JP2021082917A - Vibration-proof device, and control method and program therefor - Google Patents

Abstract

To provide a vibration-proof device, and a control method and a program therefor that can reduce a decrease in image blur correcting performance due to a difference in learning progress degree between a camera body-side learning unit and a lens barrel-side learning unit.SOLUTION: A camera system is configured to: learn, at a camera body, control over image blur correction by a camera-side learning unit 5c based upon a blur quantity detected by camera-side blur detection means 15, and then drive an imaging element 6 according to a learning result thereof to perform image blur correction; and learn, at the lens barrel, control over image blur correction by a lens-side learning unit 12c based upon a blur quantity detected by lens-side blur detection means 16, and then drive a blur correcting lens 3a according to a learning result thereof to perform image blur correction. The camera body further comprises vibration-proof control selection means 21 which selects, as learning unit data to be used to learn control over image blur correction, learning unit data higher in learning progress degree between the camera-side learning unit 5c and lens-side learning unit 12c.SELECTED DRAWING: Figure 2

Description

The present invention relates to a vibration isolator and a control method thereof, and a program, and more particularly to a vibration isolator and a control method thereof and a program for correcting image blur during photographing by driving and controlling an image sensor and a photographing optical system.

In recent years, due to the high performance of image pickup devices, many image pickup devices and lens barrels are equipped with an image blur correction mechanism for correcting image blur during shooting. The image blur correction unit makes it possible to reduce the influence of camera shake on the captured image when the user holds the image pickup device and takes a picture.

The image blur correction mechanism used in the image pickup device includes a lens side blur correction mechanism that drives a part of the lens of the photographing optical system to correct the image blur, and an image sensor in the camera body to drive the image blur correction. The camera side shake correction mechanism to be performed is known. For example, in an interchangeable lens type image pickup device, a lens-side blur correction mechanism is provided in an interchangeable lens barrel, and a camera-side blur correction mechanism is provided in a camera body. Further, there is also known a method of performing image blur correction using both a lens-side blur correction mechanism and a camera-side blur correction mechanism when shooting.

There is also known a vibration isolator that has a learning device and is adapted to the blurring characteristics that differ depending on the blurring characteristics of the user and the weight balance of the imaging device.

For example, in the image pickup apparatus of Patent Document 1, the camera body has a learner that learns the relationship between them from the time-series data of the outputs from the blur sensor and the image sensor, and based on the learning result, the lens side blur. Image blur correction is performed by both the correction mechanism and the camera side blur correction mechanism.

Japanese Unexamined Patent Publication No. 7-20547

However, in the image pickup device of Patent Document 1, since the lens barrel and the camera body are integrated, only one learning device is required, but in the case of the lens interchangeable image pickup device, the camera body and the lens used by the user. The combination of lens barrels changes. Therefore, it is necessary to provide a learning device not only in the camera body but also in the lens barrel, but in this case, the progress of learning may be biased depending on the combination of the camera body and the lens barrel. That is, there may be a difference in the image blur correction performance between the case where the image blur correction is performed using the learning device on the camera body side and the case where the image blur correction is performed using the learning device on the lens barrel side. is there.

In this case, there is a problem that the image blur correction performance may be deteriorated if both are operated at the same time.

The present invention has been made in view of the above problems, and can reduce the deterioration of the image blur correction performance due to the difference in the learning progress between the learning device on the camera body side and the learning device on the lens barrel side. It is an object of the present invention to provide an apparatus, a control method thereof, and a program.

The vibration isolator according to claim 1 of the present invention is a vibration isolator that is communicably connected to a lens barrel, and the vibration isolator includes an image pickup element and a first blur detection means for detecting the amount of blur. A first learner that learns the control of image blur correction based on the amount of blur detected by the first blur detecting means, and a drive target value of the image pickup element according to the learning result by the first learner. The first control means including the first calculation means for calculating the image blur, and the first vibration isolation that corrects the image blur by driving the image pickup element with the drive target value output from the first control means. The lens barrel includes means, a photographing optical system including a lens for blur correction, a second blur detecting means for detecting a blur amount, and the blur amount detected by the second blur detecting means. A second control including a second learning device for learning the control of image blur correction and a second calculating means for calculating a drive target value of the blur correction lens according to a learning result by the second learning device. The vibration isolator further includes means and a second vibration isolator that corrects image blur by driving the blur correction lens with a drive target value output from the second control means. The learning progress of the first and second learners is determined, and as a result of the determination, the learner data having the higher learning progress is selected as the learner data used for learning the control of the image blur correction. It is characterized by providing anti-vibration control selection means.

The vibration isolator according to claim 2 of the present invention is a vibration isolator that is communicably connected to a camera body, wherein the camera body includes an image pickup element, a first blur detection means for detecting a blur amount, and a blur detection means. The drive target value of the image pickup element is calculated according to the learning result of the first learning device that learns the control of image blur correction based on the amount of blur detected by the first blur detecting means and the first learning device. A first control means including the first calculation means, and a first anti-vibration means for correcting image blurring by driving the image pickup element with a drive target value output from the first control means. The anti-vibration device includes a photographing optical system including a blur correction lens, a second blur detecting means for detecting the amount of blur, and the image blur based on the amount of blur detected by the second blur detecting means. A second control means including a second learning device for learning correction control and a second calculation means for calculating a drive target value of the blur correction lens according to a learning result by the second learning device. A second anti-vibration means for correcting image blur by driving the blur-correcting lens with a drive target value output from the second control means, and the anti-vibration device further comprises the second. The learning progress of the first and second learners is determined, and as a result of the determination, the learner data having the higher learning progress is selected as the learner data used for learning the control of the image blur correction. It is characterized by providing a control selection means.

According to the present invention, it is possible to reduce the deterioration of the image blur correction performance due to the difference in the learning progress between the learning device on the camera body side and the learning device on the lens barrel side.

It is a block diagram which shows the central sectional view and the electric structure of the camera system including the camera body as the vibration isolation device which concerns on Example 1 of this invention. FIG. 5 is a control block diagram related to image blur correction of the camera body and the lens barrel in FIG. 1. It is a flowchart of image blur correction processing which concerns on Example 1 of this invention. It is a flowchart of the learning process of step S3005 of FIG. It is a flowchart of image blur correction processing which concerns on Example 2 of this invention. It is a flowchart of image blur correction processing which concerns on Example 3 of this invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. It should be noted that the following embodiments do not limit the present invention according to the claims, and not all combinations of features described in the present embodiment are essential for the means for solving the present invention. ..

(Example 1)
Hereinafter, a camera system including a camera body 1 as a vibration isolator according to a first embodiment of the present invention will be described with reference to FIGS. 1 to 4.

FIG. 1A is a central sectional view of the camera system 100, and FIG. 1B is a block diagram showing an electrical configuration thereof. Those having the same reference numerals in FIGS. 1A and 1B indicate the same components, respectively.

In FIG. 1, the camera system 100 includes a camera body 1, a lens barrel 2 attached to the camera body 1, and an electrical contact 11 for communicating between the camera body 1 and the lens barrel 2.

The lens barrel 2 includes a photographing optical system 3 composed of a plurality of lenses and an optical axis 4 of the photographing optical system 3. On the other hand, the camera body 1 includes a camera system control unit 5 (first control means), an image sensor 6, an image processing unit 7, a memory means 8, and a blur correction lens 3a for image blur correction. To be equipped.

The camera body 1 further includes an operation detection unit 10 that detects a signal from an operation means including a display means 9 and a shutter release button (not shown).

The display means 9 includes a rear display device 9a provided on the back surface of the camera body 1 and an EVF (electronic viewfinder) 9b provided in the finder.

Further, the lens barrel 2 further includes a lens system control unit 12 (second control means), a focus lens for adjusting the focus, a lens driving means 13 for driving a blur correction lens 3a for image blur correction, and a lens side blur. The detection means 16 (second blur detection means) is provided. Although details will be described later, the lens driving means 13 includes a lens-side blur correcting means 13a (second anti-vibration means) that drives the blur correcting lens 3a in a plane perpendicular to the optical axis 4.

Similarly, the camera body 1 further includes a camera-side blur correction means 14 (first anti-vibration means) that drives the image sensor 6 in a plane perpendicular to the optical axis 4, and a camera-side blur detection means 15 (first). It is equipped with a blur detection means).

The camera system 100 includes an imaging system, an image processing system, a recording / reproducing system, and a control system.

The image pickup system includes a photographing optical system 3 and an image pickup element 6, an image processing system includes an image processing unit 7, and a recording / reproduction system includes a memory means 8 and a display means 9 (that is, a rear display device 9a and an EVF 9b). .. The control system includes a camera system control unit 5, an operation detection unit 10, a camera side blur detection means 15, a camera side blur correction means 14, a lens system control unit 12, a lens side blur detection means 16, and a lens side blur correction means. Includes 13a.

The lens system control unit 12 and the lens driving means 13 not only drive and control the blur correction lens 3a, but also drive and control the focus lens and the aperture included in the photographing optical system 3.

The camera-side shake detecting means 15 and the lens-side shake detecting means 16 are each composed of an angular velocity sensor (gyro sensor) for detecting the rotational speed applied to the camera, an acceleration sensor for detecting the acceleration applied to the camera, and the like. The camera-side blur detecting means 15 and the lens-side blur detecting means 16 detect the amount of blur applied to the camera system 100 by the above-described sensor. The camera-side blur detecting means 15 may detect the remaining amount of blur from the time-series image (live view image) obtained on the image sensor 6.

The image pickup system described above is an optical processing system that forms an image of light from an object on the image pickup surface of the image pickup device 6 via the photographing optical system 3. Since the focus evaluation amount / appropriate exposure amount can be obtained from the image sensor 6, the photographing optical system 3 is appropriately adjusted based on this signal to expose the object light of an appropriate amount of light to the image sensor 6 and at the same time. A subject image is formed in the vicinity of the image sensor 6.

The image processing unit 7 has an A / D converter, a white balance adjustment circuit, a gamma correction circuit, an interpolation calculation circuit, and the like inside, and can generate an image for recording by image processing by these circuits. .. For example, the image processing unit 7 performs color interpolation (demosizing) processing from the signals of the Bayer array to generate a color image as an image for recording. Further, the image processing unit 7 compresses an image, a moving image, a sound, or the like by using a predetermined method. Further, since the image processing unit 7 can generate a blur detection signal based on the comparison between a plurality of images obtained from the image sensor 6, the image sensor 6 and the image processing unit 7 can generate a blur detection signal on the camera side. The detection means 15 may be configured.

The memory means 8 includes a storage unit, holds data output from the camera system control unit 5 in the recording unit, outputs data from the recording unit to the display means 9, and displays an image presented to the user on the display means 9. To do.

The camera system control unit 5 generates and outputs a timing signal or the like at the time of imaging. The imaging system, image processing system, and recording / playback system are controlled in response to external operations. For example, the operation detection unit 10 detects the pressing of the shutter release button (not shown), and the camera system control unit 5 drives the image sensor 6, performs image processing such as compression processing by the image processing unit 7, and displays by the display means 9. Control. Further, the rear display device 9a is a touch panel, and may also serve as a display means 9 and an operation means.

The adjustment operation of the photographing optical system 3 by the control system will be described.

An image processing unit 7 is connected to the camera system control unit 5, and an appropriate focus position and aperture position are obtained based on a signal from the image sensor 6. When the camera system control unit 5 issues a command to the lens system control unit 12 via the electrical contact 11, the lens system control unit 12 controls the focus lens driving means and the aperture driving means (not shown) in response to the command. Further, in the mode of performing blur correction, the camera system control unit 5 controls the camera side blur correction means 14 based on the signal obtained from the camera side blur detecting means 15, and similarly, from the lens side blur detecting means 16. The lens side blur correction means 13a is controlled based on the obtained signal.

As a specific control method, first, the camera system control unit 5 and the lens system control unit 12 detect the camera shake signal detected by the camera side shake detecting means 15 and the lens side shake detecting means 16, respectively. Based on the result, the camera system control unit 5 and the lens system control unit 12 calculate the driving amounts of the image sensor 6 and the blur correction lens 3a for correcting the image blur, respectively. After that, the calculated drive amount is sent to the camera side blur correction means 14 and the lens side blur correction means 13a as command values to drive the image sensor 6 and the lens 3a, respectively.

Further, as described above, the camera system control unit 5 and the lens system control unit 12 respond to the user operation to the operation means (not shown) provided on the camera body 1 and the lens barrel 2, and the camera body 1 and the lens mirror are used. It controls the operation of each part of the cylinder 2. As a result, it is possible to shoot still images and moving images.

(Explanation of the configuration of each learning device of the camera body and lens barrel)
Next, the configuration of the camera-side learning device 5c provided on the camera body 1 and the configuration of the lens-side learning device 12c provided on the lens barrel 2 will be described with reference to FIG.

FIG. 2 is a control block diagram related to image blur correction of the camera body 1 and the lens barrel 2.

Based on the amount of blur obtained from each of the camera-side blur detecting means 15 and the lens-side blur detecting means 16, the camera system control unit 5 generates a drive target value of the camera-side blur correcting means 14, and the lens system control unit 12 generates a drive target value of the lens side blur correction means 13a. After that, the camera system control unit 5 drives the camera side blur correction means 14 based on the drive target value generated by itself, and the lens system control unit 12 corrects the lens side blur based on the drive target value generated by itself. Drive the means 13a.

As shown in FIG. 2, the camera system control unit 5 includes a camera-side filter processing means 5a, a camera-side vibration isolation control means 5b, a camera-side learner 5c, and a camera-side target value generation means 5d.

The camera-side filter processing means 5a processes the output of the camera-side blur detection means 15. Specifically, the camera-side filter processing means 5a is composed of a high-pass filter, a gain compensator, and the like based on the characteristics of the camera-side blur detection means 15.

The camera-side vibration isolation control means 5b generates a drive target value to be sent to the camera-side blur correction means 14 based on the output of the camera-side filter processing means 5a.

The camera-side learner 5c (first learner) is a learner configured by a neural network provided in the camera-side vibration isolation control means 5b.

The camera-side target value generation means 5d is provided in the camera-side vibration isolation control means 5b, and generates a drive target value of the camera-side blur correction means 14.

Further, as shown in FIG. 2, the lens system control unit 12 includes a lens-side filter processing means 12a, a lens-side vibration isolation control means 12b, a lens-side learner 12c, and a lens-side target value generation means 12d.

The lens-side filter processing means 12a processes the output of the lens-side blur detecting means 16. Specifically, the lens-side filter processing means 12a includes a high-pass filter, a gain compensator, and the like based on the characteristics of the lens-side blur detection means 16.

The lens-side target value generation means 12d generates a drive target value to be sent to the lens-side blur correction means 13a based on the output of the lens-side filter processing means 12a.

The lens-side learner 12c (second learner) is a learner configured by a neural network provided in the lens-side vibration isolation control means 12b.

The lens-side target value generation means 12d is provided in the lens-side vibration isolation control means 12b, and generates a drive target value of the lens-side blur correction means 13a.

Further, the camera system control unit 5 further includes anti-vibration control selection means 21 for switching between the camera-side learner 5c and the lens-side learner 12c. Details of the anti-vibration control selection means 21 will be described later.

Since the camera-side blur detecting means 15 and the lens-side blur detecting means 16 detect the amount of blur applied to the camera system 100 input to the camera body 1 and the lens barrel 2, respectively, the amount of blur is basically the same. To detect. When both the image sensor 6 and the blur correction lens 3a cooperate to perform image blur correction, the image sensor 6 and the blur correction lens 3a are driven at a certain ratio to the detected blur amount. It is possible to appropriately perform image blur correction. Therefore, in the camera-side filter processing means 5a and the lens-side filter processing means 12a, the sum of the drive amounts of both the image sensor 6 and the blur correction lens 3a is the drive amount for performing the desired image blur correction. It is provided so as to share the driving amount for each ratio.

For example, when the camera side is responsible for 40% of the image blur correction amount and the lens side is responsible for 60% of the image blur correction amount, the drive target is driven by the camera side vibration isolation control means 5b and the lens side vibration isolation control means 12b. The ratio of value output is set to 2: 3.

In this embodiment, the control for sharing the drive amount for image blur correction of the image sensor 6 and the blur correction lens 3a by a certain ratio has been described, but the image sensor 6 and the blur correction lens 3a have their respective frequencies. It may be shared for each band. For example, of the amount of blur detected by the camera side blur detecting means 15 and the lens side blur detecting means 16, the blur in the low frequency band is driven by the image sensor 6, and the blur in the high frequency band is corrected by the lens side blur correcting means 13a. Image blur correction may be performed by driving.

When the anti-vibration control selection means 21 performs image blur correction, the electrical contact 11 is used to compare the learning progress of the camera-side learner 5c and the lens-side learner 12c, and which learner is used to correct the image blur. And select the learning device to be used. When the learner to be used is selected, the learner data generated by one of the camera system control unit 5 and the lens system control unit 12 is copied to the other, or the drive target value is generated on one side for driving. Notify the other of the target value. By this method, the camera system control unit 5 sends a drive target value to the camera side blur correction means 14, and the lens system control unit 12 sends a drive target value to the lens side blur correction means 13a.

In this embodiment, the anti-vibration control selection means 21 is included in the camera system control unit 5, but is not limited to such a mode. For example, the anti-vibration control selection means 21 may be included in the lens system control unit 12. In this case, the lens barrel 2 becomes the vibration isolator of the present invention. Further, the anti-vibration control selection means 21 may be included in both the camera system control unit 5 and the lens system control unit 12.

The camera-side vibration isolation control means 5b generates a drive target value for the camera-side blur correction means 14 based on the amount of blur detected by the camera-side blur detection means 15, and the camera-side blur correction means 14 generates an image. Perform blur correction. However, since the characteristics of this drive target value generation usually incorporate the characteristics set in advance at the time of shipment, sufficient image blur correction should be performed when the blur amount includes user-specific blur characteristics and the like. It may not be possible and the blurring residue may occur. The camera-side learner 5c and the lens-side learner 12c generate drive target values in response to changes in the characteristics of the camera-side blur correction means 14 and the lens-side blur correction means 13a due to user-specific characteristics and environmental changes as described above. Changes the characteristics of. Thereby, in this embodiment, it is possible to prevent deterioration of the image blur correction performance.

However, the camera system 100 is configured such that the lens barrel 2 can be attached to and detached from the camera body 1. Therefore, the combination of the camera body 1 and the lens barrel 2 used by the user changes, and the learning progress of the camera-side learning device 5c and the lens-side learning device 12c (hereinafter referred to as learning progress) is biased. Things can happen. For example, when the user purchases a new lens barrel 2 for step-up, the camera-side learner 5c has the accumulated learner data that the user has used so far, so the lens-side learner 12c It is considered that the learning progress is higher than that. On the other hand, when a user who has purchased a new camera body 1 attaches and uses the lens barrel 2 that he has owned before, the learning progress of the lens-side learning device 12c is higher than that of the camera-side learning device 5c. May be high.

When the learning progress is biased between the camera-side learner 5c and the lens-side learner 12c in this way, the accuracy of image blur correction is poor when image blur correction is performed using the trained models constructed in each case. There is a possibility of becoming. Therefore, in this embodiment, the camera body 1 and the lens barrel 2 are used by using the learner data of the trained model of the camera-side learner 5c and the lens-side learner 12c, whichever has the higher learning progress. Each of them generate a drive target value. As a result, the accuracy of image blur correction can be improved. That is, when the anti-vibration control selection means 21 detects a bias in the learning progress, the learning device data of the camera-side learning device 5c and the lens-side learning device 12c, whichever has the higher learning progress, has a lower learning progress. It is transmitted to one of the learners via the electrical contact 11. As a result, it is possible to reduce the bias in the learning progress between the camera-side learning device 5c and the lens-side learning device 12c. As a result, both the camera body 1 and the lens barrel 2 can generate a drive target value using the learner data of the learner having a high learning progress, and the image due to the bias in the learning progress. It is possible to reduce the deterioration of the blur correction performance. It is preferable that the learning progress is determined by the number of times each of the camera-side learning device 5c and the lens-side learning device 12c has learned by the learning process described later, and the learning time. Further, the image blur correction may be actually performed using each learning device at the time of the preliminary shooting operation, and the accuracy may be determined by reading out the live view image and confirming it. However, if there is a difference in the performance of the two detecting means, the camera-side blur detecting means 15 and the lens-side blur detecting means 16, the learning progress is determined according to the respective performances. This is because if the detection accuracy of the amount of blurring that is the basis of learning is not good, highly accurate learning cannot be performed.

The learning method for creating the trained model in the camera side learner 5c and the lens side learner 12c is as follows.

First, the camera system control unit 5 obtains learning data including time-series data of the amount of blur and time-series data of the remaining amount of blur based on the amount of blur and the live view image detected by the camera-side blur detection means 15. create. Similarly, the lens system control unit 12 is learning data including time-series data of the amount of blur and time-series data of the remaining amount of blur based on the amount of blur and the live view image detected by the lens-side blur detecting means 16. To create.

Of the above learning data, the time-series data of the amount of blur obtained from the camera-side blur detecting means 15 is used as the input data in the camera-side learning device 5c. Similarly, among the above-mentioned learning data, the time-series data of the amount of blur obtained from the lens-side blur detecting means 16 is used as the input data in the lens-side learner 12c.

Further, the camera system control unit 5 and the lens system control unit 12 calculate the time-series data of the correct answer value of the blur correction amount as the correct answer data, respectively. Specifically, the camera system control unit 5 is based on the amount of residual blur, the time-series data of the amount of blur detected by the camera-side blur detecting means 15, and the ratio of the drive amount of the share of the image sensor 6 described above. Calculate the correct answer data. On the other hand, the lens system control unit 12 calculates the correct answer data based on the ratio of the remaining blur amount, the blur amount detected by the lens side blur detecting means 16, and the drive amount of the share of the blur correction lens 3a described above. ..

Using these input data and correct answer data, each of the camera-side learner 5c and the lens-side learner 12c learns the control of image blur correction so that the amount of residual blur becomes 0, and creates a trained model. The trainer data of the trained model refers to the weight of each node in the neural network of the trained model. The algorithm applicable to the camera-side learner 5c and the lens-side learner 12c is not limited to the neural network. For example, if it is machine-learned based on the time-series data of each inertial sensor, it may be a model that predicts other time-series data, or a state-space model such as a linear / non-linear regression model or a Kalman filter.

The image blur correction control refers to, for example, the correction gain amount of the image sensor 6 and the phase compensation value of the blur correction lens 3a. The camera-side target value generation means 5d generates a drive target value of the camera-side blur correction means 14 by using the learning result of the control of the image blur correction by the camera-side learner 5c. Further, the lens-side target value generation means 12d generates a drive target value of the lens-side blur correction means 13a by using the learning result of the control of the image blur correction by the lens-side learner 12c. However, as described above, one of the camera system control unit 5 and the lens system control unit 12 may generate a drive target value and notify the other of the drive target value.

Further, as a method of acquiring time-series data of the remaining amount of blur, an image on the image sensor 6 is obtained from a time change of a live view image during image blur correction using the camera side blur correction means 14 and the lens side blur correction means 13a. There is a method of detecting the amount of movement of the lens as the amount of remaining blur.

(Explanation of image blur correction flow)
Next, the image blur correction process in this embodiment will be described with reference to the flowchart of FIG.

In this process, the camera body 1 and the lens barrel 2 are in a state where the learner data of the camera-side learner 5c and the lens-side learner 12c described in FIG. 2 with the higher learning progress is used by both of them. Image blur correction is performed in each case.

Further, in this process, every time the power of the camera body 1 (not shown) is turned on or the camera body 1 wakes up from the sleep state, the camera system control unit 5 reads a program held in the ROM (not shown) in FIG. Start.

First, in step S3001, the camera system control unit 5 communicates with the lens system control unit 12, confirms the connection destination that is communicably connected to the camera body 1 via the electrical contact 11, and the confirmed connection destination is Check if it is the same as the previous check. That is, the camera system control unit 5 determines in step S3001 whether or not the lens barrel has been attached or detached. As a result, if it is the same as the previous confirmation, the process proceeds to step S3002, and if it is different from the previous confirmation, the process proceeds to step S3003.

In step S3002, the anti-vibration control selection means 21 in the camera system control unit 5 acquires learning progress from each of the camera-side learner 5c and the lens-side learner 12c and compares them. From the result, the anti-vibration control selection means 21 selects a learning device having a high learning progress, reads the learning device data of the selected learning device, and reads the reading device to the other learning device via the electrical contact 11. Send data. As a result, the same trained model is set in the camera-side learner 5c and the lens-side learner 12c, and then the process proceeds to step S3003. In step S3002, when selecting the learner, the performance of the camera-side blur detecting means 15 and the lens-side blur detecting means 16 may be taken into consideration. For example, when the performance difference between the camera-side blur detecting means 15 and the lens-side blur detecting means 16 is large, the performance difference may have a greater influence on the image blur correction performance than the learning progress of the learner. Therefore, it is preferable to select the learning device in consideration of the performance of the camera-side blur detecting means 15 and the lens-side blur detecting means 16 in addition to the learning progress of the camera-side learning device 5c and the lens-side learning device 12c. The method of comparing the performances of the camera-side blur detecting means 15 and the lens-side blur detecting means 16 is not particularly limited. For example, each model may be stored in the camera system control unit 5 or the lens system control unit 12, and the models may be compared with each other via the electrical contacts 11. Further, the formats may be compared with each other by using the filter coefficients set in the camera-side filter processing means 5a and the lens-side filter processing means 12a.

In step S3003, the camera system control unit 5 determines whether or not the preparatory shooting operation has been started. As a result of this determination, if the pre-shooting operation is started, the camera system control unit 5 performs an operation (light measurement operation or AF operation) executed before the shooting of the camera body and proceeds to step S3004, otherwise the step Stand by at S3003. The preparatory shooting operation referred to here refers to a so-called half-press operation of a shutter release button (not shown) provided in the camera body 1. During the pre-shooting operation, the user is considered to hold the camera system 100 in the same manner as in the shooting state, so that the amount of blurring is also approximately the same as in the shooting state, which is suitable for executing the learning process of step S3005 described later. , Such a determination is made.

In step S3004, the camera system control unit 5 determines whether or not to start learning of the camera-side learner 5c and the lens-side learner 12c. If learning is to be started, the process proceeds to step S3005, and if learning is not started, the process proceeds to step S3005. , Step S3006. This determination may be automatically executed by the camera system control unit 5. Further, the user may be able to set in advance whether or not to start learning in step S3004. For example, the camera system control unit 5 determines that among the camera-side learner 5c and the lens-side learner 12c, the learner whose learning progress acquired in step S3002 is less than a predetermined progress is to start learning. To do. On the other hand, it is determined that learning is not started for the learners that do not.

In step S3005, if the camera system control unit 5 and the lens system control unit 12 have a learner determined to start learning in step S3004, the camera system control unit 5 and the lens system control unit 12 execute an operation process (learning process) of the learner, and then step S3006. Proceed to. This learning process will be described in detail later with reference to FIG.

In step S3006, the camera system control unit 5 determines whether or not the exposure operation has been started, and if the exposure operation has been started, the camera system control unit 5 performs the imaging operation of the camera body 1 and proceeds to step S3007. If not, the process returns to step S3003. The exposure operation referred to here refers to a so-called full-press operation of the shutter release button.

In step S3007, the camera system control unit 5 performs a shooting operation and proceeds to step S3008. During this shooting operation, the camera system control unit 5 controls the image blur correction by the image sensor 6 by the following method. First, the camera-side learner 5c in which the trained model is set in step S3002 or step S3005 outputs the correction gain amount of the image sensor 6 with respect to the blur amount detected by the camera-side blur detection means 15. Next, the camera-side target value generation means 5d generates a drive target value based on this correction gain amount. Then, the camera system control unit 5 outputs the generated drive target value to the camera-side blur correction means 14, so that the camera-side blur correction means 14 performs image blur correction by the image sensor 6. Similarly, the lens system control unit 12 controls image blur correction by the blur correction lens 3a by the following method. First, the lens-side learner 12c in which the trained model is set in step S3002 or step S3005 outputs the phase compensation value of the blur correction lens 3a with respect to the blur amount detected by the lens-side blur detection means 16. Next, the lens-side target value generation means 12d generates a drive target value based on this phase compensation value. Then, the lens system control unit 12 outputs the generated drive target value to the lens side blur correction means 13a, so that the lens side blur correction means 13a corrects the image blur by the blur correction lens 3a.

In step S3008, the camera system control unit 5 determines whether or not the shooting is completed by the shutter release button or other operation means (not shown) provided on the camera body 1. As a result of this determination, if the shooting is completed, the present process is terminated, and if not, the process returns to step S3001.

(Explanation of learning process)
Next, the learning process of step S3005 of FIG. 3 will be described with reference to the flowchart of FIG.

In this processing, the time-series data of the amount of blur detected by the camera-side blur detecting means 15 and the lens-side blur detecting means 16 with respect to the learning device that starts the learning process among the camera-side learning device 5c and the lens-side learning device 12c. Learning is performed from the time-series data of the remaining amount of blurring. Here, a case where the camera-side learning device 5c is a learning device that starts the learning process will be described.

In FIG. 4, first, the camera system control unit 5 communicates with the lens system control unit 12 and determines whether or not to stop the operation of the lens side blur correction means 13a (step S4001). As a result of the determination, if the lens side blur correction means 13a is stopped, the process proceeds to step S4002, and if not, the process proceeds to step S4003.

In step S4002, the camera system control unit 5 stops the operation of the lens side blur correction means 13a, and proceeds to step S4003.

In step S4001, whether or not to stop the lens-side blur correction means 13a is basically determined by the setting at the time of blur correction. For example, in the case of setting to drive both the lens side blur correction means 13a and the camera side blur correction means 14 to perform blur correction, it is necessary to stop the lens side blur correction means 13a, so the process proceeds to step S4002. On the other hand, for example, in the case of setting to drive only the camera-side blur correction means 14 to perform blur correction, it is not necessary to stop the lens-side blur correction means 13a, so the process proceeds to step S4003. Even when the lens barrel 2 does not have the lens-side blur correction means 13a, it is not necessary to stop the lens-side blur correction means 13a, so the process proceeds to step S4003.

In step S4003, the camera system control unit 5 acquires the learning TV, which is the period when the learning device learning in this process executes one learning, and proceeds to step S4004. At this time, the learning TV may be acquired by the camera system control unit 5 reading a preset numerical value, or may be a user-input numerical value detected by the operation detection unit 10.

In step S4004, the camera system control unit 5 determines whether or not to increase the frame rate of the live view image, and if the frame rate is increased, the process proceeds to step S4005. If not, the process proceeds to step S4006.

In this embodiment, the learning process of FIG. 4 is performed using the live view image, but the frame rate of the live view image is slow, causing camera shake or blurring of the subject. Cannot use the live view image as an image for creating correct answer data. Therefore, if the frame rate of the live view image is not sufficiently high with respect to camera shake or movement of the subject, the camera system control unit 5 determines in step S4004 to increase the frame rate of the live view image. Specifically, the determination in step S4004 is performed based on the movement of the subject in the live view image at the current frame rate acquired so far. For example, in a live view image at the current frame rate, if the subject moves quickly and is blurred, it is determined to increase the frame rate.

In step S4005, the camera system control unit 5 increases the frame rate of the live view image and proceeds to step S4006. In this case, only the frame rate of the live view image displayed on the rear display device 9a may be increased, and the frame rate of the live view image displayed on the EVF 9b may be kept constant without being changed.

In step S4006, the camera system control unit 5 reads out the number of live view images corresponding to the period set by the learning TV, and is detected by the camera side shake detecting means 15 at the time corresponding to the period in which the live view images are read out. Read the amount of blur. Then, the process proceeds to step S4007.

In step S4007, the camera system control unit 5 creates learning data based on the live view image read in step S4006 and the amount of blur detected by the camera side blur detecting means 15, and proceeds to step S4008. In step S4007, the camera system control unit 5 calculates the amount of movement of the subject image between each frame from the plurality of live view images (hereinafter referred to as each frame) read in step S4006, and blurs the amount of movement. Create time-series data for the remaining amount. In addition, the amount of blur detected by the camera-side blur detecting means 15 at the time corresponding to each frame is sampled, and time-series data of the amount of blur is created.

In step S4008, the camera system control unit 5 creates time-series data (hereinafter, correct answer data) of the correct answer value of the blur correction amount, and proceeds to step S4009. Specifically, in the correct answer data, the drive amount of the camera-side shake correction means 14 at the time corresponding to each frame is added to the remaining shake amount of each frame calculated in step S4007, and the drive amount of the above-mentioned share of the image sensor 6 is driven. Created by adding the value divided by the amount.

In step S4009, the camera system control unit 5 uses the time-series data of the amount of blur created in step S4007 as input data, and uses the correct answer data generated in step S4008 to learn the control of image blur correction in the camera-side learner 5c. To do. Then, the process proceeds to step S4010.

In step S4010, the camera system control unit 5 receives the learning result in step S4009, updates the trained model in the camera-side learner 5c, and proceeds to step S4011. At this time, the number of times of learning and the time of learning (for example, the time obtained by integrating the learning Tv with the number of learnings) are also recorded, and the number of learnings is used as the learning progress in the step described in FIG. It may be used when selecting a learning device in S3002.

In step S4011, the camera system control unit 5 determines whether or not to end the learning, and if the learning is finished, the present process is finished. If not, the process returns to step S4003.

If learning is performed while moving both the lens-side blur correction means 13a and the camera-side blur correction means 14 at the same time, each other's driving motion affects the learning of each, so the learning of the other is performed while one is stationary. It is preferable to do it. Therefore, in this embodiment, the camera-side learning device 5c performs the learning process of FIG. 4 while stopping the lens-side blur correction means 13a and driving the camera-side blur correction means 14. Further, at this time, the timing of stopping the lens-side blur correction means 13a is such that the lens-side blur correction means 13a drives the blur correction lens 3a until the center of the blur correction lens 3a coincides with the center of the optical axis 4. It is preferable to set the timing later.

Similarly, it is preferable to stop the camera-side blur correction means 14 and drive the lens-side blur correction means 13a to learn the lens-side learner 12c. Further, the timing at which the camera-side blur correction means 14 is stopped is the timing after the camera-side blur correction means 14 drives the image sensor 6 until the center of the image sensor 6 coincides with the center of the optical axis 4. It is preferable to do so.

In addition, vibration isolation control by the image sensor 6 and the blur correction lens 3a using only the learning result of the camera side learning device 5c, and vibration isolation when only the learning result of the lens side learning device 12c is used during the preliminary shooting operation. Control may be performed and the learning progress may be determined according to the accuracy thereof. Specifically, in this embodiment, the remaining amount of blurring of the live view images under vibration isolation control is compared, and it is determined that the learning progress of the learning device having a smaller remaining amount of blurring is higher.

As described above, in this embodiment, the learning progresses of the camera-side learner 5c and the lens-side learner 12c are compared, and the learner data of the learner having the higher learning progress is used as the camera-side learner 5c and the camera-side learner 12c. It is used in both the lens-side learner 12c. As a result, it is possible to reduce the deterioration of the image blur correction performance due to the difference in the learning progress between the camera-side learning device 5c and the lens-side learning device 12c.

(Example 2)
Hereinafter, the vibration isolator according to the second embodiment of the present invention will be described with reference to FIG. The second embodiment is different from the first embodiment in that the trained model is held for each focal length in each of the camera-side learner 5c and the lens-side learner 12c, and the other configurations are basically the first embodiment. Is similar to. Therefore, hereinafter, the same components as those in the first embodiment are designated by the same reference numerals, and duplicate description will be omitted. The focal length changes according to the position of the focus lens included in the photographing optical system 3 of the lens barrel 2.

FIG. 5 is a flowchart of the image blur correction process in this embodiment.

In this embodiment, each of the camera-side learner 5c and the lens-side learner 12c holds a trained model for each focal length. Then, each time the focal length is changed, the anti-vibration control selection means 21 in the camera system control unit 5 acquires and compares the learning progress of each of the camera-side learning device 5c and the lens-side learning device 12c.

Further, in the same process as in FIG. 3, the camera system control unit 5 holds the camera body 1 in the ROM (not shown) in FIG. 1 each time the power is turned on in the camera body 1 or the camera body 1 wakes up from the sleep state. The program to be loaded is loaded and started.

Hereinafter, only the steps different from those in FIG. 3 will be described.

First, in step S5001, the camera system control unit 5 determines whether or not the focal length has been changed, and if the focal length has been changed, the process proceeds to step S5002, and if not, the process proceeds to step S3003.

In step S5002, the anti-vibration control selection means 21 acquires and compares the learning progress at the focal length after the change from each of the camera-side learner 5c and the lens-side learner 12c. From the result, the anti-vibration control selection means 21 selects a learning device having a high learning progress, reads the learning device data of the selected learning device, and reads the reading device to the other learning device via the electrical contact 11. Data is transmitted, and the process proceeds to step S3003. The selection of the learner here is performed in the same manner as in step S3002.

In step S3003, as in the case of FIG. 3, the camera system control unit 5 determines whether or not the preparatory shooting operation has been started. As a result, if the pre-shooting operation is started, the camera system control unit 5 performs the pre-shooting operation and proceeds to step S3004, otherwise returns to step S5001 and determines again whether or not the focal length has been changed. ..

Further, in step S3006, the camera system control unit 5 determines whether or not the exposure operation has been started, as in the case of FIG. As a result, if the exposure operation is started, the camera system control unit 5 performs the imaging operation of the camera body 1 and proceeds to step S3007. On the other hand, if not, the process returns to step S5001 and it is determined again whether or not the focal length has been changed.

As described above, in this embodiment, the learning progress of the camera-side learner 5c and the lens-side learner 12c at the current focal length is compared each time the focal length is changed. Then, as a result of this comparison, the learner data of the learner having the higher learning progress is used in both the camera side learner 5c and the lens side learner 12c. This makes it possible to compare the learning progress between the camera-side learner 5c and the lens-side learner 12c for each focal length, that is, in more detail, and the image blur correction performance deteriorates due to the learning progress. Can be reduced more finely.

(Example 3)
Hereinafter, the vibration isolator according to the third embodiment of the present invention will be described with reference to FIG. In the third embodiment, the timing for determining which of the camera-side learning device 5c and the lens-side learning device 12c is the learning device having a high learning progress is different from that in the first embodiment. More specifically, in the first embodiment, the determination is performed each time the lens barrel 2 is attached to or detached from the camera body 1, whereas in the third embodiment, at least the camera-side learner 5c and the lens-side learner 12c are used. On the other hand, each time learning is carried out, the judgment is made. Other configurations are basically the same as in the first embodiment. Therefore, hereinafter, the same components as those in the first embodiment are designated by the same reference numerals, and duplicate description will be omitted.

FIG. 6 is a flowchart of the image blur correction process in this embodiment.

In this embodiment, each time the learning process described with reference to FIG. 4 is performed, it is determined which of the camera-side learning device 5c and the lens-side learning device 12c has a higher learning progress.

Further, in the same process as in FIG. 3, the camera system control unit 5 holds the camera body 1 in the ROM (not shown) in FIG. 1 each time the power is turned on in the camera body 1 or the camera body 1 wakes up from the sleep state. The program to be loaded is loaded and started.

Hereinafter, only the steps different from those in FIG. 3 will be described.

First, in step S3003, the camera system control unit 5 determines whether or not the preparatory shooting operation has been started. As a result of this determination, if the pre-shooting operation is started, the camera system control unit 5 performs the pre-shooting operation to proceed to step S3004, and if not, waits in step S3003.

After that, after performing the learning process in step S3005, in step S6001, the vibration isolation control selection means 21 in the camera system control unit 5 acquires the learning progress from each of the camera side learner 5c and the lens side learner 12c. Compare. From the result, the anti-vibration control selection means 21 selects a learning device having a high learning progress, reads a learned model of the selected learning device, and reads the learned model to the other learning device via the electrical contact 11. The model is transmitted, and the process proceeds to step S3006.

As described above, in this embodiment, the learning progress of the camera-side learning device 5c and the lens-side learning device 12c is compared each time the learning process is performed. Then, as a result of this comparison, the trained model constructed in the learning device having the higher learning progress is used in both the camera-side learning device 5c and the lens-side learning device 12c. As a result, the learning progress between the camera-side learning device 5c and the lens-side learning device 12c can be compared in the latest state, and the deterioration of the image blur correction performance due to the difference in the learning progress is further reduced. It becomes possible to do. In this embodiment, the timing at which learning is performed on at least one of the camera-side learning device 5c and the lens-side learning device 12c is a preliminary shooting operation, but is not limited to such a mode. For example, such timing may be at the time of user operation to execute the learning process.

[Other Examples]
Needless to say, the object of the present invention is also achieved by supplying the device with a storage medium in which the program code of the software that realizes the functions of the above-described embodiment is recorded. At this time, the computer (or CPU or MPU) including the control unit of the supplied device reads and executes the program code stored in the storage medium.

In this case, the program code itself read from the storage medium realizes the functions of the above-described embodiment, and the program code itself and the storage medium storing the program code constitute the present invention.

As the storage medium for supplying the program code, for example, a flexible disk, a hard disk, an optical disk, a magneto-optical disk, a CD-ROM, a CD-R, a magnetic tape, a non-volatile memory card, a ROM, or the like can be used.

Further, based on the instructions of the above-mentioned program code, the OS (basic system or operating system) running on the device performs a part or all of the processing, and the processing realizes the functions of the above-described embodiment. Needless to say, cases are also included.

Further, the program code read from the storage medium may be written in the memory provided in the function expansion board inserted in the device or the function expansion unit connected to the computer to realize the functions of the above-described embodiment. Needless to say, it is included. At this time, based on the instruction of the program code, the function expansion board, the CPU provided in the function expansion unit, or the like performs a part or all of the actual processing.

1 Camera body 2 Lens lens barrel 3 Shooting optical system 3a Blur correction lens 5 Camera system control unit 5b Camera side vibration isolation control means 5c Camera side learner 5d Camera side target value generation means 6 Imaging element 11 Electrical contact 12 Lens system control Part 12b Lens-side anti-vibration control means 12c Lens-side learner 12d Lens-side target value generation means 13a Lens-side blur correction means 14 Camera-side blur correction means 15 Camera-side blur detection means
16 Lens side shake detecting means 21 Anti-vibration control selecting means

Claims (18)

An anti-vibration device that connects to the lens barrel so that it can communicate with it.
The anti-vibration device
Image sensor and
The first blur detection means for detecting the amount of blur,
The drive target value of the image sensor is calculated according to the learning result of the first learning device that learns the control of image blur correction based on the amount of blur detected by the first blur detecting means and the first learning device. A first control means including a first calculation means for
A first anti-vibration means for correcting image blurring by driving the image sensor with a drive target value output from the first control means is provided.
The lens barrel
Shooting optics including a blur correction lens,
A second blur detection means for detecting the amount of blur,
A second learner that learns the control of image blur correction based on the amount of blur detected by the second blur detecting means, and a drive target of the blur correction lens according to the learning result by the second learner. A second control means including a second calculation means for calculating the value, and
A second anti-vibration means for correcting image blur by driving the blur correction lens with a drive target value output from the second control means is provided.
The anti-vibration device further
The learning progress of the first and second learners is determined, and as a result of the determination, the learner data having the higher learning progress is selected as the learner data used for learning the control of the image blur correction. An anti-vibration device characterized by comprising anti-vibration control selection means. It is a vibration isolation device that connects to the camera body so that it can communicate.
The camera body
Image sensor and
The first blur detection means for detecting the amount of blur,
The drive target value of the image sensor is calculated according to the learning result of the first learning device that learns the control of image blur correction based on the amount of blur detected by the first blur detecting means and the first learning device. A first control means including a first calculation means for
A first anti-vibration means for correcting image blurring by driving the image sensor with a drive target value output from the first control means is provided.
The anti-vibration device
Shooting optics including a blur correction lens,
A second blur detection means for detecting the amount of blur,
A second learner that learns the control of image blur correction based on the amount of blur detected by the second blur detecting means, and a drive target of the blur correction lens according to the learning result by the second learner. A second control means including a second calculation means for calculating the value, and
A second anti-vibration means for correcting image blur by driving the blur correction lens with a drive target value output from the second control means is provided.
The anti-vibration device further
The learning progress of the first and second learners is determined, and as a result of the determination, the learner data having the higher learning progress is selected as the learner data used for learning the control of the image blur correction. An anti-vibration device characterized by comprising anti-vibration control selection means. The vibration isolator according to claim 1 or 2, wherein the first and second learners generate a trained model using the learner data selected by the vibration isolation control selection means. When the learner data of the first learner is selected by the anti-vibration control selection means, the first calculation means drives the blur correction lens according to the learning result of the first learner. The vibration isolator according to claim 1 or 2, wherein a target value is calculated and output to the second vibration isolation means via the first control means. The first learner receives time-series data of the amount of blur detected by the first blur detecting means as input data, the amount of driving of the image pickup element by the first vibration isolator means, and the image formation by the image pickup element. The control of the image blur correction is learned by using the time series data of the correct answer value calculated based on the remaining blur amount, which is the motion amount of the image, as the correct answer data.
The second learning device inputs time-series data of the amount of blur detected by the second blur detecting means, the driving amount of the blur correction lens by the second vibration isolating means, and the remaining blur amount. The anti-vibration device according to any one of claims 1 to 4, wherein the time-series data of the correct answer value calculated based on the above is used as the correct answer data, and the control of the image blur correction is learned. Claims 1 to 5, wherein the anti-vibration control selection means determines the learning progress of the first and second learning devices according to the performance of the first and second blur detecting means. The anti-vibration device according to any one of the above. The first learning device is characterized in that the driving of the second anti-vibration means is stopped and learning is performed in a state where only the image sensor is driven by the first anti-vibration means. The anti-vibration device according to any one of 6 to 6. The second learning device is characterized in that the driving of the first anti-vibration means is stopped, and learning is performed in a state where only the blur correction lens is driven by the second anti-vibration means. Item 3. The anti-vibration device according to any one of Items 1 to 7. The anti-vibration control selection means any of claims 1 to 8, wherein the learning device having a larger number of learnings among the first and second learning devices is determined to have a higher learning progress. The anti-vibration device according to item 1. The anti-vibration control selection means any of claims 1 to 9, wherein the learning device having a longer learning time among the first and second learning devices is determined to have a higher learning progress. The anti-vibration device according to item 1. The anti-vibration control selection means
At the time of preparatory operation for photographing, the remaining amount of blur is measured while the image sensor and the lens for blur correction are driven by the first and second anti-vibration means using only the learning result by the first learning device. The first acquisition means to acquire and
At the time of preparatory operation for photographing, the remaining amount of blur is measured while the image sensor and the lens for blur correction are driven by the first and second anti-vibration means using only the learning result by the second learner. With a second acquisition means to acquire,
The anti-vibration device according to any one of claims 1 to 10, wherein the learning progress is determined according to the amount of residual blur acquired by each of the first and second acquisition means. .. The photographing optical system further includes a focus lens that changes the focal length.
Each of the first and second learners holds a trained model according to the focal length, and holds a trained model.
The anti-vibration control selection means learns the control of image blur correction according to the learning progress of the first and second learners according to the focal length after the change each time the focal length changes. The anti-vibration device according to any one of claims 1 to 11, wherein the learner data used in the above is selected. Further provided with a confirmation means for confirming the connection destination that is communicably connected to the vibration isolator.
When the connection destination confirmed by the confirmation means is different from that at the time of the previous confirmation, the vibration isolation control selection means learns the control of the image blur correction based on the learning progress of the first and second learners. The anti-vibration device according to any one of claims 1 to 12, wherein the learner data used in the above is selected. Each time at least one of the first and second learners is learned, the anti-vibration control selection means controls the image blur correction based on the learning progress of the first and second learners. The anti-vibration device according to any one of claims 1 to 13, wherein the learner data used for learning is selected. The anti-vibration device according to claim 14, wherein at least one of the first and second learning devices is learned every time a preparatory shooting operation is performed. It is a control method of the anti-vibration device that connects to the lens barrel so that it can communicate.
The anti-vibration device includes an image sensor, and the lens barrel includes a lens for blur correction.
In the anti-vibration device
The first blur detection step to detect the amount of blur,
A first learning step of learning the control of image blur correction based on the amount of blur detected in the first blur detection step, and
The first calculation step of calculating the drive target value of the image sensor according to the learning result in the first learning step, and
It has a first vibration isolation step that corrects image blur by driving the image sensor with a drive target value calculated in the first calculation step.
In the lens barrel
The second blur detection step to detect the amount of blur, and
A second learning step of learning the control of image blur correction based on the amount of blur detected in the second blur detection step, and
The second calculation step of calculating the drive target value of the blur correction lens according to the learning result in the second learning step, and
It has a second anti-vibration step in which the image blur correction is performed by driving the blur correction lens with the drive target value calculated in the second calculation step.
In the anti-vibration device,
The learning progress in the first and second learning steps is determined, and as a result of the determination, the learning device data having the higher learning progress is selected as the learning device data used for learning the control of the image blur correction. Anti-vibration control A control method comprising a selection step. It is a control method of the anti-vibration device that connects to the camera body so that it can communicate.
The camera body is provided with an image sensor, and the vibration isolator is provided with a blur correction lens.
In the camera body
The first blur detection step to detect the amount of blur,
A first learning step of learning the control of image blur correction based on the amount of blur detected in the first blur detection step, and
The first calculation step of calculating the drive target value of the image sensor according to the learning result in the first learning step, and
It has a first vibration isolation step that corrects image blur by driving the image sensor with a drive target value calculated in the first calculation step.
In the anti-vibration device
The second blur detection step to detect the amount of blur, and
A second learning step of learning the control of image blur correction based on the amount of blur detected in the second blur detection step, and
The second calculation step of calculating the drive target value of the blur correction lens according to the learning result in the second learning step, and
It has a second anti-vibration step in which the image blur correction is performed by driving the blur correction lens with the drive target value calculated in the second learning step.
In the anti-vibration device,
The learning progress in the first and second learning steps is determined, and as a result of the determination, the learning device data having the higher learning progress is selected as the learning device data used for learning the control of the image blur correction. Anti-vibration control A control method comprising a selection step. A program comprising executing the control method according to claim 16 or 17.

Images