JPS6118912A - Focus detecting device - Google Patents

Abstract

PURPOSE:To decide on the accurate quantity of defocusing of an objective lens to any body by calculating the quantity of evaluation from outputs of sensors which detect two images varying in relative position relation according to a focus state, and detecting a far/close contention with this quantity of evaluation. CONSTITUTION:A field lens 2 is arranged on the same optical axis with a photographic lens 1 for making a focus adjustment, two secondary image forming lenses 3a and 3b are arranged behind the field lens 2, and sensor arrays 4a and 4b for photodetection are arranged further behind it. The sensor arrays 4a and 4b are composed of plural photoelectric converting elements respectively, and when the output of the (i)th photoelectric converting element is denoted as a(i) and b(i) and P is a positive number, the quantity U of evaluation is calculated from an equation. When this quantity U of evaluation is larger than a specific value, a decision on the far/close contention is made.

Description

DETAILED DESCRIPTION OF THE INVENTION (Technical Section e) Unexploded EIQ relates to focus detection devices used in cameras and the like.

(Prior Art) A focus detection device for a camera that divides the pupil of a photographic lens and determines the focal state of the photographic lens by detecting a shift between two images is well known. For example, US Pat. No. 4,185,191 discloses a device that enables the above-mentioned focus determination by arranging a fly-eye lens group on a predetermined imaging plane of a photographing lens. In addition, there are two
A so-called secondary imaging system device that forms two images whose relative positional relationship changes according to the amount of defocus of the photographing lens by arranging two imaging lenses in parallel is known, for example, in Japanese Patent Laid-Open No. 5
It is disclosed in the newsletter No. 5-118019 and Japanese Patent Application Laid-open No. 155351/1983. Although the latter method increases the total length of the optical system, it has the advantage of not requiring a special optical system such as a fly-eye lens group, unlike the former method.

To briefly explain the principle of focus detection in the latter secondary imaging method using Fig. 1, a field lens 2 is placed on the same optical axis as the photographing lens 1 that adjusts the focus, and a field lens 2 is placed behind them. Two secondary imaging lenses 3a and 3b are arranged in parallel, and light receiving sensor rows 4a and 4b are arranged behind them, respectively. Note that 5a and 5b are secondary imaging lenses 5.
It is several seconds provided in the vicinity of a,,1. The field lens 2 forms an image of the exit pupil of the photographic lens 1 approximately on the front surface of the two secondary imaging lenses 51L and 3N+. As a result,
The light beams incident on each of the secondary imaging lenses 5a and 3b correspond to the respective secondary imaging lenses 5tL and 3b on the exit pupil plane of the photographing lens 1, and do not overlap with each other. It will be emitted from an area of Go area. When the aerial image formed near the field lens 2 is re-imaged by the secondary imaging lenses Wa, 5b onto the surfaces of the sensor rows 4m, 4b, the position in the optical axis direction where the aerial image was formed is Based on the difference, the two re-imaged images change their position.
□It will happen.

Figure 2 shows how this phenomenon occurs, and Figure 2 (a
), the sensor rows 4a,
The two images formed on the surface of tb move in opposite directions on the sensor rows (4a, 4b).The image intensity distribution is photoelectrically converted by the sensor rows 4a, 4b, and an electrical processing circuit is used to convert the The focus state is determined by detecting the amount of relative positional deviation between the two images.

As a signal processing method for the two images that have been photoelectrically converted, the amount of deviation (correlation amount) of the two images and the amount of defocus of the photographing lens 1 are
′! There is a known method of calculating the amount of lens extension by displacing one image relative to the other and finding the correlation using the relationship of approximately proportionality. For example, US Pat. No. 4.33! 007, the number of photoelectric conversion elements constituting the sensor rows 4a, 4b is N1, and the output of the first photoelectric conversion element in each sensor row 4a, 4b is a (il, b (i )
When (1=1 to N), the following equation is calculated as the correlation amount.

V (m)=zo(Ha(t)-b(ilt-m)I-l
a(1+1)-b(i-m)I)(S) V (m) obtained from equation (1) is calculated using the following equation
(il-b(1-m) + (2
This is nothing other than the amount of change in U (m) calculated by 1. This U(2))Fi is a measure of the consistency of the two images when the amount of relative displacement is m, and takes the minimum value when the amount of deviation between the two images is the smallest, that is, when they match. Therefore, at this time, the amount of change in U (m) (VCO)) should be 0.

FIG. 3(a) shows a signal a(il) showing the light intensity distribution of the two images.

In the - example of t+ (i), N=24 in this figure.

FIG. 3(b) (Q) is in the above equations (1) and (2); b tr (tn), v(ml(m--T-w
, ), it is clear from Fig. 3(0) that (, V(mo) = o, m.
．． It is possible to detect an image shift amount equivalent to 6 pixels. After this, calculate the defocus amount of the photographing lens 1 from mo and move the lens 1 out to get the focus state.
The two images coincide as shown in Figure 4 (11).
, (as is clear from 01 (U(o)=o, V
(o), = O.

In this way, the signal processing method based on Equation 1 is effective as a method for detecting the amount of deviation. When there is a subject with Itc Id, the above processing method causes the following table inconvenience. Fig. 5 (fL)
is an example of image signals a(1) and b(il) where two objects at different distances exist in the observation field, and in this figure, the area R
Since the distances of the objects in areas R1 and R2 are different,
The two image signals a(i) and b (the amount of deviation of il is m, respectively) are not equal as shown by m2. If the above-mentioned signal processing is performed on such object signals a(i) and b(i), ml (mo< I
When mo is n2, V(mo)=0. When the photographing lens 1 is driven based on this mo, the
t, (b+, (as shown by at, V(o)m
Regardless of the focus judgment o, in reality, the areas R,, R2
The focus will be on a position at an intermediate distance of the subject located within. In the following, we will refer to this type of subject condition as "near and perspective competition."
I will call it .

By the way, a method aimed at resolving the inconvenience caused by the above-mentioned "far-near competition" is disclosed in Japanese Patent Application Laid-open No. 75607/1983. In such a publication, the calculation values Y, Y, similar to the above-mentioned equations (1) and (2) are expressed as Y4 = ΣI a(1) −
11(1) I P(3) y port=Σ(l a(i)-b
Ct+ as month”-1a(t+t)-b(1)IP)
Calculate each relative displacement according to (4). When a subject signal as shown in Fig. 4(a) is obtained, the calculated value Y has multiple extreme values, and the relative displacement amount of each extreme value is the distance between each area of the subject. Accordingly, it is stated that in the case of ``r perspective conflict'', it is possible to focus on an arbitrary subject area based on the above-mentioned (3) 2. However, the calculated value Yi according to equation (3) does not actually give the result described in the publication in the case of "far-near conflict". This means that the calculated value Y obtained by equation (4) is the amount of change in the calculated value Y, and the calculated value Y, as shown in the same publication, crosses zero at only one point.
It is clear that for the near and far competing j signal, the calculated value η also takes an extreme value only at that one point. In addition, the calculated value Y
The reason why , crosses zero at a plurality of points is not due to ``perspective competition'' but due to the periodicity of the subject brightness distribution.

Therefore, even with the signal processing method according to the same publication, r
It was practically impossible to deal with near-far conflict J.

(Purpose) The present invention was made in view of the above circumstances, and its purpose is to accurately defocus the objective lens for any object by reliably detecting near and far conflicts. It is an object of the present invention to provide a focus detection device capable of determining the amount of focus.

(Example) The invention will be explained in detail based on illustrated embodiments.

The basic concept of non-invention is that in the case of table r distance competition j as shown in Fig. 6 (a), when looking at the entire observation field, compared to the normal case, the photographing lens 1 (Fig. 1 Since the coincidence of the two images, whose relative positional relationship changes depending on the focal state of the image (reference), is often low, by evaluating this coincidence, it is possible to determine whether a near-far conflict j is occurring within the observation field. The point is that it can be distinguished. The inventive feature is that this consistency is the calculated value when m = o in the above equation (2), that is, U(0) = (Ia(1)-1)(1)+
This is because we focused on the point that evaluation can be performed using (3).

For example, image signal a(1) as shown in Figure 1f14 (IL),
In the normal case where b(i) is obtained, Figure 84('l)
), (Q), in the calculation value V (m) as defined by the above formula (), V(no)
If Ino, which is
me) = 0.

Even if O, U (ol > 0. Therefore, V
(o) = 0, that is, by comparing the magnitude of the gap (0) when it is determined that the photographing lens 1 is in focus with a predetermined threshold, the current object in the observation field is in a "near and far conflict" state. This means that it is possible to determine whether or not there is.

If it is determined that the object within the observation field is in a state of ``near and far conflict'', a warning may be issued. )
.

The region b(1) is divided into regions I, n, and III so as to partially overlap, and V(m) is calculated in each region using, for example, equation (1). In the case of FIG. 7(a), each area 1,
The number of data (number of photoelectric conversion elements) of II and IH is V2 (=t
2). Further, in order to distinguish the V (m) calculation in this case from the case where the number of data is N (=24), it is written as Vi (m) tVu (m) + 7m (m).

Figure 7 (bl, (0), ((11 digits each v
In the plot of [(m)Ivll(m)IVl(m), the calculation region I contains the image signal of only the region R1 of FIG. 7(-3), so in FIG. 7(b), vI( m!ri =
m5, which is 0, is the amount of deviation between the two images in area R1, and since the image signal of only area R2 is included in the calculation area 111Kt:j, m4, which is Vm(m4) - ○ in FIG. The amount of deviation between the two images is shown. In addition, both R and R2 are partially included in the calculation area (■), so VB(m) in Figure 7 (Q) indicates the amount of deviation affected by local "far-near conflict". There is. Relative displacement town,
m4 is the subject area R1, R2 from the current shooting lens position.
Lens extension amount to focus on (defocus i)
Since it is compatible with K, for example, the closest subject area R1
If you want to focus the photographic lens 1 on an object within the town,
If you want to focus the photographing lens 1 on an object within the object region R2 on the infinity side, select m4 and control the lens drive. This selection can be set in advance or can be made by external operation. In addition, in FIG. 7(a), N discrete image signals are divided into three calculation areas with partial overlap, but the division method is not limited to this method, and other division methods may be used. Needless to say, it is also possible.

Furthermore, when considering the calculation time of the uninvented signal processing method, the number of data is 1 V (ml calculation), which is only 3 times and 4. U
Even if the calculation (0) is included, the calculation time will not exceed the calculation time in the case of a normal calculation with N data.

Next, an uninvented embodiment and operation flow for performing the above-mentioned focus determination will be described with reference to FIGS. 8 and 9.

FIG. 8 is a book showing an example of an uninvented focus detection device, so the companion optical system that forms two images whose relative positional relationship changes depending on the focus state of the photographic lens is, for example, Since it may be the same as that shown in FIG. 1, illustration is omitted. In this figure, 8 is an image signal processing device, for example, a 1-chip microcomputer having an OPU (central processing unit), memory, input/output terminals, etc.

The sensor device 4 includes sensor rows 4a and 4b and a COD (charge-coupled device).Two images are formed on the light receiving surfaces of the sensor rows 4a and 4b by the light beams that have passed through different pupil areas of the photographing lens 1, and the sensor Control signal φ from drive device 5
c, 8H, and OG perform charge accumulation and transfer according to the light intensity distribution of the image. When the image signal processing device 8 gives the start signal EITART to the sensor drive device 5, it sends out the accumulation start signal OG to the sensor device 4 along with the clock pulse φC generated by the signal OLK of the sensor drive device 5#i clock generator 6. . The sensor device 4 starts accumulating two images from this point, and when a predetermined accumulation level is reached, sends an accumulation completion signal to the sensor drive device 5. The sensor driving device 5 sends a photoelectric conversion output transfer signal BH to the sensor device 4 to transfer the accumulated charge from the sensor section to the COD section, and at the same time sends an end signal END to the processing device 8. After this,
In synchronization with the clock φoK from the sensor drive device 5, the sensor device 4 outputs two images of analog photoelectric conversion signals OSt based on the nine charges stored in time series to the A/D converter 7, which performs A/D conversion. The converter 7 performs 8-bit M/D conversion in synchronization with the conversion signal MDO from the sensor drive device 5, and the processing device 8 converts the digital time series signals DO~D7 into DBO~D.
It is input from the B7 terminal and stored in memory sequentially. Processing device 81d2 image shift amount (relative displacement amount) A/D converted photoelectric conversion signal, that is, image signal &(1), b(i) (i:1
~N) according to the flow described below. here,
The definitions of a (i) and b (1) are as described above.

Terminals RM and FM#'i of the processing device 8 are output terminals for driving the motor 12 for moving the photographing lens 1 in the direction of its optical axis.
(abbreviated as "H") means game) 1Oa.

Transistor IIIL, 110 is off, 11b via 10'b.

11a is turned on, 1 lb, 11d and diode 13a.

An electric front brake is applied to the motor 12 at t3b. When both RM and FM are at a low potential (hereinafter abbreviated as @L"), all transistors lla to 11d are turned off, and the motor 12 is electrically open. RM is 1".
, 11a when IFM is r.

1111 is off, l1m), 11G is on, and the motor 12 is energized from right to left in the figure. Also, RM is "L"1'FM is 1H0 (111), 110 is off,
11a.

11 (L is turned on, motor 121C is energized from left to right in the figure, RM is "n", FM is IL
It will be driven in the opposite direction from when ``.

Also, terminals NIF, JIF,? IF is a drive terminal of ID9 for displaying the focus state.

Next, the operation flow of the non-embodiment will be explained in order based on FIG. 9. Note that in this flow, the setting is made so that the closest subject is selected in the event of 1-near and far conflict.

(B1)...First, the control mode is set to 0. Mode 0 means normal signal processing, and as will be described later, mode 1 means "near and far conflict" state, and mode 2 means in focus in "near and far conflict" state.

(S2)...Photoelectric conversion signals of two images a(i), b(i) (i = 1 to N) by pupil division of the photographing lens
is input from the sensor.

(83)...Check the mode.

(84)...If mode 0 is displayed, normal signal processing is performed to V (
m) is calculated.

(83)...V(no) Detects the number of deviations where x=0.

(86)... Compare the absolute value of the deviation amount and the focusing threshold e1.

(S7)...1 mol) 81 Assuming that the lens is out of focus, the photographing lens 1 is driven by the lens extension amount (defocus amount) corresponding to the relative displacement amount mQK, and the process returns to (82).

(B8) If 1m01≦01, it is determined that the focus range is within the normal signal processing, and U(0) is calculated in order to check for “far-near conflict.”

(89)...U (o) is compared with the threshold value e2 for "near and far conflict", and if (Blo)...tr (o) > a2, it is determined to be "near and far conflict" and the control mode is set to 1. Set.

(818)...In the case of U(0)S02, "near and far conflict"
Press to set the focus, display focus, and return to (82).

(811)...Since it was determined that there was a "far-near conflict", the image signal area was divided as shown in FIG. 7(a), the number of data was set to 4, and the correlation amount T ,Vg
(Calculate m.

Each region obtained by dividing the original image signal into 5 parts with partial overlap represents the amount of deviation in the area.

(81!l)...Let m'o be the minimum value among m'1, m'2, and m-.

This means that when the distance conflict is 1M, the photographing lens 1
This is to focus on (1).

(1314) . . . The amount of deviation m'o is compared with the focusing threshold 135 at the time of 1 far/near conflict.

(813)...If I"of>83, it is not in focus, so the lens is driven, but check the mode again, and if it is mode 1, it is in the middle of focusing due to "far and near competition". Thinking of this, I just moved (87) to the camera lens 1.
to drive.

(816)...In mode 2, it was focused once due to "far and near competition" and then became out of focus, so I assumed that the subject had changed and returned to normal signal processing mode 0 ( Drives Sino's lens.

(81υ... If IID'ol≦e5, it is determined that the camera is in focus by the "near and far conflict" signal processing, mode 2 is set, and the focus is displayed in (131B).

In addition, in this flow, e1*e2gestJ arbitrary Kt14 clauses may be possible, and m'1. When selecting m'2 and m'5, the maximum value or the intermediate value may be selected as m6.

In addition, the evaluation t U (0) in non-inventiveness is not limited to formula (3), but for example, by setting P as a positive number and using U (ol = zo 1-1) - true 1llP (6). Also good. Furthermore, the correlation amount V(ml is V(ml= f(1
a(il-b(1+1-m) 1P-la(i+1)-
'b(1-m month') (71, m1n(x,
y), maw(c, y) to indicate the smaller or larger of 2 real numbers -m))
-min(a(i+1), b(1-m))) (a
)V(m)=((max(a(1),b(1+1-m)
)-max(a(100, b(1-m))1 (9
) may be used.

(Effects) As described in detail above, according to the invention, it is possible to reliably determine the "r-far-near conflict" state, and therefore it is possible to perform highly accurate focus determination in any state.

[Brief explanation of the drawing]

Figure 1 is a plan view showing an example of a secondary imaging type focus detection optical system, Figure 2 (IL) * (1)), (0) is 2
g A plan view for explaining the focus detection principle in the optical system in Figure 1, Figure 3 (a) I (b) I (Image signal for a two-dimensional bag object when QIFi is not aligned, based on this image signal Figure 4 shows the evaluation amount U (m) and the correlation amount V (m), respectively. Fig. 5 shows the evaluation quantity (m) and the correlation quantity V (ml) based on (&). (b) j ((1) is the image signal for a three-dimensional object when out of focus , a diagram showing each of the evaluation quantity U(m) and correlation i V (m) based on this image signal,!6 Figure (a
l, (t)l, (0) is an image signal for a three-dimensional object at the time of focusing, and the evaluation 1 based on this image signal is
), a diagram showing each of the correlation amount V (m), ! Figure 7 (a
) * (b) a ('l * (d) is a diagram showing the image signal division method in the invention and the amount of correlation in each division area, and FIG. 8 shows an embodiment of the focus detection device of the invention. The circuit diagram and FIG. 9 are flowcharts showing the signal processing method of the non-embodiment. 1...Photographing lens 4a 4b...Sensor array 8...Image signal processing device 12...Motor C [Yes]

Claims (4)

[Claims] (1) An optical system that forms first and second images whose relative positional relationship changes depending on the focal state of the objective lens, and a plurality of photoelectric conversion elements that form each of the first and second images. A focus detection device that detects the focus state of an objective lens by determining the relative displacement amount of the first and second images based on the output of the sensor. When the output of the first photoelectric conversion element that detects an image is a(i), b(i), and P are positive numbers, U=Σ_i|a(i)−b(i)|＾P A focus detection device characterized by determining an evaluation amount U and determining the reliability of the relative displacement amount based on the evaluation amount when the relative displacement amount becomes less than or equal to a predetermined value. (2) When the evaluation amount is greater than or equal to a predetermined value, the output of the sensor is divided into multiple regions for each of the first and second images, and the relative displacement of the first and second images is calculated for each region. The focus detection device according to claim 1, wherein the focus detection device calculates the amount. (3) Claim (2) characterized in that when the output of the sensor is divided into a plurality of regions for each of the first and second images, the division is performed so that a portion of each region overlaps.
The focus detection device described in . (4) The focus detection device according to claim (3), wherein the threshold value for determining the evaluation amount is adjustable.