JPH021808A - Focal point detecting device - Google Patents

Abstract

PURPOSE:To improve the responsiveness when an image signal is filtering processed by selecting which is used among the filtering means in response to the condition of a photoelectric conversion signal of a luminous flux passed through an objective lens. CONSTITUTION:When a signal processing means 13 instructs a sensor device 10 to start the accumulation of photoelectric conversion output, and when the accumulation is completed, the signal processing means 13 instructs the selecting operation to the selecting means 14. The sensor device 10 receives the luminous flux passed through the plural regions which have the different injection pupils on a photographing lens and forms the image signal. The selecting means 14 fetches the image signal from the sensor device 10 and operates the filtering selection, and either means among no filtering, process a filtering means 11 and a filtering means 12 is selected. The signal processing means 13 detects the focusing condition of the photographing lens by the relative positional relationship of the two image signals which are processed by non-filtering processing or filtering processing.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Application of the Invention) The present invention relates to an improvement in a focus detection device used in a camera or the like.

(Background of the Invention) Conventionally, as one type of focus detection device for a camera, the focus state of the photographic lens is detected by observing the shift between two subject images formed by a beam of light that has passed through the divided pupil area of the photographic lens. What it does is known. For example, U.S. Pat. No. 4,185,191 discloses an apparatus in which a group of prior-eye lenses is arranged on a predetermined image formation plane of a photographic lens, and two images are generated shifted in accordance with the amount of defocus of the photographic lens.

In addition, the aerial images formed on the predetermined imaging plane by two parallel imaging optical systems are guided to two photoelectric conversion element rows, and the relative positional shift of the two images is detected. The following imaging methods are disclosed in Japanese Patent Application Laid-open No. 55-118019 and Japanese Patent Application Laid-open No. 5
It is disclosed in JP 5-155331 and the like. Although this method requires a slightly longer overall length, it has the advantage of not requiring a special optical system.

FIG. 8 shows the optical system of this secondary imaging type focus detection device. A field lens 3 is placed on the same optical axis 2 as the photographing lens 1 whose focus is to be detected. Behind that,
Two secondary imaging lenses 4 are placed at symmetrical positions with respect to the optical axis 2.
a and 4b are arranged. Furthermore, photoelectric conversion element rows 5a and 5b are arranged behind it. Apertures 6a and 6b are provided near the secondary imaging lenses 4a and 4b. The field lens 3 connects the exit pupil of the photographing lens 1 with two secondary imaging lenses 4.
The images are approximately formed on the pupil planes a and 4b. As a result, the light beams incident on the secondary imaging lenses 4a and 4b are emitted from regions of equal area on the exit pupil plane of the photographic lens 1 that correspond to the respective secondary imaging lenses 4a and 4b and do not overlap with each other. It becomes something. When the aerial image formed near the field lens 3 is re-imaged by the secondary imaging lenses 4a and 4b onto the surfaces of the photoelectric conversion element arrays 5a and 5b (it may be out of focus), the optical axis Based on the displacement of the aerial image position in the direction, the two images on the plane of the light beam 1y and the conversion element arrays 5a and 5b change their positions. FIG. 9 shows the displacement of the relative positions of the two images.

Figures 9(b) and () centering on the focused state of Figure 9(a)
As shown in C), the two images formed on the surfaces of the photoelectric conversion element arrays 5a and 5b in the rear focus and front focus respectively move in opposite directions. This image intensity distribution is converted into a photoelectric conversion element array 115.
The focal state of the photographic lens 1 can be known by photoelectrically converting the images by a and 5b and detecting the relative positional shift between the two images using a signal processing circuit.

As a method for processing a photoelectrically converted signal, for example, the method disclosed in US Pat. No. 4,250,376 is known. This means that when the image signals output from the two photoelectric conversion element arrays are respectively A(i) and B(i) (where i=1 to N), for an appropriate constant k in the above example, The correlation amount V is calculated by an analog calculation circuit or digitally, and the direction of movement of the photographing lens is determined based on the sign or negative of the value of the correlation amount V.

Furthermore, the applicant of the present application has previously proposed a method in which the moving direction of the photographing lens is determined by calculating 1"1 or
No. 313) Furthermore, in a focus detection device that determines the in-focus state from the deviation between two images, the relationship between the deviation amount of the two images and defocus -1 is approximately proportional, and one image is detected relative to the other image. A method of calculating the amount of movement of a photographic lens by relatively displacing it is known.

This method has been known for a long time for baseline distance meter type focus detection devices. TTL type focus detection devices are also known from Japanese Patent Application Laid-open No. 56-75607, Japanese Patent Application Laid-Open No. 57-45510, and the like.

For example, by moving the image signal represented by B (i) relative to the image signal represented by A (i) in terms of circuit processing, V (m) = Σ l A (i) - B ( i
ll-+s) lΣ I A (ill) -B
(i-m)...(4) The correlation amount V (m) is repeatedly calculated for each integer value in the range m1≦m≦m2 of the set relative displacement im. FIG. 10 is a diagram in which the value of the correlation iV(m) is plotted against the relative displacement amount m. When the two images match,
Since the correlation ff1V (m) should be 0, the first
In diagram O, there is an image shift amount equivalent to 1.5 bits.

In addition, using equation (2) (or equation (3)), calculate the correlation MV (m) using the following equation for each integer value in the range ml≦m≦m2 of a relative displacement of 1 m, which is set in the same way, and calculate the image The amount of deviation can be determined.

V (m) = Σ win (A (i), B (i+
ni-m) )-Σ sin (A(i+k), B(i-
m) )...(5) However, these calculations can accurately detect a phase shift when only the phases of two identical images are shifted; If there are frequency components higher than the frequency or low frequency components generated by the optical system or shooting conditions, errors will occur in the correlation calculation. In order to remove frequency components that cause these errors, a filtering method is disclosed in US Pat. No. 4,561,749.

In this conventional example, a method for determining the filter to be used when a plurality of filters are used will be explained with reference to FIG. First, photoelectric detection power accumulation (#l) is performed to obtain an image signal. First, a correlation calculation (#2) is performed using a signal that has not been subjected to filtering processing. Next, reliability calculation (#3)
The amount of noise and ghost contained in the image signal is quantified based on the contrast of the images and the degree of coincidence between the two images. Compare this reliability number with the predetermined reliability threshold #4
), if there is sufficient reliability, a focus judgment is made (#5), the lens is driven when out of focus, and release permission is performed when in focus.

If it is determined in #4 that the reliability value is lower than the reliability threshold, filtering process 1 using the first filter (#6
), and the same correlation calculations, reliability calculations, and reliability judgments as in #2 to #4 are performed in #7 to #9. Again, if the reliability value is lower than the reliability threshold, filtering process 2 (#10) is performed using the second filter, and the same correlation calculations, reliability calculations, and reliability judgments as in #2 to #4 are performed. 1
Do it again in 3. Finally, if it is determined in #13 that the reliability value is lower than the reliability threshold, distance measurement is determined to be impossible. Here, filtering process 2 may be performed on the image signal after filtering process 1 is performed, or may be performed on the original image signal without filtering process 1.

In addition to the method of processing by sequentially switching filters as described above, there are methods of switching filters depending on the type of photographic lens or switching based on the previous defocus state.

As explained above, in the conventional focus detection apparatus having a plurality of filters for image signals, filtering selection is performed by sequentially switching filters until a good result is obtained. With this method, filtering processing, correlation calculation, and reliability determination must be repeated many times.

Moreover, due to system constraints or processing methods, there may be cases where it is necessary to start over from accumulating photoelectric detection power.
The responsiveness of the focus detection calculation becomes extremely poor.

Furthermore, the method of switching filters in advance based on the type of photographic lens or previous defocus condition does not take into account the condition of the subject or the effects of noise, ghosts, etc.
Optimal filtering processing may not always be performed.

(Objective of the Invention) An object of the present invention is to solve the above-mentioned problems, improve responsiveness when performing image signal filtering processing, and perform appropriate filtering selection in consideration of noise, ghosts, etc. An object of the present invention is to provide a focus detection device that can perform the following functions.

(Features of the Invention) In order to achieve the above object, the present invention determines which of the filtering means to use or not to use, based on the photoelectric conversion signal (including the image signal) of the light flux passing through the objective lens. The present invention is characterized in that a selection means is provided to make a decision depending on the state, and the filtering selection is first performed using a photoelectric conversion signal (including an image signal) of the light beam that has passed through the objective lens.

(Embodiment of the Invention) FIG. 1 is a block diagram showing the main parts of a focus detection device which is an embodiment of the invention.

Reference numeral 10 denotes a sensor device, which has a plurality of accumulation-type photoelectric conversion element arrays 10a and 10b composed of CCD sensor arrays and the like. The photoelectric conversion element arrays 10a and 10b are the photoelectric conversion element array 5a shown in FIG.

Similarly to 5b, the light beams that have passed through a plurality of different areas of the exit pupil of the photographic lens are received. Reference numeral 11 and 12 denote filtering means that performs filtering processing on the two image signals output by the sensor device 10. Reference numeral 13 denotes a signal processing means that detects the focal state of the photographing lens from the relative positional relationship between the two filtered image signals.

Reference numeral 14 denotes a selection means that determines which of the filtering means 11 and 12 to use, or none of them, depending on the state of the image signal. In order to control accumulation of the photoelectric conversion element arrays 10a and 10b without using image signals, peak signals taken out from the photoelectric conversion element arrays 10a and 10b or output from separately provided photoelectric sensors and filtering control can be used. It is also possible to use a photoelectric conversion signal output from a photoelectric sensor provided exclusively in the sensor device 10 for this purpose. Note that in reality, the filtering means 11
,12. The signal processing means 13 and the selection means 14 are constituted by one microcomputer.

The operation of the embodiment illustrated in Figure 51 will be explained with reference to the flowchart in Figure 2.

At #20, the signal processing means 13 instructs the sensor device 10 to start accumulation, so that the sensor device 10 accumulates the photoelectric conversion outputs of the photoelectric conversion element arrays 10a and 10b. When the accumulation is completed, the signal processing means 13 instructs the selection means 14 to perform a selection operation in #21. The selection means 14 takes in the image signal from the sensor device 10, performs a calculation for filtering selection, and the details of the 0 selection calculation for selecting one of no filtering processing, filtering processing 1, and filtering processing 2 in #22 will be described later. When no filtering processing is selected, the image signal is directly input to the signal processing means 13, and the signal processing means 13 performs a correlation calculation (#
When filtering processing is performed (23), the range of i becomes narrower, and the range of detectable defocus amounts becomes smaller.

Furthermore, focus detection accuracy may deteriorate depending on the state of the image signal, and it may be desirable not to perform filtering processing.

When filtering process 1 is selected, the image signal is input to filtering means 11, and filtering process 1
In the filtering process 1 in which (#24) is performed, the number of pixels used in the filtering calculation formula is small, so the calculation time is relatively short.

When filtering process 2 is selected, the image signal is input to filtering means 12, and filtering process 2 (
In filtering process 2, in which step 25) is performed, the number of pixels used in the filtering calculation formula is large, so the calculation time is relatively long, but the amount of noise and ghosts that can be removed is superior to filtering process 1. After each process, the signal processing means 13 performs a correlation calculation in #23,
Reliability calculation is performed in #26, reliability judgment is performed in #27,
In step #28, the focus is determined and the subsequent processing is performed.

Next, referring to FIG. 3, we will describe which filtering means to select when the image signal is in a certain state.

FIGS. 3(a), 3(b), and 3(d) are image signals that have not been subjected to filtering processing. (a) is a good image signal without noise, ghost, or level difference between the two images. (
b) is an image signal in which a level difference occurs between two images,
(d) is an image signal in which ghosts with opposite inclinations occur between two images. If correlation calculation is performed without filtering, only (a) can accurately detect deviations, while (b) and (d) will cause errors in calculations. For example, if the image signal A (i) For filtering means 11
A' (i)=-A(i-3)/4-A(i-2)/2
-A(i-1)”A(i)+A(isu)/2+A(i+
2)/4...(6) Filtering means 12 A''(i)=-0,0825A(i-5)-0,125
^(i-4)-0,5825A(i-3) 10,125
A(i-1)+1.25A(i)+0.125A(i+
1) -0,5B25A(i-3)-0,125A(i+
4) -0,0825A(i+5)...If configured as shown in (7) and similarly configured for the image signal B (i), the impulse response of the filtering means 11 will be as shown in FIG. 4(a). The impulse response of the filtering means 1z is as shown in FIG. 4(b). Figure 4 (a
) has an antisymmetric type with i and (i-1) as centers. Hereinafter, filtering means whose impulse response is antisymmetric will be referred to as antisymmetric filtering means.

The feature of this filtering means is that it calculates the difference (differentiation) between A(i) and A(i-1) for adjacent pixels, and in order to further increase the certainty, it reduces the weighting for the outside of the adjacent pixels. By calculating the difference between (i + 1) and A (i - 2) and the difference between A (i + 2) and A (i - 3), the DC component can be removed, that is, even if there is a level difference between the two images. It is possible to remove it.

The impulse response in FIG. 4(b) is symmetrical about i. Hereinafter, filtering means whose impulse response is symmetrical will be referred to as symmetrical filtering means. The symmetrical filtering means is particularly designed to exhibit characteristics equivalent to performing antisymmetrical filtering twice. Therefore, the symmetrical filtering means 12 can handle image signals containing ghosts, for which level differences remain even after passing through the antisymmetrical filtering means 11.
works effectively.

For example, when the filtering means 11 processes an image signal including a level difference as shown in FIG. 3(b),
An image signal containing no level difference as shown in FIG. 3(C) can be obtained. However, even if the image signal including a ghost is processed by the filtering means 11 as shown in FIG. 3(d), a level difference remains as shown in FIG. 3(e), and even if correlation calculation is performed, it will not be accurate. Focus detection is not possible. In the case of an image signal as shown in FIG. 3(d), the filtering means 12
When processed by , the result is as shown in FIG. 3(f), and the level difference disappears. The image signal subjected to the symmetric filtering process by the filtering means 12 has some noticeable noise, but the error is smaller than that in focus detection using the image signal in FIGS. 3(d) and 3(e).

That is, in the case of an image signal with no level difference or ghost, no filtering processing is selected, in the case of an image signal with only a level difference, the filtering means 11 is selected, and in the case of an image signal containing a ghost, the filtering means 11 is selected. is filtering means 1
If 2 is selected and the correlation calculation is performed, the detection error will be the smallest.

Next, the operation of the selection means 14 will be explained.

(I) As shown in Figure 3(a), the shapes of the two images are equal and only the phase difference is different. (II) As shown in Figure 3(b), the shapes of the two images are equal, but
In addition to the phase difference due to the focus state, there is a level difference (m) When the level of the B image is high at the left edge of the image signal, and the level of the A image is high at the right edge, as shown in Figure 3 (d). The selection means 14 discriminates between the above three cases.

The degree of coincidence between the two images is used for these determinations. Regarding the degree of coincidence of the z images, first, the amount of deviation between the two images is detected, one of the image signals is shifted by the amount of deviation, and the degree of coincidence is calculated at the image position where the deviation becomes zero.

The amount of deviation can usually be calculated using the above equation (4) or (5). However, when calculating the amount of deviation using this method, filtering selection takes time, and if filtering processing is not performed, the same calculation will be performed again.

Therefore, in order to detect the amount of deviation, the maximum value and minimum value of the image signal are used. This method will be explained with reference to the flowchart in FIG.

In #30, the maximum value A (jmax) and minimum value A (i
1lIin). i w the position showing them
Let ax and i win be, and let the difference between them be Δi. #3
1, calculate the B image in the same way. When the respective extreme values correspond to each other, the difference Δi and the difference Δj become equal as shown in FIG. 6(a) (#32), and the deviation between the two images ffi p
is determined by p = in + ax - jmax (#33
). However, as shown in FIG. 6(b), when the extreme values do not correspond, the difference Δi and the difference Δj are not equal. in this case#
In 34, the maximum value B (j'
mat), and furthermore, the minimum value of image A is 11m.
2t (i lln+Δj
) to find the local maximum value A (i'wax). If either local maximum value exists, the process branches to #36 at #35. If not, in #37, a search is made to see if there is a minimum (fi B (j'ff1in)) near a position (jmax-Δi) that is −Δi away from the maximum value position jrmax of the B image.

Furthermore, a minimum value A (i'
If either of the minimum values is found in #38, the process branches to step 39. If there is no minimum value, it is determined that the degree of coincidence between the two images is very poor, and in #40, the filtering means is used. Select 12. In #36, using the newly found maximum value position Fa j 'saw or i'1Ilax, the deviation M
P=i-layer a-j′ma! or p = i'wa
Calculate using a formula that allows calculation of x − j wax. #3
Similarly, in 9, the deviation amount p is p=i win −j 'w
It is calculated by a computable formula of in or p=i'vertexin-jmin. In this way, the deviation np can be easily determined.

Once the shift amount P is determined, the degree of coincidence between the two images is determined by the degree of coincidence determination value S, , S2 calculated by the following formula.

S, = Σ l A(i) −B(i-p) l
...(8) [S2 = 1 Σ (A(i) -B(i-p)
) l...(9) That is, if the two images match, the -growth judgment values S, , S2 will both become O, and if there is a level difference, the -growth judgment value SL + 32 will both become large. .

When a ghost occurs, only the -growth determination value Sl becomes large, and the -growth determination value S2 becomes a small value.

From this, filtering selection is performed as follows using the -growth determination values S, , S2.

S, +O, 32ki Q=s No filtering process S+=
Large, S2 = large groan filtering means 11 (S+ Ki52
) Sl=large, S2=small, medium filtering means 12 (S+>
32) A certain threshold value is used to determine whether the matching degree determination value S, , S2 is O, large, or small.

As described above, since the optimal filtering is selected first before performing the correlation calculation, there is no need to repeat unnecessary correlation calculations before selecting the optimal filtering means, and responsiveness can be improved. In addition, since filtering selection is determined according to the state of the image signal (or according to the state of the photoelectric conversion signal of the light flux passing through the objective lens), appropriate filtering selection can be made taking into account noise, ghosts, etc. be able to. Therefore, focus detection accuracy can be improved overall.

As another method for determining the deviation ff1p, it is possible to thin out the image signal and perform the calculation to shorten the calculation time. It is publicly known that this method reduces the calculation time of the correlation calculation itself for focus detection (Japanese Patent Application Laid-Open No. 107312/1982).
However, a method for calculating the amount of deviation using an image signal thinned out by 0 using this method as a part of the filtering selection calculation will be described below.

Calculating the differential charge using the procedure of calculating the correlation ffiV (m) in the above equation (4) or (5) for different relative displacement amounts m takes a significantly long calculation time when the number of data N is large. . For example, in order to calculate the correlation amount V(m) of equation (0) for one m, 2N differences in the absolute value sign, 2N sums during integration, and a total of 4N additions and subtractions are required. On the other hand, for the upper and lower limits of m, ml, and m7, approximately ±N/2 can be obtained, so the number of V (m) to be calculated is approximately N. Therefore, the total number of calculations is approximately 4N2 additions and subtractions. If the calculation is performed by reducing the number of data N, for example, the image signal A (i) of the original data as shown in FIG. The number of data used for the image signal A' in FIG. 7(b) is reduced.
(i) is a selection of every other piece of original data,
The image signal A''(i) in FIG. 7(c) is the sum or average of the photoelectric change detection powers of two adjacent pixels.

To write it mathematically, when the image signals of the original data obtained by the photoelectric conversion element arrays 10a and 10b are A(i) and B(i) (however, B(i) is not shown), FIG. In the case of b), the value obtained by the following formula is used as calculation data.

A' (i) = A (2i-1) ・・
・(11)B' (i)=B (2i-1)
...(12) However, i=l to N/2 In the case of FIG. 7(c), equations (13) and (14) are used.

A” (i)=A (2i-1)+A (2i)
...(13) B ″ (i)=B (2i-1)+B (2i)
...(14) In addition, in equations (11) to (14), the conversion from the original data is shifted by 1 bit, and A' (i) = A (2i
), A″(i)=A (2i)+A (2i+1), etc. Of course, Mffi is Mffi.These operations reduce the number of data N by half, so the amount of operation n to find the amount of deviation is approximately Reduced to 1/4.

Using the value of m when the correlation ff1V (m) calculated as described above is closest to O as the deviation amount p, the above-mentioned matching degree determination values Sl and S2 can be calculated. At this time, if the calculations of equations (8) and (9) are also performed using the reduced data, the amount of calculation can be further reduced.

(Modification) In the illustrated embodiment, the filtering means 11 is of an antisymmetric type, and the filtering means 12 is of a symmetric type, but the invention is not limited to this. Both may be of an antisymmetric type, or both Both may be symmetrical. Further, the number of filtering means may be three or more. The filtering means may be an electrical analog filter.

(Effects of the Invention) As explained above, according to the present invention, which of the filtering means to use or not to use is determined depending on the state of the photoelectric conversion signal of the light flux that has passed through the objective lens. Since the selection means is provided and the filtering selection is first performed using the light beam (including the image signal) of the light beam that has passed through the objective lens, the filtering process of the image signal is performed. It is possible to improve responsiveness and to perform appropriate filtering selections that take noise, ghosts, etc. into account.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a flow chart showing the operation of an embodiment of the present invention, FIG. 3 is a waveform diagram showing an image signal, and FIG. 4 is a block diagram showing an embodiment of the present invention. FIG. 5 shows the impulse response of the filtering means according to the embodiment.
6 is a flowchart showing the operation of the selection means according to an embodiment of the present invention, FIG. 6 is a waveform diagram showing the operation of the selection means according to the deviation of the image signal, and FIG. 7 is a waveform diagram showing the operation of the selection means according to the embodiment of the present invention. 8 is a schematic diagram showing the optical system of a conventional focus detection device, FIG. 9 is a diagram illustrating the principle of image shift in a conventional focus detection device, and FIG. FIG. 11 is a flow chart showing the filtering selection operation of the conventional focus detection device. 1...Photographing lens, 3...Field lens, 4a, 4b...Secondary imaging lens, 5
FL, 5b...Photoelectric conversion element array, 6a, 6b.
...Aperture, 10...Sensor device, loa
, lOb... photoelectric conversion element array, 11.12...
... filtering means, 13 ... signal processing means, 14 ... selection means. Figure 1 Figure 2 Focus display on lens drive release O Distance measurement impossible (a) i-3i-2i-1i i+1 i+2i-5i
-4i-3i-2i-1 i Selection Figure 7 Figure 8 Figure 6 (a) jmax jm+n Jmln JmaX Figure 9 V (m) Lens drive release to O

Claims (1)

[Claims] (1) a plurality of photoelectric conversion element arrays that receive light beams that have passed through a plurality of regions with different exit pupils of an objective lens; and a plurality of filtering means that performs filtering processing on image signals obtained by the photoelectric conversion element array; and a signal processing means for detecting a focus state of the objective lens using a filtered image signal, the objective lens determines which of the filtering means to use or not to use. What is claimed is: 1. A focus detection device comprising a selection means for making a decision according to the state of a photoelectric conversion signal of a light beam passing through the focus detection device.