JP3162100B2 - Focus detection device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a focus detection device, and more particularly to a focus detection device that detects an in-focus point by selecting an optimum interpolation calculation method based on a video signal.

[0002]

2. Description of the Related Art A system that focuses on a change in the frequency component of a video signal and detects a focal point is disclosed in NHK Technique Vol. 17 No. 1 by Ishida et al. Autofocus Adjustment "(referred to as hill-climbing AF). FIG.
It shows a voltage value of a lens defocus specific frequency component (hereinafter referred to as an MTF curve). It can be seen from the figure that the sharpness is high during focusing, that is, the specific wave component is large. Conversely, when out of focus, the sharpness is reduced and the edge is blurred, that is, the specific wave component is reduced. This phenomenon occurs particularly when the specific frequency component is a high frequency component. As described above, the hill-climbing AF drives the photographing optical system to the peak position of the frequency component, focusing on the change of the specific frequency component.

A method of detecting the focal point based on three frequency signals sandwiching the focal point by changing the optical path length is disclosed in Japanese Patent Application No. Hei.
No. 22027 and Japanese Patent Application No. 2-109876.

[0004]

However, in the case of the hill-climbing AF, it takes time to detect the focal point while driving the photographing optical system at intervals related to accuracy.

[0005] To solve these problems, Japanese Patent Application No. 1-222 / 1990.
No. 027 and Japanese Patent Application No. 2-109876, based on frequency information of three screens having different optical path lengths across a focal point,
An attempt is made to detect the focal point using a linear equation. However, depending on the magnitude relationship between the frequency information of the three screens having different optical path lengths, when a single method is used, for example, when only the linear expression is used, the influence of noise on the signal increases as shown in FIG. There are problems such as a reduction in focusing accuracy.

Here, FIG. 11 will be described. FIG.
(A), (b) and (c) show the relationship among three points A, B and C on the MTF curve. Assuming that the MTF output difference between A and B is Z ₁ and the output difference between B and C is Z ₀ , when the relationship between the three points changes from FIG. As shown, the characteristics related to the focus detection are obtained (approximately 5% random noise is added to each point, Z ₁ is fixed,
By changing Z ₀ , the relationship between the three points of the MTF curve is changed as shown in FIG. 11 (a) → (b) → (c)).

The curve shown by the solid line in FIG.
The characteristic obtained by interpolation using the following equation is indicated by a broken line. The MTF curve can be mathematically approximated by a Bessel function, and the focus detection accuracy improves when the interpolation operation itself is approximated by a higher-order function. However, when the approximation is performed using a higher-order function, the influence of noise is lower depending on the relation among the three points than when the interpolation is performed using a lower-order function (see the range of A in FIG. ).

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and switches to a more accurate detection algorithm in accordance with the magnitude relationship between three frequency information, thereby detecting a focus point with high speed and high accuracy. It is an object to provide a focusing device.

[0009]

That is, the present invention provides a photographing optical system, a photoelectric conversion element group for photoelectrically converting a light distribution of a subject image passing through the photographing optical system and outputting an image signal, and the image signal. Frequency detection means for detecting a specific frequency component from the image data; storage means for storing outputs of at least three frequency detection means sandwiching the focal point while changing the focus position of the imaging optical system; Discriminating means for discriminating the magnitude of the output of the means, selecting means for selecting an interpolation calculation method according to the discrimination result of this discriminating means, and using the outputs of the three frequency detection means and the selected interpolation calculation method. And a photographing optical system driving means for driving the photographing optical system to a focal position in accordance with an output of the focusing point determining means. Further, the selection by the selection means of the present invention.
In the selected interpolation calculation method, interpolation calculation is performed by a linear expression.
It is characterized by including the method.

[0010]

In the in-focus point detecting apparatus according to the present invention, when the light distribution of the subject image that has passed through the photographing optical system is photoelectrically converted by the photoelectric conversion element group and an image signal is output, the frequency signal is output by the frequency detecting means. A specific frequency component is detected from the image signal. Also, while changing the focus position of the photographing optical system, the magnitudes of the outputs of at least three frequency detection means detected across the focal point are determined by the determination means. Then, according to the result of the determination, an interpolation calculation method is selected,
The focal point is determined using the outputs of the three frequency detecting means and the selected interpolation calculation method, and the photographing optical system is driven to the focal position.

[0011]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram for explaining the concept of a focus detection apparatus according to the present invention. In FIG. 1, an in-focus detection device includes an imaging optical system 1, a photoelectric conversion element 2 that converts a light distribution from the imaging optical system 1 into a distribution of electric signals, and a pixel output signal of the photoelectric conversion element 2. A frequency detection circuit 3 for extracting a specific frequency from the image data, a drive circuit 4 for driving the photographing optical system 1, a control circuit 5, a first focus detection circuit 6, and a second focus detection circuit 7.

The control circuit 5 includes first and second in-focus detection circuits based on signals from the frequency detection circuit 3 and information (lens position, aperture, focal length, MTF characteristics, etc.) of the photographing optical system 1. 6 and 7 are switched to detect the focal point. Then, the drive amount of the photographing optical system 1 is calculated, and a drive signal is output to the drive circuit 4. The control circuit 5 drives the photographing optical system 1 to determine that three points (output of the frequency detection circuit 3) sandwiching the top of the MTF curve have been obtained, and determines the magnitude relationship between the three points. Then, the detection method is changed based on the magnitude relation of the above three points. That is, the first focus detection circuit 6 and the second focus detection circuit 7 are switched to detect the focus from the three points.

FIG. 2 is a functional block diagram of the control circuit 5. The control circuit 5 extracts the frequency of the focus area from the frequency detection circuit 3 in accordance with the focus area signal from the detection control circuit 51, and A window circuit 52 for selecting a BPF (band-pass filter) of the frequency detection circuit 3, and an A / D converter for A / D converting a signal of the window circuit 52 substantially in response to pixel reading of a CCD (charge coupled device) Conversion circuit 53
And Further, the control circuit 5 adds the data A / D-converted by the A / D conversion circuit 53 to generate a frequency signal of a focus area and outputs the current signal based on the lens information from the imaging optical system 1. A lens information processing circuit 55 for transmitting the relationship between the position of the photographing optical system, the frequency to be selected, and the driving amount of the photographing optical system 1 to the detection control circuit 51 is provided.

The detection control circuit 51 transmits the driving amount of the photographing optical system 1 to the driving circuit 4 based on the information from the lens information control circuit 55, while the lens information processing circuit 55
The detection frequency is selected based on this information. This allows
The extraction area is determined corresponding to the focus area, the magnitude relation of the three points with different optical path lengths is determined, and the interpolation method is switched.

FIG. 3 is an actual block diagram of the in-focus point detecting device. The photographing lens 11, CCD 12, BPF 1
3, a motor (M) driver 14, a CPU 15 for detecting the first and second states and the lens state, controlling the driving of the CCD 12, and detecting the focal point, and a motor 16 for driving the photographing lens 11. . That is, the photoelectric conversion element 2 is constituted by a CCD 12, the frequency detection circuit 3 is constituted by a plurality of BPFs 13, and the control circuit 5 is constituted by a CPU (central processing unit) 15. Note that this CPU 15
It comprises a control circuit 5, a first focus detection circuit 6, and a second focus detection circuit 7 shown in FIG. In addition, the first
And the second state is a first release switch (1s
tSW) and second release switch (2ndSW)
Required by Next, referring to FIG. 4, the focal point (M
From three frequency signals sandwiching the TF curve vertex), 1
The manner in which the focal point is obtained by the following equation will be described.

FIG. 4A shows the relationship between the MTF curve and three points, and FIGS. 4B and 4C show the interpolation. Now, as shown in FIG. 4A, three points (A,
B and C) show different lens positions when the lens is driven.
Represents the relationship of the MTF at the distance between the three lens positions.
Each is set to 1. When integration is performed during driving, the interval is the median value of the movement amount during integration. The frequency signal is at each point A,
F (A), f (B), f (C) corresponding to B and C.

FIG. 4B shows that f (B) ≧ f (A)> f
(C) or the case of f (B)> f (A) = f (C), which has a straight line y ₂ connecting points B and C2 and a slope opposite to y ₂ , the origin intersection (B point y ₁ passing, {(l / 2) ( f (C) -f (a))} / (f
(B) -f (C))) is determined as the focal point.

FIG. 4C shows that f (B) ≧ f (C)> f
The case of (A) is shown. Here, the intersection of a straight line y ₁ connecting two points A and B and y ₁ passing through point C with a slope opposite to y ₁ (（(l / 2) (f
(C) −f (A))｝ / (f (B) −f (A))) is determined as the focal point. When the quadratic interpolation is obtained by a quadratic equation (quadratic curve approximation), the focal point ((l / 2) ｛(f
(A) -f (C)) / (f (A) + f (C) -2f
(B))｝). Next, a description will be given of a process from the first release to the detection of the focal point with reference to the flowchart of FIG.

When the sequence is started, first, the first release is checked (step A).
1) If the first release is off, the process stands by; if it is on, the lens information is read (step A2). Next, the lens driving amount is calculated (driving amount l ₁ ,
l ₂ , l ₃ ) (step A3), and the extraction frequency is calculated (frequency fa, fb, fc) (step A).
4). The optimization of these steps A3 and A4 is described in Japanese Patent Application Nos. 1-222027 and 2-1098.
No. 76, etc.

Then, the first detection frequency fa is set (step A5), and subsequently, an in-focus direction detection subroutine program FBH is executed (step A6). Thereafter, a flag GM for determining whether or not the current position of the lens is near the focal point is determined (GM = 0: the current position of the lens is not near the focal point) (step A7).

In step A7, if GM = 1 (the lens is near the focal point), the flow proceeds to step A13. If GM = 0 (the lens is not near the focal point), the process proceeds to step A8 and the second detection frequency fb
(Fb is higher than fa; fa ≦ fb).

Next, a subroutine program CAF for in-focus detection and interpolation calculation data detection is executed (step A).
9) The flag RM for detecting the release state is determined (RM = 1: the first release is in the off state) (step A10). In this step A10, if RM = 1 (first release off), each state is reset (step A20) and the process returns to step A1. On the other hand, if RM = 0 (first release is in the on state) in step A10, the in-focus point detection interpolation calculation subroutine program HAF is executed (step A1).
1).

Based on the result of step A11, the lens is driven (step A12), and the third detection frequency fc (fc is higher than fa and fb; fc ≧ f)
b ≧ fa) is set (step A13). And
After the in-focus confirmation subroutine program MH is executed (step A14), the flag Z indicating the relationship between the three points near the in-focus point is determined (Z = 1: the relationship between the three points is inclined in one direction, and the out-of-focus state is established). (Step A15).

In step A15, if Z = 1, the process returns to step A5. On the other hand, if Z = 0, 3 near the focal point
The flag Q indicating the point relationship is determined (step A).
16). In step A16, if Q = 1 (the peak of the MTF curve is sandwiched, but the relationship is not optimal), the process returns to step A11. If Q = 0, the process proceeds to step A17, where the frequency signal is taken into F (1).

In step A18, F (0) and F (1)
The absolute value of the difference, whether greater than the predetermined amount epsilon ₃ is determined. At this step A18, F than the predetermined amount epsilon ₃
If the absolute value of the difference between (0) and F (1) is not small (the focus state has changed), the process returns to step A11. on the other hand,
Absolute value of the difference between F (0) and F (1) than a predetermined amount epsilon ₃ is small (is followed by a focus state) when the determination of the second release is followed (step A19). Then, in step A19, if the second release is off, the process returns to step A17, and if the second release is on, the process exits this sequence. Next, the in-focus direction detection subroutine program FBH will be described with reference to FIG.

When the program FBH is started, first, a frequency signal is taken into F (0) (step B).
1). Next, the flag FC indicating the driving direction is set to FC = 1 for driving the lens (driving amount l ₂ ; determined by the lens position and the frequency fa) (step B2).
Thereafter, the frequency signal is taken into F (1) (step B3).

Next, in step B4, a comparison is made between the frequency signals F (0) and F (1) (| F (0)
−F (1) | <ε ₁ ). In this step B4, F
When the absolute value of the difference between (0) and F (1) is smaller than ε ₁ (in a focused state), the flag GM is set to GM = 1, and the routine goes to Step B5. The absolute value of the difference between F (0) and F (1) is when not smaller than epsilon _1, the magnitude relation of F (1) is determined F (0) proceeds to step B6 (F
(0)> F (1)).

In step B6, F (0)> F
If (1), the flag FC is set to FC = 0 and the process proceeds to step B7. On the other hand, if F (0)> F (1) is not satisfied, the flag GM is set to GM = 0 (step B8).
After that, the process exits the focus direction detection subroutine program FBH. Next, with reference to FIG. 7, the in-focus point detection interpolation calculation data detection subroutine program CAF will be described.

When the program CAF is started, first, memory initialization (i = 1, RM = 0) is performed (step C1), and then a first release state is determined (step C2). If the first release is OFF in step C2, the flag RM is set to RM = 1, and the routine goes to step C3. On the other hand, if the first release is on, the flag FC indicating the driving direction of the lens is determined (FC = 0) (step C).
4).

In step C4, if FC = 1, the process proceeds to step C5, where the motor is driven (drive amount l ₁ ) in the direction of FC = 1 (the same direction as the initially driven direction).
If FC = 0 in step C4, the flow advances to step C6 to drive in the direction of FC = 0 (the direction opposite to the initial driving direction) (driving amount l ₁ ; lens position, detection frequency fb, and initial value). Determined by the driving amount l ₂ ).

Next, the frequency signal is taken into f (i) (step C7). Then, after i is incremented (step C8), the condition of i is determined (i = 4) (step C9). In this step C9, i = 4
If not, the process returns to step C2. If i = 4, the magnitude relationship between the frequency signals f (1) and f (2) captured while driving is determined (step C10).

In step C10, f (1)> f
In the case of (2), the process proceeds to step C11 to shift the memory of the frequency signal (f (1) → f (2), f (2) → f
(3), F (1) × (l ₁ / l ₂ ) × (fa / fb) →
f (1)). On the other hand, if f (1)> f (2) is not satisfied in step C10, the process proceeds to step C12, where the magnitude relationship between f (2) and f (3) is determined.

At step C12, f (2)> f (3)
If so, the program exits the in-focus point detection interpolation calculation data detection subroutine program CAF. On the other hand, f (2)> f
If not (3), the process proceeds to step C13 to perform a memory shift (f (2) → f (1), f (3) → f (2)), and after i = 3, Return to C2.

If the frequency signal memory is shifted in step C11, the magnitude relationship between the frequency signals f (1) and f (2) is determined in step C14. If f (1) <f (2), the process exits the in-focus detection / interpolation calculation data detection subroutine program CAF. On the other hand, unless f (1) <f (2), the MTF
In order to shift the frequency signal memory from the opposite direction in the curve, the process proceeds to step C15, where the frequency signal memory shift (F (0) × (l ₁ / l ₂ ) × (fa /
fb) → f (1)), the process exits from the in-focus point detection interpolation calculation data detection subroutine program CAF.

FIG. 8 shows the flow of the in-focus detection interpolation subroutine program HAF. In the figure, when the program HAF is started, in step D1,
f (2) −f (1) → δ ₁ , f (2) −f (3) → δ ₂
It is said. Then, in step D2, the absolute value of the difference between δ ₁ and δ ₂ is determined (| δ ₁ −δ ₂ | <δ; δ is a predetermined amount).

In this step D2, | δ ₁ −δ ₂
If not | <δ, linear interpolation is performed (step D3). On the other hand, | δ ₁ -δ ₂ | if <[delta], 2 quadratic interpolation is performed (step D4). The data interval of the interpolation in steps D3 and D4 is 12. Next, when the drive amount Δl is calculated (step D 5 ), the process exits the in-focus point detection subroutine program HAF. Next, the flow of the focus confirmation subroutine program MH will be described with reference to FIG.

When the program MH is started, first, memory initialization (P = 0, Q = 0, Z = 0) is performed (step E1), and a frequency signal is taken into F (0) (step E2). Next, after the lens is driven (driving amount l ₃ ; determined by the detection frequency fc) (step E3), the frequency signal is taken into F (1) (step E4). Then, after the lens is driven again (driving amount-2l ₃ ) (step E5), the frequency signal is taken into F (2) (step E6).

Here, the determination of the absolute value of the difference between the frequency signals F (0) and F (1) (| F (0) -F (1) | <ε ₁ ;
epsilon ₁ is a predetermined amount) is performed (step E7). At step E7, | F (0) -F (1) | If <not epsilon _1, the flag P is set to P = 1 indicating the state of transition to the magnitude comparison to step E8, the flow proceeds to step E9. On the other hand, in step E7 | F (0) -F ( 1) | < If epsilon _1,
The absolute value of the difference between the frequency signals F (0) and F (2) is determined (step E9).

In step E9, | F (0) -F (2)
| <Ε ₂ (ε ₂ is a predetermined amount) is determined, and | F (0) −F
₍₂₎ | <ε 2, then the determination of the flag P proceeds to step E10 (P = 1) is performed. Then, step E1
At 0, if the flag P is P = 1, the flag Q is Q = 1.
(Step E11), the frequency signal memory shift (F (1) → f (1), F (0) → f (2), F
(2) → f (3)) (step E1)
2). After that, the focus confirmation subroutine program MH
Through.

In step E10, if P = 0,
The lens is driven (drive amount l ₃ ) (step E13)
Thereafter, the process exits the focus confirmation subroutine program MH.
In step E9, | F (0) −F (2) |
<If not epsilon _2, the determination of the flag P (P = 1) is made (step E14). In this step E14, P
If not, the process proceeds to step E11. If P = 1, the flag Z is set to Z = 1 (step E15).
The process exits the in-focus confirmation subroutine program MH. According to the above-described embodiment, a high-speed and high-precision in-focus detection device is provided by using an optimal interpolation method in relation to the magnitude of three points used for interpolation.

In this embodiment, a CCD is used as a sensor. However, a MOS type sensor which can be accessed at random, for example, SIT (Static Induction Tra) is used.
nsistor) type solid-state imaging device, CMD (Charg)
An eModulation Device) type solid-state imaging device or the like may be used. Further, although the optical path length is changed by the photographing optical system, the optical path length may be changed physically. Further, the optical path length may be changed by having a plurality of sensors or by shaking the sensors. The same can be said for a system (TTL-SLR) that divides an optical path from an imaging optical system. The frequency detecting means is a digital filter.
Each frequency may be detected at once. Also, if the detection method is 1
Next, although the switching is performed by using a quadratic function expression, the switching may be performed by using a ROM or the like.

[0043]

As described above, according to the present invention, by switching to a more accurate detection algorithm in accordance with the magnitude relationship of the three points of frequency information, it is possible to quickly and accurately detect a focal point. A focus detection device can be provided.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating the concept of a focus detection device according to the present invention.

FIG. 2 is a functional block diagram of a control circuit 5 of FIG.

FIG. 3 is an actual block diagram of the focal point detection device.

4A is a diagram showing a relationship between an MTF curve and three points, and FIG. 4B is a diagram showing f (B) ≧ f (A)> f (C) or f (B).
FIG. 4B shows a state of interpolation when (B)> f (A) = f (C), and FIG. 4C shows a state of interpolation when f (B) ≧ f (C)> f (A). FIG.

FIG. 5 is a flowchart illustrating an operation from first release to focus detection.

FIG. 6 is a flowchart of an in-focus direction detection subroutine program FBH of FIG. 5;

7 is a flowchart of a focusing point detection interpolation calculation data detection subroutine program CAF of FIG. 5;

FIG. 8 is a flowchart of a focus detection interpolation subroutine program HAF of FIG. 5;

9 is a focus confirmation subroutine program MH shown in FIG. 5;
It is a flowchart of FIG.

FIG. 10 is a diagram illustrating a voltage value of a lens defocus specific frequency component.

11 (a), (b) and (c) are diagrams showing a relationship between three points A, B and C on the MTF curve, and (d) is a characteristic interpolated by a linear expression and a quadratic expression It is a figure showing a curve.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Imaging optical system, 2 ... Photoelectric conversion element, 3 ... Frequency detection circuit, 4 ... Drive circuit, 5 ... Control circuit, 6 ... First focus detection circuit, 7 ... Second focus detection circuit, 11 ... Photographing lens, 12 ... CCD (charge coupled device), 1
3 BPF (Band Pass Filter), 14 Motor Driver, 15 CPU (Central Processing Unit), 16 Motor
51: detection control circuit, 52: window circuit, 53: A
/ D conversion circuit, 54 ... addition circuit, 55 ... lens information processing circuit.

Continuation of the front page (56) References JP-A-2-37313 (JP, A) JP-A-60-54152 (JP, A) JP-A-64-54408 (JP, A) JP-A-63-150057 (JP) JP-A-3-251967 (JP, A) JP-A-3-71104 (JP, A) JP-A-3-504419 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB G02B 7/28-7/40

Claims (2)

(57) [Claims] 1. An imaging optical system, a photoelectric conversion element group that photoelectrically converts a light distribution of a subject image passing through the imaging optical system and outputs an image signal, and a frequency detection unit that detects a specific frequency component from the image signal Means, storage means for storing the outputs of at least three frequency detection means sandwiching the focal point while changing the focus position of the imaging optical system, and discrimination for judging the magnitude of the output of the three frequency detection means. Means, selecting means for selecting an interpolation calculation method in accordance with the result of the determination by the determination means, focus determination for determining a focal point using the outputs of the three frequency detection means and the selected interpolation calculation method. And a photographing optical system driving means for driving the photographing optical system to a focusing position in accordance with an output of the focusing point determining means. 2. An interpolation calculation method selected by the selection means.
Is characterized by including a method of performing an interpolation operation by a linear expression.
The focus detection device according to claim 1, wherein: