JPH1010418A - Automatic focusing device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately track a moving subject which comes close to and goes away from a camera by predicting image surface data after switching a state based on the image surface data obtained until switching from the state where the image surface of a photographing lens goes to a close range from a far range to a state where it goes from the close range to the far range. SOLUTION: When the moving subject comes close from the far range, a defocusing amount detecting device 6 inputs defocusing amount, its time and the position of a focusing lens and stores them in a storage device 9, and image surface velocity is calculated from the data. Such operation is repeated to continue to form this table till reaching a turning point. In the case of exceeding the turning point, tracking driving is performed by using the table. Namely, when the turning point is passed and the data is inputted from the device 6, corresponding time is calculated by subtracting the table from the image surface position in the data and complementing the table data. The image surface velocity is accurately outputted from the table in such a way, accurate acceleration correction and overlap correction are executed.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an automatic focusing device for driving a taking lens such as a camera.

[0002]

2. Description of the Related Art In an automatic focusing apparatus for a camera, there is known a tracking method in which a moving subject is tracked, exposure is performed at a certain time, and a focused picture is taken. The first tracking method is the simplest method, and many cameras still use this method. In this tracking method, a difference is obtained from a set of image plane positions of a plurality of subjects obtained in the past and their detection times, an amount of change in the image plane position, that is, an image plane speed is obtained, and the position is extended by a slope of the image plane velocity. Then, the image plane position at a predetermined time is predicted. This method can be said to be a Newton-like first-order approximation method for the position in which the first derivative is obtained as an alternative to the difference.

A second tracking method is a method in which the first method is further deepened by a first order to perform a secondary correction on an image plane position. Specifically, a set of image plane position and time data of three generations is taken, a parabola in the space of image plane position-time passing through these three points is determined, and this is extrapolated to obtain an image plane position at a predetermined time. To predict. This tracking method is a function that is not linear on a predetermined image plane even if the subject moves at a constant speed.

(Equation 1) (Where f is the focal length of the lens, v is the object speed, h is the distance at the time of closest approach, t is time, and t = 0 is the closest approach.) Because of the fact that there is, an attempt is made to reach the correct answer by evaluating the image plane acceleration by a quadratic equation. However, actually, the measured value of the defocus amount has a considerable amount of fluctuation, and the influence of this fluctuation becomes larger as the derivative coefficient becomes higher. This is fateful as it uses discrete data sets, which makes it very questionable if this method can correctly capture velocity changes.

As a further improved tracking method, there is a method of performing a statistical operation. The purpose is to try to minimize the effects of measurement fluctuations. The third tracking method is a method of predicting the position of the image plane by a statistical method. That is, a pair of the image plane position and the measurement time is stored over the past several generations, and a linear regression equation is derived from these data groups. Since the slope of this linear regression equation changes over time in the image plane position, it represents the image plane velocity representing the data group. Although this method is resistant to fluctuations in measured data and can provide highly reliable data, it is difficult to deal with rapidly changing subjects because it represents the amount of change in a data group having a certain time interval.

As a fourth method improved from the third method, a regression equation is obtained in the space of the position of the image plane and the square root of the image plane velocity, and the image plane velocity and the image plane acceleration are obtained from the regression equation. There is a way. In this method, since the movement of the non-linear image plane is basically captured by mathematical expressions, it is possible to obtain an image plane acceleration which is less affected by the fluctuation.

[0006]

The above-described conventional tracking method has a problem that the more the method of predicting the image plane position of a moving subject is improved, the more the number of data involved in the prediction decision increases. FIG. 1 shows the relationship between the image plane position (shown by a solid line) and the image plane speed (shown by a broken line) of a photographic lens when a subject moving at a constant speed v approaches and moves away from a camera from a distance. The image plane position becomes maximum when the moving subject comes to the position of the closest approach distance h in front of the camera. If the moving speed when the moving subject approaches the camera is positive, the image surface speed increases as the moving subject approaches the camera, and the polarity of the image surface speed is inverted at a position directly in front of the camera. Then, as the moving subject moves away from the camera, the image plane speed decreases.

In this specification, the point at which the moving subject comes closest to the camera and the polarity of the image plane speed is reversed is called a turning point, and the moving object approaches the camera and the image plane position and the image plane velocity increase. The range in which the moving subject moves away from the camera and the image plane position and the image plane speed decrease is called a downhill.

When tracking such a moving subject,
When the vehicle approaches a downhill, it predicts the image plane position and the image surface speed on the downhill based on data such as the image plane position and the image surface speed on the previous uphill. Cannot follow the movement of the image plane. That is, a function that partially approximates the motion of the image plane assuming a monotonous increase becomes unpredictable at the local maximum point or the turning point of the function representing the actual motion of the image plane. In addition, at the turning point, not only the speed-first derivative-changes drastically, but also the driving direction is reversed, causing backlash to occur in the mechanical members such as gears transmitting the driving force, and the control becomes nonlinear. It is assumed that. The movement immediately after the turning point is difficult to predict, and the tracking drive itself is difficult even if the movement can be predicted, so that the tracking near the turning point has been thrown until now.

An object of the present invention is to provide an automatic focusing apparatus for accurately tracking a moving subject approaching and moving away from a camera.

[0010]

[Means for Solving the Problems]

(1) The invention of claim 1 is a focus detecting means for detecting a focus adjustment state of a photographing lens, a lens position detecting means for detecting a position of a focusing lens of the photographing lens, and a focus detected by the focus detecting means. Image plane position calculating means for calculating and storing the image plane position of the taking lens based on the adjustment state and the position of the focusing lens detected by the lens position detecting means;
Image plane state determining means for determining, based on the image plane position calculated by the image plane position calculating means, a first state in which the image plane of the photographing lens moves from a distant position to a close position and a second state in which the image plane moves from a close position to a distant position; If it is determined by the image plane state determining means that the state has been switched from the first state to the second state, the second state is determined based on the image plane data of the photographing lens until the state is switched to the second state.
Image plane data prediction means for predicting the image plane data of the photographing lens after switching to the state, and lens drive control means for driving and controlling the focusing lens based on the predicted image plane data by the image plane data prediction means. Is provided. (2) In the automatic focusing apparatus according to the second aspect, the image plane data predicting unit is configured to switch to the second state based on image plane position data of the photographing lens until the state is switched to the second state. The image plane speed of the taking lens is calculated by interpolation. (3) In the automatic focusing device according to the third aspect, the image plane position calculating means stores the image plane position of the current calculation result by a predetermined amount or more from the image plane position of the previous calculation result. It was made. When the image plane of the taking lens is in the first state in which the image plane faces from a distance to the closest, the focus adjustment state of the taking lens is detected, and the position of the focusing lens at that time is detected. Then, the image plane position of the photographing lens is calculated and stored based on the focus adjustment state and the focusing lens position. When the image plane of the photographing lens is switched from the first state to the second state, the detected focus adjustment state and focusing lens position of the photographing lens, and image plane position data until the state is switched to the second state. The image plane speed of the photographing lens after the switching to the second state is interpolated based on the stored value and the focusing state is driven based on the calculated image plane speed.

[0011]

FIG. 2 shows an embodiment of the present invention.
The lens unit 1 includes a photographing lens 2, a lens position detecting device 3, and a driving gear 4. The lens position detecting device 3 detects a position of a focusing lens (not shown) included in the photographing lens 2. The driving gear 4 includes a camera body 5
The driving force is transmitted from the camera via a transmission mechanism (not shown) to drive the focusing lens.

On the other hand, the camera body 5 contains a defocus amount detecting device 6, a clock 7, an arithmetic device 8, a storage device 9, an image plane state judging device 10, and a drive control device 11. . The defocus amount detecting device 6 passes a pair of light beams from the subject, which have passed through the photographing lens 1 and the main mirror 12 and are reflected by the sub-mirror 13, onto an image sensor (not shown) by a focus detection optical system (not shown). Then, a signal corresponding to the light intensity distribution of the subject image is generated by forming a pair of subject images. Then, a well-known focus detection calculation is performed based on the signal, and a defocus amount of the photographing lens 1 from a predetermined image forming plane is detected. The clock 7 measures the time when the defocus amount is detected by the defocus amount detection device 6. The arithmetic unit 8 performs various arithmetic operations of the camera. The storage device 9 stores the data calculated by the calculation device 8. The image plane state determination device 10 monitors the movement of the image plane based on various data such as the amount of defocus, and detects the turning point.

The drive control device 11 focuses via the drive gear 4 in accordance with the drive data calculated by the calculation device 8, ie, the designated drive straight line based on the image plane speed and the offset amount at a predetermined time as shown in FIG. Drive lens.

Now, as shown in FIG. 1, the operation of the embodiment will be described by assuming a case where a moving subject is passing at a constant speed v at the closest approach distance h and the moving subject is tracked by a camera. explain. Tracking in this embodiment follows the following process. First, a moving subject approaches the camera from a distance. At this time, first, as usual, the defocus amount bf0, the time to, and the position lp0 of the focusing lens at to are input from the defocus amount detection device 6, and the image plane position (df0 = bf0 + lp)
0) and its detection time are stored in the storage device 9 and the focusing lens is driven based on the data. This autofocusing cycle is repeated several times, and when data sets of the image plane position and the detection time are stored for a plurality of generations, the movement of the image plane per unit time, that is, the image plane speed is calculated from the data. Then, the automatic focusing device enters a tracking state.

As for the method of calculating the image plane speed, 2
From the method of taking the difference between the two image plane positions and the detection time data,
There are various methods up to the method using the slope of the regression equation of the statistical operation based on a plurality of data, but a detailed description is omitted because it is not directly related to the present invention. However, it is necessary to simultaneously calculate the image plane position yi and the time ti corresponding to the calculated image plane velocity vimi. When the image plane speed is simply obtained by the difference, the intermediate position and the time between the two data are representative thereof. In the case of the statistical operation, the average position and the time of the adopted data are represented by this. Would be equivalent. This set of data is a first set of numerical values in the drive table. Similarly, in the next cycle in which the amount of defocus is detected and various parameters are calculated, the image plane position calculated this time is shifted by a predetermined amount d from the image plane position previously stored in the table.
It is checked whether it has changed, and only when it has changed, it is stored in the table as the next data set.
This is because the auto focus cycle is much faster than the image plane movement of the subject, so storing all the data as a table may make the table huge, and the table may increase monotonically due to data fluctuation. This operation is performed in the sense of preventing the property from being lost. It is a problem in software production that the monotonous increase property is broken due to the relation of searching the tables by magnitude comparison. This operation is repeated to form a table. This situation is shown in FIG. 4, and the image plane position and the image plane velocity are stored at the times of the storage 1, the storage 2 and the storage 3. In this way, a table is formed when the image plane is on an uphill.

On the other hand, while repeating such operations, the image plane state determination device 10 continues to determine whether or not the turning point has been exceeded. As described above, the formation of the table is continued until the turning point is reached. Image plane state determination device 1
0 determines the turning point based on various parameters and the characteristics of the predictive driving being performed. For example, a system using a statistical operation that overlaps the acceleration correction when exceeding a certain image plane speed, and the third and fourth tracking methods described in the section of the prior art, are as follows. The current image plane position is closer than the first image plane position by a certain amount to the close side. (2) Acceleration correction overlap has started. (3) The maximum image surface speed up to that time has exceeded a certain value. (4) The focal point is on the closest side. (5) The current image plane speed is close to 0 or negative, or the previous and current measurement data have a relationship corresponding thereto. (6) There is no judgment that the turning point has been reached yet. When all of these conditions are satisfied, the image plane state determination device 10 determines that the turning point has been reached.

Among these conditions, (1) to (3) also have a factor for judging the state. If these conditions are not satisfied, the closest distance is long and the change in image surface speed is small, so the turning point There is also an aspect that there is no particular problem even if driving is continued without judging. (4) is a judgment factor derived from the property that the third and fourth tracking methods described above have a tendency that the acceleration correction tends to be excessive near the top of the mountain. About (6),
It may be understood that the detection of the turning point is performed only once in one pass.

The table should be as shown in Table 1 for example.

[Table 1]

When the turning point is exceeded, a tracking drive is performed using this table. Next, a description will be given of a tracking drive method when the vehicle is approaching a downhill after the turning point. When the data is input from the defocus amount detection device 6 after the turning point, the table is subtracted from the image plane position in the data, and the corresponding time is calculated by complementing the table data. The simplest complementation is a method of obtaining an answer by proportional distribution. For example, assuming that the image plane position yt is 3.0, in Table 1, the first data y1 and the second data y2 are selected, and these proportional distributions are selected. You can ask for it.

[Mathematical formula-see original document] ratio = (y2-yt) / (y2-y1)

[Mathematical formula-see original document] time = (t2-t1) * ratio + t1 A virtual time time is easily calculated from the above equation.
-3.9.

This time is the time at which the measurement was made, and should have a slight difference td from the current time. Since the speed before the turning point is opposite to the speed, if this value is subtracted from time, the virtual time time0 corresponding to the present time is calculated. Here, assuming that this time is 0.1, this virtual time is -4.0. If the image plane speed and the image plane position corresponding to the virtual time are reversed, the image plane velocity and the image plane position at the present time are calculated. If the same calculations as in Equations 2 and 3 are performed again, the image plane position is 2.92 and the image plane velocity is 0.75. When calculating the image plane speed, half of the measurement interval may be further added only when calculating the image plane speed in consideration of the measurement interval because it is a phase having acceleration. When outputting to the drive control device 11, the current lens position is subtracted from the image plane position for the image plane position.
Needless to say, the output must be output as the remaining drive amount.

In this way, while the table can be adapted, if the overlap correction is repeatedly performed from the learned data, the image plane speed can be correctly output from the table, and the behavior of the image plane is known. Accurate acceleration correction and overlap correction can be immediately performed.

In the configuration of the above embodiment of the present invention,
The out-of-focus amount detector 6 is a focus detector, the position detector 3 is a lens position detector, the calculator 8 is an image plane position calculator and an image plane data predictor, and the image plane state determiner 10 is an image plane state detector. The drive control device 11 constitutes the lens drive control means as the determination means.

[0023]

As described above, according to the present invention, on a downhill after the turning point of the image plane movement curve, which has been difficult to track, a high-precision tracking drive can be performed immediately after the turning.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating an image plane position and an image plane speed of a photographing lens when tracking a moving subject approaching and moving away from a camera.

FIG. 2 is a diagram illustrating a configuration of an embodiment.

FIG. 3 is a diagram illustrating driving data of a photographing lens and actual movement of the photographing lens.

FIG. 4 is a diagram showing an example of creating a table on an uphill.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 lens unit 2 photographing lens 3 lens position detecting device 4 drive gear 5 camera body 6 defocus amount detecting device 7 clock 8 arithmetic device 9 storage device 10 image plane state determining device 11 drive control device

Claims (3)

[Claims] A focus detection unit configured to detect a focus adjustment state of the imaging lens; a lens position detection unit configured to detect a position of a focusing lens of the imaging lens; a focus adjustment state detected by the focus detection unit; An image plane position calculating unit that calculates and stores an image plane position of the photographing lens based on the position of the focusing lens detected by the lens position detecting unit; and an image plane position calculated by the image plane position calculating unit. Image plane state determining means for determining a first state in which the image plane of the photographing lens moves from distant to close and a second state in which the image plane moves from close to distant based on the first state by the image plane state determining means. If it is determined that the state has been switched to the second state, the photographing after switching to the second state is performed based on image plane data of the photographing lens until the state is switched to the second state. Image plane data prediction means for predicting the image plane data of a lens, and lens drive control means for driving and controlling the focusing lens based on the predicted image plane data by the image plane data prediction means. Automatic focusing device. 2. The automatic focusing apparatus according to claim 1, wherein the image plane data predicting unit is configured to calculate the image plane data based on image plane position data of the photographing lens until switching to the second state.
An automatic focusing device, wherein an interpolation operation is performed on an image plane speed of the photographing lens after switching to the second state. 3. The automatic focusing apparatus according to claim 1, wherein said image plane position calculating means determines that an image plane position of a current calculation result is a predetermined amount from an image plane position of a previous calculation result. As described above, the automatic focusing apparatus is characterized in that the information is stored only when it changes.