JPS62280708A - Interpolating method for focus detecting device - Google Patents

Abstract

PURPOSE:To reduce variance in error as to all object luminance distribution and to improve detection accuracy by finding a point S, deviation quantity, at which the angles for vlewing F(M)F(M+1) or F(M-1) and F(M) are equal approximately. CONSTITUTION:An object light image is received by the 1st and the 2nd photoelectric converting element groups and a focus detecting device detects the distance to the object from the correlation between the output signals of the element groups. Outputs of the 1st photoelectric converting element group are denoted as a1-an, outputs of the 2nd photoelectric converting element group are denoted as b1-bn, and a function which evaluates deviation quantities of a1-an, and b1-bn is F(S). Here, when the function is shown by an equation I and the minimum value of the F(S) is F(M), the deviation quantity S between the 1st image by the 1st photoelectric converting element group and the 2nd image by the 2nd photoelectric converting element group is interpolated by an equation I.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a method for supplementing a focus detection device, and more specifically, to a method for supplementing a focus detection device, and more specifically, to detect the amount of deviation in a focus detection device that receives a subject light flux and detects a focus state in an optical device such as a camera. Regarding the completion method.

[Prior art] A first part and a second part of a photographic lens sandwiching the optical axis of the photographic lens.
The photographing lens is equipped with an exit pupil split type focus detection device and a baseline long range meter type focus detection optical system to determine the focus state by detecting the correlated position of the two images created by the limb photographing light beams that have passed through the respective parts. A focus detection device that detects the in-focus position by moving a mirror and using a double edge matching method is already well-known, and the principle (including its optical system) was developed in Japanese Patent Application Laid-Open No. 1986-
It is disclosed in detail in Japanese Patent Publication No. 126517 and Japanese Patent Publication No. 60-5925.

In these focus detection devices, the photoelectric conversion element group includes, for example,
Pixel row A consisting of CCD line sensors.

A column B is arranged, and output signals 21 of the pixel columns A and B,
, a2-a, , b, , b2'''bn is evaluated using a function F (S).

That is, F (S) - Σ1ak - bk + s I (S is shift 2), and in the focus detection device of the above-mentioned Japanese Patent Application Laid-Open No. 126517/1982, F (S) is F (M),
As shown in FIGS. 5(a) and 5(b), the amount of deviation is F (M
-1)<F (M+1) [Problem to be Solved by the Invention] By the way, the complementing means in the focus detection device of the above-mentioned Japanese Unexamined Patent Publication No. 59-126517, as shown in FIG. This is based on the assumption that the value of F (S) increases at equal angles θ to the left and right with a peak near the minimum value of F (S) with respect to the First, the error varies widely depending on the brightness distribution of the subject. That is, in this interpolation method, when the value of F (S) is near the matching state of columns A and B, one pixel column is replaced by another pixel column.
This is based on the assumption that the amount of change in the value of F (S) is equal when shifting left and right relative to This assumption is not strictly correct.

Furthermore, it is clear that the idea itself is not valid for the following reasons. That is, the intersection of the straight lines shown in FIGS. 5(a) and 5(b) corresponds to a point where the amount of deviation is zero. Originally, the correlation output F (S) corresponding to this point should be zero, but this is not the case.

On the other hand, in the automatic focus detection device disclosed in Japanese Patent Publication No. 60-5925, the amount of complementation is obtained by approximating F (S) to a quadratic curve, but according to the results of a simulation of this method of complementation, , maximum 0.1
A pitch error occurs, and it is not as good as the previous one. This is also due to the reason mentioned above.

However, these conventional techniques do not have large errors for all wave object brightness distributions. In fact, there are cases in which the accuracy is very high depending on the limb being photographed. Tata problem is 12;
The difference is that the variation is large.

Therefore, an object of the present invention is to eliminate the problems in the conventional interpolation method described above, and to eliminate the error λ for all subject brightness distributions.
An object of the present invention is to provide a method for supplementing a focus detection device with less variation in the focus detection device.

[Means and effects for solving the problem] The point corresponding to S shown in FIGS. 5(a) and 5(b) is F(S)-0. The most straightforward and reasonable approach is to start from the idea of what kind of relationship S has with F (M) and l'i' (M+l) or with F () and F (M-1). It's a way of thinking. Figure 6 above.

As suggested from FIG. 7, even if one of the pixel columns A and B is slightly changed left and right relative to the other from the in-focus position, the amount of change in the correlation output F (S) is almost the same. The present invention is based on these four ways of thinking, and approximates S to F(M)F (M+1) or S as shown in FIG. 1(a) (11).
From F (M-1) F (M) is assumed f9 or equal point S
We seek. That is, when F(M-1)≧F(M+1), when F(M-1)<F(M'+1), the present invention will be described below with reference to an illustrated embodiment.

FIG. 2 is a block diagram of the configuration of an electric circuit when the method of the present invention is applied to a focus detection device for a camera.

Note that many methods have already been proposed for detecting the distance to an object from the correlation between the two images that have passed through the two imaging lenses L t and L 2; Since the method is described in detail in Japanese Patent Publication No. 59-126517 and Japanese Patent Publication No. 60-5925, the optical system of the present invention uses these already known methods, and their explanation will be omitted.

The imaging lens L1. The transmitted light L2 is made to enter the CCD line sensor 1, and the object structure that has passed through the imaging lens L is reflected on the light receiving surface of the first photoelectric conversion element group IA forming the pixel row A, and The limb photographic image that has passed through the imaging lens L2 is respectively formed on the light receiving surface of the second photoelectric conversion element group IB forming the pixel row B. Above column A,
For example, each of the pixels on the light receiving surface of column B is formed of 64 pixels.

The drive control circuit 2 includes a COD drive circuit and an A-D converter 3.
This is a circuit for controlling the timing of CCD transfer lock φ1. φ2, and an integral output signal MO3 of the monitor light receiving element for determining the integration time of the CCD light receiving section is input manually.

When this integral output signal MO3 reaches the integral level set by the drive control circuit 2, the drive control circuit 2 outputs the signal to the CCD.
An integral control signal OFG is output to the line sensor 1, and the CCD line sensor 1 thereby completes the integration. That is, the integral control signal OFG is a signal that controls the integral operation of the CCD light receiving section. Further, φT is a signal for transferring the charge generated in the light receiving section to the COD transfer line, S) is a signal for temporarily holding the output signal of the CCD, and O8 is the output signal of the CCD.

Further, RDY is a signal for notifying the CPU (central processing unit) 4 that the AD conversion has been completed, and φ0 is a signal that is the base zero of all timing signals.

In addition, the lens ROM (read-on memory) 5 stores in advance data necessary for focus detection, such as the F number of the lens and conversion coefficients for determining the amount of defocus of the lens from the amount of deviation of t&. It is manually powered by the CPU 4 one after another. Then, the lens drive control circuit 6 operates based on the lens defocus amount calculated by the CPU 4,
Move the lens to the in-focus position.

Next, the operation of the focus detection apparatus configured as described above to which the method of the present invention is applied will be explained with reference to the flowchart shown in FIG.

Each of the pixel columns A and B consists of 64 pixels, and the outputs are sequentially a1, a, . . . a64, b1,
b2...b64, each pixel output that has been A-D converted by the A-D converter 3 is sequentially stored in the RAM (random access memory) of the CPU 4.

Next, 0≦S≦32 F(S)=:11ak+8−b, l・・・・・・(1
)F(S)=, Mi, 1ak-bk+81 """(2
Perform operation 1. The above equation (1) calculates the sum of the differences between each pixel in column A and column B while sequentially shifting column A by one shift with column B as a reference. -1] Equation (2) calculates the sum of the differences between each pixel in column A and column B while sequentially shifting column B by 1 nft based on column A.

Next, F (-32) F (-31)...F
(0) F (1)...F (31) F (3
The minimum value MINF (S) is determined from the 65 correlation values in 2). The shift at this time, B5, is assumed to be M. That is, F
(M) −MIN F (M). It goes without saying that this M is the closest integer value of the deviation to the in-focus point. However, as mentioned above, this alone will result in extremely poor focusing accuracy, so it is necessary to supplement it.

Therefore, complementation is performed by the method of the present invention. That is, when F(M-1)≧F(M+1), when F(M-1)<F(M+1), and when the image shift amount S is determined, then the defocus amount -Ig(S-
S'') is calculated. Here, S' is the amount of image shift when in focus,
K is a coefficient that varies for each lens (S' is usually 0
).

In the above explanation, F(S)=Σ' ``k+s ''l. 1 or F(S
)=Σ"k-bk+s', but it goes without saying that the same effect can also be obtained.

Furthermore, p(s)=kx, 1 飄+s ``k'' or F
Although t1ak-bk+s' is set for (S)=, 'a1ak+s-bIIN+S≦kk for F(S)= may be the number of pixels in column A and column 8.

Finally, we show the simulation results of three types of interpolation methods.

This is the result of determining the difference δ (see FIG. 4) between the amount of relative shift of the limb image on the CCD surface and the shifted claw determined by three types of interpolation methods.

The three types of interpolation methods are as follows: A is the conventional method shown in FIG. 5(a), and B is the conventional method in which the peak value is determined by quadratic curve approximation. C is the method according to the present invention.

171 (XIO-3) MAX (xto-3)M
IN (xto')A 10th place 326 2
B・083 0・9498B 70・
8 100・87 41 fist 90C3・15
10.11 0.3976131:
Average absolute value of error in deviation amount calculation (unit pitch) Paste X: Maximum value of error in deviation amount calculation MIN: Minimum value of error in deviation amount calculation The number of data is 22.

[Effects of the Invention] As described above, the method of the present invention provides a complementation method for a focus detection device that has less variation in errors for all subject brightness distributions and can therefore improve detection accuracy. can do.

[Brief explanation of the drawing]

1(a) and 1(b) are correlation output characteristic diagrams for explaining the interpolation method of the present invention, FIG. 2 is a block configuration diagram of a focus detection device to which the method of the present invention is applied, and FIG. 4 is a flowchart of the method of the present invention, FIG. 4 is a diagram showing the simulation results, and FIG. 5(a) (b)
) is a correlation output characteristic diagram for explaining the conventional interpolation method, and FIGS. 6 and 7 are diagrams showing pixel outputs when pixel columns are shifted. 1... CCD line sensor IA...
...First photoelectric conversion element group 1B...Second photoelectric conversion element group 2...Drive control circuit 3...A-D converter 4....... CPU 5... Lens ROM 6... Lens drive control circuit type 10 (a) (b) 20 3F) 3 [12 Mi 4 Garden O 5 [ (a) (b) Mi 67
A70 Procedural Amendment (Voluntary) July 1986
Month 110 1, Incident Display 1986 Patent Application No. 1243
48 No. 2, Title of the invention, 1. (Supplementary method for υn detection device 3, relationship with the case of the person making the amendment Patent applicant location 2-43-2 Habagaya, Shibuya-ku, Tokyo Name (037) Olympus Optical Industry Co., Ltd. 4, Agent address Address: 52-14 Matsubara 5-chome, Tomodaya-ku, Tokyo
No. "Column for title of invention in specification, column for scope of claims. Column for detailed explanation of invention 1 Column for brief explanation of drawings"<1)
"1. Title of the invention: Interpolation method for a focus detection device" written on page 1 of the specification has been corrected to "1. Title of the invention: Interpolation method for a focus detection device." (2) “Claims” stated on page 1 of the book
amended as shown in the attached sheet. (3) "Complement" written in lines 9 and 12 of page 2 of the specification will be corrected to "suruma" respectively. (4) The words ``r hli completed'' written in lines 1, 7, and 19 of page 4 will be corrected to ``interpolation.'' (5) Change the "minimum value" stated at the beginning of the third line of page 4 to "
Correct it to "minimum value". (6) "Completion" written in lines 1, 10, and 13 of page 5 will be corrected to "interpolation." (7) "Minor" written in the first line of page 6 is corrected to "minor." (8) "Minimum value" written in the third line from the bottom of page 9 of the same is corrected to "minimum value". (9) "Minimum value" written in lines 4 and 5 from the bottom of page 10 of the same Complement”
Correct each to "interpolation". (10) "Completion" written in lines 9, 12, and 15 of page 11 will be corrected to "interpolation." (11) "Completion" written in lines 6 and 3 from the bottom of page 12 will be corrected to "interpolation" respectively. (12) "Complement" written in the 4th line of page 13 of the same is replaced with "
``Tsurima,'' he said, correcting himself. Attachment "2. Claims In a focus detection device that receives an optical image of a subject into a first photoelectric conversion element 7ff and a second photoelectric conversion element group and detects the distance to the subject based on the correlation of the output signals. , the outputs of the first photoelectric conversion element group are sequentially a1, a2゜n, the outputs of the second photoelectric conversion element 7+1 are sequentially b1, a3, etc.
・bn When N+ l S l≦n, a, P are integers, and the 1zJs value of F (S) is F (M), the first image by the first photoelectric conversion element jff and the first The amount of shift S of the second image due to the second photoelectric conversion probe Iff is expressed as F(M
-1) An interpolation method for a focus detection device, characterized in that an interpolation method is performed when ≧F(M+1) or when F(M−1)<F(M+1). ”

Claims (1)

[Scope of Claims] In a focus detection device that receives an optical image of a subject into a first photoelectric conversion element group and a second photoelectric conversion element group, and detects a distance to the subject based on the correlation between the output signals, the first The outputs of the photoelectric conversion element group are a_1, a_2, a
_3...a_n The outputs of the second photoelectric conversion element group are sequentially b_1, b_2, b
When _3...b_n, F(S)=Σ＾N_k_=_α|a_k_+_s-b_
k｜＾P or F(S)=Σ＾N_k_=_α｜a_k−
b_k_+_s|＾PN+|S|≦n, α, and P are integers, and when the minimum value of F(S) is F(M), the first image by the first photoelectric conversion element group and the second photoelectric conversion element group are The amount of shift S of the second image due to the conversion element group is calculated as follows: When F(M-1)≧F(M+1), S=M+[F(M)]/[F(M)+F(M+1)] or F( A method for complementing a focus detection device, characterized in that when M-1)<F(M+1), complementing is performed using S=M-[F(M)]/[F(M-1)+F(M)].