JPH04133016A - Overlap system autofocusing device - Google Patents

Abstract

PURPOSE:To prevent getting in and out of a tracking state, to stably correct the motion of an object and to smooth the driving of a lens by applying hysteresis to the condition of a moving body decision test once tracking servo is started. CONSTITUTION:A defocusing quantity is obtained by AF algorithm based on the focus detection signal of a photoelectric conversion element 101 from a focus detection means 103 by a defocusing quantity arithmetic means 104, and the image-plane speed of the object is arithmetically operated by an equation (14) based on an image plane moving quantity in two cycles of range-finding by an object image surface speed arithmetic means 105. In the case, the speed is obtained based on the range-finding data at this time and the range-finding data at least before two cycles so as to improve accuracy. A tracking correcting quantity is sometimes obtained by an equation (15) by neglecting a lens driving quantity calculated by a means 107 from (tn) to (tm). Then, a lens driving quantity is obtained by an equation (16), and a photographing lens 102 is sometimes driven PD(n) longer than an object position tn+1 by a driving means 109 so as to perform the servo so that the lens 102 is in a focusing state. Then, the decision of a moving body is performed by a moving body decision means 201, and when it is recognized to be the moving body, the tracking action is stably continued by applying the hysteresis. Thus, the lens is driven with good responsiveness without giving the sense of incongruity to actuations.

Description

DETAILED DESCRIPTION OF THE INVENTION A. Industrial Application Field The present invention relates to an automatic focus adjustment device for a camera, etc. More specifically, when a moving subject is moving, the moving subject is always kept in focus. The present invention relates to an autofocus device that performs tracking according to the current state, and also overlaps the charge storage of a charge storage type photoelectric conversion element even while driving a photographic lens.

B1 Background of the Invention The applicant has developed a charge storage type photoelectric conversion element (AF sensor).
As a control method for the AF servo system using the AF servo system, the servo time and A so-called "overlap servo" control method that improves accuracy was proposed in Japanese Patent Application Laid-Open No. 2-146010. On the other hand, in recent years, AF servo technology has been developed to detect not only when the subject is stationary but also when the subject is moving, and to use the servo while predicting the position of the moving subject (hereinafter referred to as a moving subject). It has been proposed that some things be done, and some of them have been implemented.

The present invention describes a method for applying an overlap servo control method to such a "moving object".

C0 Prior Art FIG. 11 is a general block diagram of an automatic focus adjustment (autofocus) device that drives a photographic lens by a motor and servos it to a focused state.

In FIG. 11, the focus detection light flux from the subject that has passed through the photographic lens 1 forms an image on a charge storage type photoelectric conversion element 2 such as a CCD provided in the camera body, and a light image from the photoelectric conversion element 2 is formed. The signal is sent via an interface 3 to a controller (hereinafter referred to as CPU) 4 consisting of a microprocessor that controls the entire system.

The optical image pattern of the focus detection light beam projected onto the photoelectric conversion element 2 is AD converted by the interface 3 and sent to the CPU 4.
or is amplified to an appropriate signal level by the interface 3 and directly AD-converted by the AD converter built in the CPU 4. The CPU 4 processes the optical image pattern thus converted into a digital signal using a predetermined algorithm to calculate the so-called defocus amount, and based on this calculates the amount of movement of the photographing lens required to bring the image into focus. . Here, since the optical principles and algorithms for specifically detecting the amount of defocus are already well known, their explanation will be omitted.

The photographic lens 1 is provided with an encoder 6 to monitor its movement, and the encoder 6 generates a pulse every time the photographic lens 1 moves a certain amount along the optical axis. C
The PU 4 instructs the motor driver 5 about the calculated lens movement amount, drives the servo motor 7, and drives the photographing lens 1 in the focusing direction. The movement of the photographic lens 1 is monitored by the CPU 4 using feedback pulses from the encoder 6, and when the pulses from the encoder 6 are counted by the number of pulses corresponding to the amount of defocus, the driving of the motor 7 for driving the photographic lens is stopped. Normally, the encoder 6 includes a photointerrupter attached to a rotating shaft of the motor 7 or a part of a reduction gear, and detects the rotation of the motor 7 for driving the photographic lens.

FIG. 12 is an explanatory diagram of the amount of defocus detected by the automatic focus adjustment device.

Here, the defocus amount refers to the surface (imaging surface) on which the focus detection light beam transmitted through the photographic lens 1 forms an image, and the film surface (
It is expressed in terms of the amount and direction of deviation relative to the planned imaging plane).

A conventional example of performing tracking servo on a moving object using the automatic focus adjustment device shown in FIG. 11 will be explained with reference to FIG. 13.

In the figure, the horizontal axis is time, and the vertical axis is the position of the image plane. The diagram indicated by Q in FIG. 13 shows how the imaging position of the subject moves with time as the subject moves. This is called a subject image plane movement curve. The line diagram shown as
This shows the image plane position when moving the photographic lens to focus on this subject. This is the image plane movement curve due to the photographing lens.

Alternatively, it is called a lens image plane movement curve. Therefore Q and L
The difference between them indicates the amount of defocus. Also tn, tn-
The time indicated by 1, etc. is approximately the center time of each accumulation period of the photoelectric conversion element, and the period sandwiched between two vertical lines drawn upward to Q or L across this point is the accumulation period. From then on,
In other figures, the vertical axis, horizontal axis, diagram Q, and diagram have the same meaning, for example, times t n-1, t
The defocus amount at n is D (n-1), D (
n).

[Conventional Method of Tracking Servo] An example of a conventional tracking servo for a moving object will be explained with reference to FIG.

The conventional AF servo method alternately repeats distance measurement and lens bouncing.
It is called "discrete servo".

Note that distance measurement here refers to charge accumulation in the photoelectric conversion element,
It is used to include calculation of the amount of defocus based on the focus detection signal output from the photoelectric conversion element and calculation of the driving amount of the photographing lens based on the amount of defocus.

In FIG. 13, the portion where the line diagram rises and the position of the image plane changes is the lens driving period. If the subject is moving, the imaging position of the subject will also change over time like Q, so the defocus amount D (n-1) obtained by distance measurement at time tn-■ is servoed and then set at time tn. Even if the distance is measured again, the defocus amount D (n
) is detected. As is clear from the figure, in the discrete servo, the defocus amount D(n) at time tn is
This is detected because the subject moved between time tn-1 and time tn. If the subject is stationary and the distance measurement accuracy and servo accuracy are sufficiently high, D (n) is D (n
-1), and it is presumed that the reason why D(n) has a considerable size is because the subject is moving.

If the ranging period t (n) - t (n-1) is approximately the same each time and the image surface velocity of the subject (the slope of Q) is also constant, the defocus amount D (n) is also approximately the same each time. . conventionally,
If the subject is determined to be a moving object, its movement is predicted, and the correction amount is added to the raw defocus amount D (n), so that, for example, the next defocus amount is zero. If the subject is determined to be a moving object for the first time when the distance measurement result at time tn in the figure is calculated, then in order to make the next defocus amount zero, the lens drive amount will be 2D(n). . This is the correction amount (hereinafter referred to as "correction amount") that should be added to the raw defocus amount D (n).

C(n)), that is, the amount by which the subject is predicted to move by the next distance measurement time tn+1 is also D(n).

Although this is sufficient as the initial correction amount, FIG. 14 shows an analysis of how subsequent correction amounts are generally calculated.

In the figure, the movement of the subject has already been detected, and a correction value C(n) that takes into account the image plane speed of the subject is added to the raw defocus amount each time, and the servo is adjusted so that the defocus amount becomes zero in the next distance measurement. It shows how to do it.

As shown in the figure, the correction amount C (n-1) is calculated for the raw defocus amount D (n-1) which is the distance measurement result at time tn-1.
Suppose that servo is performed by adding . Assuming that the actual movement of the lens image plane from time tn-1 to time tn until the photoelectric conversion element accumulates is M(n-1), in the discrete servo, no new photoelectric conversion is performed until the lens servo is completed. Because the conversion element does not accumulate.

M (n-1) = D (n-1) + C (n-1)
−(1). Defocus amount D (n) obtained by distance measurement at time tn and previous defocus amount D
(n-1) and the amount of movement of the image plane during this time M(n-1)
From the same figure, the movement amount P (n) of the image plane of the object from time tn-1 to time tn is P (n) = D (n) + M (n-1) -
D (n-1,) = D (n) ten C (n-1)
'= (2). Also, from this, the image plane speed S (n) of the subject is 5(n)”P(n)/ (tn-tn-1) ・
...(3). Therefore, the period from the current distance measurement time tn to the next distance measurement time tn+1 is the previous distance measurement time t
Since it is expected to be approximately the same as from n-1 to the current distance measurement time tn, the current correction amount C(n) is C(n) =
P (n) ... (4) is appropriate. Therefore, the total amount to be servoed, that is, the lens drive amount X(n) for the defocus amount at time t (n) is the raw defocus amount D (n) obtained by distance measurement at time tn and the correction amount. Sum of C(n), D(n)+C(n)=2D(n)+C(n-1)-(5
). According to this formula, the previous correction amount C(n-1)
=0, that is, if no tracking servo was performed before then, the servo amount is 2 D (n), which agrees with the analysis in FIG. 13.

If you use this correction method to perform tracking servo for a moving object,
As shown in Figure 15, the trajectory of the lens's image plane position traces the trajectory of the subject's image plane up and down, and at this time, the polarity of each distance measurement value D (n) can be either. I get it, but
In either case, feedback is applied using the correction amount C(n) so that the defocus amount at the next distance measurement timing becomes zero, so that the photographing lens is servoed so as to be in focus macroscopically. In other words, at each accumulation time tn, the focus is almost on, so the amount of defocus each time is almost zero, and the amount of lens image plane movement M (p) due to the servo is the correction amount C (n
) = P (n) becomes dominant.

The above are examples of basic conventional techniques of analysis and servo methods for tracking a moving object using discrete servos.

[Conventional method of motion detection] Next, a method for determining a moving object will be described.

In the explanation of FIG. 13, if the subject is a moving object, D (
n) should be sufficiently smaller than D (n-4),
We analyzed that the reason why D(n) is so large is presumed to be because the subject is moving, but even though this alone is a condition for entering the tracking servo, it is conversely true that the subject is moving. This is insufficient as a detection condition when the vehicle stops. In JP-A-63-14821.8, the ratio of the amount of movement P (n) of the imaging plane of the object, P (n)/P (n-1)
When γ takes a value between γ for a certain threshold value, that is.

g < P (n)/ P (n-1) < y
・-(6), or more precisely, when P
It has been proposed that (n) be divided by the distance measurement period (interval) and determined as the ratio of the speed of the subject.

In addition, in Japanese Patent Application Laid-open No. 62-253107, simply P
The condition for entering the tracking servo is that the amount corresponding to (n) is greater than a certain value.

[Conventional Method 1 of Aligning Exposure Timing with Focusing Time Next, a conventional method will be described in which the amount of lens drive is determined so that the lens is brought into focus at the timing when the shutter opens and exposure starts after the release operation.

It is important that the purpose of the tracking servo is to focus at the timing of film exposure when actually photographing.

As an example of servo after release, JP-A-63-148
In Publication No. 218, the time Td from the start of the servo (start of lens drive) to the start of the next distance measurement is kept constant, and the release is set so that the exposure time coincides with the next accumulation time that would be performed if the release was not performed. There are examples of how to control timing. This will be explained with reference to FIG.

Generally, even after the photoelectric conversion element has finished storing data, it takes several tens of meters to transfer the data (input time) and calculate the amount of defocus.
It takes approximately 5 ec to 100 m5 ec to start the servo, and in the figure, the servo start time based on the distance measurement result at time tn is expressed as tm. This is also the reason why each of the figures used so far has been drawn with a little time between the end of accumulation and the start of servo. In the method disclosed in Japanese Unexamined Patent Publication No. 63-148218, the time Td from the servo start time tm to the start of the next accumulation is controlled to be constant every time. This time is set to a time that ensures lens drive time for tracking servo for a subject moving at a certain high speed, while not sacrificing responsiveness too much, and is, for example, about 100 m5ec. Even if the servo ends earlier than this, the storage of the photoelectric conversion element for the next distance measurement is not started until this time Td has elapsed. The tracking servo method itself is the same as that explained in FIG. 14, and the method of calculating the correction amount C(n) is also the same. This has the effect of stabilizing the distance measurement cycle each time, but it is also expected that the in-focus state will be reached after Td time has passed after the servo starts, regardless of the speed of the subject. It is possible. Therefore, the release start time tr is set after Td-Tr has elapsed after the servo start, taking into consideration the mechanical delay time Tr until exposure actually starts.

Problems to be Solved by the D8 Invention However, the servo method, moving object detection method, and control method for adjusting the focus timing to the exposure time during release, which are used in conventional tracking servos using discrete servos, have been replaced by overlap servo type auto. If applied as is to a focus device, there are various inconveniences and many irrational aspects, as explained below, and it is necessary to fundamentally reconsider.

(1) Problems with the conventional tracking servo system Let us now consider applying the above tracking correction to an overlap servo as previously proposed by the applicant. 1st
FIG. 7 is a diagram showing overlap servo on a moving object Q, and the symbols are the same as in FIG. 14 and the like. As is clear from the figure, the amount of defocus D obtained by distance measurement at time tn
(n), the previous defocus amount D (n-1), and the image plane movement amount M (n-1) during this time, 1 time tn-1
The amount of movement P (
n) is P(n)=D(n)+M(n-1) D(n
-1) -(8). From this, the image plane movement speed S (n) of the subject is S (n) = P (n) / (t n -t n-
1) −(9). Both equations (8) and (9) are the same as in the case of discrete servo. However, with overlap servo, the next distance measurement starts before the servo ends, so the amount of lens image plane movement between time tn-1 and time tn is calculated by the following formula, M (n -1) = D (n-1) + C(n-1)
...(1) does not apply in general. Even with overlap servo, if we predict that the distance measurement period tn+1-tn is approximately equal each time, then from time tn to next distance measurement time tn+1
The amount of movement of the subject up to can be expressed as P (n) in equation (8) above. P (n) for discrete servo
It was good to use the correction amount C(n), but as explained earlier, with overlap servo, it takes CCD data transfer time and calculation time from the completion of accumulation in the photoelectric conversion element to the calculation of the defocus amount, and during this time, As shown in FIG. 17, the photographing lens is normally driven, so if the time indicated by t in FIG. 17 is the defocus calculation end time, the amount PD(n) the lens has moved from time tn to time 1++ P
It is appropriate to set the value subtracted from (n) as the correction amount at time tm.

Therefore, considering this, the amount of total lens movement to be servoed is calculated by subtracting PD(n) from the sum of the raw defocus amounts D(n) and P(n) obtained by distance measurement at time tn. So, D(n)+P(n)-PD(n)=2D(n)+M(n
-1) -D (n-1) -P D (n) (10).

However, through experiments conducted by the inventors, it has been found that the following problem occurs when formula (10) is applied to overlap servo. This will be explained in detail with reference to FIG.

Assume that the focus detection signal obtained by accumulation at time tn is processed to calculate the defocus amount D (n) at time tm, and the imaging position by the photographing lens at that time is defined as point A. If the amount of lens movement from time t, n to time tm is PD(n), the amount C(n) to be corrected at point A is: C(n) = D(n) + M(n-1)-' D(n-1)
−PD(n) ...(11),
The lens drive amount of equation (10) is obtained by adding the defocus amount D (n) at time tn to this. If the lens drive motor is sufficiently powerful and can instantaneously move the amount of drive expressed by equation (10), the lens will move from point A as shown in diagram L1.

At the next accumulation time tn+1, the image is almost in focus.

However, in reality, the motor power is limited, and the advantage of overlap servo is that the servo has good responsiveness.In order to not lose this, the lens must be driven as soon as the distance measurement result is obtained at time ti+. At the same time, I would like to start measuring the distance of the blowing. Therefore, in general, the photographing lens must be driven during the next distance measurement time as shown in the diagram L2 indicated by a broken line.

However, the correction amount that we have considered so far is the current defocus amount D (n ), so in a drive model like diagram L2, even if the correction amount in equation (4) is used as is, it will converge so that the defocus amount becomes almost zero in each distance measurement. Therefore, a servo as shown in diagram L2 cannot catch up with diagram Q and will always be located at the lower side.

As described above, if overlap servo is performed by following the servo method adopted in conventional tracking servo based on discrete servo, it will not be possible to focus.

(2) Problems with the conventional method for determining moving objects In the conventional example of discrete servo, the amount of movement P (n) of the object imaging plane and its ratio, P (n) /
P (n-1) or (P (n)/ (tn-
tn-1) )/(P (n-1)(tn-1-tn-
2)) etc. can be said to basically examine the moving speed of the subject and its stability. Even in the case of overlap servo, the subject speed can be determined using equation (9), so for example, S (n) > δ ... (12) or a<S (n) /S (n-1 ) <β −(1
3) The following conditions can be considered. Equation (12) requires that the speed of the subject is greater than a certain threshold value δ[am/5ecl, (
13) Equation is the ratio α of the continuously measured object speed,
This means that it is within the range of β. In other words, it is verified that the subject speed is stable.

All of these conditions are reasonable to some extent, but with overlap servo, the distance measurement cycle is extremely short and the amount of image plane movement of the subject during that period is small, so the subject speed S (n) calculated using equation (9) is Since the distance measurement data is obtained from two consecutive distance measurements, we cannot expect much detection accuracy.

(3) Problems with the conventional method for adjusting the exposure timing to the focusing time As mentioned above, in the conventional method, the time Td from the start of the servo to the start of the next distance measurement is kept constant, and the exposure timing by the release is set to the focusing time. Therefore, even if the release operation is performed, the release is not necessarily permitted immediately. Therefore, if such a method is applied to the overlap serve method in which the photographing lens is driven to track the subject even during accumulation, the advantages of the overlap method will be lost.

SUMMARY OF THE INVENTION An object of the present invention is to provide an overlap type autofocus device that enables a new tracking servo method, moving object determination method, and release control using an overlap method.

89 Means for Solving the Problem The invention of claim 1, which characterizes the present invention corresponding to the claim correspondence diagram in FIG. Conversion element 101
a focus detection means 103 that converts the focus detection light beam incident thereon into a focus detection signal and outputs it;
a defocus amount calculating means 104 that calculates a defocus amount consisting of the amount of deviation and the direction of deviation between the image forming plane of the focus detection light beam by the photographing lens 102 and the planned image forming plane based on the focus detection signal from; a predetermined time interval; A subject image plane speed calculation means for calculating a subject image plane velocity from two defocus amounts calculated by the two defocus amounts and the amount of movement of the image plane by the photographing lens 102 that moves toward the point where these two defocus amounts are calculated. 105; a correction amount calculating means 106 for calculating a correction amount correlated to the object image surface velocity calculated by the object image surface speed calculating means 105; and; adding this correction amount to the latest defocus amount to control the tracking servo lens driving amount calculating means 107 for calculating the lens driving amount for; photographing lens 1;
Lens movement amount detection means 1 for detecting the actual movement amount of 02
08; a driving means 109 for servoing the photographing lens 102 to the focusing position according to the lens driving amount calculated by the lens driving amount calculation means 107 and the detected actual lens movement amount; and; Control means 1 for controlling the charge accumulation type photoelectric conversion element 101 to overlap the charge accumulation type photoelectric conversion element 101 even during servo at the focus position, and to control the accumulation of the photoelectric conversion element 101 and the servo operation to be overlapped one after another.
The present invention is applied to an overlap type autofocus device having the following.

The above-mentioned problem of accuracy of the object image plane velocity is solved by using the object image plane velocity calculating means 105 according to the first aspect of the present invention, using two defocus amounts obtained from the latest focus detection signal and the previous focus detection signal for two or more cycles. In the invention of claim 2, the sampling time interval of the two focus detection signals used to determine the object image surface velocity is calculated based on the charge storage type photoelectron. The shorter the storage time of the conversion element 101, the longer it is.

According to a third aspect of the invention, the focus detection means 103 is provided.

A defocus amount calculating means 104, a subject image plane speed calculating means 105, and a moving object determining means for determining whether or not the subject is a moving object based on the subject image plane velocity calculated by the subject image plane velocity calculating means 105. 201, and the correction amount calculation means 106, when it is determined that the object is a moving object, calculates the lens drive amount based on this correction amount and the defocus amount, and when it is not determined that the object is a moving object, the lens drive amount is calculated based on the defocus amount. The present invention is applied to an overlap type autofocus device comprising a lens movement amount calculating means 107A for calculating the amount of lens movement, the lens movement amount detection means 108, a driving means 109, and a control means 110.

The above-mentioned problems in the moving object determination can be solved by configuring the moving object determining means 201 as follows. When a plurality of consecutive object image surface velocities calculated at certain time intervals have the same polarity and the object image surface velocities are larger than respective thresholds, the object is determined to be a moving object.

Once it is determined that the object is a moving object, it is determined that the object is not a moving object when the image plane velocity of the object becomes equal to or less than a threshold value, which is smaller than the threshold value.

The invention according to claim 4 is the invention according to claim 3, further comprising defocus amount determining means 202 for determining whether the defocus amount is less than or equal to a predetermined value, and even if the moving object is determined to be a moving object by the moving object determining means 201, the defocus amount is determined to be a moving object. The invention according to claim 5, in which the amount of lens drive is characterized using the correction amount only when it is determined that
01 is the photographing lens 1 within the sampling time interval of at least two focus detection signals used to calculate the object image plane velocity.
According to 02, when the ratio of the object image plane speed to the image plane moving speed is equal to or higher than a predetermined threshold value, the object is determined to be a moving object.

The invention of claim 6 is similar to claim 4, in which the moving object determination means 2
Even if the object is determined to be a moving object based on 01, the lens drive amount is calculated using the correction amount only when the defocus amount is determined to be less than or equal to a predetermined value.

According to a seventh aspect of the present invention, in the overlap type autofocus device according to the fifth or sixth aspect, once the moving object is determined and the tracking servo is activated, the ratio of the object image surface speed to the moving speed of the image forming surface by the photographing lens 102 is determined. This is to prevent judgment.

The invention of claim 8 provides a focus detection means 103, a defocus amount calculation means 104, and a subject image surface velocity calculation means 105.
The present invention is applied to an overlap type autofocus device comprising a lens drive amount calculation means 107, a lens movement amount detection means 108, an active means 109, and a control means 110.

Then, the activation of the release is always permitted, and when the lens drive amount calculation means 107 detects the release signal, it calculates the latest object image surface speed calculated by the object image surface speed calculation means 105. Focusing is performed at the exposure time based on the elapsed time from the accumulation time of the focus detection signal used until the release signal is detected, the amount of lens drive during this time, and the predetermined delay time required from the release signal to the actual exposure. This is to calculate the amount of lens drive to make the lens move.

The invention according to claim 9 is the invention according to any one of claims 1 to 8, wherein the acceleration/deceleration calculation calculates the acceleration/deceleration of the object image surface velocity based on a plurality of object image surface velocities calculated at a certain time interval. The correction amount correlated to the image plane velocity of the subject is calculated by taking into account its acceleration/deceleration.

F0 Effect According to the invention of claim 1, the object image plane velocity is determined from the current distance measurement data and the distance measurement data from at least two periods before, thereby improving accuracy.

In the second aspect of the invention, the shorter the storage time of the charge storage type photoelectric conversion element 101, the longer the sampling time interval of the two focus detection signals used to obtain the object image plane velocity. As a result, the object image plane velocity at high brightness can be determined with high accuracy.

In the invention of claim 3, the moving object determination means 201 determines whether a plurality of consecutive object image surface velocities calculated at a certain time interval have the same polarity, and whose object image surface velocities are larger than the respective threshold values. When the object is moving, the object is determined to be a moving object, and once the object is determined to be a moving object, it is determined that the object is not a moving object when the image plane velocity of the object becomes less than a threshold value, which is smaller than the threshold value, so that stable moving object determination is possible. .

In the invention of claim 4, the defocus amount determining means 202 determines that the defocus amount is equal to or less than a predetermined value, and
The object is determined to be a moving object only when the following conditions are met, and the amount of lens drive is calculated using the correction amount to perform tracking servo.

In the invention of claim 5, the moving object determining means 201 determines the object image surface velocity relative to the moving speed of the imaging surface by the photographing lens 102 within the collection time interval of at least two focus detection signals used for calculating the object image surface velocity. When the ratio is greater than or equal to a predetermined threshold value, the object is determined to be a moving object. Therefore, even if the lens is driven at high speed, erroneous detection of a moving object is prevented.

In the invention of claim 6, moving object determination is performed in the same manner as in claim 4.

In the seventh aspect of the present invention, once the moving object is determined and the tracking servo is started, the ratio of the object image plane speed to the moving speed of the image plane by the photographing lens 102 is not determined.

In the invention of claim 8, activation of the release is always permitted. When the lens drive amount calculation means 107 detects the release signal, it calculates the latest object image surface velocity, the elapsed time from the accumulation time of the focus detection signal used to calculate this to the detection of the release signal, and the time during this time. In the invention according to claim 9, the lens driving amount for focusing at the exposure time is determined based on the lens driving amount and a predetermined delay time required from the release signal to the actual exposure. Since the correction amount is determined by taking this into consideration, even more accurate tracking servo becomes possible.

In the tenth aspect of the invention, the correction amount depending on the image plane speed of the subject takes into account the amount of lens movement between the charge accumulation time and the time when the defocus amount is calculated from the focus detection signal obtained by this charge accumulation. You are asked to do it without doing it. Therefore, even if overlap servo is applied to the tracking servo, accurate tracking and focusing can be achieved.

In the invention of claim 11, the correction amount depending on the object image plane velocity is the defocus amount at time tn and the time t(n-
It is determined by subtracting the amount of defocus at time t(n-1) from the sum of the amount of movement of the imaging plane by the photographing lens 102 between 1) and times t and n.

G5 Embodiment An overlap type autofocus device according to the present invention will be explained with reference to FIGS. 2 to 10.

The overall configuration is the same as that in FIG. 11, and different processing procedures will be explained.

First, FIG. 2 is a flowchart of autofocus processing executed by the CPU 4.

In step S1, accumulation of the photoelectric conversion element 2 is started,
In step S2, it is determined whether the accumulation has ended. The accumulation time is controlled to be shorter as the subject brightness becomes brighter.

When it is determined that the accumulation has ended, the accumulation of the photoelectric conversion element is ended in step S3. Next, in step S4, the focus detection signal output from the photoelectric conversion element 2 is AD converted,
In step S5, a defocus amount D (n) is calculated according to a well-known AF algorithm. and step S
6, the object image surface velocity S (n) is calculated as follows.

[Calculation of subject image plane speed S (n)] The subject image plane velocity S (n) is calculated using the following equation (14).

In this equation (14), the object image surface velocity is calculated from the amount of image surface movement in two periods of distance measurement. Here, as shown in FIG. 5, D (n-2), D (n-1), D(
n> represents the defocus amounts at times tn-2, tn-1, and tn, respectively, and M2 (n) represents the defocus amount from time tn-2 to t
It is the amount of lens movement up to n, and equation (14) is the amount of lens movement up to time tn-
is the average image plane movement speed from 2 to tn.

Of course, more stable detection accuracy can be obtained by calculating over a wider span of three or more cycles.

As the base time for calculating the image plane movement speed increases, stability increases, but responsiveness to accelerated movement of the subject decreases, so an appropriate number of accumulations must be selected. Furthermore, when the subject brightness is low, the storage time becomes longer and the cycle of distance measurement becomes longer, so if the speed is detected in the same number of cycles as when the brightness is high, the responsiveness deteriorates. Therefore, it is preferable to select the number of cycles used for speed detection according to the accumulation time so that the base time for detecting the speed of the subject is approximately constant. Of course, if the distance measurement period is longer than the base time, a change in defocus amount for only one distance measurement period is used as in equation (9).

Furthermore, in step S7, the tracking correction amount C(n) and the lens drive amount X(n) are calculated, and the process proceeds to step S8.

[Calculation of tracking correction amount C(n) and lens drive amount D(n)+M(n-1)-D(n-1
) -(15), and the lens drive amount X(n) is: X (n) = D (n) + c (n) = 2
D (n) 10M (n-1) -D (n-1)
- Determine from (16). As a result, the time tn+ is changed from the broken line diagram L2 to the solid line diagram L3 in FIG.
Since the photographing lens is activated an additional distance PD(n) from the subject position at point 1, the subject position approaches line Q. This lens drive can be continued until time tm+1 when the next distance measurement result is obtained, but the servo may be completed before then.

In addition, if the defocus amount is large, the driving may not be completed yet at time 2 tai1, but in this case, as shown in step S8, the memory is filled with the lens driving amount determined at time t■+1. updated (refreshed),
The next servo takes over and the lens continues to be activated.

As a result of this method, as shown in FIG. 8, the line drawing moves up and down the line Q, and macroscopically, the photographing lens is truncated so that it is in focus. The advantage of overlap tracking servo over discrete tracking servo is that the distance measurement cycle is short, so servo data can be refreshed quickly and the lens can be driven more smoothly.

[Procedure for moving object determination In co-step Sll, check that the polarity of the object velocity is the same three times in a row, that is, that the determination of whether the object is approaching or moving away from the lens is stable, and all must match. For example, in step S18, it is determined that the subject is a stationary subject. This first ensures that the subject is detected as moving in one direction.

Next, step S12. S13. In S14, the subject image plane velocity 5(n) obtained after a predetermined time interval is compared with each threshold value. That is, in step S12, S
(0) >2+++@/See in step S13 (1) >1.5mm/Sec in step S14
Then, it is determined that S (2) > 1 mm/Sec. At this time, unless it is determined in step S12 that the object's speed has exceeded 2 mm/see at least once, it will not be determined that it is a moving object, and the history of the object's speed up to that point will be recorded in steps 312 to S14 when the object has been accelerated to a certain degree. It is said that it is good as long as it is targeted.

When the object is recognized as a moving object from the non-tracking state and the tracking servo is entered, it exits and becomes the normal servo at step S15.
and 5(0)≦1. This is a case where it is determined that mm/See. In other words, conditions have been set that make it difficult to exit tracking once it has started tracking. In this example, at least once 2mm
The tracking servo for a subject whose velocity is detected to be greater than /Sec is maintained unless the velocity becomes less than 1 mm/Sec. During tracking, detection speed is 1mm/See≦S (0)
< 2mm/Sec, step S1 is completed by then.
If the object is recognized as a moving object and the tracking servo is activated in step 7, the determination in step S16 is affirmative, and the tracking servo is continued.If it is not determined in step 516 that the object is being tracked, the object is recognized as a stationary object in step 318. If half-press is determined in step 519, a motor start command is output in step S20, and if half-press is not determined,
In step S21, a motor stop command is output, and in step S21, a motor stop command is output.
Return to 1.

This allows hysteresis to be applied to moving object recognition or entering and exiting the tracking servo. As long as a moving object is determined based on a threshold value, it is unavoidable that the servo becomes jerky to some extent at speeds near the threshold value, but this method significantly improves stability.

The above-described moving object determination method has the following effects.

Even if the image plane movement speed determined by the above equation (14) is judged using the moving object judgment conditional equations (6) and (7), there is a possibility that the object speed may decelerate when it is around the threshold value or during tracking servo. When this happens, the moving object detection becomes unstable every time, and the lens moves in and out of tracking mode, causing awkward movements. This also occurs when the distance measurement accuracy and subject speed detection accuracy are insufficient.

In contrast, according to the moving object determination method of this embodiment, the history of the obtained object image surface velocity S (n) several times is checked to determine whether to enter or exit the tracking servo mode. In other words, since hysteresis is provided in the movement in and out of the tracking servo motor, the responsiveness and stability of moving object determination are improved.

FIG. 3 is an operation flowchart of the AF motor.

When the motor start command is output in step S20 shown in FIG. 2, this program is started, and the lens drive amount X,(n) updated in step S8 shown in FIG. 2 is stored in the memory in step S31. set. Then, in step S32, the number of feedback pulses from the encoder is counted, and when the amount of lens drive set in the memory is reached, the process proceeds to step S33.
Stop the F motor.

[Release Control for Focusing at Exposure Timing] Next, a lens driving method after the start of release will be described.

Unlike the conventional example of discrete servo, the release timing is not limited to once in each distance measurement cycle, but the release is always possible.

FIG. 7 shows the case where the camera is released at an arbitrary time tr during overlap servo with tracking correction as described above. As mentioned earlier, the delay time Tr from the start of camera release to the start of exposure is almost fixed depending on the model, and usually ranges from 50II sec to 100 + 1ls.
A typical value is ec. At normal brightness (indoor brightness around Ev8), the storage time of the photoelectric conversion element is 10m5
ec to several tens of m5ec, and the calculation time is several tens of m5
Since it is about ec, it can be generally considered that the delay time Tr is longer than the ranging period. In the future, the calculation speed will be DSP.
(Digital Signal Processor
) and other dedicated hardware are expected to become shorter and shorter, so it is thought that the distance measurement cycle will also become shorter.

The servo after the release starts drives the lens so that the defocus amount becomes zero at the exposure time tx.
Since the defocus amount D(n) is an instantaneous value at the distance measurement time tn, the reference time for calculation is not tr but the immediately preceding accumulation time tn. Analysis using Figure 5 shows that the object image plane velocity is (1
4) From the equation, the amount that the subject image plane moves during the time Tr from the release time tr to the exposure time tx is 5.
(n) This is the current Tr, but time tn has already passed since the previous distance measurement time tn.
The amount by which the subject moves from to exposure time tx is S (n) * (tr - tn + Tr).

Therefore, the amount of lens movement from time tn to time tr is PD (n), and the amount of defocus at time tn is D
Expressed as (n), the lens drive amount X (n) from the time tr when the release holder is attached to the exposure time Lx is, from FIG. 7, X (n) = S (n) * (tr - tn + T r)
+ D (n) - P D (n) What must be noted here is that if the calculation for the previous distance measurement has not been completed at time tr in the figure, the distance measurement time of - generation ago It is necessary to regard tn-1 as the previous distance measurement time and use the defocus amount D(n-1) based on this distance measurement.

Naturally, the amount of lens movement from time tn-1 to tr is used as P D (n).6 FIG. 4 shows an interrupt program that is activated when the release button is fully pressed.

In step S41, it is determined whether the subject is recognized as stationary or moving. When the object is recognized as a moving object, the lens drive amount X (n) is calculated by the above equation (17) in step S42. Next, in step S43, the lens drive amount is updated and a motor start command is output. Then, the process proceeds to a photographing sequence in step S44, and a well-known photographing operation is executed. When it is determined in step S41 that the subject is stationary, a photographing process is executed in step S44. Note that in this case, the driving of the photographic lens is performed in a different flow shown in Figure 3, so even if the subject is stationary, if the servo is in progress towards the subject at the time of release, the lens will continue to be driven. be done.

Furthermore, there is a problem unique to overlap servo with respect to detection accuracy of the object image plane velocity, which will be explained.

Subject image plane average moving speed 5(n
) is the defocus amount D obtained by distance measurement while driving the lens.
(n), D (n-2),
The faster the relative speed between the subject and the lens during distance measurement, the worse the accuracy. Therefore, the subject image plane velocity calculated from this is also the same, and even if the subject image plane velocity S (n) is the same, the accuracy will differ depending on the lens speed at that time, (12
) and (13), and the flow shown in FIG. 2, the reliability when determining a moving object varies.

A processing procedure for solving this problem is shown in FIG.

FIG. 9 shows the ratio r of the object image surface speed S (n) to the average image surface speed 5l(n) of the lens during the period in which it was detected, as an index of the measurement accuracy of the object image surface speed in equation (14). This should be taken into consideration. Hereinafter, only the steps different from those in FIG. 2 will be explained.

Steps S51 and S52 are added before step S]-1. That is, it is determined in step S51 whether tracking is being performed, and when tracking is not in progress, it is determined in step S52 whether S (n)/SL (n)>λ, where λ≦1. Judgment step Sl
Proceed to after l. During tracking, step S52 is skipped.

Adding such a procedure has the following advantages.

Since M2 in FIG. 5 is the amount of image plane movement of the lens from time tn-2 to tn, the image plane moving average speed S1 of the lens is Sl (n) 2M2 (n) / (tn - tn −
2) −(18). Since the speed ratio r is the ratio between the image plane speed of the object and the lens image plane speed, when r≧1, the object is faster than the lens, and when r≦1, the opposite is true. When r=1, it means that both move at approximately the same speed, and in this case, the image on the photoelectric conversion element is almost stationary, and it can be considered that the distance measurement accuracy is the best. In other words, it can be said that the closer r is to 1, the higher the accuracy of the object image surface velocity S (n) is.

After entering the tracking servo, the subject image plane velocity and the lens velocity are approximately the same on average, so r is also close to 1, and there is little adverse effect on the detection accuracy of the subject image plane velocity. In general, the biggest problem is that when the lens speed increases during simple overlap servo before the tracking servo starts, the distance measurement accuracy deteriorates, so even though the subject is stationary, the moving object test flow shown in Figure 2 This is a case where the object is mistakenly recognized as a moving object. In order to prevent this, if r deviates significantly from 1, even under conditions where a moving object is determined according to the flow diagram shown in Fig. 23 described above, by determining the speed ratio r in advance, it is possible to To avoid desired moving object determination.

In other words, when calculating the image surface speed of a stationary object while performing overlap servo, it is highly unlikely that the speed will be faster than the lens speed no matter how bad the distance measurement accuracy is. Therefore, in step S52, the speed ratio r is By adding the following restriction for moving object recognition: r≧λ (λ≦1) (19), it is possible to avoid misrecognizing a stationary object as a moving object.

λ is determined depending on the distance measurement accuracy and the accuracy of each parameter, but
In general, it appears experimentally that about λ=0.5 is good. When λ=0.5, this means that when the object image plane speed S (n) is less than 50% of the lens speed 5L (n), it is not determined to be a moving object and the tracking servo is not activated. As a result, when the maximum lens driving speed is considerably faster than the object image plane speed (approximately twice or more), the tracking servo is activated after the lens approaches the object considerably at the maximum driving speed. Once the tracking servo is activated, in order to stabilize the operation, it is better to remove the moving object recognition condition based on equation (19) and simply determine the moving object using the flow chart shown in FIG.

In other words, even if equation (19) is no longer satisfied during servo, if tracking servo is already in progress, tracking servo will continue to be performed. Or switch to normal overlap servo for stationary subjects.

FIG. 1o is an example of a processing procedure for further improving servo performance. Only the points that are different from FIG. 9 will be explained.

As a condition for moving object determination, as shown in step S61,
Add that the defocus amount D(0) is less than or equal to the predetermined value Da. Further, as a condition for exiting from tracking, as shown in step S62, the defocus amount D(0) is set to a predetermined value of 9.
Add that it is 0 or more. For example, the predetermined value Dα=
2 mm, and the predetermined value Dβ = 6 mm.

If such a method is adopted, it is possible to prevent the subject from being tracked while it is approaching from a highly defocused and blurry state, and to prevent the camera from accidentally tracking a stationary subject. Conversely, when exiting from tracking, the threshold value is widened to provide hysteresis to stabilize moving object determination.

Finally, by detecting changes in the speed of the subject and taking this into consideration in the amount of correction during overlap tracking servo, tracking performance (tracking ability) can be further improved, and the amount of servo correction after release can be taken into account when exposure is made. We will explain how to improve time focusing performance.

The average image surface velocity S (n) of the moving object was determined in Figure 5, but this is updated one after another each time distance measurement is performed, and if the servo device has some storage area, it can be used for several generations. Since the previous speed of the moving object can be saved, the acceleration state of the object (imaging plane) can be determined by examining these changes. For example, if you memorize three generations of S (n), if the subject has already increased its speed, it will change from the previous speed of 5 (2) to S (1).
, S (0,). Even if the subject itself is not moving at an accelerated rate, this situation will always occur optically if the subject approaches the lens at a short distance. It is clear from the simple lens formula % that when a subject at a distance of 2 m from the lens moves by 1 m, the amount of movement of the imaging plane is much greater than when an object at a distance of 10 m from the lens moves by 1 m. In Fig. 5, assuming this, the locus Q of the object imaging plane is drawn not as a straight line but as a curved line in an accelerating manner.

There are several possible correction methods for the servo, but apart from the degree of precision, all of them depend on the degree of acceleration (16)
The servo performance is improved by driving the lens somewhat more than the lens drive amount calculated assuming uniform motion of the subject as determined by the formula.

To examine the accelerated motion of the object image plane, it is easy to check that the speed of the object image plane increases monotonically over several generations up to that point. For example, when the object image plane velocities of the three generations are s (2) < s (1) < S (0), it is assumed that the object image plane is moving at an accelerating rate. At this time, the 6th
The secondary correction amount Ca (n) that compensates for the subject acceleration at time tm in the figure is linearly predicted (interpolated) by calculating the next speed increase as (S (n) - 5 (n - 1)) and Ca(n)= (S (n) −S (n-1)) *
(tn-tn-1)... (20) It is easy to do so. The amount of lens drive can be determined by adding this to equations (15) and (16).

Also, using the same principle regarding the correction amount after release, 5(n) used for the lens drive amount in equation (17) can be changed to S (n) + (S (n) - 8 (n - 1)) =
22 tree S (n)-5(n-1) and replace it with X (n) ” (2*S (n) -3(n-1)
) ning (tr-tn ten Tr) + D
(n)-P D (n) (21). In particular, when the delay time Tr after release becomes longer than the distance measurement cycle, the correction amount is still insufficient even in equation 21), so for example, the number S (n) -8(n-1) in equation (21) is An appropriate count η
Multiply by (〉1) and
) ) *(tr-tn×Tr) + D (n)
It is better to add some correction, such as -P D (n).

However, these are simple calculations, and since the correction amount is obtained by approximating the accelerated motion of the subject by a polygonal line, it is also possible to correct using a more advanced approximation function, although the calculations become more complicated.

As explained in detail about the H0 invention, according to the present invention, it is possible to detect with high reliability whether or not the object is a moving object while performing alternating servo, so that tracking servo is not erroneously entered during servo for a stationary object. In addition, once the tracking servo is activated, hysteresis is applied to the conditions of the moving object detection test to avoid entering and exiting the tracking state, and to maintain a stable tracking state, which may cause discomfort to the operation. In addition to this, the lens can be driven smoothly and responsively to the movement of the subject because the servo can be corrected for the movement of the subject without slowing down the distance measurement cycle even during tracking servo. .

In addition, since there is no restriction on the timing of accepting a release, the release can be permitted at any time, so the shutter opportunity during shooting is not lost, and after the release, the amount of movement of the subject from the release to the exposure is calculated, Since the lens is driven as much as possible, it is possible to provide operability that does not give the photographer any discomfort that is different from when photographing a normal still subject.

Furthermore, instead of simply applying the conventional overlap servo method to moving object tracking, when calculating the amount of correction to be added to the latest defocus amount, we intentionally ignore the amount of lens drive that is being accumulated by the photoelectric conversion element. This makes it possible to reliably achieve a focused state.

[Brief explanation of the drawing]

FIG. 1 is a complaint correspondence diagram. 2 to 10 show an embodiment of the overlap type autofocus device according to the present invention, and FIGS. 2 to 4 are flowcharts showing an example of its processing procedure. FIGS. 5 to 8 are diagrams illustrating moving object tracking in overlap servo. FIGS. 9 and 10 are flowcharts showing two other embodiments of moving object determination. FIG. 11 is a diagram showing the overall configuration of the autofocus device. FIG. 12 is a diagram illustrating defocus. FIGS. 13 to 17 are diagrams illustrating moving object tracking using conventional discrete servos. 1: Photographing lens 2: Photoelectric conversion element 4: CPU 6: Encoder 101: Photoelectric conversion element 102: Photographing lens 103:
Focus detection means 104: Defocus amount calculation means 105: Subject image plane velocity calculation means 106: Correction amount calculation means 107: Lens drive amount calculation means 108: Lens movement amount detection means 109: Driving means 110: Control means 201: Moving object determination Means 202: Defocus amount determining means 203: Acceleration/deceleration calculation means

Claims (1)

[Scope of Claims] 1) A focus detection means having a charge accumulation type photoelectric conversion element and converting a focus detection light beam incident on the photoelectric conversion element through a photographing lens into a focus detection signal and outputting the same; defocus amount calculation means for calculating a defocus amount consisting of the amount of deviation and the direction of deviation between the imaging plane of the focus detection light beam by the photographing lens and the expected imaging plane based on the focus detection signal from the focus detection means; Subject image plane velocity calculation that calculates the subject image plane velocity from two defocus amounts calculated with a time interval and the amount of movement of the imaging plane by the photographing lens that moves while these two defocus amounts are calculated. correction amount calculation means for calculating a correction amount correlated to the object image surface velocity calculated by the object image surface velocity calculation means; A lens drive amount calculation means for calculating the lens drive amount; a lens movement amount detection means for detecting the actual movement amount of the photographing lens; and a lens drive amount calculated by the lens drive amount calculation means and the detected actual lens. a driving means for servoing the photographing lens to the focusing position according to the amount of movement; and a driving means for servoing the photographing lens to the focusing position in an overlapping manner while the driving means is servoing the photographing lens to the focusing position; In an overlap type autofocus device, which is equipped with a control means for controlling the accumulation of elements and controlling servo operations to be performed in an overlapping manner one after another, the object image plane velocity calculation means is configured to detect the latest focus detection signal and the previous focus detection signal for two or more cycles. An overlap type autofocus device that calculates a subject image surface velocity based on two defocus amounts obtained from a focus detection signal. 2) In the overlap type autofocus device according to claim 1, the sampling time interval of the two focus detection signals used for determining the object image plane velocity is set to be longer if the storage time of the charge storage type photoelectric conversion element is shorter. An overlap type autofocus device characterized by: 3) a focus detection means having a charge storage type photoelectric conversion element and converting a focus detection light beam incident on the photoelectric conversion element through a photographing lens into a focus detection signal and outputting the signal; and a focus detection means from the focus detection means. a defocus amount calculating means for calculating a defocus amount consisting of the amount of deviation and the direction of deviation between the imaging plane of the focus detection light beam by the photographing lens and the expected imaging plane based on the detection signal; a subject image plane velocity calculating means for calculating a subject image plane velocity from two defocus amounts and an amount of movement of an imaging plane by a photographing lens that moves while the two defocus amounts are being calculated; a moving object determination means for determining whether the object is a moving object based on the object image surface speed calculated by the object image surface speed calculation means; and a correction correlated to the object image surface speed calculated by the object image surface speed calculation means. a correction amount calculation means for calculating the amount of correction, and a lens driving amount is calculated based on the correction amount and the defocus amount when the object is determined to be a moving object, and a lens driving amount is calculated based on the defocus amount when the object is not determined to be a moving object. A lens driving amount calculating means for calculating, a lens moving amount detecting means for detecting the actual moving amount of the photographing lens, and photographing according to the lens driving amount calculated by the lens driving amount calculating means and the detected actual lens moving amount. a driving means for servoing the lens to the focusing position; and a driving means for servoing the photographing lens to the focusing position, the storage of the charge storage type photoelectric conversion element is overlapped with the storage of the photoelectric conversion element; In an overlap type autofocus device comprising a control means for controlling the servos to overlap one after another, the moving object determination means is configured to determine whether a plurality of object image surface velocities calculated at a certain time interval have the same polarity. , and when the image plane velocity of the subject is greater than the respective threshold, the subject is determined to be a moving object, and once it is determined that the subject is a moving object, the image plane velocity of the subject is equal to or less than a threshold smaller than the threshold. An overlap type autofocus device characterized in that it determines that a moving object is not a moving object. 4) The overlap type autofocus device according to claim 3, further comprising defocus amount determining means for determining whether or not the defocus amount is less than or equal to a predetermined value, and the moving object determining means includes, in addition to the condition, An overlap type autofocus device characterized in that a moving object is determined when a focus amount is determined to be less than or equal to a predetermined value. 5) a focus detection means having a charge accumulation type photoelectric conversion element and converting a focus detection light beam incident on the photoelectric conversion element through a photographing lens into a focus detection signal and outputting the signal; and a focus detection means from the focus detection means. a defocus amount calculating means for calculating a defocus amount consisting of the amount of deviation and the direction of deviation between the imaging plane of the focus detection light beam by the photographing lens and the expected imaging plane based on the detection signal; a subject image plane velocity calculating means for calculating a subject image plane velocity from two defocus amounts and an amount of movement of an imaging plane by a photographing lens that moves while the two defocus amounts are being calculated; a moving object determination means for determining whether the object is a moving object based on the object image surface speed calculated by the object image surface speed calculation means; and a correction correlated to the object image surface speed calculated by the object image surface speed calculation means. a correction amount calculation means for calculating the amount of correction, and a lens driving amount is calculated based on the correction amount and the defocus amount when the object is determined to be a moving object, and a lens driving amount is calculated based on the defocus amount when the object is not determined to be a moving object. A lens driving amount calculating means for calculating, a lens moving amount detecting means for detecting the actual moving amount of the photographing lens, and photographing according to the lens driving amount calculated by the lens driving amount calculating means and the detected actual lens moving amount. a driving means for servoing the lens to the focusing position; and a driving means for servoing the photographing lens to the focusing position, the storage of the charge storage type photoelectric conversion element is overlapped with the storage of the photoelectric conversion element; In an overlap type autofocus device comprising a control means for controlling servos to overlap one after another, the moving object determination means collects at least two focus detection signals used for calculating the object image surface velocity. An overlap type autofocus device characterized in that a moving object is determined when a ratio of a subject image plane speed to a moving speed of an image forming plane by a photographing lens within a time interval is equal to or higher than a predetermined threshold value. 6) The overlap type autofocus device according to claim 5, further comprising defocus amount determining means for determining whether the defocus amount is less than or equal to a predetermined value, and the moving object determining means is configured to perform defocusing in addition to the condition. An overlap type autofocus device characterized in that a moving object is determined when the amount of the object is determined to be less than or equal to a predetermined value. 7) In the overlap type autofocus device according to claim 5 or 6, once the moving object is determined and the tracking servo is activated, the ratio of the object image surface speed to the moving speed of the image forming surface by the photographing lens is not determined. An overlap type autofocus device characterized by: 8) A focus detection means having a charge storage type photoelectric conversion element and converting a focus detection light beam incident on the photoelectric conversion element through a photographing lens into a focus detection signal and outputting the same; and a focus detection means from the focus detection means. a defocus amount calculating means for calculating a defocus amount consisting of the amount of deviation and the direction of deviation between the imaging plane of the focus detection light beam by the photographing lens and the expected imaging plane based on the detection signal; a subject image plane velocity calculating means for calculating a subject image plane velocity from two defocus amounts and an amount of movement of an image plane by a photographing lens that moves while the two defocus amounts are calculated; Lens drive amount calculation means that calculates the lens drive amount for tracking servo based on the amount of defocus and the correction amount that correlates to the image plane speed of the subject, and lens movement amount detection that detects the actual amount of movement of the photographing lens. means, a driving means for servoing the photographing lens to a focusing position according to the lens driving amount calculated by the lens driving amount calculating means and the detected actual lens movement amount, and the driving means servoing the photographing lens to the focusing position. An overlap type autofocus comprising control means for controlling the storage of the charge storage type photoelectric conversion element to overlap with each other during servo, and to control the storage of the photoelectric conversion element and servo to be performed in an overlapping manner one after another. In the apparatus, activation of the release is always permitted, and when the lens drive amount calculation means detects the release signal, the lens drive amount calculation means calculates the latest object image surface velocity calculated by the object image surface speed calculation means. The exposure time is adjusted based on the elapsed time from the accumulation time of the focus detection signal used in An overlap type autofocus device characterized by calculating the amount of lens drive for focusing. 9) In the overlap type autofocus device according to any one of claims 1 to 8, the acceleration/deceleration of the object image surface velocity is calculated based on a plurality of object image surface velocities calculated at a certain time interval. An overlap type autofocus device comprising a speed calculation means, wherein the correction amount correlated to the image plane speed of the object is calculated by taking into account its acceleration/deceleration. 10) A focus detection means having a charge storage type photoelectric conversion element and converting a focus detection light beam incident on the photoelectric conversion element through a photographing lens into a focus detection signal and outputting the same; and a focus detection means from the focus detection means. a defocus amount calculating means for calculating a defocus amount consisting of a deviation amount and a deviation direction between an imaging plane of a focus detection light beam by the photographing lens and a planned imaging plane based on a detection signal; a lens drive amount calculation means for calculating a lens drive amount for tracking servo based on a correction amount correlated with the object image plane speed; a lens movement amount detection means for detecting an actual movement amount of the photographing lens; a driving means for servoing the photographing lens to the focusing position according to the lens driving amount calculated by the lens driving amount calculation means and the detected actual lens movement amount; In the overlap type autofocus device, the overlap type autofocus device is provided with a control means for controlling the storage of the charge storage type photoelectric conversion element to be performed in an overlapping manner, and the storage of the photoelectric conversion element and the servo to be performed in an overlapping manner one after another. The correction amount that depends on the image plane speed of the object is determined without considering the amount of lens movement between the charge accumulation time and the time when the defocus amount is calculated from the focus detection signal obtained by this charge accumulation. Overlapping autofocus device. 11) In the overlap type autofocus device according to claim 10, the correction amount depending on the image plane speed of the subject is determined by combining the amount of defocus at time tn and the amount of correction by the photographing lens between (n-1) and time tn. An overlap type autofocus device characterized in that the time is set to a value obtained by subtracting the amount of defocus at (n-1) from the sum of the amount of movement of the image plane.