JPS63267907A - Focus detector - Google Patents

Abstract

PURPOSE:To improve the focus detection precision by receiving the output of a maximum value detecting means and controlling the output level of an image sensor means so that said output is a prescribed value. CONSTITUTION:The image formed by an optical system is converted to an electric signal by the image sensor means. Though the output of the image sensor means includes a DC component, a filter means 21 eliminates the DC component. The output of the filter means 21 is subjected to the operation processing by an operating means 22 to detect position information of the image formed on the image sensor means, and the output of the filter means 21 is converted to the absolute value by a maximum value detecting means 23 to detect the maximum value. A sensor control means 24 which receives the output of the maximum value detecting means 23 controls the output level of the image sensor means so that said output has a prescribed value. Thus, correct focus control data of high reliability and high focus detection precision is outputted.

Description

[Detailed description of the invention] "Industrial application field" The present invention relates to an improvement in a focus detection device for a camera.

``Prior Art'' As a focus detection device used in cameras, a microcomputer performs arithmetic processing on data obtained by AD converting the output of an image sensor such as a CCD, which converts an image produced by an optical system into an electrical signal, and forms the image on the image sensor. It is known that focus detection is performed by detecting information regarding the position of a focused image.

In this type of focus detection device, if the output of the image sensor exceeds the AD conversion range, the AD conversion data will become saturated and will no longer correspond to the output of the original image sensor, resulting in a decline in focus detection accuracy. Resulting in.

Therefore, in order to prevent this situation from occurring,
A technique was used to control the image sensor so that the output of the image sensor reached a maximum value or a predetermined value.

This will be explained in detail using FIGS. 8 to 10.

In FIGS. 8 and 9, the image sensor output is indicated by a dashed line, and the solid vertical line indicates AD conversion data.

Assuming that the input range of AD conversion is 0 to 5V and AD conversion is performed in 8 bits (0 to 255), if the image sensor output exceeds 5V as shown in Figure 8, the AD of the part exceeding 5V will be All conversion data will be clamped to 255.

Therefore, by controlling the image sensor, as shown in Fig. 10, the maximum value of the image sensor output or the voltage Vp within the range of O to 5V is set.
If this is done, the AD conversion data will all fall between 0 and 255, and will not be clamped.

"Problems to be Solved by the Invention" However, the above-mentioned conventional techniques have the following problems.

Conventionally, when processing AD-converted data to obtain information regarding the position of an image formed on an image sensor, in order to improve accuracy, DC conversion is performed from AD-converted data.
The components were removed to create data containing only AC components, and then predetermined calculations were performed to obtain information regarding the position of the image.

However, even if the AD-converted data falls within a predetermined range, when the DC component is removed, the ACJi data may be clamped beyond the predetermined range. As a result, in such cases, even if the AC component data is used to obtain information regarding the position of the image, it is unreliable.

The above drawbacks will be explained in more detail.

For example, AD converted data X (n): n = 1-N.

The process of removing the DC component is performed by a digital filter that takes differences of the first order or higher.

As an example, the processing of the second-order difference filter is expressed as an arithmetic expression in equation (1).

Y (m) = (-X (m) + 2X (
m + 1) - = 1 is selected as the minimum coefficient value for preserving the information of the least significant bit of the AD conversion data.

At this time, if the data range of the AD-converted data
~51O) will be required. Therefore, if AC component data is handled in 8 bits (-128 to 127), data exceeding that range will be clamped.

FIG. 9 shows the clamping state of data when second-order difference data obtained by the second-order difference filter of equation (1) is used as AC component data.

As described above, the focus detection apparatus 3 using the conventional technology has the drawback that data clamping occurs and focus detection accuracy deteriorates.

"Means for Solving the Problems" The gist of the present invention for solving the above problems is to provide: an image sensor means for converting an image formed by an optical system into an electrical signal; a filter means (21) for removing DC components from the output; a calculation means (22) for calculating position information of the image formed on the image sensor means (4) by processing the output of the filter means; maximum value detection means (23) for converting the output of the filter means (21) into an absolute value and further detecting and outputting the maximum value; and receiving the output of the maximum value detection means so that the output becomes a predetermined value. The focus detection device comprises sensor control means (24) for controlling the output level of the image sensor means.

"Operation" The image produced by the optical system is converted into an electrical signal by the image sensor means.

If the output of the image sensor means includes a DC component, the filter means removes the DC component.

The output of the filter means is subjected to calculation processing by a calculation means to detect the positional information of the image formed on the image sensor means, and the output of the filter means is converted into an absolute value by a maximum value detection means, and the maximum value thereof is detected. be done.

The sensor control means that receives the output of the maximum value detection means controls the output level of the image sensor means so that the output becomes a predetermined value, and outputs appropriate focus control data with high reliability and high focus detection accuracy. It is something to do.

"Example" Hereinafter, various embodiments of the present invention will be described based on the drawings.

1 to 3 show a first embodiment of the present invention.

As shown in FIG. 2, a camera to which the present invention is applied includes a photographing optical system 1, a main mirror 2 that is constrained by the photographing optical system 1, and a sub-mirror 3 for extracting focus detection information.
It includes an image sensor 4 that converts the image taken out by the sub-mirror 3 into an electrical signal, a central processing unit 5 that processes the output of the image sensor 4, and a display means 6 and a lens driving means 7 that are controlled by the central processing unit 5. ing.

FIG. 3 shows an example of a focus detection optical system used in a focus detection device, in which an exit l111 having areas 13a and 13b is shown.
Field lens 10 installed on the focus detection plane after 2
It consists of a pair of re-imaging lenses 11 and a pair of image sensors 4a, 4b, and the subject image formed on the focus detection surface is transmitted to the pair of image sensors 4a, 4b.
It is configured to be re-imaged on top.

FIG. 1 shows the configuration of the central processing unit 5 of the camera shown in FIG.

The central processing unit M5 includes an AD conversion means 2o and a second-order difference means 2.
1. Correlation calculation means 22. It is composed of a maximum value detection means 23 and a sensor control means 24.

Next, the effect will be explained.

A part of the light flux from the subject that has passed through the photographic optical system 1 is guided to the finder optical system by the main mirror 2.
The other part passes through the main mirror 2, is further reflected by the sub-mirror 3, and is guided to a focus detection optical system in which a focus detection surface is set on or near the film surface.

The pair of image sensors 4 incorporated in the focus detection optical system are charge accumulation type image sensors, and require a charge accumulation time corresponding to the amount of light in order to obtain an appropriate output signal.

The output signal from the image sensor 4 is taken into the central processing unit 5 and subjected to processing to be described later, and the defocus direction and defocus amount are detected. The central processing unit M5 controls the display means 6 based on the defocus direction and defocus amount to make the user aware of the focus adjustment state.
At the same time, the lens driving means 7 is controlled to drive the photographing optical system to the goldfish state.

The light beam received by the image sensor 4a is transferred to the re-imaging lens 1.
The light flux passes through a region 13a projected onto the exit pupil 12 of the photographing optical system by the aperture field lens 10 of No. 1. On the other hand, the light flux received by the image sensor 4b is
Similarly, the light flux passes through the area 13b on the exit pupil 12.

In this focus detection optical system, when the object image plane formed by the photographing optical system coincides with the focus detection plane, a predetermined position determined by the optical system arrangement is placed on the pair of image sensors 4a and 4b. The two subject images are re-imaged in relation to each other.

Further, when the object image plane is deviated from the focus detection plane in the optical axis direction, the pair of image sensors 4a, 4
The pair of subject images formed on the image sensor b also deviate from the predetermined positional relationship in the lateral direction on the image sensor.

Therefore, by determining the positional relationship between a pair of subject images from the outputs of a pair of image sensors and comparing that positional relationship with the predetermined positional relationship, it is possible to determine the direction of deviation of the subject image plane from the focus detection plane (defocus direction). and the amount of deviation (defocus amount) can be detected.

Note that the defocus amount in the optical axis direction and the deviation amount in the image sensor arrangement direction are linked by the opening angle α between the optical axes of the pair of reimaging optical systems shown in the figure.

In the central processing unit 5, a pair of image sensors 4a and 4b
The output signals from the AD conversion means 2° are converted into n pieces of primary data A I (1) to A 1 (n), respectively.

It is AD converted into B1(1) to B1(n).

The AD conversion means 20 converts the output signal of the image sensor 4 into 8 bits of voltage levels 0 to 5vftO to 255, for example.
It is configured to perform D conversion, rounding values below Ov to 0, and rounding values above 5V to 255.

Next, AD converted primary data A I (1) ~ A 1
(n), B 1 (1) ~ B 1 (n) is 2
The secondary data A 2 (1) to A2 (■) are sent to the floor difference means 21 and subjected to second floor difference calculation processing as shown in equation (2).
.

It is converted into B2(1) to B2(s).

A 2(h)--A I(h)÷2* A 1(h◆2
)-A 1(h+4)B 2(h)-81(h)421
81 (h ◆ 2) - 81 (h + 4)・- (2) h depth 1~
l, 1 stomach n-4 By performing second-order difference processing as in equation (2), 1
Next data AI(1) ~Al(n), Bl(1) ~B
1(n), it is possible to remove the C component and the first-order linear component, which are harmful to focus detection.

If the primary data is 8 bits, the output range of the secondary data is -510 to +510 according to equation (2), but it is better to treat the secondary data as having the same number of bits as the primary data.
The output range of secondary data is also 8 bits - 128 to 8 bits because it saves memory and simplifies arithmetic processing, and because the fluctuation range of secondary data is usually smaller than that of primary data. As +127, data below -128 becomes -128, and data above +127 becomes +12.
It's wrapped up in 7.

The correlation operator stage 22 outputs the secondary data A2(1) to A2(m), B2(1) to B2(■
), a correlation calculation process to be described below is performed to obtain the relative shift amount between the two images on the image sensor 4, and further, the defocus amount and defocus direction on the focus detection plane are detected from the shift amount.

The correlation calculation process first calculates the correlation amount C(L) by shifting one set of secondary data by a relative shift amount by calculating equation (3).
seek.

C(L)-Σl A 2(i) -82(i+L) l
-(3) Indeed, in equation 3), the range of Σ is within the range where secondary data exists.

FIG. 4 shows a general example of a graph of the correlation amount C(L) when the relative shift amount is treated as the horizontal axis.

In FIG. 4, the correlation 1c(L) becomes the minimum value in a relative shift amount where the correlation between a set of secondary data is high.

However, since the relative shift ML is discrete and the correlation amount C(L) is also obtained only at discrete times, the minimum value C(u) of the correlation amount C(u) when the continuous relative shift amount is 2 xm
) is determined by interpolation.

FIG. 5 is a diagram showing an example of interpolation.

First, the relative shift amount χ that gives the minimum value of the correlation amount C(L)
, relative shift amounts X-1, X+1 before and after that, and correlation amounts C(x-1), C(x).

When C(x◆l), the minimum correlation amount C(x) and the remaining 2
The larger correlation amount among the correlation amounts (C(x
◆Draw a straight line Q connecting 1)).

Next, the remaining correlation amount (C(x-1) in Figure 5)
Draw a straight line P that has the opposite slope to the straight line Q through
The intersection point of and Q is found, and the intersection coordinates C (x layer) and x grave are the minimum correlation amount for the continuous relative shift amount sentence and the relative shift amount that gives the minimum correlation amount.

Equation (4) is an equation that gives the minimum correlation amount C(xm) and its relative shift amount×I by the above-mentioned interpolation method.

X■= x + D / E C(X@) = (C(X)-1D I ) /EHowever, D=1/2 X (C(x-1)-C(x+1))E
= wax((:(x+1)-C(x), C(x-
1) −C(x)l −(4) xm obtained by equation (4) represents the relative shift amount of a set of secondary data, and if the pitch of the secondary data is S, then 2 on the image sensor The relative lateral shift amount of the image is S x xm, and the defocus amount d on the focus detection plane is finally determined as in equation (5).

d=kXSXxv ---(s
) In equation (5), K is a coefficient determined by the configuration of the focus detection optical system (opening angle α in FIG. 3, etc.).

As described above, the defocus amount and defocus direction are determined by the correlation calculating means 22, and further processing such as lens driving and display is performed based on the outputs thereof.

On the other hand, the output data A 2 (1) of the second-order difference means 21 ~
A 2 (@) and B 2 (1) to B2 (Otori) are also sent to the maximum value detection means 23.

The maximum value detection means 23 first detects data A 2 (1) to A
2 (■) and B2 (1) to B2 (■) are converted into feed values as shown in equation (6) to obtain tertiary data A 3 (1) to A3 (■).

Obtain B3 (1) to B3 (■).

A 3(i) = l A2(i) l , B 3(i
)=I B2(i) l・-(6)i=1 ~m Next, tertiary data A 3 (1) ~A3(謹), B3(1
)~B 3 (m), the maximum value F in the tertiary data is (7
) is calculated using the formula.

F = max(A3(i), B3(i)) i
= 1 to m - (7) The maximum value detection means 23 sends the obtained maximum value F to the sensor control means 24.

The sensor control means 24 controls the charge accumulation type image sensor 4.
means for controlling the charge accumulation time of the maximum value F and the charge accumulation time Ts in the sequence in which the maximum value F is determined.
And from the data of the target value Fa of the maximum value F, the charge accumulation time Tn in the next sequence is determined as shown in equation (8).

Tn=FaxTs/F−(8
) The sensor control means 24 controls the charge accumulation time of the charge accumulation type image sensor 4 in the next sequence according to the determined charge accumulation time.

In the first embodiment of the present invention, since the configuration is as described above, two images can be taken regardless of the contrast of the subject.
Next data A2 (1) ~ A2 (s), B 2 (1) ~ B
The value of 2(■) can be kept within the range of -Fa to +Fa as shown in FIG.

FIG. 7 shows a second embodiment of the present invention, which differs from the first embodiment in that a second maximum value detection means 25 for primary data is provided, and the sensor control means 24 has two maximum value detection means 23. 25
The difference is that sensor control is performed based on the output of the sensor.

In the second embodiment, the AD conversion means 2 of the central processing unit 5
The operations of the 0.2-order difference means 21, the correlation operator stage 22, and the first maximum value detection means 23 are the same as those in the first embodiment, and will therefore be omitted.

The second maximum value detection means 25 receives primary data A I (1) to A 1 (n) from the AD conversion means 20 .

Upon receiving B 1 (1) to B 1 (n), the maximum value G in the primary data is determined as shown in equation (9) and the sensor control means 24
Output to.

G = max(A I (i), B 1 (+)
) i-1~n = (9) The sensor control means 24 detects the maximum value F sent from the first maximum value detection means 23, its target value Fa, and the maximum value sent from the second maximum value detection means 25. From the charge accumulation time Ts in the sequence in which the value G, its target value Ga, the maximum value F, and the maximum value G were determined, the charge accumulation time Tn in the next sequence is determined as shown in equation (io).

T n = win(TI, T 2) TI=Ga
XTs/G, T2-Fa xTs/F =・(l
O) The sensor control means 24 controls the charge accumulation time of the charge accumulation type image sensor 4 in accordance with the determined charge accumulation time in the following sequence.

In the second embodiment, since the configuration is as described above, the secondary data A 2 (1) to A2 (■), B2 (1
) ~B2(m) can not only be kept within the range of -Fa to +Fa, but also the primary data A I
(1) ~A 1 (n), B 2 (1) ~B
Since the value of 2 (s) can also be kept within the range from 0 to Ga and neither the primary data nor the secondary data is clamped, the focus detection accuracy is further improved than in the first embodiment.

In the above embodiment, the primary data is the second difference means 21
is subjected to second-order difference calculation processing as shown in equation (2), but is not limited thereto, and may be any filter means that performs calculation processing to remove DC components from primary data.

Further, the sensor control means 24 is a charge accumulation type image sensor 4.
The magnitude of the output is controlled by controlling the charge accumulation time of the image sensor, but it may also be configured to control the gain of the amplification of the image sensor's output, or the amount of light incident on the image sensor may be reduced. It may also be configured to be controlled by a filter or a density filter.

Further, in the above description, the image sensor output is processed digitally, but of course it may be processed analogously.

Further, in the above embodiment, the sensor control means 24 controls the accumulation time of the charge accumulation type image sensor 4 in the next sequence, but it is possible to control the output in real time using an image sensor capable of non-destructive continuous operation. may be configured.

"Effects of the Invention" According to the focus detection mN according to the present invention, the data used for correlation calculation is not clamped, so it is possible to provide a focus detection device with high focus detection accuracy.

[Brief explanation of drawings]

1 to 6 show a first embodiment of a focus detection device according to the present invention, FIG. 1 is a block diagram of the main part of the focus detection device, and FIG. 2 is a configuration when applied to a camera. Explanatory diagram,
Figure 3 is a perspective view of the focus detection optical system used in the focus detection device, Figures 4 and 5 are illustrations of correlation calculation, Figure 6 is an illustration of signal processing of image sensor output, and Figure 7 is Block diagram of the main part configuration of the focus detection device according to the second embodiment, No. 8
Figures 1 to 1O are explanatory diagrams of conventional image sensor output signal processing. 4.4a, 4b--Image sensor 20--AD conversion means 21...Second-order difference means 2
2... Correlation calculation means 23. 25... Maximum value detection means image Fig. 1 [(L) Fig. 5 Fig. 6 Fig. 7 Fig. S

Claims (2)

[Claims] (1) image sensor means for converting an image formed by an optical system into an electrical signal; filter means for removing DC components from the output of the image sensor means; and arithmetic processing of the output of the filter means to be applied onto the image sensor means. calculation means for detecting positional information of the formed image; maximum value detection means for converting the output of the filter means into an absolute value and further detecting and outputting the maximum value; and receiving and outputting the output of the maximum value detection means. 1. A focus detection device comprising sensor control means for controlling the output level of the image sensor means so that the image sensor has a predetermined value. (2) The focus detection device according to claim 1, wherein the filter means performs second-order difference processing on the output of the image sensor means.