JP2910102B2 - Focus detection device - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a focus detection device such as a camera.

B. Conventional technology Conventionally, a re-imaging optical system that separates an optical image of a target object obtained through a photographing lens into a pair of same optical images and re-images the same, and a pair of re-imaged optical images There is known a focus detection device that includes a pair of image sensors that respectively perform photoelectric conversion and outputs the respective signals, and a detection circuit that detects a relative position of the pair of light images based on output signals from these image sensors.

C. Problems to be Solved by the Invention However, in this conventional focus detection device,
When the target object has a periodic pattern, a focus detection result having a high correlation with false focus was obtained, and it was difficult to discriminate from a true focus.

 This will be described in more detail.

For example, as disclosed in Japanese Patent Application Laid-Open No. 60-37513, in a conventional detection device, when the outputs of individual pixels constituting a pair of image sensors are expressed as a ₁ ... An, b ₁ . Assuming that the mutual shift amount of both images is L, the correlation amount C (L) of both images is defined as follows: C (L) is obtained for the number of consecutive shifts L. Here, when a periodic pattern as shown in FIG. 14 is projected on a pair of image sensors, the correlation amount C (L) for the pair of image outputs at that time is as shown in FIG. As is clear from the figure, there are a plurality of image shift amounts Q ₁ , Q ₂ , Q ₃ , Q ₄ . If Q ₂ is selected, it corresponds to the true focusing position, but if another image shift amount is selected, false focusing will occur. Note that in FIG. 14, the asterisks correspond to the same part on the subject. In FIG. 12, the scale at the top represents the pixel shift amount L, and the scale at the bottom is the defocus amount of the corresponding photographing lens.

In general, in a focus detection device using the conventional image shift detection method, an erroneous focus determination can be made for a subject having such a periodic pattern because the correlation is improved even when the image is completely shifted by one cycle. There is.

Further, in the conventional focus detection device, when the defocus amount cannot be obtained by one focus detection operation, it is necessary to scan the photographing lens and perform the focus detection operation again, which may take time for focus detection. Was.

Furthermore, in the conventional focus detection device, if the distance between the axes (base line length) of the pair of re-imaging optical systems is unnecessarily increased in order to increase the detection accuracy, the detectable defocus amount becomes small. Therefore, the distance between the axes of the pair of re-imaging optical systems is determined at a compromise between the detection accuracy and the detectable defocus amount, and a desired detection accuracy may not be obtained.

An object of the present invention is to provide a focus detection device capable of accurately performing focus detection even for a periodic pattern.

D. Means for Solving the Problems The present invention will be described with reference to FIG. 1 which is a diagram corresponding to the claims. The focus detecting device according to the first aspect of the present invention provides a focus detection device having at least three substantially identical light images having parallax. Optical system A for separating and forming an image, and at least 3
At least 3 for photoelectrically converting and outputting each of the three light images
Photoelectric conversion means B, detection means C for detecting relative positions of at least two pairs of optical images based on output signals from these photoelectric conversion means B, and relative positions of each pair obtained by the detection means C Means for comparing the detection results with each other to eliminate a focus detection result having a high correlation associated with false focusing by a periodic pattern, and a signal for focus adjustment based on the focus detection result after the elimination. And a signal forming means E for forming the signal.

According to a second aspect of the present invention, l ₁ and l ₂ are set such that there is no integer ratio between the axial distances l ₁ and l ₂ of at least two pairs of pupils of the imaging optical system according to the _first aspect. It has been decided.

E. Operation In the invention of claim 1, the subject image is separated into at least three substantially identical light images having parallax and projected onto the corresponding photoelectric conversion means B. Based on the output from the photoelectric conversion means B, the detection means C calculates the relative positions of the two pairs of light images obtained by the imaging optical systems having different axial distances. Further, the elimination unit D eliminates a focus detection result having a high correlation associated with false focusing by comparing the two. As a result, a focus detection result indicating true focusing is obtained. In this case, the distance between the axes of the two pupil pairs is defined as in claim 2.
By setting l ₁ and l ₂ , false focusing can be eliminated even more accurately.

F. Embodiment An embodiment of the present invention will be described with reference to FIGS.

In FIG. 2A, the interchangeable photographic lens barrel 10 is shown.
0 receives the driving force from the lens driving device 201 of the body 200 by the coupler 101 and moves the movable lens through the gear train 102.
This is a well-known configuration for moving the 103. In a storage circuit 104 provided in the taking lens barrel 100, the open F value of the taking lens and exit pupil position information are stored. Control unit 202
Read the data.

Part of the light that has passed through the photographing lens is guided to the optical system 210 of the focus detection device via the center semi-transparent portion of the click return mirror 203 and the sub-mirror 204. Here, as shown in FIG. 2B, the focus detection optical system 210 in this embodiment is separated from the optical system Y disposed on the optical axis of the photographing lens by a predetermined distance from the optical axis as in the related art. It has optical systems X and Z arranged symmetrically, respectively. Each optical system Y, X, Z has a field stop 211 that cuts off extra light outside the focus detection area.
, A field lens 212, an aperture plate 213 for determining a pupil of the re-imaging lens, a re-imaging lens 214, and an IC substrate 215 provided with a plurality of image sensors. FIG. 2B is a view of the focus detection optical system 210 viewed from the film surface side.

Each image output related to the optical image formed on each of the image sensors of the optical systems X, Y, and Z is stored in the memory unit 206 via the interface unit 205 in FIG. The arithmetic and control unit 202 calculates a relative image shift amount for a pair of image outputs having different parallaxes by a known method, and based on this, drives the lens driving device 201 by a predetermined amount to achieve focusing,
At this time, the display device 207 is turned on.

3 and 4 show FIG. 2 (b) in more detail. FIG. 3 is a side view of each focus detection optical system Y, X, Z.
4A is a front view of the field stop 211, and FIG.
13 is a front view of the IC board 215. FIG. The field stop 211 has openings 211Y, 211X, 211Z for three optical systems Y, X, Z.
Having. The aperture plate 213 also has apertures (pupils) ya, yb, xa, xb, xc, za, zb, and zc for three optical systems Y, X, and Z. As can be seen from the figure, the three pupils xa, xb, xc are arranged such that their pupil centers are on a straight line. Similarly, IC board 215
Image sensors Ya, Yb, Xa, Xb, Xc for three optical systems Y, X, Z
And Za, Zb, Zc. Note that the optical systems X and Z are arranged symmetrically with respect to the optical axis of the photographing lens, and the optical system Z is the same except for the point symmetrical to the optical system X, so that the description is omitted.

As shown in FIG. 3, the re-imaging lens of the optical system Y has a pair of lenses 214ya and 214yb as in the past, and the re-imaging lens of the optical system X has three lenses 214xa, 214xb and 214xc. The field lens 212 of this embodiment forms a conjugate image of each aperture of the re-imaging lens substantially at the position of the exit pupil position of 100 mm, and the spread α of the detection light beam is about F7.

Furthermore, a focus detection area (a) on the optical axis of the taking lens
The images (b) to (c) are formed on the image sensors Ya and Yb, and the images of the focus detection areas (d) to (e) to (f) have three re-imaging lenses 214xa, 214xb, 214xc and the same. It is formed on the image sensors Xa, Xb, Xc via the three previously placed pupils xa, xb, xc. Reference numerals 215a and 215b in FIG. 3 denote light-shielding plates for preventing stray light from coming in.

Next, the relationship between the exit pupil of the mounted photographing lens and the vignetting of the detected light beam will be described in detail with reference to FIGS.

FIG. 5 is a view for explaining in which range of the exit pupil position the detection light beam of the optical system X is vignetted with respect to the F5.6 lens. FIG. 6 is a view showing the exit pupil of the taking lens by the field lens 212. FIG. 4 is a diagram projected on a pupil position of an aperture stop 213.

 The following can be seen from FIGS. 5 and 6.

For taking lens exit pupil position is in the range of about 100mm of L ₂ are re-imaging lens 214xa, 214xb, either detecting light beam La of 214xc, Lb, Lc also without shielding and thus the image sensor
All three image outputs Xa, Xb and Xc can be used for focus detection.

For taking lens exit pupil position is in the range of L ₁ of about 50mm~80mm are detecting light beam Lc is for shading occurs, the image output of the image sensor Xc is not available. However, since the detection light beams La and Lb do not cause vignetting, the image sensor X
A focus can be detected by detecting an image shift from the image outputs of a and Xb.

For taking lens exit pupil position is in the range of L ₃ of about 120~200mm are detecting light beam La can not be used to cause vignetting. However, since the detection light beams Lb and Lc do not cause vignetting,
Focus detection is possible by detecting an image shift from the image outputs of the image sensors Xb and Xc.

Since vignetting does not always occur in the optical system Y, focus detection can be performed from the image outputs of the image sensors Ya and Yb. That is, in the focus detection optical system Y along the optical axis, focus detection is always performed by the pair of image sensors Ya and Yb.

Here, in the focus detection optical systems X and Z, which image output of which image sensor is used as a pair of image outputs having different parallaxes is determined by the arithmetic and control unit 202 as shown in Table 1 below. You.

In the focus detection optical systems X and Z described above, depending on the exit pupil position PO and the open F value of the photographing lens to be mounted, as shown in Table 1, the conditions under which none of the detected light beams La, Lb, and Lc cause vignetting. A, the detection light flux Lc is vignetting condition B, the detection light flux La is vignetting Condition C is known in advance. For example, the exit pupil position PO is 90 to 109.
If the aperture F value is 9 mm and the open F value is 5.6 or less, condition A, if the exit pupil position is 50 to 59.9 mm and the open F value is more than 2.8 and 5.6 or less, condition B, the exit pupil position is 110 to 129.9 mm and the open F Value 5
If it exceeds 5.6 and is less than 5.6, it is determined as in condition C.

In the case of the condition A, as a method of taking an image pair, as for the optical system X, any of the image sensors Xa and Xb, the image sensors Xb and Xc, and the image sensors Xa and Xc can be used. Regarding the image sensor Za
And Zb, the image sensors Zb and Zc, and the image sensors Za and Zc. Further, in the case of the condition B, as a method of taking an image pair, the optical systems X and Z are paired with the image sensors Xa and Xb and the image sensors Za and Zb. Further, in the case of the condition C, the image sensors Xb and Xc and the image sensors Zb and Zc
Becomes a pair.

In the case of condition A, three choices are possible, but it is advantageous in terms of detection accuracy to use a pair of image sensors Xa and Xc and a pair of Za and Zc having a large base line length. This will be described in detail later.

Next, the focus detection operation will be described with reference to the flowchart of FIG.

In step S1, the arithmetic and control unit 202 reads the open F value and the exit pupil position PO from the memory unit 104 of the taking lens. In step S2, the arithmetic and control unit 202 determines which of the conditions A, B, and C the photographed lens attached roughly corresponds to based on Table 1. Thereby, an image sensor in which vignetting occurs is identified. Next, accumulation of the image sensor is started in step S3, and when a predetermined amount of charge is accumulated, image data is transferred in step S4 and stored in the memory unit 206. Here, the accumulation time is determined by the calculation / control unit 202 and the interface unit 2
05 in a known manner. In this case, since the transfer takes time, it is efficient to omit the transfer of the output of the image sensor that is determined to cause vignetting. For example, when a CCD is used as an image sensor, transfer of only a CCD determined to cause vignetting among three CCDs is not performed.

Next, in step S5, it is determined whether or not vignetting has occurred. If there is vignetting, the process proceeds to step S6, and if there is no vignetting, the process proceeds to step S8.

If there is vignetting, the process proceeds to step S6, focus detection calculation is performed using a pair of image sensor outputs without vignetting, and based on the result, the movable lens 103 is driven in step S6 and displayed on the display device 207. Perform

If there is no vignetting, the procedure proceeds to the procedure after step S8, and an optimal defocus amount is obtained. In the processing after step S8, the defocus amount is roughly calculated as follows.

The aperture of the re-imaging optical system to be used (xa, xb,
Regarding xc), if the interval between the openings of the pair of re-imaging optical systems to be used is called an inter-axis distance (base line length), an image output from the image sensor of the pair of re-imaging optical systems having a long inter-axis distance is used. First, focus detection is performed.

When only one image shift amount having a high correlation is obtained from the focus detection result, the image shift amount or the defocus amount is obtained.

When a highly correlated image shift amount cannot be obtained from the image sensor outputs of the pair of re-imaging optical systems having a long inter-axis distance, focus detection is performed using the image sensor output of the re-imaging optical system having a short inter-axis distance.

If a highly correlated image shift amount is obtained from the focus detection result by the image sensor output of the re-imaging optical system having a short inter-axis distance, the defocus amount is calculated based on the obtained image shift amount.

When a plurality of highly correlated image shift amounts are obtained from the image sensor outputs of a pair of re-imaging optical systems having a long inter-axis distance, they are converted into defocus amounts, and the image of the re-imaging optical system having a short inter-axis distance is obtained. A comparison is made with the defocus amount obtained from the sensor output to discriminate true focus.

 The details will be described below.

In the case of the above condition A, that is, when vignetting does not occur in any of the three re-imaging optical systems, three types of image sensor combinations are possible when detecting a shift between two images. That is, with respect to the optical system X, image shift detection can be performed for each pair of the image sensors Xa and Xb, Xb and Xc, and Xa and Xc. Here, the larger the inter-axis distance is, the smaller the defocus amount corresponding to the predetermined image shift is, so that the focus detection accuracy is high. Conversely, the shorter the inter-axis distance is, the larger the defocus amount can be. There is a characteristic that becomes.

Therefore, first, in step S8, based on the output image data of the image sensors Xa and Xc relating to the re-imaging lenses xa and xc having a long inter-axis distance, the image shift amount Qi (i = 1,2
…). Here, the number of the obtained image shift amounts Qi is stored in I. If only one image shift amount is obtained, step S9
There is affirmative, to convert the image shift amount Q ₁ of the defocus amount in step S10.

If the defocus amount is large, the image shift cannot be detected from the outputs of Xa and Xc, so that I = 0, and step S9
Is rejected, and the re-imaging lenses xa and xb with a shorter inter-axis distance
For the pair of image sensors Xa and Xb (or Xb and Xc) corresponding to, an image shift amount Rj (j = 1, 2,...) Having a good correlation is obtained in the same manner as in step S8 described above. And step S12
To determine I = 0, and if I = 0, the image shift amount obtained from the outputs of the image sensors Xa and Xb in step S13
Converting the R ₁ on the defocus amount Z.

Of course, after first performing the image misregistration operation on a plurality of pairs of (Xa, Xc) and (Xa, Xb) or a plurality of pairs of (Xa, Xc), (Xa, Xb), and (Xb, Xc), When the defocus amount is large, the distance between the axes is short (Xa, Xb) or (Xb,
The magnitude and the direction of the defocus amount may be determined from the image shift amount of the pair Xc).

As described above, for a bright photographing lens, expansion of the defocus determination area or the front / rear focus determination area and expansion of focusing accuracy in the vicinity of focusing are achieved.

The above is the specific procedure of the method described in the above-mentioned items.

FIG. 8 shows an image output when the periodic pattern is projected on the image sensors Xa, Xb, Xc, where the asterisk corresponds to the same part on the subject. FIG. 9 shows the mutual shift amount L with respect to the image output of the image sensors Xa and Xc.
Shows the above-described correlation amount C (L) in the case where is changed.
In FIG. 9, the scale at the top represents the pixel shift amount L, and the scale at the bottom is the defocus amount of the corresponding photographing lens. As is clear from the figure, a plurality of image shift amounts Q ₁ , Q ₂ , Q ₃ , Q ₄ , etc. for which the correlation is judged to be good.
Exists. On the other hand, FIG. 10 shows the same as above with respect to the outputs of the image sensors Xa and Xb, and the scale shown in the upper part of FIG. 9 is different from that of FIG. This is because the distance between the axes for detection (the distance between the pair of re-imaging lenses) is different. When the re-imaging lenses are arranged at equal intervals as shown in FIG. This is because the amount differs twice in the case of FIG. 9 (using xa and xc) and in the case of FIG. 10 (using xa and xb).

As can be seen from the comparison FIG. 9 and FIG. 10, it is possible to eliminate the false focus of Q ₃ and Q ₁ in comparison with the scale of defocus.

The flow of processing for eliminating erroneous detection due to such a periodic pattern will be described with reference to the flowchart of FIG.

When a plurality of correlation positions Qi are obtained for the outputs of the image sensors Xa and Xc in step S8, if the target object has a periodic pattern, the number I of correlation peaks is plural, so that I ≠ 1 and I ≠ 0
Then, the process proceeds to step S11, and an image shift amount Rj having a good correlation with respect to the image data of the image sensors Xa and Xb is calculated. Next, since I ≠ 0, the process proceeds to step S14, and the defocus amounts ZQi and ZRj corresponding to Qi and Rj are calculated. In step S15, in order to eliminate erroneous detection, it is determined whether or not | ZQi−ZRj | <predetermined value, and a ZQi satisfying this condition is selected.

That is, | ZQ ₁ -ZR ₁ |, | ZQ ₁ -ZR ₂ |, | ZQ ₂ -ZR ₁ |, | ZQ ₂ -ZR
_{2 |, | ZQ 3 -ZR 1} |, | ZQ 3 -ZR 2 |, | ZQ 4 -ZR 1 |, | when the computed respectively, in the illustrated _{example, | | ZQ 4 -ZR 2 ZQ} 2 -ZR 1 |, And | ZQ ₄ -Z
Only R ₂ | becomes smaller than the predetermined value. So ZQ ₂ and Z
Q ₄ is selected, ZQi except when there is ZRj substantially the same value as ZQi are removed. Then, in step S16, determines a ZQ ₂ absolute value of the defocus amount is smallest among the ZQi selected as the defocus amount.

According to the configuration described above, false focusing can be eliminated even for a periodic pattern that has been impossible in the past. Note that if the apertures xa, xb, xc of the re-imaging optical system are arranged at equal intervals as described above, the possibility of selecting ZQ ₄ as false focusing remains, but the probability that the lens is in a very defocused position In step S16, there is no problem if the absolute value is large or small by selecting ZQ ₂ having the minimum absolute value. Step S15 is not limited to the above-described method as long as it is a method for excluding the defocus amounts ZQi and ZRi calculated by the detecting means having two different base line lengths except for the defocus amounts that are substantially the same.

Next, a method for completely eliminating false focusing will be described.

FIG. 11 is similar to FIG. 4, but the apertures xa, xb, xc of the re-imaging lens are arranged at non-equidistant lab @ lbc. In this case, when the same periodic pattern as described above is projected on the image sensors Xa, Xc corresponding to the openings xa, xc, a graph of the correlation regarding the image output is as shown in FIG. FIG. 13 is a graph showing the correlation between the image outputs of the image sensors Xa and Xb corresponding to the openings xa and xb. As is clear from the figure, the distance l ₁ between the axes of the openings xa and xc = (lab + lb
When there is no simple integer ratio relationship between c) and the axial distance l ₂ = (lab) between the apertures xa and xb, the defocus amounts ZQi and ZRj match at positions other than the true correlation positions Q ₂ and R _1. I will not do it. Here, as the degree of non-equidistant spacing, the larger axis distance l ₁ , the smaller axis distance l ₂ , Here, the closest integer value is a value obtained by rounding off the decimal part, and || represents an absolute value.

In this case, t> 0.02 is necessary, t> 0.04 is preferable, and t> 0.1 is sufficient.

Thus, the distance between the axes of the re-imaging lenses (xa, xb) and (x
A slight difference in the distance between the axes (b, xc) can reliably eliminate false focusing.

Note that the present embodiment is also applicable to a so-called external light triangle distance measuring device, in which case xa, xb, and xc are lenses that directly image an object on an image sensor rather than re-imaging lenses. .

In the above embodiment, the target object is set to the same 3D only for the focus detection devices of the focus detection areas provided above and below the axis.
The two light images are separated and received by the corresponding image sensors. However, the focus detection device of the focus detection area provided at the position of Y in FIG. it can. Alternatively, the present invention may be applied to a device that does not have an off-axis focus detection device and detects focus only around the optical axis.

In the above description, the number of split pupils is set to three. However, the number may be set to four or more if inconveniences such as a decrease in the pupil area, an increase in the low luminance limit, and an increase in the scale of the apparatus are ignored. However, in the case where the image height is in the range of about 5 to 10 mm in the focus detection area, the number of divided pupils is optimally three.

G. Effects of the Invention According to the invention of claims 1 and 2, while separating the subject image into at least three substantially identical light images having parallax,
Focus detection with a high correlation associated with false focusing by a periodic pattern by detecting the relative positions of two sets of light images each of which is a pair and comparing and calculating the detection results regarding the relative positions of each set. Since the result is excluded, the focus of a subject having a periodic pattern can be detected.

[Brief description of the drawings]

FIG. 1 is a diagram corresponding to claims. FIGS. 2 to 12 illustrate one embodiment. FIG. 2 (a) is a block diagram showing the overall configuration, and FIG. 2 (b) is a front view of the focus detection optical system viewed from the film side. FIGS. 3 and 4 are enlarged views, FIGS. 4 (a) to 4 (c) are front views showing a field stop, an aperture plate, and an IC substrate, respectively. FIG. 5 is a vignetting detection beam for the focus detection optical system X. An optical path diagram to explain the
FIG. 6 is a diagram for explaining a case where the exit pupil of the photographing lens at various positions is projected onto the pupil via a field lens,
FIG. 7 is a flowchart showing a processing sequence of focus detection calculation, FIG. 8 is a diagram for explaining a periodic pattern projected on three image sensors, and FIG. 9 is obtained from a pair of image sensor outputs having a long base line length. FIG. 10 is a graph of the correlation amount C (L) obtained from a pair of image sensor outputs having a short base length. FIG. 11 is a front view of each optical element of the re-imaging optical system in which apertures are arranged at unequal intervals. FIG. 12 is a diagram corresponding to FIG. 9 by the re-imaging optical system of FIG. FIG. 13 is a diagram corresponding to FIG. 10 by the re-imaging optical system of FIG. FIG. 14 is a diagram for explaining a periodic pattern projected on two image sensors. 100: Interchangeable lens, 200: Camera body 202: Operation / control unit, 210: Focus detection optical system 211: Field stop, 212: Field lens 213: Aperture plate, 214: Re-imaging lens 214xa to 214xc: Re-imaging lens 215: Ic substrate, xa to xc: pupil Xa to Xc: image sensor, La to Lc: detection light beam A: imaging optical system, B: photoelectric conversion means C: detection means, D: rejection means E: signal formation means, F: Re-imaging optical system G: Photoelectric conversion means, H: Detection means I: Signal formation means

Claims (2)

(57) [Claims] An imaging optical system for separating and forming an image of a subject into at least three substantially identical light images having parallax, and at least three for photoelectrically converting and outputting the formed at least three light images. Photoelectric conversion means, detection means for detecting the relative positions of at least two pairs of the light images based on the output signals from these photoelectric conversion means, and detection results relating to the relative positions of each pair obtained by the detection means Elimination means for comparing and calculating the focus detection results with each other and eliminating a focus detection result having a high correlation due to false focusing by a periodic pattern, and a signal formation for forming a signal for focus adjustment based on the focus detection result after the exclusion. And a focus detecting device. The apparatus according to the claim 1, the center distance between the pupils pair of at least two pairs of the imaging optical system when a l _1, l _2, integer ratio between l ₁ and l ₂ Wherein the distance between the axes is determined so as not to have a relationship.