JPS5860710A - Focusing device for optical system - Google Patents

Abstract

PURPOSE:To detect the out-of-focus with a high precision, by converting the first image, which is formed by a luminous flux passing through a certain region in or near a pupil, and the second image, which is formed by a luminous flux passing through a different region, in a photoelectric converting device. CONSTITUTION:An optical system is provided with a focusing lens 1, a light shielding plate 2, an image sensor 3, etc. When the optical system is focused, an image Q is formed on the focus plane, namely, the image sensor 3, and images Q and Q' which are formed when an aperture 2a is in an upper position and a lower position respectively are in the same position in the direction vertical to an optical axis O; but when the optical system is not focused, an image Q1 or Q2 is formed in a position other than the focus plane. In this case, the optical system is so constituted that the light intensity distribution of the image is converted to an photoelectric signal corresponding to the quantity of light received by the image sensor 3, and thus, the optical system is focused with a high precision.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention 1c is a method that temporally or spatially splits a part of the light flux passing through the pupil of an optical system such as a microscope or the vicinity of the pupil, and splits it into two parts except when in focus. The present invention relates to a focusing device that detects out-of-focus by utilizing the fact that images produced by two light beams are shifted in a direction perpendicular to the optical axis.

As a conventional focusing device of this type, there is one described in Japanese Patent Publication No. 56-13929, for example, but this is not based on the idea of detecting the lateral deviation of the image itself, but rather The idea was that the observed light intensity at one point changes as a result of this, and when it becomes equal to the observed light intensity at the other conjugate point, it is determined that the state is in focus, so the information of the entire image is valid. There were problems in that it could not be used for many purposes, had poor focusing sensitivity, and was prone to malfunctions. Also, like conventional methods that photoelectrically detect the contrast of the subject image, only the peak value is detected, so it is not possible to distinguish between front and rear focus, and the amount of out-of-focus is also detected. The problem was that it couldn't be done. Furthermore, if there is unevenness in the division of light, the light amounts of the two images will be different, resulting in the problem of incorrect focusing.

In view of the above-mentioned problems, the present invention provides photoelectric conversion of a first image formed by a light beam that has passed through a certain area in or near the pupil and a second image formed by a light beam that has passed through a different area. Receive light on the light receiving surface of the device, converting each light intensity distribution into a photoelectric signal, and calculating the photoelectric signal according to a predetermined evaluation function to detect the relative positional relationship of the two images on the light receiving surface. The purpose of this invention is to provide a focusing device that can detect the direction and amount of out-of-focus with high precision.This will be explained based on an example shown below. The basic optical system of the alignment device is shown. 1 is an imaging lens, and 2 is an aperture 2a that is rotatably arranged around the optical axis 0 at or near the pupil in front of the lens 1 and located inside the pupil. The provided light-shielding plate 3 is an image sensor such as a solid-state image sensor provided at the focus position, and when focusing, the image Q
is formed on the focus plane, that is, the image sensor 3, and the aperture 2a is located above as shown in (a), and (b).
), the images Q and Q are located at exactly the same position in the direction perpendicular to the optical axis 0 as in the case where the aperture 2a is located downward.
' is formed, but when the image is out of focus, that is, when the front focus or the back focus, the image Q1 or Q2 is formed at a position other than the focus plane, so the optical axis is 0 with respect to the image Q on the image sensor 3. Zipped images are formed at positions shifted to opposite sides in the vertical direction, and in case (a) and case (b), Q1
→Q+'Father's image position moves as in Q2→Q2'.

Therefore, by adjusting the image position so that it does not move even when the light shielding plate 2 is rotated, the lens 1 can be focused correctly. In addition, since the moving directions of images Q1 and Q2 are different in the case of front focus and rear focus, if this is detected, it is possible to determine whether the images are front focus or rear focus.
Since the amount of focus shift can be calculated from the amount of shift in -
It is possible to adjust the focus in just one step.

FIG. 2 shows a block diagram of the electric circuit of this focusing device, and 2' is the central axis 2' as shown in FIG.
A is parallel to the optical axis O, can rotate around the axis 2'a, and is bounded by a dividing line 2'b cut from the circumference of a radius R that coincides with the center of the pupil 4, that is, the optical axis 0. It is a light-shielding plate whose inner and outer parts are formed transparent alternately at every rotation angle of 90 degrees, and has the same function as the light-shielding plate 2 shown in FIG.

For example, 3 is a 512-bit image sensor, 15 is a sample and hold circuit, and 6 is an analog-to-digital conversion circuit. 7 is a switch, 8 and 9 are memories, 10 is a shift register, 11 is a shift value setting circuit, 12 is a subtracter, 13
14 is an absolute value circuit, 14 is a memory, 15 is an addition circuit, 16 is a memory, and 17 is a central processing unit, which constitute a central processing circuit. 18 is a display device, 19 is a lens drive circuit, and 20 is a lens drive device. 21 is a light source,
22 is a light receiving element arranged on the opposite side of the light source 21 with respect to the light shielding plate 2'; 23 is a light shielding plate rotation detection circuit; 24 is a start nozzle generating circuit; 25 is a counter; 26 is an image sensor drive circuit; 27 is a master clock, 28 is a counter, 29 is a gate, and 30 is a gate control circuit, which constitute an image information processing circuit.

Since the focusing device according to the present invention is constructed as described above, first we will discuss the operation centered on the central processing circuit.
(See FIG. 3) When the image sensor 3 receives an image formed by the light beam passing through the light beam, the light intensity distribution of the image is converted into a photoelectric signal corresponding to the image. This photoelectric signal is held in a sample and hold circuit 5, then converted into a digital signal in an analog-to-digital conversion circuit 6, and stored in a memory 8 via a switch 7.

Here, the memory 8 is 8-3.・・・－・・・・・・、8−
This is a general term for 512 memories up to 512.

That is, in this example, the photoelectric output from the image sensor 3 is 51
Since two data are obtained, these are sequentially converted from analog to digital and sequentially stored in memory 8-7. ......, 8-
Husband at 512? 'A(1), ・・・・・・・・・, fA
(512). Note that the photoelectric signals may be analog-to-digital converted in parallel without successive analog-to-digital conversion.

Next, the light shielding plate 2' is rotated to a second state, that is, a state in which the light flux passes through the transparent part B of the light shielding plate 2' (see Figure 3), and the light intensity distribution of the image formed by the light flux is determined. Image sensor 3, sample hold circuit 5.
The analog-to-digital conversion circuit 6 and the switch 7 are used to store the signal. At this time, the switch 7 is switched to the memory 9 side, and the analog-to-digital converted photoelectric signal is stored in the memory 9.
-+v...-+ fB(1) at 9-512,...
...Stored as fB (512). At this time,
The light intensity distribution of the stored image is as shown in FIG. 4(a), for example, and since the light intensity distributions in cases A and B are shifted, it can be seen that the image is not in focus.

Next, the amount of deviation between images A and B is calculated using correlation. For example, 128 bits (4th
Paying attention to Figure (b)), calculate the correlation between fA(n) and fB(n) of the images in the case of A and B. For example, the image in case of A is 1
It is fixed at 91 to 319 bits, and the image in case of B is set to 128 to 319 bits.
25.5, 129~256°130~257.・・・・・・
..., 255 to 382, 256 to 383 bits when the correlation is calculated, the amount and direction of out-of-focus can be determined from the amount of image shift in case B when the correlation value reaches its peak. A specific example of the calculation formula is shown in Figure 4 (b
), R(δ)−Σ fA(X)fB(X+δ
)・・・・・・・・・ (1)-192 Or R(δ)=zeABS(fA(X)-fB(X+δ))・
...(2)X==I 02 etc. can be considered. However, calculations are made from δ=-64 to δ-64, and ABS is an absolute value. Then, δ' at which R(δ) has a peak (the maximum value in the case of equation (1) and the minimum value in the case of equation (2)) becomes the image shift amount (FIG. 4(C)).

To explain this in detail using the above equation (2) based on the central processing circuit as an example, first, the central processing unit 17 issues an address designation signal and the address X=192, that is, the memo '17
The data stored in 8-192 is taken into the subtracter 12.

On the other hand, this address designation signal is also input to the shift register 10, where it is shifted by a predetermined value δ (in this case = 64) by the shift value setting circuit 11 to designate the address of the memory 9. Therefore, The contents of the memo ') 964 are taken into the subtractor 12. ? ! t1. The calculator 12 calculates the difference between both data, that is, fA(129)-fB(64). This value is sent to the 6' color pair value circuit 13, converted into an absolute value, and then stored in the memory 14. Then, when a signal indicating that the data has been stored is input to the central processing unit 17, the central processing unit 17 uses the memories 8 and 9 based on this signal.
Shift the address designation signal of l f corresponding to X=193
A<X>-fB<X+δ)1 is calculated in the same manner as above and stored in the memory 14. Thereafter, the data is stored in the memory 14 repeatedly until X=319. At this point, all the contents of the memory 14 are taken into the adder circuit 15 and R(-64
)=,5,2ABS(fA(X)-fB(X-64))
is calculated and stored in memory 't6+. Then,
Based on a command from the central processing unit 17, δ is set to -63, X is changed again from X=192 to 319, R(-63) is calculated in the same manner as above, and it is stored in the memory 16-2. put. Thereafter, in the same manner, n and (64)i are sequentially calculated and stored in the memory 16.

Once the above has been completed/MoIJ16. The contents of ~16.2°, that is, R(-64) ~ R(64) are processed by the central processing unit 1.
7 and compare, and if the minimum value R(δ.) is detected, the image shift between the first state and the second state is δ. It can be seen that it is. Therefore, as shown in Figure 4(c),
and δ. The magnitude of the shift can be used to determine the amount of shift, and the sign can be used to determine the direction of the image shift (whether the image is shifted forward or backward relative to the normal focal position).These signals are input to the lens drive circuit 19 and the lens , drive device 2
If the lens 1 is driven by 0, automatic focusing can be performed. Further, the signal is displayed on the display device 18,
Of course, it is also possible to manually adjust the focus based on the display. Note that the above δ corresponds to each bit in the image sensor 3 and is an integer, but R
(δ) has been calculated, it is possible to obtain a more precise value of δ by performing appropriate curve fitting or the like (FIG. 4(d)). That is, it is possible to detect image shifts finer than the pitch interval of the image sensor.

Next, among the images captured by the image sensor 3, unnecessary images, that is, overlapping images that occur when transitioning from one of the above-mentioned first and second states to the other (this is caused by a certain amount of accumulation during image capture) This is due to the existence of time.The principle of automatically removing ((This is caused by the existence of time) is explained based on the image information processing circuit. Fig. 5 is a time chart, and
The (,a,) indicates a change in the image. First, as the light shielding plate 2I rotates, the amount of light received by the light receiving element 22 changes over time as shown in FIG. 5(b). This light receiving element 2
Based on the output of 2, the light-shielding plate rotation detection circuit 23 detects the pulse I generated every time the brightness changes and the pulse I generated every time the darkness changes, as shown in FIGS. 5(c) and 5(d). Create two types of ξruses of ξruses H that occur in .

Next, the pulse I is input to the start nozzle generating circuit 24 and the counter 25 at the same time. Then, the start value ξ
The pulse generating circuit 24 generates a start pulse ξ as shown in FIG. Based on this signal, the drive circuit 26 drives the image sensor 3 according to the clock pulse (FIG. 5(f)) generated by the master clock 27, and sequentially outputs the photoelectric output stored in the image sensor 3. Read sample hold circuit 5
However, this sample-and-hold circuit 5 holds the transferred data only when the sample-and-hold signal (Fig. 5 (g)), which will be described later, is input, and otherwise discards it without holding it. I end up.

Therefore, since the sample and hold ξ pulse is input at this time, this denta is not held, but this is an unnecessary photoelectric signal accumulated when changing from the first state to the second state. On the other hand, the counter 25 counts the clock pulses of the master clock 27 for a predetermined image accumulation time due to the input of the noggles I, and the master clock pulse ξ reaches a predetermined number (at least 512 in this example). Give a signal when This signal is simultaneously input to the start pulse generating circuit 24 and the counter 28.

Since the start noise generating circuit 24 operates in the same manner as when receiving the signal from the detection circuit 23, data from the image sensor 3 is sequentially transferred to the sample hold circuit 5 as described above.

On the other hand, the counter 28 starts counting the master clock signal ξ based on the signal from the counter 25, and sends the sample and hold signal to the sample and hold circuit 5 at every count. Since the timing of supplying this sample and hold ξ pulse and the supply of data from the image sensor 3 exactly match, all data from the image sensor 3 is held in the sample and hold circuit 5.

The count number of this counter 28 is the image sensor 3
The number of pixels is the same as the number of pixels, which is 512 in this embodiment.

The data held by the sample hold circuit 5 is converted into a digital signal by the analog-to-digital conversion circuit 6.
Based on the principle described above, the data is taken into the central processing unit 17 and used for later processing. Next, when a pulse ξ is output from the detection circuit 23, the start pulse generation circuit 24 operates again.
Thereafter, data is taken into the central processing unit 17 in the same manner as described above. As is clear from the above, the detection circuit 2
The data obtained immediately after operation 3 is discarded, and the data obtained next is bolded. Therefore, the image signal accumulated when the photoelectric force obtained from the image sensor 3, that is, the transparent part of the light-shielding plate 2' changes from A to B or from B to A immediately after the detection circuit 23 operates, is divided into two images. This is inappropriate for focus detection and is discarded.In contrast, image accumulation starts when pulse I is generated and ends after a predetermined time measured by counter 25. The resulting image signal is an image signal that has only one of the state B and the state B, and is suitable for focus detection, so it is saved and not discarded. In this way, only necessary image information can be saved.

Note that in this embodiment, focus detection cannot be performed unless information on the states of A and B is received bare (information on two images is required to detect image deviation). be.).

Furthermore, it would be a problem if image information were automatically sent to the central processing unit 17 even when it is not needed, so it is necessary to prevent this from happening. For this purpose, the above-mentioned norm ξ is used. That is, the gate control circuit 30
When the read signal from the central processing unit 17 is inputted to the input signal ξ, and later the signal ξ is inputted from the detection device 23, an output is output and the gate 29 is turned on to maintain that state. Therefore, in this state, the operations from the generation of the start noculus to the generation of the sample hold pulse are performed as described above, and B.
Image information is sent to the central processing unit 17 one after another in the order of A, B, A, . . . . On the other hand, central processing unit 1
7, when the image information is no longer needed, the central processing unit 17 sends a signal to the gate control circuit 30, and the gate control circuit 3
0 turns off the gate 29 and maintains that state when the signal ξ is input. Therefore, in this state, no sample-and-hold pulse is sent to the sample-and-hold circuit 5, so no data enters the central processing unit 17. If you do the above, B, A, etc.
, B, and A, and the central processing unit 17 always receives the same number of image information for B and A in sequence, so the image information for B and A is paired. By using it as a lens, focus detection becomes possible.

FIG. 6 shows an optical system that is slightly different from the optical system shown in FIG. 1, and 31 is a prism that is provided near the pupil and plays a role corresponding to the light shielding plate 2 in FIG. According to this, the images formed in the device (2) when in focus are formed at positions P1 and P2, and when out of focus, the images are formed at positions p, /, p2/ or positions p, l/. ,
Images are formed on p and #. This is the position p when in focus, p
The optical axis of the image of 0. , 02 as a standard, it can be seen that the positions P, l, P2I or the positions p, II, p2# are laterally shifted in the direction perpendicular to the optical axis 01°02, respectively.

That is, the principle is the same as the optical system shown in FIG.
While the method shown in the figure observes the image shift as vibration in a time-division manner, the method shown in FIG. 6 can be thought of as a space-division method in which focusing is observed at two locations. Even with the method shown in Fig. 6, if the image positions p,, p2 are close, two images can be taken with - image sensors, but the periphery of the image formed at position P overlaps with the image at position P2. Therefore, it is a good idea to set up a field diaphragm and narrow down the field of view in advance so that it does not become too narrow. If you compare the time-division method and the space-division method, the time-division method forms images in the same place, so the point where the images overlap is considered to be in focus, and the criterion for determining in-focus or out-of-focus is Although it has the advantage of being easy to determine, it also has the disadvantage of requiring a mechanically moving mechanism such as a rotating light-shielding plate, whereas the space division method has the advantage of not requiring a mechanically moving varnish structure, but has the same disadvantage. The disadvantage is that it is difficult to determine the criteria for determining in-focus and out-of-focus conditions because the location (l) and image cannot be formed.

FIG. 7 shows an optical system of a microphotograph apparatus using this focusing device, where 41 is a light source, 42 is a collector lens, and 43 is a pupil position S of the optical system. 44 is a condenser lens, 45 is a specimen, 46 is an aperture stop of the illumination system placed in
is an objective lens, 47 is an eyepiece lens, 48 is a beam splitter, 149 is a film surface, 50 is a relay lens, 51
52 is a focusing mirror, 52 is a viewer, and each pupil position S
. ,S.

, S2, a light shielding plate 2 shown in FIG.
' is placed.

Note that the above-mentioned image information processing circuit can be applied not only to focusing but also to various other applications. For example, when performing spectrophotometry with two beams using an image sensor, it can also be used to remove unnecessary information when the standard beam and sample beam are input alternately using Chotsuno ξ-, etc. I can do it. or,
It can also be used when transferring movie film to a VTR.

That is, movie film has black partitions between pictures, and the data resulting from this can be discarded as unnecessary data obtained when moving from one screen to another.

As described above, the optical system focusing device according to the present invention uses an image sensor to detect focus by detecting output changes accompanying movement of images using a plurality of light receiving elements. Since the lateral shift of the image itself is detected by obtaining information about the entire image using -, it is possible to detect focus more accurately and with high sensitivity.

Furthermore, the amount and direction of out-of-focus can be detected just by capturing an image once. Furthermore, since the correlation value can be calculated from the image, the amount of deviation can be calculated with finer accuracy than the pitch of the image sensor. Furthermore, since focus detection is performed by calculating the correlation between the two images, even if the light meters of the two images are slightly different, it can be corrected and erroneous focusing will not occur. Furthermore, since the light intensity distribution of the entire image is obtained as a photoelectric signal, the brightness of the image can be determined by using their sum or average value.
Based on this, it is also possible to perform the necessary photometry for exposure control at the same time.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the basic optical system of an embodiment of the focusing device according to the present invention, FIG. 2 is a block diagram of the electric circuit of the above embodiment, and FIG. 3 is a front view of the light shielding plate of the above embodiment. Fourth
The figure is a graph showing the light intensity distribution and correlation on the image sensor of the above embodiment, Fig. 5 is a time chart of the above embodiment, Fig. 6 is a diagram showing another basic optical system, and Fig. 7 is a graph showing the above implementation. 1 is a diagram illustrating an optical system of an example microphotographic device; FIG. DESCRIPTION OF SYMBOLS 1... Imaging lens, 2'... Light shielding plate, 3... Image sensor-14... Pupil, 5... Sample hold circuit, 6... Analog-digital conversion circuit, 7...・
Switch, 8.9... Memory, 10... Shift register, 11... Shift value setting circuit, 12... Subtractor,
13...Absolute value circuit, 14...Memory, 15...
Addition circuit, 16... memory, 17... central processing unit,
18... Display device, 19... Lens 1 drive circuit, 2
0... Lens drive device, 21... Light source, 22...
Light receiving element, 23... Light shielding plate rotation detection circuit, 24...
Start noise ξ pulse generation circuit, 25...carunter,
26... Image sensor drive circuit, 27... Master clock, 28... Counter, 29... Gate, 30... Dart control circuit. Figure 4 Figure 16 Figure 5 (c) Pal entry I (9) Su〉F5 Rehold no V Rusuoff diagram procedural amendment (voluntary) December 22, 1980 Director General of the Japan Patent Office 1, Patent application for indication of the case Publication No. 56-159720 Publication No. λ Name of the invention Optical system focusing device 3, person who makes correction Patent applicant 43-2-4, Hatagaya 2, Shibuya-ku, Tokyo, Agent Address: 105 Shin, Minato-ku, Tokyo! 85-19 6. Contents of amendment (1) Letter of specification, page 4, lines 2-6, “Image Q, father is Q2
is...formed," is corrected to the following sentence. ``The image Q is formed at a position other than the focus plane, that is, at a position other than the image sensor 3.Therefore, a blurred image Q1 or Q2 is formed on the image sensor 3 in the direction perpendicular to the optical axis 0 with respect to the image Q. (2) Change "191-319" on page 7, line 16 of the specification letter to "
192-319”. (3) [Memory 9-64J] on page 9, line 1 of the specification
Corrected to memory 9-128Jl. (4) r fA (129) on page 9, line 3 of the specification letter
−fB(64) J is “fA(192) −fB(128
)” is corrected. (5) Change “Memory 14” on page 9, line 12 of the specification to “
Memory 14" is corrected. (6) [... is placed in the first line of page 18 of the detailed letter]. ], insert the following sentence. ``Figure 7 has a configuration that can be observed with the naked eye, but if the image sensor 3 is placed at the position of the focusing mirror 51, the present focusing device can be used.''

Claims (1)

[Claims] The light-receiving surface of the photoelectric conversion device is disposed on or near the regular image-forming surface of the image-forming optical system, and the first light beam formed by the light beam that has passed through the pupil of the image-forming optical system and a certain area in the vicinity thereof is provided. and a second image formed by a light flux that has passed through another area different from the certain area are received by the light receiving surface of the photoelectric conversion device, and each light intensity variation distribution is converted into a photoelectric signal. , these photoelectric signals are calculated according to a predetermined evaluation function to detect the relative positional relationship between the first image and the second image on the light receiving surface, and determine the imaging state by the optical system. Optical focusing device.