JPS6146805B2 - - Google Patents

Description

[Detailed description of the invention]

The present invention relates to an automatic focus detection device for an optical system.
Conventionally, various automatic focus detection devices have been proposed. For example, in Japanese Patent Application Laid-open No. 15432/1983, it is proposed that if the first and second light beams at two positions away from the optical axis are separated from the light beam immediately after passing through the imaging optical system, the separated first and second light beams will be separated. The two light beams utilize the movement of the imaging optical system depending on the in-focus and out-of-focus states to place the first and second photoelectric element arrays at the fixed imaging positions of the first and second light beams, respectively. Disclosed is a technology in which the outputs of the individual photoelectric elements of the first element array are individually compared with the corresponding outputs of the photoelectric elements of the second element array, and focus detection is performed based on the individual comparison results. ing. However, this technique has the drawback that it is not possible to determine whether the object is in focus or in focus, or to determine the degree of out-of-focus, only by determining whether the object is in-focus or out-of-focus. On the other hand, a device that discriminates between front focus, focus, and back focus, and also detects the degree thereof, requires a mechanical vibrating part, which increases the overall size of the device and increases power consumption. have. Furthermore, Japanese Patent Laid-Open No. 49-90529 discloses that the phase difference between the output signals to a pair of photoelectric element arrays is
A device is disclosed that also detects the back pin and its degree. However, general phase discriminators can detect only when the input signal is a simple waveform signal such as a sine wave or a rectangular wave. However, in the case of a camera focus detection device that detects the focus of an arbitrary subject, the optical images vary greatly in shape, contrast, and spatial frequency components contained therein.
Therefore, the output waveform of the photoelectric element array is not a simple waveform such as a sine wave or a rectangular wave, but is extremely complex. A phase discriminator cannot detect with high accuracy. In this respect, the focus detection devices of cameras that are in practical use today perform so-called correlation calculations (sequentially shifting the time-series signal of one photoelectric element array by one photoelectric element and comparing it with the time-series signal of the other photoelectric element array). This can be exemplified by the fact that the phase difference is detected using an extremely complicated method of calculating the correlation and calculating the phase difference from the shift amount when the correlation is the strongest. In order to solve this problem, the present invention extracts specific spatial frequency components in optical images from the output signals of a pair of photoelectric element arrays and compares the phases thereof. Specifically, when the photoelectric element outputs of the photoelectric element array are f ₁ , ..., f _N in the order of arrangement of the photoelectric elements, these outputs f ₁ , ..., f _N have vector quantities a ₁ e ⁱ 〓 ⁱ , ..., a _N e ⁱ 〓 ⁱ (ψ _i is the vector deviation angle and after multiplying by ψ ₁ < ψ ₂ <...< ψ _N ), these are added to obtain a composite vector representing the specific spatial frequency.

[Formula] is calculated and the phase difference between these composite vectors is detected. This means converting the complicated waveform of an optical image, which has a wide variety of variations, into a specific sine wave (specific spatial frequency component) with a certain amplitude and phase, and then detecting the phase of this sine wave. There is. Therefore, the present invention enables highly accurate focus detection even with a phase comparison means that is relatively simpler than the conventional one. Prior to explaining the present invention, a photoelectric conversion device for optical images, which is one important element of the present invention, will be explained in principle. In FIG. 1, when an optical image with a certain brightness distribution is formed by an imaging optical system on an element array 1 consisting of N photoelectric elements P ₁ to P _N arranged on a plane, each element P ₁ ~P _N generate photoelectric outputs ₁ to _N depending on the intensity of light incident thereon, respectively. The vectorization means 3 serves to vectorize the photoelectric outputs ₁ to _N , that is, to add phase information to each of the outputs, and has N multiplication means 3A ₁ to 3A _N.
Each multiplication means 3A ₁ to 3A _N has a respective associated photoelectric output.
₁ to _N are multiplied by vector multipliers of a ₁ e ^{i 1} to a _N e ^{i i} to generate vector outputs ₁ a ₁ e ^{i 1} to _N a _N e ^{i N.} However, the argument angle of each vector _{is 1} ~ _N is ₁
The addition means 5, which is determined to satisfy < ₂ <...< _N , adds the vector outputs of the multiplication means 3A ₁ to 3A _N and obtains a composite vector thereof.

Hereinafter, the properties of the composite vector I, which is the output of the adding means 5 thus obtained, will be considered. To simplify the explanation, all the absolute values of the vectors a ₁ ^{i 1} to _{a N} ^{i N} to be multiplied are equal and the value is 1, so a _o = 1
(n=1,...N) and the argument _o of the vector multiplied by the output of the n-th element P _o from the leftmost element is _o
=2π×n/N, that is, if the vector sequence e ^{i 1} to ^{e i N} is set to have exactly one period (one rotation in terms of phase),

[Formula] Next, a composite vector I' is obtained when the optical image on the element array 1 is displaced by one element to the left in the arrangement direction of the elements in FIG. At this time, the photoelectric output of the element P _o becomes _o+1 , which is multiplied by the vector e ⁱ² 〓×n/N. For example, the output of P ₁ is ₂ and P _N
The output of will be _N+1 . In addition, this _N+1 is before displacement,
This is the output due to the light intensity that was incident immediately to the right of P _N. accordingly _N+1 in the second term of equation (2) is due to the light that newly entered the array due to the displacement of the optical image, and ₁ is due to the light that has left the array. If the term is negligible with respect to the first term, that is, if _o+1 and ₁ are almost equal, then by comparing equations (1) and (2), the resultant vector I' after displacement is the same as before displacement. It can be seen that e ^-2 〓×1/Ni is added to the composite vector I of . This shows that when the image is displaced by exactly the width of one element in the element arrangement direction, the phase of the composite vector I uniquely increases or decreases by 2π/N depending on the displacement direction. Note that the second term in equation (2) above can be ignored if the light newly entering the array due to the displacement and the light exiting the array are approximately equal, but otherwise. In order to ensure that light images with arbitrary luminance distributions can be sufficiently ignored, the element array 1 may be configured as shown in FIG. 2a.
In the figure, the photoelectric element array 1 is composed of the above-mentioned set of photoelectric elements P ₁ to P _N , that is, one cycle of photoelectric elements,
In the illustrated example, three sets are connected in series, and
One end of the element at a corresponding position in each set is connected to a corresponding one of the N external terminals T ₁ to T _N of the array 1, respectively. With this configuration, the total photoelectric output of the elements P ₁ in each element set is output to the external terminal T ₁ .
The same applies to the other external terminals T ₂ to _TN . Then, the photoelectric outputs appearing at the external terminals T ₁ to _TN may be multiplied by the vectors a ₁ e ^{i 1} to _{a N} e ^{i N} as in FIG. 1. By configuring the array in such a way that several sets of elements for one period are connected in series, the influence on the composite vector I by the output of the photoelectric elements located at both ends of the array is
Compared to a structure consisting of only one period of elements,
Since it is small, the second term in equation (2) above can be ignored. In order to further ensure this,
For example, a density filter may be placed in front of an array consisting of a plurality of element sets, and the light transmittance characteristics of the filter may be determined such that the transmittance near both ends of the array gradually decreases relative to the center of the array.
It is preferable that the light-receiving surfaces of the elements near both ends of the array are gradually reduced compared to the light-receiving surfaces of the elements at the center of the array. As a result, the outputs of the photoelectric elements near both ends of the array 1 become relatively small compared to the outputs of other photoelectric elements, so that the influence on the composite vector I can also be reduced. The present invention utilizes an imaging optical system using at least two devices for converting the photoelectric output of each element of an array into an electrical signal whose phase uniquely changes according to a relative change of an optical image with respect to the array, as described above. This is to detect the focus of the system. A detailed explanation will be given below based on examples. FIG. 3 is a schematic diagram of an embodiment of the automatic focus detection device of the present invention, particularly showing the relationship between the photoelectric element array and the optical system. Among the light beams from an object (not shown) converged by the imaging lens _L0 , the light beams that have passed through two spatially different parts of the lens _L0 are reflected by a pair of mirrors M _a M _b . and separate it from other light beams. In this illustrated example, both mirrors M _a M _b are located at symmetrical positions with respect to the optical axis of the lens L ₀ , and each reflects the light rays incident thereon in a direction perpendicular to the optical axis. As shown in FIG. 2a, a photoelectric element array 1A composed of element sets corresponding to a plurality of periods is in the optical path of the light reflected by the mirror M _a , and is located in the optical path of the light reflected by the mirror M _{a .}
is placed in a plane conjugate to the fixed focal plane of Further, the same photoelectric element array 1B is arranged at the same position in the optical path of the light reflected by the mirror _Mb . Both arrays 1A and 1B are arranged so that when the lens _L0 is focused, images are formed on corresponding elements of both arrays as shown in FIG. 4b. FIG. 4 shows an array in which the number N of elements in one period is 4 and the number of element sets is 3. This array 1
A and 1B respectively process the photoelectric outputs of their outer end output terminals T ₁ to T _N and generate the composite vector I described in FIG.
Phase information containing output generation circuits 10A and 10B that generate electrical signals are connected. This array 1
The vector a _o ⁱ which multiplies the connection relationship between A and the circuit 10A by the output of any photoelectric element P _o of this array.
If the polarization _angle of the vector multiplied by the output of the photoelectric element P _o+1 adjacent to the left of the element P _o is set to be larger than the polarization angle o of ⁿ , that is, in FIG. 1A is connected to elements P ₁ , P ₂ ... from right to left.
...Array 1B and circuit 10 if arranged as P _N
In other words, the arrangement of the elements in FIG. 3 is reversed and the elements P ₁ . . . P _N are arranged from left to right. With this configuration, as shown in the figure, when lens L ₀ is in focus, the image of the object is reflected in both arrays 1 as described above.
Images are formed at the same position on A and 1B (Fig. 4b)
Therefore, each circuit 10A, 10B has the same argument α
Output composite vectors _{I a} and I _b having a and α _b .
If the front focus is achieved from this state, the image of the object will be formed in front of the fixed focal plane.
-15432, the images on arrays 1A and 1b both move leftward in FIG. 3, and conversely move rightward in the case of rear focus. The image changes to the left in Figure 3 due to this front focus.
When the array directions of the elements of both element arrays 1A and 1B are aligned as shown in FIG. 4, the image moves to the right in array 1A and to the left in array 1B, as shown in FIG. At this time, the output vector _I'a of the phase information containing output generation circuit 10A has an argument α'a larger than the argument _α at the time of focusing, and conversely, the argument α of the output vector _I'b of the circuit 10B ′ _b becomes smaller. In the case of a rear pin, the situation is reversed as shown in Figure 4c. Therefore, as shown in FIG.
3, both phase information containing output generation circuits 10
If we measure the phase difference α _a - α _b between the outputs of A and 10B, when α _a - α _b = 0, it is in focus, and α _a -
When α _b > 0, it is front pin, and α _a −α _b < 0
You can tell that it is a rear pin when . The absolute value |α _a −α _b | of this phase difference increases as the focus shifts from the focused state. Therefore, this phase difference measuring device 13
When the output of is connected to the display 15, the front pin,
The distinction and degree of focus and back focus can be displayed. Of course, automatic focusing can be achieved by operating the drive mechanism of the lens L ₀ based on the output of the phase difference measuring device 13, for example. Figure 5 shows that the light beams other than those near the optical axis that have passed through the imaging lens _L0 are collected by two converging lenses L arranged approximately symmetrically with respect to the optical axis behind the fixed focal plane F.
Each lens L is separated into two directions by _a and L _b .
1 placed on the plane conjugate to F with respect to _a and L _b
An example will be shown in which images are formed on a pair of element arrays 1A and 1B. The above figures 3 and 5 show one imaging lens.
This is an example of separating the transmitted light of L ₀ into two, but next, an example of an optical system different from that will be explained with reference to FIG. 6. In FIG. 6, a pair of imaging lenses create an image of the same object S as the lens L whose focus is to be detected.
Mirrors M ₄ and M ₅ are arranged behind each of L ₄ and L ₅ , respectively. This lens is related to L ₅ mirror M ₅
is fixed, but the mirror _M4 associated with the lens _L4 is interlocked with the imaging lens L and swings in the direction of the arrow in accordance with the movement of L in the optical axis direction. Prism PL is a mirror
The reflected light from M ₄ and M ₅ is further transmitted to each photoelectric element array 1.
It is reflected towards A and 1B. Of course, the photoelectric element arrays 1A, 1B are arranged at fixed focal planes of the lenses _L4 , _L5 . The image position of the object S by the lens L ₅ formed on the element array 1B is the lens L ₅
It is uniquely determined by the position of the object S relative to the object S.
On the other hand, since the image position of the object S by the lens _L4 is displaced in the arrangement direction of the elements of the array 1A depending on the rotation angle of the interlocking mirror _M4 , when the imaging lens L is in the focused state, as shown in FIG. When the image of the object S is in the same position for the arrays 1A and 1B as shown in FIG. If the relationship between the lens L, the interlocking mirror, and the element arrays 1A and 1B is determined so that the output changes in the opposite direction to Automatic focus detection of the lens L can be performed. Next, the phase information containing output generation circuit 10 of the present invention
An example of the configuration of A and 10B, that is, the vectorization means and the addition means will be explained. In this first configuration example, each photoelectric element output _o is multiplied by a vector a _o e ⁱ⁴ⁿ , and the vector x, y
component, and then add the outputs of each photoelectric element multiplied by the x component to obtain the x component
Then, the outputs of each photoelectric element multiplied by the y component of each vector are added together to determine the y component Y of the composite vector I. That is, in this configuration example, the composite vector I is calculated in the form of its x and y components X and Y. In FIG. 7, the number of elements constituting one period is four, and the number of periods in the array, that is, the number of element sets, is three. Each photoelectric element P ₁ to _{P 4} of each element set
are the external output terminals T ₁ to _{T 4} of the array as in Figure 2.
are connected to each other. Therefore, the total photoelectric output of P ₄ of each element set is at terminal T ₁ , and the terminal T 1 is also connected to terminal T 1.
A similar photoelectric output appears at T ₂ to _{T 4} as well. This T ₁
~ For each output of T ₄ , vector e ⁱ² 〓 ^×1/2 , e ⁱ² 〓 ^×2/
4 , e ⁱ² 〓 ^×3/4 and e ⁱ² 〓 ^×4/4 are multiplied by the xy components of each vector (cos2π×1/4, sin2π×1/
4) = (0,1), (cos2π×2/4, sin2π×2/4)=
(-1, 0), (cos2π×3/4, sin2π×3/4)=(0,-
1), Multiply by (cos2π×4/4, sin2π×4/4) = (1,0). In the figure, terminals T ₁ to _{T 4} are connected to this x component sequence 0, -
The multipliers 3x ₁ to 3x ₄ for multiplying by 1, 0, 1 are similarly connected to the multipliers 3y 1 to 3y ₄ for multiplying by the y component sequence 1, ₀ , -1, 0, respectively. adder 5x
is the x component X of the composite vector I by summing the outputs of multipliers 3x ₁ to 3x ₄ ;
The outputs of y ₁ to 3y ₄ are summed, and the y component Y of the composite vector I is output, respectively. Such a composite vector I is defined as its components (X,
When output generation circuits containing _{phase information} that calculate in the form of (X
_a , Y _a ), (X _b , Y _b ). The phase difference between I _a and I _b can be calculated from (X _a , Y _a ) (X _b , Y _b ) using various methods, and a particularly preferred calculation method will be described below. In Fig. 8, I _a = (X _a , Y _a ), I _b = (X _b , Y _b ), the absolute value of each vector changes greatly depending on the average brightness of the optical image, so the average brightness X _a , Y _a , X _b , Y _b are connected to arithmetic circuits 20x and 20, respectively, so as not to depend on
Standardized by y, 22x, and 22y, respectively. Convert to These normalized vectors I _a , I _b
cross product of

[Formula] is determined by the arithmetic circuit 24. As is well known, this cross product Q has a correspondence with the angle formed by the two vectors I _a and I _b , so based on the output of this circuit 24 both vectors I
It is possible to know the difference in the declination angles of a _{and Ib} , that is, the size of the phase difference, including its sign, and _therefore , from this Q value, it is possible to know the distinction between front focus, in-focus, and rear focus, and the degree thereof. I can do it. In the above example, one period is made up of four elements, and the absolute values of the vectors to be multiplied are all equal, so if a light image with a uniform brightness distribution is formed on the array, the composite vector I will be zero, which is preferable. However, in order to obtain such a result, for example, one period may be composed of an even number of elements, and the absolute values of vectors shifted by a half period may be set to be equal to each other. Next, a specific circuit example of this first configuration example will be explained with reference to FIG. Since N=4, the photoelectric element array 1A has four external output terminals _T1 to _T4 as specifically shown in FIG. amplified. As mentioned above, the x component sequence by which each of these outputs is multiplied is
0, -1, 0, 1, so to find X _a ,
Is it okay to multiply the outputs of T ₂ and T ₄ by -1 and 1, respectively, and then add them? In this circuit example, T ₂ and T ₄
The outputs of are multiplied by 1 and 1, respectively, and then the difference between the two is calculated. That is, T ₂ and T ₄ are connected to equal resistances via the amplifier 26, multiplied by the above-mentioned x component 1, and then connected to both inputs of the differential amplifier 28n, so that the output of T ₄ after amplification is amplified. The output of T ₂ is subtracted. Also, the y component sequence is 1, 0, -1, 0
Therefore, as _in the _{case of}
Y _a can be obtained by inputting to 8y. The other element array 1B and phase information containing output generation circuit 10B have the same configuration, so _that
Y _b can be obtained. From these (X _a , Y _a ), (X _b , Y _b ), the phase difference between the composite vectors I _a and I _b can be determined from the configuration shown in FIG. Note that the external common terminal T _c of each element array is connected to other terminals of each photoelectric element in each array, and the external common terminal T _c is connected to a power source (not shown). Of course, in the above example, the x and y components other than zero that should be multiplied by the photoelectric output are all equal, so the resistance values are the same, but if the x or y components are different, The value of the resistor must be selected according to the Next, the above-mentioned improved modification will be explained with reference to FIG. In this modification, the photoelectric element is a photodiode, the common terminals _Tc of the photodiode arrays 1A and 1B are connected to each other, and this terminal is connected to the ground.
A FET switching element sw is connected. this
FETsw is turned on and off by the control circuit 30. Therefore, when this FET sw is on, each photodiode in the array 1A instantaneously generates a photoelectric current corresponding to the length of the off time of sw and the intensity of incident light during that time. Since the AC outputs from each photodiode are all converted into AC by a common FETsw, they have the same period and the same phase, and are inputted to the resistors and differential amplifiers 28x and 28y as described above.
The sample-and-hold circuits 32x, 32y rectify the outputs of the differential amplifiers 28x, 28y based on a signal from the control circuit 30 that corresponds to the on/off period of the FETsw, and convert them into DC outputs X _a , Y _a . still,
The control circuit 30 inputs the above-mentioned X _a and Y _a and when the sum of their absolute values is small, that is, when the average brightness of the optical image is low, the control circuit 30 increases the on/off period of FETsw1 according to the degree and turns off the FET. By increasing the time and increasing each photoelectric current, the magnitude of the photoelectric current is kept approximately within a certain range regardless of changes in the average brightness of the optical image. Of course, the photoelectric current from the element array 1B is also converted into X _b and Y _b by the same process. This specific circuit example is this vector I _a = (X _a ,
The circuit for determining the phase difference of Y _a )I _b =(X _b , Y _b ) also has a different configuration from that of FIG. The details are explained below. First, I will explain the principle. Multiplying X _a by AC signal sin wt, Y _a by AC cos wt, and adding both results in √ _a ² + _a ² sin (w ^t +
A signal α _a ) is obtained. Here, α _a is the argument angle of vector I _a . Similarly, cos wt for X _b and sin for Y _b
Multiplying wt and subtracting the latter from the former gives √ _b ² + _b ²
Obtain cos(wt+α _b ). α _b is the declination angle of I _b In this way, the declination angle of the vectors I _{a and} I _b can be expressed as the phase of an AC signal of a certain frequency w, but the declination angles of I _a and I _b are the same. In other words, when focusing, a phase difference of π/4 is given to both AC signals, or in other words, one vector is made into a sine signal and the other is made into a cosine signal. By integrating sin(α _a −α
This is to find _b ). Expressed by the formula, 2/T _p ∫ ^To _p sin (wt + α _a ) × cos (wt + α _b ) dt = sin (α _a − α _b ) Therefore, this integral output becomes zero at the time of focusing, and the detection at the time of focusing is It becomes easier. To explain the above with reference to FIG. 10, the DC outputs X _a and X _b are turned on and off at the same frequency w by the FET switching elements 36x and 38x, and the DC outputs Y _a and Y _b are turned on and off by the same FET switching elements 36
By y and 38y, it is turned on and off at the same frequency w as above, but with a phase shifted by π/4. These four FET36x, 36y, 38x, 38y
is controlled by the output from the controller 34. In this way, X _a and Y _a have the same frequency w, but are converted into rectangular wave signals with a phase shift of π/4, and then,
Added by adder 40. The w bandpass filter 41A extracts the w frequency component from the output of the adder 40. Therefore, the extracted output of this filter 41 is the above-mentioned √ _a ² + _a sin (wt + α _a )
It will contain a lot of. Similarly, for X _b and Y _b , by using the subtracter 42 instead of the adder 40 described above, √ _b ² + _b ² cos (wt+α _b ) can be obtained as the output of the bandpass filter 41B. The waveform shaping circuit 44A converts the output of the bandpass filter 41A into a rectangular wave signal of uniform amplitude while keeping the phase and frequency the same. The waveform shaping circuit 44B similarly converts the output of the bandpass filter 41B into a rectangular wave signal of uniform amplitude. A multiplier 46 multiplies the rectangular wave outputs of both circuits 44A and 44B, and an integral smoothing circuit 47 integrates and smoothes the output. As a result, the output of the integral smoothing circuit 47 is the vector I _a
Since the value is related to the difference in the declination angle between Ib and _Ib , front focus, rear focus, the extent thereof, and in-focus can be detected with zero output, depending on the positive/negative of this output and its magnitude. In the configuration shown in FIG. 10, since the output of the photodiode is converted into AC by a single FET sw1, the spike noise generated when the FET sw1 is turned on is generated equally for each AC photoelectric output. Therefore, those noises can be canceled by the differential amplifiers 28x and 28y. In addition, the luminance distribution of the optical image is output as a time-series electrical signal using a photodiode array or C, C, D, etc. as a photoelectric element array, and this photoelectric output is multiplied by cos wt, integrated, and sin wt is calculated. By multiplying and integrating vector I, its components X, Y
It can be found in the form of That is, X=∫I(t)・cos wt dt, Y=∫I(t) sin
wt dt. Here, I(t) is the time-series photoelectric output. Next, another configuration example of the phase information containing output generation circuit of the present invention will be explained. In FIG. 11, the configuration of the photoelectric element array 1 is exactly the same as that in FIG. 7, and external output terminals T ₁ to _{T 4}
Let g ₁ to _{g 4} be the photoelectric outputs appearing in . These photoelectric outputs g ₁ to _{g 4} are supplied with alternating currents e ^{i (wt+} 〓 ^/2 , e ^{i (wt+ 〓 ) , e i (wt + 3/2} 〓 ⁾ whose frequencies are equal to each other but whose phases are sequentially ^{advanced by 2π/4 .} and e ^{i (wt+2} 〓 ⁾ by the multipliers 51 to 54. That is, the respective photoelectric outputs g ₁ to g ₄ have the same frequency, but the phase sequentially advances by 2π/4 and the amplitude AC output proportional to the magnitude of its photoelectric output
Convert g ₁ e _i(wt+ 〓 ^/ ₂₎ to g ₄ e ^i(wt+2 〓 ⁾ . Adder 55 adds the AC outputs of each multiplier 51 to 54 to generate a composite AC output I _p . Of course, in this example 4
Since one period is made up of one photoelectric element, the phase of each AC output is shifted by 2π/4, but in general, when one period is made of N elements, the phase of each AC output is shifted by 2π × n/N. This becomes the output. Now, in the general case, if we calculate the composite output I _p of the adder 55, we get: As is clear from comparing this equation with equation (1), when the optical image moves relative to the array 1 in the arrangement direction of its elements, the phase of the synthesized output I _p becomes unique. Change. Next, a specific circuit example of this second configuration example and a specific circuit example of the phase comparator 13 will be explained with reference to FIG. In this example, a photodiode is used as a photoelectric element, and one period is composed of six elements, that is, N=6.The specific structure of the pair of element arrays 1A and 1B is shown in FIG. 2c. External output terminals T ₁ to _{T 6} of arrays 1A and 1B
are connected to switching FETs ₁ to _s6 and _s10 to _s60 , respectively. As shown in FIG. 13, these switching elements s ₁ to s ₆ are sequentially turned on for a short time at the same period T by pulses from the timing pulse generator 60, but the timing at which each switch s ₁ to s ₆ is turned on is different. are sequentially delayed in phase by T/6. Similarly, the switching elements s ₁₀ to s ₆₀ related to the array 1B are turned on by the pulse generator 60 at the same period T and with a phase delay of T/6. The relationship between the turning on timings of both switching element pairs s ₁ to s ₆ and s ₁₀ to s ₆₀ is determined such that the turning on of switching element s ₁₀ is delayed by T/4 (π/2 as a phase) than the turning on of switching element s ₁ . It is being Therefore, the external common terminal _Tc of arrays 1A and 1B is
_The photoelectric current _of the photodiode (Fig _{. 2c}
In the array, there are element sets for 3 periods, so 3
(total output of two photodiodes) appear in sequence. This photoelectric current has a peak value that is related to the light intensity and the off time of the switching FET, as described above. A pulse shaping circuit 62 is connected to the common terminal _Tc of the arrays 1A and 1B, which includes an OP amplifier 62a and a feedback capacitor 62b.
and an FET 62c connected in parallel to this and turned on and off by a circuit 60, and shapes the photoelectric current from _Tc into a pulse signal having a constant pulse width and an amplitude related to its peak value. do. Therefore, apart from the pulse amplitude, the output of the shaping circuit 62 has s ₁ , s ₆₀ , s ₂ , s ₁₀ , s ₃ , s ₂₀ , s ₄ , s over time at exactly the timing shown in FIG. ₃₀ ...
A pulse output is generated according to the on state. FETs ₁ ′ to s ₆ ′ connected in series to the output of the shaping circuit 62,
s ₁₀ ′ to s ₆₀ ′ are each synchronized with each of FETs ₁ to s _R and s ₁₀ to s ₆₀ (Sn corresponds to Sn′), and the pulse generator 6
It is turned on and off by 0. Therefore, the photoelectric current generated by turning on _s1 is waveform-shaped and becomes the 13th
The shaping circuit 6 generates pulses with a waveform as shown in Figure _s1 .
However, since only _s1 is on at this time, this pulse output is output from peak hold circuit 6.
4, the peak is held for T/2 time and converted into a rectangular wave of length T/2 as shown in FIG. 14a. When s ₆₀ continues to turn on, the photoelectric current is peak hold circuit 6 because s ₆₀ ' is on at this time.
9, it is similarly converted into a rectangular wave. Furthermore, by turning on s ₂ , s ₁₀ , s ₃ , s ₂₀ ..., the corresponding photoelectric currents are caused by the corresponding FETs ₂ ′, s 2 ′, s 20 , etc.
It is converted into a rectangular wave by peak hold circuits 65, 67, 66, 68, . . . selected according to s ₁₀ ′, s ₃ ′, s _{20 ′} . Furthermore, on this circuit side, the switch element pairs s ₁ and s ₄ , s ₂ and s ₅ , and s ₃ and s ₆ have on-periods that are different from each other by T/2, so the FETs ₁ ′ and s ₄ ′ Connect s ₂ ′ and s ₅ ′ and s ₃ ′ and s ₆ ′ to each other in parallel,
Common peak hold circuits 64, 65, and 66 are used to simplify the configuration. FETs ₁₀ ′～
The same goes for s ₆₀ ′. With this configuration, the peak hold circuit 64 is configured as shown in FIG. 14a.
A rectangular wave caused by turning on s ¹ and a rectangular wave caused by turning on s ₄ appear alternately, and similarly, peak hold circuits 65 and 66 receive signals s 2 and s ₂ as shown in FIGS. 14b and 14c, respectively. Square waves caused by s ₅ , s _6, and s ₃ , respectively, appear alternately. Addition circuit 7
0A adds the outputs of these circuits 64-66 to generate the output shown in FIG. The bandpass filter 72A with a center frequency of 2π/T extracts a frequency component of 2π/T from the added output and outputs a sine wave shown in FIG. This sine wave is the equation (3) stated in Figure 11.

It corresponds to [Formula]. Similarly, the photoelectric output from array 1B is converted into a sine wave by addition circuit 70B and 2π/T bandpass filter 72B. This sine wave is generated at the time of focusing, that is, when a light image is formed on array 1B exactly the same as on array 1A.

[Formula]. In this equation, π/4 is due to the fact that the turn-on of FETs ₁₀ in array 1B is phase-wise delayed by π/4 from the turn-on of FETs ₁ in array 1A. And these two filters 72A,
72B output to the waveform shaping circuit 4 similar to that shown in FIG.
4A and 44B, and then multiplied and smoothed by a multiplier 46 and an integral smoothing circuit 47. When focusing, the output of the integral smoothing circuit 47 becomes zero, and when focusing on the front and back, outputs with different signs are obtained. According to the present invention, by using two photoelectric element arrays, front focus, focus, back focus, and the degree thereof can be detected with high precision from the phase difference between the two outputs caused by the photoelectric outputs of both arrays.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing a photoelectric conversion device for optical images, which is the basis of the present invention, and FIG. 2 is a diagram showing an example of a photoelectric element array used in the photoelectric conversion device of FIG. FIG. 3 is a diagram showing an embodiment of the automatic focus detection device according to the present invention, and FIG.
FIG. 6 is a diagram showing the image formed on the photoelectric element and its output vector in the apparatus shown in the figure; FIG. 5 is a diagram showing the arrangement of the lens and the photoelectric element in another embodiment of the present invention; The figures are diagrams showing the arrangement of the optical system in other embodiments of the present invention, FIG. 7 shows a first configuration example of the output vectorization means, and FIG. 8 shows the arrangement of the two output vectors. FIG. 9 is a diagram showing a specific circuit of the configuration example of FIG. 7, and FIG. 10 is a diagram showing an improved modification of the circuit example of FIG. 9. FIG. 11 is a diagram showing a second configuration example of the phase information containing output generation circuit of the present invention, FIG. 12 is a diagram showing a specific circuit of the configuration example of FIG. 11, and FIG. 13 is a diagram showing the pulses that drive the switching FET in FIG. 12, and FIG. 14 is a diagram showing the output of the peak hold circuit in FIG.
5 is a diagram showing the output of the adder circuit of FIG. 12, and FIG. 16 is a diagram showing the output of the bandpass filter of FIG. 12.

【table】

Claims (1)

[Scope of Claims] 1. A first imaging optical system and a second imaging optical system into which light rays from one subject are incident through first and second paths spaced apart from each other; 1
and first and second photoelectric element arrays are respectively provided at or near the fixed focal plane of the second imaging optical system, and the first and second photoelectric element arrays are arranged on the first and second photoelectric element arrays according to the in-focus and out-of-focus states. In an apparatus for detecting the focal point of an optical system from the relative positional relationship of the first and second images formed by the first and second imaging optical systems moving in the arrangement direction of the photoelectric elements of the array with respect to the respective photoelectric element arrays. , generating a first phase information-containing output that generates a first electrical output that causes a phase change in response to a relative displacement of the first optical image with respect to the first element array based on the output of each photoelectric element of the first element array; and second phase information containing means for generating a second electrical output that produces a phase change in response to a relative displacement of the second optical image with respect to the first array of elements based on the output of each photoelectric element of the second array of elements. output generating means, the first and second electrical outputs are f 1 , the outputs of the photoelectric elements of the photoelectric element array corresponding to the first and second phase information containing generating means, respectively, in the order of arrangement of the photoelectric _elements ; ..., f _N , these outputs f ₁ , ..., f _N have vector quantities a ₁ e ⁱ 〓 ¹ , ...
a _N e ⁱ 〓 ^N (ψ _i is the vector argument angle and ψ ₁ < ψ ₂ <...
... < ψ _N ) and then adding these to calculate a composite vector [Formula], which is generated as an electrical output, and further includes phase comparison means for comparing the phases of the first and second electrical outputs. What is claimed is: 1. A focus detection device comprising: a focus detection device that performs focus detection based on the output of the phase comparison means. 2. The focus detection device according to claim 1, wherein the vector deviation angle ψ _i is 2π×i/N. 3. Each of the phase information-containing output generating means modulates the output of the photoelectric element into an alternating current signal having an amplitude corresponding to the magnitude thereof and a phase corresponding to the arrangement position of the photoelectric element, and adding the respective alternating current signals. A focus detection device according to claim 1, characterized in that: 4. Each of the phase information-containing output generating means multiplies the photoelectric element output f by the x component and the y component of the vector amount, respectively, as the vector amount a _i e ⁱ 〓 ⁱ , adds the multiplication results regarding the x component, and similarly. above y
2. The focus detection device according to claim 1, wherein the x component and the y component of the composite vector I are calculated by adding multiplication results regarding the components. 5. Each of the phase information-containing output generating means modulates the x component and the y component of the composite vector into two AC signals having amplitudes corresponding to the magnitudes of the components and having a phase shift of 90 degrees from each other, The focus detection device according to claim 4, characterized in that both AC signals are added.