion vector detection means)

Claims (14)

An image shake correction device that corrects a shake of an image captured by the imaging device based on information on an angular shake amount due to rotation of the imaging device and a parallel shake amount due to parallel movement of the imaging device,
A first acquisition unit configured to acquire information on a first parallel shake amount based on information on an angular velocity and an acceleration of the imaging device;
Subject vector selection means for selecting a motion vector of the subject based on information of the motion vector in the image,
Second acquisition means for acquiring information on a second parallel shake amount based on the motion vector of the subject;
A focus range detection unit that detects a focus range of the subject in the image;
Multiplying means for multiplying the second parallel shake amount by a correction coefficient;
Setting means for setting a correction amount for correcting the image shake by adding the first parallel shake amount and the second parallel shake amount multiplied by the correction coefficient by the multiplying unit. And
The multiplying unit is configured such that the correction coefficient when the ratio of the focusing range detected by the focusing range detection unit is less than a threshold is larger than the correction coefficient when the ratio of the focusing range is equal to or more than a threshold. An image blur correction device characterized by being set large.   The apparatus according to claim 1, wherein the multiplying unit sets the correction coefficient to zero when the ratio of the focusing range is equal to or greater than a threshold. The second acquisition means,
Vector calculation means for converting the motion vector of the subject into angular velocity,
Integration means for converting the angular velocity converted by the vector calculation means into angle information;
The image blur correction device according to claim 1, further comprising: Further comprising a motion vector detecting means for detecting a motion vector in the image,
4. The apparatus according to claim 1, wherein the subject vector selection unit selects a motion vector of the subject based on information on the motion vector detected by the motion vector detection unit. 5. An image blur correction device according to claim 1. Further comprising a brightness value information obtaining means for obtaining information of the brightness value of the image,
The multiplying unit sets the correction coefficient when the luminance value acquired by the luminance value information acquiring unit is less than a threshold, smaller than the correction coefficient when the luminance value is equal to or greater than a threshold. The image blur correction device according to claim 1, wherein:   The apparatus according to claim 4, wherein the focus range detection unit detects the focus range based on a frequency distribution of a vector detected by the motion vector detection unit.   The said focusing range detection means detects the said focusing range by the imaging magnification calculated from the focal length of the said imaging device, and a subject distance, The Claims any one of Claim 1 thru | or 5 characterized by the above-mentioned. Image blur correction device.   The image blur correction apparatus according to claim 4, wherein the subject vector selection unit selects a motion vector of the subject based on a frequency distribution of a vector detected by the motion vector detection unit. An image shake correction device that corrects a shake of an image captured by the imaging device based on information on an angular shake amount due to rotation of the imaging device and a parallel shake amount due to parallel movement of the imaging device,
Subject vector selection means for selecting a subject motion vector based on the motion vector in the image,
Acquiring means for acquiring information on the amount of parallel shake based on the motion vector of the subject selected by the subject vector selecting means;
A focus range detection unit that detects a focus range of the subject in the image;
Multiplying means for multiplying the parallel shake amount obtained by the obtaining means by a correction coefficient,
Have a,
The multiplying unit is configured such that the correction coefficient when the ratio of the focusing range detected by the focusing range detection unit is less than a threshold is larger than the correction coefficient when the ratio of the focusing range is equal to or more than a threshold. An image blur correction device characterized by being set large . Further comprising a brightness value information obtaining means for obtaining information of the brightness value of the image,
The multiplying unit sets the correction coefficient when the luminance value acquired by the luminance value information acquiring unit is less than a threshold, smaller than the correction coefficient when the luminance value is equal to or greater than a threshold. The image blur correction device according to claim 9 . Imaging apparatus characterized by having an image blur correction apparatus according to any one of claims 1 to 10. An image shake correction method for correcting a shake of an image captured by the imaging device based on information on an angular shake amount due to rotation of the imaging device and a parallel shake amount due to parallel movement of the imaging device,
Acquiring information on a first parallel shake amount based on information on an angular velocity and an acceleration of the imaging device;
Selecting a motion vector of the subject based on information of the motion vector in the image;
Obtaining information on a second parallel shake amount based on the motion vector of the subject;
Detecting a focus range of the subject in the image;
Multiplying the second parallel shake amount by a correction coefficient;
Setting a correction amount for correcting the image shake by adding the first parallel shake amount and the second parallel shake amount multiplied by the correction coefficient;
Has,
An image blur correction method, wherein the correction coefficient when the ratio of the focusing range is less than a threshold is set to be larger than the correction coefficient when the ratio of the focusing range is equal to or more than a threshold. An image shake correction method for correcting a shake of an image captured by the imaging device based on information on an angular shake amount due to rotation of the imaging device and a parallel shake amount due to parallel movement of the imaging device,
Selecting a motion vector of the subject based on the information of the motion vector in the image,
Acquiring information on the parallel shake amount based on the motion vector of the subject;
A focus range detecting step of detecting a focus range of the subject in the image;
A multiplying step of multiplying the obtained amount of translational shake by a correction coefficient,
Have a,
In the multiplying step, the correction coefficient when the proportion of the focusing range detected in the focusing range detection step is less than a threshold is larger than the correction coefficient when the proportion of the focusing range is greater than or equal to a threshold. An image blur correction method characterized by setting a large value . Program for executing the image blur correcting method according to claim 12 or 13 into the computer.