JPH087319B2 - Automatic focus detector for camera - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an automatic focus detection device for a camera, which detects the focus state of an image pickup lens by receiving subject light that has passed through the image pickup lens of the camera to perform focus detection.

2. Description of the Related Art The subject light fluxes that have passed through the first and second regions of the imaging lens, which are in a symmetric relation to each other with respect to the optical axis, are re-imaged to form two images. Focus detection that obtains the positional relationship and obtains the amount of deviation of the image formation position from the planned focus position and its direction (whether the image formation position is on the front side or the rear side of the planned focus position, that is, whether it is the front focus or the rear focus). The device has already been proposed. The optical system of such a focus detection device has a structure as shown in FIG. 8, and this optical system has a planned focal plane (4) behind the imaging lens (2).
Alternatively, a condenser lens (6) is provided further rearward from this surface, and reimaging lenses (8), (10) are further provided behind the condenser lens (6), and the respective reimaging lenses (8), (10) are provided.
The image plane of the image sensor (12) such as CCD,
(14) is arranged. Each image sensor (12), (14)
As shown in FIG. 9, the upper image is close to the optical axis (1) in the case of a so-called front focus in which the images (9) and (11) of the object to be focused are formed in front of the planned focal plane. In the case of the rear pins, on the contrary, they become far from the optical axis (1). When in focus, the distance between two corresponding points of the two images (9) and (11) is a specific distance proposed by the configuration of the optical system of the focus detection device. Therefore, in principle, the focus state can be known by detecting the interval between two corresponding points of the two images.

The distance between the two corresponding points is the image sensor (1
2), the correlation is obtained for the two light distributions on (14), and the positions of the light receiving elements forming the image sensors (12), (14) are relatively shifted to obtain the shift position (that is, the best correlation). , The distance between the two images). The principle of this detection is described in detail, for example, in JP-A-60-4914.

In an automatic focus adjustment device for a camera that incorporates this type of focus detection optical system, the CCD image sensor integrates the light amount of the object, the focus state detection calculation (defocus amount calculation) using the CCD image sensor output, and the defocus amount A sequence of driving the lens in accordance with the operation and stopping at the focus position (shutter release ... When the shutter button is pressed) is program-controlled by a control circuit including a microcomputer.

Then, this automatic focus adjustment device continuously performs the above-described sequential automatic focus adjustment control even when the subject image is near the in-focus position, so that the in-focus position can be finally set accurately. Perform continuous autofocus (AF) on.

Problems to be Solved by the Invention When focus detection is performed, image sensors (12), (1)
4) The distance between the upper images must be detected. Therefore,
A correlation is obtained for two light distributions on the image sensors (12) and (14). By shifting the positions of the light receiving elements constituting both image sensors (12) and (14) relatively, a shift position (ie, an interval between two images) at which the best correlation is obtained is obtained.

By the way, this correlation calculation requires a considerable amount of time because the positions are relatively displaced in a wide range. For this reason, there is a problem that the responsiveness and the object tracking property of the automatic focusing device are deteriorated.

Now, to consider the tracking ability of the subject, the defocus amount is detected by one distance measurement when the subject is approaching or moving away from the camera with the above-described automatic focus adjustment device. When the imaging lens is moved to the in-focus position based on this defocus amount, the subject is moving during that time, so the subject is no longer in focus.

Figure 10 shows the situation. The horizontal axis is the time axis, and the vertical axis is the defocus amount on the film surface. First
In FIG. 10, a curved line 1 indicates the degree of increase in the defocus amount on the film surface when the subject approaches at a constant speed, and a straight line m traces the position where the image pickup lens is about to form an image. It is a thing. The product of the time axis is CC
D represents the integration time, and “act” represents the calculation time of the defocus amount. Since the subject approaches even during the calculation time, even if the lens is driven based on the calculation result, the subject has already moved when the lens is stopped. As the subject approaches, the amount of defocus grows larger and larger, leaving the depth of field immediately and becoming out of focus.

Therefore, it is necessary to shorten the calculation time of the defocus amount as much as possible.

In Japanese Patent Laid-Open No. 59-126517, the detection zone of the image sensor is divided into three blocks, the entire area of each block is shifted, the best correlation position in each block is determined, and the best correlation position of each block is calculated. Among them, the method of adopting the value with the highest degree of correlation has been proposed. However, this method ultimately results in calculation over the entire correlation range for each focus detection, and does not provide an effective time reduction.

In Japanese Patent Application Laid-Open No. 56-75607, the maximum shift amount is adjusted using the first shift amount because the shift amount in the second and subsequent focus detections does not become larger than the shift amount in the first focus detection. Although a method has been proposed,
When the shift amount is large, the calculation still needs to be performed over a wide range, which requires time for the calculation, and the shift amount must be changed every time, which complicates the control. In addition, it is not possible to deal with the case where focus detection becomes impossible, and continuous automatic focus adjustment may not be possible.

An object of the present invention is to provide a focus detection device for a camera that shortens the time required for focus detection calculation.

Means for Solving the Problems In order to achieve the above object, the present invention is to detect a focus state with respect to a subject in a plurality of areas in a photographing screen, and to output focus detection data for each area; Selection means for selecting the area based on a plurality of focus detection data output from the focus detection means, and focus adjustment means for performing focus adjustment based on the focus detection data corresponding to the area selected by the selection means ,
An automatic focus detecting device for a camera, comprising: a repeating unit that repeats the operations of the detecting unit, the selecting unit, and the focus adjusting unit, and a storage unit that stores an area selected last time; The focus state is detected for the previously selected area, the reliability of the result is determined, and if it is determined that the area is reliable, the focus detection operation in other areas is omitted and that area is selected. If it is determined that the focus detection operation in the other area is performed, the control means for selecting the other area is provided.

Action Once a subject is detected in an area, it is highly likely that it will be detected in that area thereafter. Therefore, the present invention first detects the previously selected area, if the detection result is not reliable, detects other areas,
If the detection result is reliable, the detection operation in the outside area is omitted. This shortens the time required for the focus detection operation.

Embodiment An embodiment of the present invention will be described below with reference to the accompanying drawings.

FIG. 1 shows the circuit configuration of an automatic focus detection device for a camera according to the present invention.

FIG. 1 is a circuit diagram of a focus detection device in the case where a line sensor (15) made of a CCD as shown in FIG. 2 is used as a light receiving element and an automatic focus adjustment device using the same. (20) is a photoelectric conversion circuit including the above line sensor (15) and a light receiving element for monitoring, and a shift pulse (SH),
Transfer clock (φ ₁ ), (φ ₂ ), clear pulse (IC
G) is input, and the time-sequential pixel signal (OS), monitor output (AGCOS), and reference voltage output (DOS) are output. Here, the clear pulse (ICG) is a pulse for setting each pixel in the line sensor (15) to the initial state, whereby each pixel in the line sensor (15) discharges the accumulated charge, and a new light integration, That is, charge accumulation is started.
Also, with this pulse, the integration of the output of the monitor light receiving element is started in the photoelectric conversion circuit (20), and the monitor output (AG
COS) changes over time with respect to the reference voltage output (DOS) at a speed according to the brightness of the object. Shift pulse (S
H) is a pulse for shifting the accumulated charge from the pixel unit of the line sensor (15) to the shift register unit. When this pulse is input, the light integration in the pixel unit ends. The transfer clocks (φ ₁ ) and (φ ₂ ) are pulses 180 ° out of phase with each other for sequentially outputting the accumulated charges shifted in the shift register unit from the shift register unit in time series.
The accumulated charge thus output is converted into a negative voltage signal in the photoelectric conversion circuit (20) and output as a pixel signal (OS).

(22) is reference voltage output (DOS) from each pixel signal (OS)
Of the pixel signal (DOS ′) output from the subtraction circuit (22), and a pixel circuit (DOS) that outputs a pixel signal (DOS ′) as a positive voltage signal. In FIG. 2, a peak hold circuit that peak-holds pixel signals corresponding to (a few pixels further left than (l ₁ )) and outputs a voltage (VP) corresponding to the maximum value of those pixel signals, (26) Is the pixel signal (DO
S ') to the output voltage (VP) of the peak hold circuit (24)
Is a variable gain amplifier that amplifies the pixel signal by subtracting the dark current component. The dark current component included in each pixel signal (DOS ′) is removed by the subtraction in the amplifier circuit (26). (28) is an A / conversion circuit that converts the amplified pixel output (DOS ″) from the amplification circuit (26) into a digital value of a predetermined bit, the output of which is called a microcomputer (30) and hereinafter a microcomputer.) (32) is a monitor output (AG
When the change amount of the COS) with respect to the reference output (DOS) is detected, and the change amount reaches the predetermined threshold value within a predetermined time from the start of the change of the monitor output (when it is bright), it notifies the microcomputer (30) to that effect. Output signal (TINT) and amplifier (26)
The gain signal for setting the gain of "1" is output. Further, when a predetermined time has elapsed from the start of the monitor output (AGCOS) output, the forced shift signal (SHM) output from the microcomputer (30) is output to the gain control circuit (32). 32) shows that the gain of the amplifier (26) is "1 time" or "2" according to the amount of change of the monitor output (AGCOS) with respect to the reference voltage output (DOS) at the time of inputting the signal (SHM).
Outputs a gain signal that is set to "times", "4 times", or "8 times". In this case, the smaller the amount of change, the greater the gain that is set. (AN) and (OR) are AND circuits, respectively.
It is an OR circuit, and the above-mentioned signal (TINT) from the gain control circuit (32) and the signal (SHEN) from the microcomputer (30) are input to the AND circuit (AN), and the AND circuit is input to the OR circuit (OR). The output signal of (AN) and the above signal (SHM) from the microcomputer (30) are input. Here, the signal (SHEN) from the microcomputer (30) is a signal for permitting the generation of the shift pulse by the shift pulse generation circuit (34), and the generation of the shift pulse (SH) is prohibited (for example, the photoelectric conversion circuit). (During the data dump from (20) to the microcomputer (30) and during the data calculation by the microcomputer (30)) become “Low”, but then become “High” to open the AND circuit (AN). Therefore, when this signal (SHEN) is "high", the signal (TINT)
Occurs, the AND circuit (AN) outputs the “High” signal (TIN
T) is output. The OR circuit (OR) outputs this signal (TINT) or signal (SHM) to the shift pulse generation circuit (34), and in response, the shift pulse generation circuit (34) generates a shift pulse (SH). (36) is a transfer clock generation circuit that generates a transfer clock (φ ₁ ) and (φ ₂ ) in response to a clock pulse (CL) from the microcomputer (30), and a signal (TINT) from the OR circuit (OR) or When it receives (SHM), it is reset to the initial state, and it starts to generate new (φ ₁ ) and (φ ₂ ) regardless of the phase of the transfer clock (φ ₁ ) and (φ ₂ ) before that. Is to synchronize the shift pulse (SH) with the transfer clocks (φ ₁ ) and (φ ₂ ). The signal (SH) output from the microcomputer (30) is a sample hold signal for designating the pixel signal (DOS ') taken in by the peak hold circuit (24).

The microcomputer (30) is circuit-connected to the display circuit (38) and the lens driving device (40), and displays the focus adjustment state of the photographing lens (2) obtained by calculation as described later.
, And the lens driving device (40) drives the photographing lens based on the display. In this embodiment, the focus adjustment state of the photographing lens (2) calculated by the microcomputer (30) is represented by the defocus amount and the defocus direction. Therefore, the photographing lens (2) by the lens driving device (40) is used. The driving amount and driving direction in 2) are determined. The lens driving device (40) drives the taking lens (2) according to the driving amount and the driving direction, and outputs a signal indicating the executed lens driving amount to the microcomputer (30),
The microcomputer (30) outputs a signal for stopping the lens driving to the lens driving device when the executed lens driving amount reaches the driving amount obtained by the calculation.

In FIG. 1, S1 is an AF start signal switch (AF switch). The focus detection operation is started when the AF start signal switch S1 is closed, and the focus detection operation is stopped when the AF start signal switch S1 is opened. Is done.

S2 is a release signal switch.When the release signal switch S2 is closed, the camera itself shifts to the release operation and exposure control is performed, but the focus detection circuit temporarily closes the release signal switch S2 for mirror operation during release. Along with that, the operation is stopped.

Soc is one-shot AF mode, continuous AF
Mode changeover switch So, mode changeover switch So
c sets the one-shot AF mode when closed and the continuous AF mode when opened. The continuous AF mode has already been described.In the one-shot AF mode, the lens position adjustment is completed once the in-focus state has been detected during the focus detection operation, and the AF mode that performs AF lock is effective for still subjects. AF operation.

The switch Swn is a focus detection sensitivity range switching switch according to the present invention. When the switch Swn is closed, the focus detection sensitivity range is the central spot-shaped sensitivity range B in the field frame 42 of the camera shown in FIG. 3, and when it is open. The focus detection sensitivity range A is wide.

Next, before explaining the operation of the automatic focus detection device for a camera of FIG. 1, the correlation calculation used in the first and second inventions of the present application with reference to FIGS. 2 and 4 above. The interpolation calculation will be described.

The correlation calculation used here is this photoelectric conversion element array l _{1 to} l ₄₀ ,
It is calculated using r _{1 to} r ₄₈ . The 8-bit data of l _{1 to} l ₄₀ and r _{1 to} r ₄₈ is first set to difference data Dln, Drn (in order to limit the spatial frequency to be detected as described in JP-A-60-4914). However, Dl _n = l _n −l _{n + 4} , Dr _n = r _n −r _{n + 4}
Dl _{1 to} Dl ₃₆ , Dr _{1 to} Dr ₄₄ ). Subsequently, calculation of all contrast values is performed. The contrast value C is
It is calculated as the sum of the absolute values of the differences between the adjacent difference data in the reference portion (L) as follows.

Finally, the calculation is performed by shifting the correlation value YM / C (I) of the standard portion (L) and the reference portion (R) by one pixel by the following equation.

However, I = 1 to 9.

As a result, the correlation value YM / C (I) decreases as the degree of correlation increases. And find the point where the degree of correlation is maximum,
When the I value is 5, focusing is achieved. When I is smaller than 5, the camera is in a front focus state, and when I is larger than 5, the camera is in a rear focus state.

The correlation value obtained in this manner is a line sensor (1
Since it is calculated discretely by the output of 5), interpolation calculation is performed to calculate the in-focus point in order to obtain it with fine accuracy.

Various kinds of interpolation calculations have been proposed in Japanese Patent Laid-Open No. 59-126517, etc., but here, they are performed by steps # 100 to # 106 of the flowchart shown in FIG.

First, the smallest YM (1M) among the correlation values is detected.

Next, the value YM (lM) is standardized by the contrast value C, and the value YM (lM) / C is less than a predetermined value (YM (lM) / C <0.
Different methods are used for the case 1) and the case where the value is equal to or more than a predetermined value (YM (1M) /C≧0.2). First, in the former case, the minimum value YM (lM) / C
Is determined to be highly reliable, and when YM (lM + 1) ≧ YM (lM-1), When YM (lM + 1) <YM (lM-1) Becomes

This method shows that the correlation curves are 45
It is based on having an inclination of ° and being orthogonal. In the ideal state, the standard part (L) and the reference part (R) are l _n
= R _{n + 4} is output exactly the same value. Thus the difference data thereof Dl _n = Dr _{n + 4.} Where the contrast value C is Is.

On the other hand, the minimum correlation value is The correlation value on both sides is As long as the value of | Dl ₁ −Dr ₄ |, | Dl ₃₆ −Dr ₄₁ | is sufficiently small with respect to the overall contrast C, the correlation curves are orthogonal at 45 ° to each other. This interpolation calculation method has hitherto been a highly accurate interpolation method as compared with the interpolation method proposed in Japanese Patent Laid-Open No. 59-126517. However, the value YM (1M) obtained by normalizing the minimum correlation value YM (1M) with the contrast value C
When / C becomes larger than a predetermined value, the minimum correlation value becomes a large value due to some influence, here, a near-far competing object or electrical noise, resulting in a poor reliability value. Therefore, if this YM (lM) / C is greater than or equal to the specified value, use the correlation values on both sides of it instead of using the minimum value as the basis. The interpolated value is obtained by In addition, this value YM (lM) / C has hysteresis considering that the interpolated value at the threshold level varies, and once the former interpolation method is used, until this YM (lM) / C becomes 0.2 or more. The former interpolation method is used.

Next, the basic operation of the first and second inventions of the present application will be described with reference to the flowchart shown in FIG.

When the power switch (not shown) is turned on, the flow of FIG. 3 starts, the microcomputer (30) executes step # 1, and waits for the AF operation switch S ₁ to be turned on. The AF switch S ₁ is normally turned on by pressing the first step of the release switch to start the AF operation.

First, the microcomputer (30) is equipped with a photoelectric conversion circuit (2
After the initialization of 0) is performed, the integration of the image brightness information is started by the photoelectric conversion circuit (20) in step # 2. After the integrated charge reaches a predetermined level, in step # 3 (hereinafter, “step” is omitted), the microcomputer (3
0) stores the A / D conversion value of each pixel output in the RAM (not shown) in the microcomputer (30) in synchronization with the transfer clock after the shift pulse is generated. After inputting the required 8-bit digital data of each pixel data l _{1 to} l ₄₀ , r ₁ to ₄₈ , the microcomputer (30) sets the above-mentioned difference data Dl _{1 to} Dl ₃₆ , Dr _{1 to} Dr ₄₄ at # 4. Create.

After the preparation for the correlation calculation is completed in this way, the microcomputer (30) confirms the focus detection sensitivity range switch Swn at # 5-φ. If the focus detection sensitivity range switch Swn is turned on and the narrow focus detection sensitivity range is requested by the photographer, the microcomputer (30) determines the number of correlation calculation data in # 5-2. The number of data 16 in the narrow focus detection sensitivity range and the start difference data 11 are set to D and the start difference data number S, and the predetermined lower limit contrast value CS is set to a predetermined value CS1. On the contrary, when the focus detection sensitivity range switch Swn is turned off and a wide focus detection sensitivity range is requested by the photographer, the microcomputer (30) determines the number D of correlation calculation data in # 5-1. As the difference data number 36 in the entire sensitivity range, 1 as the start difference data number, and a predetermined value CS ₂ is set as the focus detection lower limit contrast value CS. Next, # 5-
At 3, the microcomputer (30) sets S, D set by turning on and off the focus detection sensitivity range changeover switch Swn.
Correlation values for the respective k values of 1 to 9 are calculated according to the respective values. Here, the focus detection sensitivity range is limited.

The minimum value is detected in # 6 with respect to the correlation value thus obtained, and the above-described interpolation calculation is performed in # 7 to calculate the image interval XM.

The difference between the image distance XM and the reference focused image distance XJ is calculated as the image distance deviation amount P at # 8. This image interval shift amount P
Is a value proportional to the defocus amount.

Next, the reliability is confirmed for this value in # 9.
Details thereof are described in JP-A-60-7687. That is, the contrast value C has a predetermined value or more (here, the CS value is already set or more), the maximum value of each pixel output is a predetermined value Ps or more, and the minimum correlation value is not at both ends. Further, when the YM / C value obtained by performing the same interpolation calculation as the image interval XM value in the interpolation calculation is less than or equal to a predetermined value YMs, when the four points are both satisfied, the image distance deviation amount P obtained by the calculation is Converted to defocus amount in # 13 as highly reliable data. Further, in # 1, the defocus amount is converted into the lens driving amount. After that, it is determined in # 15 whether the focus state or the lens drive is required based on the lens drive amount, and in the former case, the focus display is performed in # 17. In the latter case, the lens is driven at # 16 and the photoelectric conversion circuit (20) again
To focus confirmation operation by reintegration and refocus detection. After the focus is displayed once again, the microcomputer (30)
At # 18, the ON / OFF of the mode changeover switch Soc is determined, and during the one-shot AF in which the mode changeover switch Soc is ON, the AF operation is terminated with the lens stopped.
If the continuous AF mode is selected with the mode change switch Soc turned off, the photoelectric conversion circuit (20) is reintegrated to repeat this routine.

On the other hand, if it is determined in # 9 that the data of any one of the contrast value, the maximum pixel output value, and the correlation value is unreliable, the lens is driven in # 11 and the entire lens drive range is changed. At least at least once perform a low-contrast scan LO-CONSCAN, and if it is determined in # 10 that reliable data cannot be obtained even after this is finished, the lens drive is terminated, and in # 12, a low contrast is obtained. The LO-CON display indicating that focus detection is impossible is displayed.

In the flow of FIG. 5, the microcomputer (30) confirms whether the focus detection sensitivity range is wide or narrow in the flow of # 5-φ, but it is either continuous AF mode or one-shot AF mode. Alternatively, sensing may be performed. If the mode selector switch So is OFF and the continuous AF mode is selected, #
Set the focus detection sensitivity range wide toward the flow of 5-1.
When the one-shot AF mode is selected with the mode changeover switch So turned on, the focus detection sensitivity range is narrowed toward the flow of # 5-2. Here, it is assumed that the focus detection sensitivity range changeover switch Swn shown in FIG. 1 is deleted.

If you do this, continuous AF mode
The wide focus detection sensitivity range makes it easier to follow and shoot a moving subject, and it is easy to capture the subject in the limited focus detection sensitivity range during one-shot AF, and accurate AF can be achieved for stationary subjects.

Next, the operation of the embodiments of the first and second inventions of the present application when the line sensor (15) is divided into blocks as shown in FIG. 2 and the following table, FIG. 7
This will be described with reference to the flowcharts of FIGS. 7A and 7B.

The line sensor (15) has a reference portion (L) composed of pixels l _{1 to} l ₄₀ and pixels r _{1 to} r with an intermediate separation band in between.
_{It is divided} into a reference section (R) consisting of ₄₈ . Reference part (L)
The first block to the pixel l ₁ ~l ₂₀ (I), the pixel l ₁₁ ~
second block up to l ₃₀ (II), is divided into blocks by overlapping so that the third block (III) to the pixel l ₂₁ to l _40. Interval between the two image formed respectively reference unit (R) and the reference portion (L), when the focus is achieved, a predetermined distance L _2.

In FIG. 6, after starting the operation, the microcomputer (30) first designates the start block (zone) of the focus detection calculation as the second block (II) in # 20 and limits the shift amount of the defocus amount calculation. Reset flag ZF. Then wait for the input of the AF operation start signal (# 1). When the AF switch S1 is closed and the start of the AF operation is instructed, first, the line sensor (15) is initialized, and then the image information accumulation of the line sensor (15) is started (# 2). When the charge accumulation of the line sensor (15) is completed to a predetermined level, the accumulated charge is transferred in parallel to the analog shift register by the shift pulse, and thereafter, the charge transfer and the data dump are synchronized with the transfer clocks φ ₁ and φ _2. (# 3) is performed. When the data dump of all pixels required for focus detection calculation is completed,
The microcomputer (30) first converts the pixel data into difference data (# 4). This is to remove a spatial frequency component which is not necessary for the focus detection calculation and is an obstacle.

After the calculation of the difference data is completed, the microcomputer (30) senses the focus detection sensitivity range changeover switch Swn to determine whether or not the focus detection sensitivity range is limited. Focus detection sensitivity range switch Swn
When it is in the OFF state, at # 21, the entire focus detection sensitivity range, that is, the first to third blocks (I), (I
In order to perform the focus detection calculation in I) and (III), 3 is set to the zone calculation counter j. Conversely, when the focus detection sensitivity range changeover switch Swn is in the ON state, in # 22, Second
In order to perform the focus detection calculation for the focus detection sensitivity region of only the block (II), 1 is set to the zone calculation counter j, and n = to set the zone designation to the second block (II).
Set to 2.

Focus detection calculation is started in this state, but zone designation n
= 2, and the defocus amount restriction flag ZF = 0, so after going through # 23 to # 27, first at # 29, the second block (II)
Correlation function calculation is performed over the entire shift range of, and in # 30, the image interval that minimizes the correlation function YM2 / C2 (k), that is, the image interval 1M2 having the highest degree of correlation is obtained. This lM2, YM2 / C2 (lM2) and the correlation value before and after it are interpolated at # 31, and the image interval XM2, YM2 / C2
(XM2).

Next, LO-CON discrimination is performed at # 32 for this value. If the obtained XM2 is reliable, from the image interval,
In # 33, the image shift amount P2 is calculated and stored in memory. Further, in order to speed up the subsequent correlation calculation of the block, only the image shift larger than this image shift amount is calculated. Therefore, at # 34, the defocus limit flag ZF is set to 1, and 1M2-15 is stored as the limit value Pmin. This is because the correlation value of the pitch shift on both sides is necessary for the interpolation operation with a value smaller by one than the line sensor pitch image interval at which the degree of correlation is maximum. The interpolation calculation and the LO-CON discrimination are the same as those described in the flowchart in FIG.

Thereafter, the routine counter is checked. At this stage, if the focus detection sensitivity range switch Swn is set to ON and the focus detection sensitivity range is set only in the second block (II), it is set to j = 1 in advance. The counter j becomes φ by passing the step of j = j−1 of # 35, and at that time point when the focus detection calculation of the second block (II) is completed, the process proceeds from # 36 to the lens driving routine of # 38.

On the other hand, a wide focus detection sensitivity range is selected, the focus detection sensitivity range switch Swn is preset to j = 3 in the OFF state, and at the end of this routine, j = j-1 =
2 ≠ 0, n = n + 1 = 2 + 1, and the third block (III) is selected as the next focus detection zone, and the third block (I
Return to the calculation of the correlation function for (II) (# 25, # 41 ~ # 4
Four). In the correlation calculation of the third block (III), the second
If it is determined that the block is not LO-CON in block (II), the second block (I
Correlation calculation is performed only with an image shift amount larger than the image shift amount detected in I). When the LO-CON is determined in the second block (II), the correlation operation is performed in the third block (III) under the whole image shift amount. Thus the second block (II)
Calculation of lM3 (# 45) and X by interpolation calculation as done in
Calculation of M3, YM3 / C3 (XM3) (# 46), LO-CON discrimination (# 4
7) is performed, and if it is not LO-CON, the image shift amount P3 is stored in memory (# 48), the defocus limit Pmin = lM3-25 is stored, and the flag ZF = 1 is set (# 49). It is conceivable that P3 obtained here is smaller than P2 by a difference within one pitch, but Pmin at that time becomes the same value, and Pmin cannot be reduced.

Next, in # 50 to # 59, n = n + 1 in exactly the same manner.
= 4, and the correlation operation of the first block (I) is performed.

When the wide focus detection sensitivity range is thus selected, an image shift range larger than the image shift amount of the block previously obtained by the correlation operation in the first to third blocks (I), (II), and (III). After the focus detection calculation is performed in step (3), the zone counter becomes j = 0, and the flow shifts to the lens drive routine of # 38.

Here again, the focus detection sensitivity range switch Swn turns ON and O
The same routine as in FIG. 5 is executed regardless of FF.

First, it is determined whether there is a zone that is not LO-CON (# 39). Here, the focus detection sensitivity range switch Sw
It is needless to say that when the narrow focus detection sensitivity range of only the second block (II) is selected with nON, the determination is made only for the second block (II). When the first to third blocks (I), (II), and (III) are LO-CON, the taking lens (2) moves the entire area at least once from the closest state to ∞, and the focus detection is performed. Move to LO-CON SCAN for calculation and drive the lens. During this operation, focus detection calculation is performed several times, and all become LO-CON, and when LO-CON SCAN is completed, the lens operation is stopped at the closest or infinity lens end point and the lens operation is left as it is. AF switch S1 ON
As long as the condition continues, the LO-CON display is performed while driving the image sensor and performing focus detection calculation (# 12), and waits until the subject state is appropriate for focus detection.

When a zone other than LO-CON is detected, a zone having the largest image interval among them, that is, an image shift amount and a zone of a zone including the closest subject among the subjects are extracted (# 40). The image shift amount Pi is converted by multiplying the defocus amount by a constant (# 13), and n = zone number i
As a memory. The calculated defocus amount is converted to the lens drive amount by taking the product of the defocus amount and the lens drive amount coefficient which differ for each photographing lens (2) (# 14). When the lens drive amount is extremely small, Focus display (# 17), lens drive in other cases (# 16)
After that, reintegration of the image sensor and refocus detection calculation are performed.
When the one-shot AF mode is selected by sensing the mode switch Soc after the focus display is displayed, the microcomputer (30) stops while the focus display is displayed, and the continuous AF mode is set. When is selected, reintegration and refocus detection calculation of the image sensor are performed as in the non-focus state.

As described above, in the first invention of the present application, the block having the largest image interval is stored in the previous focus detection calculation, and in the case of the refocus detection calculation, the first zone calculation routine is performed on the block. Focus detection calculation is performed at. By performing the first focus detection correlation calculation in the flow branched by the set n, the focus detection calculation after the focus detection operation is once performed is quick.

As described above, in this block subdivision focus detection calculation, the block with the largest image interval is extracted in the first zone calculation routine, and the block with the largest image interval is given priority in the subsequent focus detection calculation routine. In this case, the number of correlation calculations for subsequent focus detection of the block is reduced, and the responsiveness of focus adjustment is improved.

Next, the flow of the embodiment of the second invention of the present application is shown in FIGS. 7 (a) and 7 (b).

In this embodiment, in order to make the operation quicker, the block having the largest image interval is stored in the previous focus detection calculation, and in the case of the refocus detection calculation, the limited shift amount near the focus is set. The focus detection calculation is performed in the block memorized only by the lens, and in the case of reliable data, not the LO-CON, the focus detection calculation is not performed at all in the other block and the lens is driven. .

In FIGS. 7A and 7B, the first focus detection operation after the AF operation start switch S ₁ is turned on is completely the same as that in FIG. After the defocus amount is calculated here and the lens is driven according to the defocus amount (#
13 to # 16), the lens driven flag SF which is cleared at # 20 at the start of AF operation is set to 1 at # 72. This lens driven flag SF indicates that the probability that the subject exists near the in-focus position is high. After this, the focus detection device re-integrates the image sensor to obtain new image information,
The focus detection calculation is repeated. Here, in the case of SF = 1, the image interval having the highest correlation is predicted in advance because the probability that the subject exists near the in-focus position is high. Therefore, the correlation value is obtained only for the image intervals of five points -2, -1, 0, 1, 2 from the focused image interval (# 62). As a result, when the image interval for obtaining the highest correlation is any of −1, 0, and 1, the interpolation calculation is performed according to the value to perform the LO-CON determination (# 63, # 64, # 6
5) When the value is determined to be highly reliable, the lens drive is repeated according to the interpolated image interval XMn. That is, after calculating the image distance shift amount Pn at # 66 in FIG. 7A using this interpolated image distance XMn, the image distance shift amount Pn is defocused at # 13 in FIG. 7B. Converted to quantity. on the other hand,#
If LO-CON is determined in 65, the maximum correlation image interval is 2
Alternatively, the case of -2 is included, but at that time, the same correlation calculation as in the first time is performed. In addition, the focus detection calculation finally
If it is judged to be CON, the lens driven flag SF is returned at # 76.
Will not follow this routine after being cleared,
Until the focus is detected again, the focus detection calculation is repeated for the entire defocus amount range and all blocks as in the first focus detection calculation. When focus is detected, one-shot AF
In the case of, the lens driven flag SF is set at # 74.

Here, an example is shown in which refocus detection is performed after lens driving is completed.However, it is also possible to repeat focus detection while driving the lens, and in such an example, the lens drive completion flag SF indicates the focus detection. It can be set only at the time.
Here, the correlation value calculation image shift amount at the time of simple focus detection is
Five points of 2,1,0, -1, -2 are set, but this is a value that depends on the adjustment accuracy between the AF sensor and the re-imaging lens. If sufficient accuracy is secured, 1,0,- It goes without saying that the three points of 1 are also possible.

Finally, in FIG. 6, FIG. 7 (a) and FIG. 7 (b), the ON / OFF discriminating portion of the focus detection sensitivity range changeover switch Swn is made to discriminate ON / OFF of the mode changeover switch Soc. As described above with reference to FIG. 5, it is possible to switch between a wide focus detection sensitivity range and a narrow focus detection range by continuous AF and one-shot AF.

EFFECTS OF THE INVENTION The present invention first detects the previously selected area and omits the detection operation in other areas if the detection result is reliable, so that the time required for the focus detection operation is shortened. , Followability to the subject is improved.

[Brief description of drawings]

FIG. 1 is a block diagram of an automatic focus adjustment circuit, FIG. 2 is an explanatory diagram showing a pixel array of a line sensor, FIG. 3 is an explanatory diagram of switching of a focus detection sensitivity region within a field frame of a camera, and FIG. Is a flowchart of the interpolation calculation, FIG. 5 is a flowchart of the basic operation of the first and second inventions of the present application, and FIGS. 6, 7 (a) and 7 (b) are the first and the second of the present application, respectively. 8 is a flowchart of the operation of the automatic focus detection device for a camera according to the invention of FIG. 2, FIG. 8 is an explanatory view of the focus detection optical system, FIG. 9 is an explanatory view showing the positions of two images in focus detection, and FIG. It is explanatory drawing which shows the change of a focus amount. Reference numeral 1 denotes a subject light beam, 2 denotes a photographing lens, 9,11 denotes two images, 12,14 denotes a photoelectric conversion element array. 15 ... Line sensor, 20 ... Photoelectric conversion circuit, 30 ... Microcomputer.

 ─────────────────────────────────────────────────── ─── Continuation of front page (72) Inventor Toshihiko Karazaki 2-30 Azuchi-cho, Higashi-ku, Osaka, Osaka Prefecture Osaka Osaka Building Minolta Camera Co., Ltd. (72) Nobuyuki Taniguchi 2-chome, Azuchi-cho, Higashi-ku, Osaka, Osaka No. 30 Osaka International Building, Minolta Camera Co., Ltd. (56) References JP 59-9611 (JP, A) JP 59-123810 (JP, A)

Claims (1)

[Claims] 1. A focus detecting means for detecting a focus state for a subject in a plurality of areas in a photographing screen and outputting focus detecting data for each area, and a plurality of focus detecting data outputted from the focus detecting means. Selecting means for selecting the above area, focus adjusting means for performing focus adjustment based on the focus detection data corresponding to the area selected by the selecting means, and operation of the detecting means, the selecting means, and the focus adjusting means An automatic focus detection device for a camera, comprising: a storage unit that stores a previously selected area; and a focus state for the previously selected area in the focus detection this time. Judge the reliability of the result, and if it is judged to be reliable, skip the focus detection operation in other areas and select that area. When it is determined that no reliable other performs focus detection operation in area and control means for selecting another area, automatic focus detection device for a camera, characterized in that it comprises.