JPH05308556A - Method and device for automatic focusing - Google Patents

Abstract

PURPOSE:To enable surely and quickly focusing by estimating appropriate focus displacement amount from the focus displacement amount calculated by a focal voltage through a band pass filter and a membership function corre sponding to the band pass filter based on fuzzy inference. CONSTITUTION:The three values of the focus displacement amount are obtained from the focal voltage detected through band pass filters 2, 3, and 4. These three values are A/D converted 44, a fuzzy set is found from the membership functions by an arithmetic processing circuit 20, and fuzzy inference for deciding the centroid of the set as the focus deviation amount is performed. The frequency components for a plurality of different high frequency bands a are extracted from video signals from a video element 30 as a plurality of focusing voltages, the moving distances of a part of an optical system up to respective focal positions are calculated based on them, adaptation is decided by the moving distances and the membership functions having the high frequency bands the as variables, the representative value of the set of the adaptation is defined as the distance to the focal position, so that the appropriate estimate can be performed based on the fuzzy inference and the focusing can be performed surely and quickly.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an autofocus method and device, and more particularly to extracting a high frequency component as a focus voltage from a video signal obtained from an image pickup device such as a CCD so that the focus voltage becomes the maximum level. The present invention relates to an autofocus method and apparatus for automatically focusing on a subject.

[0002]

2. Description of the Related Art Focusing on the fact that the voltage level of the high frequency component of a video signal obtained by photographing a subject conventionally corresponds to the definition of the image, the high frequency component from the video signal is focused. There is an autofocus device that extracts the voltage and drives the focus lens so that the focus voltage becomes the maximum level, and adjusts the lens position to the in-focus position.

In this type of autofocus device, the focus voltages at two points are successively compared in level, and the lens is moved by detecting the presence or absence of focus deviation and the lens moving direction depending on the magnitude of the focus voltage. There is a method of moving the lens, and a method of calculating the distance to the in-focus position from the differential value of the focus voltage at an arbitrary position and driving the lens based on the calculated value. The former is known as a mountain climbing servo system, and NHK Technical Research Report 40. Volume 17: Vol. 1, No. 86, No. 21
Page, or the literature such as the Technical Report of the Television Society of Japan published in November 1982, ED page 675, page 7 and the like. The latter is disclosed in JP-A-62-272217.
JP-A-62-272218 and the like.

That is, a high frequency component is extracted from a video signal obtained from an image pickup device such as a CCD, and a voltage (hereinafter referred to as a focus voltage) obtained by detecting the high frequency component is set as Es, and a distance to a focus position P of a focus lens is set. Where x is the focal voltage E
The relationship between s and the distance x is expressed by the following equation: Es (x) = b · exp {− (ax) ² } (1) where a: a coefficient that determines the curve slope of the focus voltage Es b: focus voltage Es It is known that this can be approximated by a coefficient that determines the maximum value of the curve of, and the distance x to the focus position is calculated based on this equation (1).

[0005]

In the case of the former hill-climbing servo system, the focus voltage gradient is sequentially detected, and the peak (focus position) of the focus voltage is searched while moving the lens. There is a problem that it takes a long time to reach the position. On the other hand, the latter extracts a high frequency component from the video signal as a focus voltage and calculates the distance to the in-focus position based on this focus voltage, but the calculation result depends on the actual subject state. Therefore, depending on the calculation result based on the focus voltage obtained from one frequency component, until the distance to the in-focus position falls within the permissible circle of confusion,
There may be a case where the movement of the focus lens and the calculation of the distance to the in-focus position after the movement of the lens have to be repeated many times.

The present invention has been made in view of the above circumstances, and an autofocus method and apparatus capable of estimating an appropriate defocus amount based on fuzzy reasoning and thereby performing focusing reliably and quickly. The purpose is to provide.

[0007]

In order to achieve the above-mentioned object, the present invention forms a subject image on an image pickup device through an optical system, and outputs a plurality of different high frequencies from video signals obtained from the image pickup device. A frequency component for each band is extracted as a plurality of focus voltages, a distance to move a part of the optical system to a focus position is calculated based on the plurality of focus voltages, and the optical system is moved by the calculated distance. Is an autofocus method for moving a part of the optical axis in the optical axis direction to focus on a subject, and the calculation of the distance to be moved should be performed based on a plurality of focus voltages. A distance is calculated, a plurality of conformances are determined from the calculated plurality of distances to be moved and a membership function having the plurality of high frequency bands as variables, and a representative value of a set of the plurality of conformances is focused. Distance to position It is characterized in that calculated as the.

[0008]

According to the present invention, frequency components for a plurality of different high frequency bands, which are different from each other, are extracted from a video signal obtained from an image pickup device as a plurality of focus voltages, and based on the plurality of focus voltages, a focus position up to a focus position is obtained. The distance to be moved of a part of the optical system is calculated. Then, a plurality of goodness-of-fits is determined from the calculated plurality of distances to be moved and a membership function having the plurality of high-frequency bands as variables, and a representative value of a set of the plurality of goodness-of-fits is set to a focus position. It is calculated as a distance. In this way, the distance to the focus position is calculated, and based on the calculation result, a part of the optical system is moved in the optical axis direction to focus on the subject.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT A preferred embodiment of an autofocus method and apparatus according to the present invention will be described in detail below with reference to the accompanying drawings. FIG. 1 is a block diagram showing an embodiment of an electronic camera equipped with an autofocus device according to the present invention.

As shown in the figure, the photographing lens 10 is a rear focus type zoom lens, and the front lens group 1
It has 0A and a rear lens group 10B. Zooming and focusing are performed by appropriately moving these lens groups by the zoom motor 12 and the focus motor 14. The positions of the front lens group 10A and the rear lens group 10B are detected by the zoom position detector 16 and the focus position detector 18, respectively, and these position data are added to the arithmetic processing circuit 20. The diaphragm 22 is driven by an iris motor 24, and the diaphragm position is detected by a diaphragm position detector 26.

The light from the subject is incident on the light receiving surface of the image pickup device (CCD) 30 through the taking lens 10 and C
Each sensor of the CD 30 converts into a video signal having a signal charge according to the intensity of light. The video signal converted by the CCD 30 is sequentially read out, appropriately amplified by the preamplifier 32, and then added to a signal processing circuit (not shown) and the AF circuit 40. The CCD 30 outputs a video signal for one field in every predetermined sampling period (1/60 second) in synchronization with the sync signal from the sync signal generation circuit 34, and the arithmetic processing circuit 20 outputs the sync signal as well. The arithmetic processing is executed in synchronization.

The AF circuit 40 includes a gate circuit 42 and n band-pass filters (BPF1, BPF2, ..., BPF).
n), n focus voltage detection circuits (DET1, DET2,
... DETn) and an A / D converter 44. The gate circuit 42 allows only a video signal within a predetermined measurement range to pass among the video signals for one field to be input. When a signal indicating a predetermined measurement range is input from the arithmetic processing circuit 20, the input signal is input. The video signal is output to BPF1, BPF2, ..., BPFn only during the period. BPF
1, BPF2, ..., BPFn pass the frequency components of different high frequency bands in the input video signal and output them to DET1, DET2 ,.

DET1, DET2, ... DETn are composed of detection circuits, for example, and output a voltage signal obtained by detecting an input high frequency component signal as a focus voltage. still,
FIG. 4A shows a characteristic curve of each focus voltage when the coefficient a changes due to the frequency component of the object, the lens diaphragm, etc., and FIG. 4B shows the characteristic curve of the object due to the contrast of the object and the gain of the amplification system. The characteristic curve of each focus voltage when the coefficient b is changed is shown.

The n focus voltages thus detected are converted into digital signals by the A / D converter 44 and then output to the arithmetic processing circuit 20. Next, the operation of the arithmetic processing circuit 20 will be described with reference to the flowchart shown in FIG. First, the arithmetic processing circuit 20
N focus voltages at the current movement position of the rear lens group 10A (hereinafter referred to as focus lens) are input (step 100).

Next, the focus voltage is verified according to rule 1 (step 102). [Rule 1] [Verification of focus voltage] If the focus voltage value detected from the frequency component of the video signal passing through each BPF is equal to or higher than a predetermined value, step 104 is performed. If the focus voltage value is equal to or lower than the predetermined value, step 104 is performed. Let it run.

In step 104, R ₊ according to rule 2 is used.
Perform a test of + R ₋ <0. “Rule 2” [verification of R ₊ + R ₋ <0] That is, as shown in FIG. 5, the focus voltage at a position away from the in-focus position P by distance x _{0 is set} to Es _0, and a small distance forward and backward from position x _0. Let Es ₊ and Es ₋ be the focus voltages at positions (x ₀ + dx) and (x ₀ −dx) separated by dx, respectively.
s ₊ , Es ₀ and Es ₋ are Es ₊ = b · exp- {a (x ₀ + dx)} ² (2) Es ₀ = b · exp {-(ax ₀ ) ² } (3) Es ₋ = b · exp− {a (x ₀ −dx)} ² (4) When the ratios of the focus voltages Es ₊ and Es ₋ to the focus voltage Es ₀ are obtained from the above equations (2) to (4), the following equation is obtained: Es ₊ / Es ₀ = exp {−a ² dx (2x ₀ + dx)} (5) Es ₋ / Es ₀ = exp {a ² dx (2x ₀ −dx)} (6) (5) and (6), taking the natural logarithm of both sides, these respectively R _+, R _- if that, R _+,
R ₋ is the following equation: R ₊ = LN (Es ₊ / Es ₀ ) = − a ² dx (2x ₀ + dx) (7) R ₋ = LN (Es ₋ / Es ₀ ) = a ² dx (2x ₀ -Dx) (8) Then, from the expressions (7) and (8), the sum and the difference of R ₊ and R ₋ are calculated. R ₊ + R ₋ = −2 · (adx) ² (9) R ₊ −R ₋ = −4 · a ² dxx ₀ (10) As is clear from the above equation (9), (R ₊ +
R ₋ ) is always negative, but the calculated value of (R ₊ + R ₋ ) may be positive depending on the measured value of the focus voltage. In this case, it is determined that the measured value of the focus voltage has an error.

Therefore, when R ₊ + R ₋ <0, step 106 is performed, and when R ₊ + R ₋ ≧ 0, step 116 is performed.
To run. In step 106, the test of the calculated value according to rule 3 is executed. "Rule 3" [Test of calculated value] Above (9) and (10)
When the coefficient a is deleted from the equation and x ₀ is summarized, x ₀ = dx (R ₊ −R ₋ ) / 2 (R ₊ + R ₋ ) ... (11)

Therefore, the focus voltage Es₊, Es₀And Es
_-Is measured, and these focal voltage Es ₊, Es₀And Es
_-Based on equations (7) and (8)₊, R_-Calculate
Out, this R₊, R_-By substituting
The amount of movement of the focus lens to the in-focus position P (focus
Deviation amount x₀Can be calculated. It ’s out of focus
Amount x₀Is calculated for each focus voltage obtained through each BPF.
Will be issued.

The defocus amount calculated as described above may exceed a reliable value for each BPF. In step 106, if the defocus amount is within a reliable value for each BPF, the process proceeds to step 108 for each BPF.
If the value exceeds the reliable value in step 118, the process proceeds to step 118 where the calculated value is corrected to a predetermined value (maximum reliable value). The number of BPFs is n = 4, and each BPF 1, 2,
An example of types 3 and 4 and their maximum confidence values is shown in the following table. Next, in step 108, the driving amount of the lens is determined by fuzzy reasoning according to Rule 4.

[Rule 3] [Driving amount is determined by fuzzy reasoning] BPF 2, 3,
The defocus amount can be obtained in three values from the focus voltage detected via 4. Then, a fuzzy set is obtained from these three values and the membership function, and fuzzy reasoning is performed to determine the center of gravity of the set as the defocus amount.

That is, as shown in FIG.
The membership functions corresponding to 3 and 4 are μ2 and μ, respectively.
Assuming that the defocus amount calculated from the focus voltage detected through BPFs 2, 3, and 4 is a2, a3, and a4, the fuzzy set S is given by the following equation: S = ∨ (∧a2∩μ2) ∪ (∧a3∩μ3) ∪ (∧a4∩μ4) (12), and the center of gravity A of this set S is determined as the defocus amount (FIG. 3 (B)).

The defocus amount A is defined by the area of the membership function of the inference result obtained from the membership function μ2 and the defocus amount a2 as S2 and the center of gravity thereof as A2. Let S3 be the area of the inference result membership function obtained from the defocus amount a3, and let A3 be its center of gravity, and let S be the area of the inference result membership function obtained from the membership function μ4 and the defocus amount a4.
4. If the center of gravity is A4, the center of gravity A can be calculated based on the following equation: A = (S2.A2 + S3.A3 + S4.A4) / (S2 + S3 + S4).

Then, in step 110, a focus test is performed from the defocus amount (center of gravity A) thus calculated. That is, whether the calculated defocus amount is within the permissible circle of confusion,
That is, the absolute value of A is equal to or less than the radius of confusion circle δ _min (| A | ≦ δ
_min )), and if | A | ≦ δ _min , it is determined to be in focus, and the autofocus operation ends.

On the other hand, if _│A│ > δ _min , it is determined that the lens is out of focus, the focus lens is driven by the calculated defocus amount A (step 112), and the process returns to step 100. The calculation of the amount of defocus etc. is executed again. If the focus voltage value verified in step 102 is less than or equal to the predetermined value, the ratio of error components such as noise increases, which may adversely affect the subsequent verification. Therefore, the process proceeds to step 114. Then, the calculated value error 1 according to rule 5 is dealt with.

[Rule 5] [Treatment of calculation value error 1] This is a case where the slope of the focus voltage cannot be detected by the detection of BPF1. In this case, the search drive is performed in a fixed direction. In step 104, B
If any of the calculated values obtained with PF2-4 are inappropriate,
Since it becomes impossible to drive the lens based on the calculated value, in this case, the process proceeds to step 116, and the calculated value error 2 according to rule 6 is dealt with.

[Rule 6] [Remedy for calculation value error 2] In this case, the unit feed amount and the focus position are moved based on the inclination of the focus voltage detected by the BPF 1. The method of calculating the defocus amount based on the focus voltage is not limited to this embodiment, and various methods can be applied.

[0027]

As described above, according to the autofocus method and apparatus of the present invention, a plurality of different frequency components for different high frequency bands are extracted from a video signal obtained from an image sensor as a plurality of focus voltages, Based on the plurality of focus voltages, the distance to move a part of the optical system to the in-focus position is calculated, and the calculated plurality of distances to move and a membership function having a plurality of high frequency bands as variables. Since a plurality of goodness-of-fits are determined from this, and the representative value of the set of these goodness-of-fits is calculated as the distance to the in-focus position, the distance to the in-focus position can be estimated properly based on fuzzy inference. Therefore, focusing can be performed surely and quickly.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of an electronic camera equipped with an autofocus device according to the present invention.

FIG. 2 is a flowchart used to explain the operation of the arithmetic processing circuit of FIG.

3A is a diagram showing an example of a membership function corresponding to each BPF and an out-of-focus amount calculated from each BPF, and FIG. 3B is a diagram showing a membership function of an inference result. It is a figure showing a set.

FIG. 4A is a characteristic curve of each focus voltage when a frequency component of a subject, a lens diaphragm, and the like change.
(B) is a graph showing a characteristic curve of each focus voltage when the contrast of the subject and the gain of the amplification system are changed.

FIG. 5 is a graph showing a characteristic curve of a focus voltage used for explaining the principle of the present invention.

[Explanation of symbols]

 10 ... Shooting lens 10B ... Focus lens 14 ... Focus motor 20 ... Arithmetic processing circuit 30 ... Imaging device (CCD) 40 ... AF circuit 42 ... Gate circuit BPF1 to BPFn ... Band pass filter DET1 to DETn ... Focus voltage detection circuit

Claims (3)

[Claims] 1. A subject image is formed on an image pickup device through an optical system, and frequency components for a plurality of different high frequency bands are extracted as a plurality of focus voltages from a video signal obtained from the image pickup device, Calculate a distance to move a part of the optical system to a focus position based on a plurality of focus voltages,
An autofocus method for focusing a subject by moving a part of the optical system in the optical axis direction by the calculated distance, wherein the calculation of the distance to be moved is based on a plurality of focus voltages. A plurality of distances to be moved to the position are calculated, and a plurality of fitness values are determined from the calculated plurality of distances to be moved and a membership function having the plurality of high frequency bands as variables. The autofocus method is characterized in that the representative value of the set is calculated as the distance to the focus position. 2. A subject image is formed on an image pickup device through an optical system, and frequency components of a plurality of different high frequency bands which are different from each other are extracted as a plurality of focus voltages from a video signal obtained from the image pickup device. Calculate a distance to move a part of the optical system to a focus position based on a plurality of focus voltages,
An autofocus method for focusing a subject by moving a part of the optical system in the optical axis direction by the calculated distance, wherein the calculation of the distance to be moved is based on a plurality of focus voltages. A plurality of distances to be moved to the position are calculated, a membership function of a plurality of inference results is obtained from the calculated plurality of distances to be moved and a membership function having the plurality of high frequency bands as variables, and the plurality of The barycenter value of the membership function of the inference result is Ai (i = 1,2, ... N), and the area of the membership function of the multiple inference results is Si (i = 1,2, ... N)
Then, the distance A to the in-focus position is calculated based on the following formula: A = Σ (Ai × Si) / ΣSi. 3. An optical system for forming a subject image on an image pickup device, and a plurality of high frequency components for extracting, as a plurality of focus voltages, frequency components for a plurality of different high frequency bands from a video signal obtained from the image pickup device. Band frequency component extracting means, first computing means for computing a distance to be moved to a focus position based on a plurality of focus voltages obtained from the high frequency component extracting means, and the first computing means A plurality of goodnesses of fit are determined from a plurality of distances to be moved calculated by the above and a membership function having the plurality of high frequency bands as variables, and a representative value of a set of the goodnesses of fit is set as a distance to a focus position. Second computing means for computing, driving means for moving a part of the optical system in the optical axis direction by a motor, and distance calculated by the second computing means so as to focus on a subject. Autofocus apparatus characterized by and a control means for controlling said drive means based on.