JPH11119087A - Focus detector for camera - Google Patents

Abstract

PROBLEM TO BE SOLVED: To minimize an detection error of an image moving amount by making the charged level of sensor data related to the predicting arithmetic operation of a moving body nearly the same. SOLUTION: This focus detector for a camera is equipped with a focus detection part 1 outputting a signal for focus detection, and an integration level control part 2 controlling so that the integration level of the focus detection signal used for the image moving amount arithmetic operation may be nearly the same is connected to the focus detection part 1. An image moving amount detection part 3 arithmetically calculating the image moving amount of a subject by performing correlation arithmetic operation to the focus detection signal controlled to be at the same integration level is connected to the control part 2. Then, a moving body predicting part 4 arithmetically calculating and predicting the image moving amount at the time of exposure based on the detected image moving amount is connected to the detection part 3.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a focus detection device for a camera having a moving object predicting function for focusing a photographing lens on a moving subject.

[0002]

2. Description of the Related Art Conventionally, a movement of a photographic lens of an object in an optical axis direction is detected, an image plane position of the object after a predetermined time has elapsed is predicted, and the photographic lens is driven to the predicted image plane position. Various techniques have been proposed for a focus detection device of a camera having a so-called moving object prediction function that also focuses on a moving subject.

For example, in Japanese Patent Publication No. Hei 8-27425, a release time lag, that is, a movement amount of a subject image which is predicted to move during a time lag such as lens driving or mirror up is corrected based on a change in defocus amount. There is disclosed a technique relating to an "autofocus device".
Japanese Patent Application Laid-Open No. 5-93850 discloses an "autofocus device" that detects a moving amount of an image on a focus detection sensor surface and predicts a future image moving amount based on the detected image moving amount.
Is disclosed.

[0004]

However, the conventional camera focus detection device having the above-described moving object prediction function has the following problems. That is, in general, a focus detection sensor mounted on a conventional camera focus detection device is a charge storage type photoelectric conversion element, and a moving object prediction is performed based on the latest sensor data and past sensor data. there were. Therefore, if the latest charge accumulation level of the photoelectric conversion element (hereinafter, integration level) is significantly different from the integration level of the past photoelectric conversion element, the photoelectric conversion element may capture the same subject. Also, there is a disadvantage that the contrast of the subject image is different and the error of the image movement detection becomes large.

[0005] In particular, the technique disclosed in the above-mentioned Japanese Patent Application Laid-Open No. 5-93850 is a technique for calculating the image movement amount by performing a correlation operation between the latest sensor data and the past sensor data. Therefore, when the difference between the integration level of the latest photoelectric conversion element and the integration level of the past photoelectric conversion element is large, the reliability of the correlation calculation is significantly reduced, and the error of image movement detection increases. Was.

FIG. 2 is a diagram conceptually showing the difference between the integration level and the difference between the sensor outputs.

When a photoelectric conversion element (AF sensor) captures the periphery of a black line in a line chart in which a vertical black line is drawn on a white background as shown in FIG. 2A, the output is, for example, as shown in FIG. As shown in FIG. Here, the vertical axis of the figure indicates the sensor output, and the horizontal axis indicates the order of the sensor pixels. That is, in the case of a sensor having 64 pixels, the left end of the horizontal axis in the drawing indicates the first pixel, and the right end indicates the 64th pixel.

More specifically, FIG. 2B shows the current (latest) sensor data, in which the output of the white portion is A and the output of the black portion is A '. FIG. 2C shows the previous (past) sensor data, in which the output of the white portion is B and the output of the black portion is B '.

From these FIGS. 2B and 2C, A <B
Since A '<B', it can be seen that the integration level of the previous photoelectric conversion element is higher than the integration level of this time, and that the integration amount of the previous time is larger. This is considered to be due to variation in integration control that governs the integration level and a slight difference in luminance between subjects.

Therefore, as shown in FIGS. 2D and 2E, if the previous sensor data can be corrected to the same integration level as the latest sensor data, the image movement amount Detection errors can be reduced.

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and an object of the present invention is to provide a focus detection apparatus for a camera having a moving object predicting function, in which a charge accumulation level of sensor data relating to a moving object predicting operation is set to the same level. Accordingly, an object of the present invention is to reduce the error of the image movement amount detection.

[0012]

In order to achieve the above object, a focus detection apparatus for a camera according to a first aspect of the present invention includes a charge storage type photoelectric conversion element and repeatedly detects a focus adjustment state of a photographing lens. Detection means for outputting a focus detection signal in time series; control means for controlling the charge accumulation level of the latest focus detection signal and the charge accumulation level of the next focus detection signal to be substantially the same; In response to the time-series signal output from the detecting means, the focus adjustment prediction calculation for focusing on the subject moving in the optical axis direction of the photographing lens is repeatedly performed, and the photographing lens is calculated based on the result. Focus adjusting means for performing the focus adjustment of (1).

Further, a focus detection device for a camera according to a second aspect is a charge storage type photoelectric conversion means for repeatedly photoelectrically converting an input subject image and outputting detection signals in time series, and a latest detection signal. Storage means for storing a plurality of recent detection signals including: a control means for controlling the charge accumulation levels of the plurality of detection signals output in time series to be substantially constant; and And a calculating means for performing a calculation for focus adjustment based on the detection signal in order to focus on a subject moving in the optical axis direction of the taking lens.

According to a third aspect of the present invention, in the camera focus detection device, the control means controls the charge accumulation time of the photoelectric conversion means to be substantially constant, or corrects at least one of the detection signals. By performing the correction, the level of the detection signal is corrected to be substantially constant.

According to the first to third aspects, the following operations are provided.

That is, in the focus detection apparatus for a camera according to the first aspect of the present invention, the focus adjustment state of the photographing lens is repeatedly detected by the detection means including the charge storage type photoelectric conversion element, and the focus detection signal is time-sequentially detected. The charge control level is controlled by the control means so that the charge accumulation level of the latest focus detection signal is substantially the same as the charge accumulation level of the next focus detection signal, and is output from the detection means by the focus adjustment means. In response to the series signal, the focus adjustment prediction calculation for focusing on the subject moving in the optical axis direction of the photographing lens is repeated, and the focus adjustment of the photographing lens is performed based on the result.

Further, in the camera focus detection apparatus according to the second aspect, the input subject image is repeatedly photoelectrically converted by the charge storage type photoelectric conversion means, detection signals are output in time series, and the storage means A plurality of the latest detection signals including the latest detection signal are stored, and the control means controls the charge accumulation levels of the plurality of detection signals output in time series to be substantially constant, and the calculation means On the basis of the plurality of stored detection signals, calculation for focus adjustment is performed so that a subject moving in the optical axis direction of the photographing lens is also focused.

In the focus detection device for a camera according to a third aspect, the control means controls the charge accumulation time of the photoelectric conversion means to be substantially constant, or corrects at least one of the detection signals. By doing
Correction is performed so that the level of the detection signal is substantially constant.

[0019]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a conceptual diagram showing the configuration of a camera focus detection device according to a first embodiment of the present invention. As shown in the figure, the focus detection device for a camera according to the present embodiment includes a focus detection unit 1 that outputs a signal for focus detection including an AFIC 240 described later, and the focus detection unit. 1 is connected to an integration level control unit 2 for controlling the integration level of the focus detection signal used for calculating the image movement amount so as to be the same. This integration level control unit 2
Is connected to an image movement amount detection unit 3 that performs a correlation operation on the focus detection signal corrected to the same integration level to calculate the image movement amount of the subject. A moving object predicting unit 4 for predicting and calculating an image moving amount at the time of exposure based on the detected image moving amount is connected.

FIG. 3 is a block diagram showing in detail a control system of a camera system to which the camera focus detection device according to the first embodiment is applied.

As shown in the figure, a camera system to which the focus detection device of the first embodiment is applied has a CPU
201, interface IC 202, power supply unit 2
03, strobe unit 204, mirror shutter unit 205, winding unit 206, lens unit 2
07, finder unit 208, display unit 20
9, each unit of the AF unit 210 and the like.

The CPU 201 controls the entire camera system. The CPU 201 controls the interface IC 202 and the LCD IC 23 via a serial communication line 211.
5. Data is transmitted / received to / from the AFIC 240 and the EEPROM 237.

The CPU 201 and an interface IC
(IFIC) 202, another communication line 251
For inputting various analog signals, inputting a signal after shaping the waveform of a photo interrupter, and the like. The analog signal is input to an A / D conversion input terminal of the CPU 201 and is converted into a digital signal. In addition, the CPU 201 includes various calculation units (not shown), a data storage unit, and a time measurement unit.

The interface IC 202 is a Bi-CMOS IC having a mixture of digital and analog circuits, such as a motor and magnet drive, photometry, battery check, a backlight LED, an auxiliary light LED lighting circuit, and a photo interrupter waveform shaping circuit. , And a digital processing unit for input serial communication data conversion of a switch (SW).

The power supply unit 203 supplies two systems of power. In the power supply unit 203, one is a power supply used for a driver requiring power such as a motor or a magnet, and the voltage of the battery 212 is always supplied. The other one is a DC / DC converter 2
13 is a power source for small signals stabilized by CP
It is controlled by the U201 through the interface IC202.

The strobe unit 204 includes a strobe charging circuit 214, a main capacitor 215, a strobe light emitting circuit 216, a strobe light emitting tube 217, and the like. When the strobe light needs to be emitted in a low-luminance or backlight state, the interface I
Via C202, the flash charging circuit 214 boosts the battery voltage and charges the main capacitor 215. At the same time, the charging voltage divided from the flash charging circuit 214 is input to the A / D conversion input terminal of the CPU 201, whereby the CPU 201 controls the charging voltage.

When the charging voltage reaches a predetermined level, a charging stop signal is transmitted from the CPU 201 to the flash charging circuit 214 via the interface IC 202, and the charging of the main capacitor 215 is stopped. Also,
The CPU 201 controls the flash light emitting tube 21 via the flash light emitting circuit 216 at a predetermined timing during film exposure.
The control of the start of light emission and the stop of light emission of 7 are performed.

The mirror shutter unit 205 includes a mirror shutter motor 218, two shutter magnets 219 for controlling the traveling of the front curtain and the rear curtain, a front curtain traveling completion switch included in the sequence switch group 244, and the like. I have. This mirror shutter motor 218 is
Controlled by the CPU 201 via the interface IC 202 and the motor driver 241, the forward rotation of the main mirror 102 causes the main mirror 102 to move up and down, the aperture of the shooting aperture to be reduced, and the opening shutter to be charged, that is, the operation to close the front curtain and open the rear curtain. It is.

The shutter magnet 219 is controlled by the CPU 201 via the interface IC 202. At the start of exposure, the mirror shutter motor 218 first retracts the main mirror and narrows down the photographing aperture immediately before the exposure starts. Next, the shutter magnet 219
When the shutter magnet 219 of the front curtain is released at the same time as the start of exposure to attract the magnet, the front curtain is opened. First curtain advance completion switch 24
After a desired exposure time elapses from the input of No. 4, the rear curtain is closed by releasing the suction of the shutter magnet 219 of the rear curtain. Thus, the film is exposed between the opening of the front curtain and the closing of the rear curtain. Next, the mirror is lowered by the forward rotation of the shutter motor 218, and the photographing aperture is opened.
At the same time, the shutter is charged. The shutter motor 218 reverses the film to rewind the film.

The winding unit 206 includes a winding motor 220 and a film detection photo interrupter 221.
Etc. The hoist motor 220 is controlled by the CPU 201 via the interface IC 202 and the motor driver 241. The output of the film detection PI 221 is waveform-shaped by the interface IC 202 and transmitted to the CPU 201 to generate a winding amount feedback pulse. The CPU 201 controls the winding amount for one frame of the film by counting the number of pulses.

The lens unit 207 includes a photographing lens 222, a zoom motor 223, a zoom gear train 224,
F motor 225, AF gear train 226, AFPI 227,
It comprises a zoom encoder 228, an aperture PI 229, an aperture magnet 230 and the like. This zoom motor 22
3. The AF motor 225 is an interface IC 20
2. Controlled by the CPU 201 via the motor driver 241. The rotation of the zoom motor 223 is reduced by the zoom gear train 224, thereby
The second zoom system is driven. Also, the zoom encoder 2
Reference numeral 28 denotes an encoder including six switches provided around a lens frame that supports the photographing lens 222. The ON / OFF data of the six switches is input to the CPU 201, and the absolute position of the zoom lens is detected. It has become.

The CPU 201 obtains the focal length from the absolute position of the zoom lens and stores it in the focal length storage unit 247. The rotation of the AF motor 225 is controlled by the AF gear train 226.
As a result, the focusing lens of the photographing lens 222 is driven. On the other hand, the output of the AF photointerrupter 227 is extracted from the middle of the AF gear train 226. The output of AFPI 227 is waveform-shaped by interface IC 202 and transmitted to CPU 201,
An AF lens drive amount feedback pulse is generated. C
The PU 201 counts the number of pulses so that A
The driving amount of the F lens is controlled. The protrusion amount of the AF lens from the mechanical stopper or the infinite reference position is AFPI2
The pulse amount is 27 and is stored in the lens ejection amount storage unit 246 in the CPU 201.

The aperture magnet 230 is controlled by the CPU 201 via the interface IC 202. At the same time as the mirror-up is started, a magnet to which a current is supplied is attracted. At the same time as the above-described mirror-up operation of the mirror shutter motor 218 of the mirror shutter unit 205, the aperture of the shooting aperture is mechanically started by a spring. Then, when the desired aperture value is reached, the attraction of the aperture magnet 230 is released and the aperture operation is stopped to set. Aperture PI229
Is shaped by the interface IC 202 and transmitted to the CPU 201 to generate a narrowing-down amount feedback pulse. The CPU 201 controls the stop-down amount of the shooting aperture by counting the number of pulses.

The finder unit 208 includes an LCD panel 231 in the finder and a backlight LED 23.
2 and a photometric 8-division photodiode element 233 and the like. The LCD panel 231 in the finder is made of transmissive liquid crystal.
The display is controlled by the LCD IC 235 in accordance with the display content sent to the. And the backlight LED 232
Is an interface IC 202 by the CPU 201.
Is turned on through the LCD panel 23 in the viewfinder.
Light 1 The photometric element 233 is controlled by the CPU 201 via the interface IC 202. The photocurrent generated by the photometric element 233 is sent to the interface IC 202 every eight elements, and is subjected to current / voltage conversion therein. Only the output of the element specified by the CPU 201 is transmitted from the interface IC 202 to the CPU 2.
01, which is sent to an A / D input conversion terminal, converted into digital form, and used for photometric calculation.

The display unit 209 includes an external LCD panel 234, an LCD IC 235, and a key switch (SW).
Group (1) 236 and the like. And the LCD panel 23
Reference numeral 4 denotes a reflective liquid crystal, which is transmitted from the CPU 201 to the LCD IC 2.
The display is controlled by the LCD IC 235 according to the display content sent to the LCD 35. Key switch group (1) 236
Is mainly for setting the mode of the camera.
Mode selection switch, camera exposure mode selection switch,
Switches such as a strobe mode selection switch, an AF / PF switch, and a macro mode switch are included. The states of these switches are read into the CPU 201 via the LCD IC 235, and the respective modes are set.

The AF unit 210 is composed of the EEPROM 23
7, condenser lens 238, separator lens 239,
The photo sensor array 250 includes an AFIC 240 and the like.

A part of the light image of the subject is
38, divided into two images by the re-imaging lens 239,
The light is received by two rows of photoelectric conversion elements on the FIC 240.
The AFIC 240 generates an analog output corresponding to the light intensity for each element. The analog output is sent to an A / D conversion input terminal of the CPU 201 and converted into a digital signal.
It is stored in the element output storage unit 245 in U201.

The CPU 201 calculates an image interval between two divided images or a moving amount of each image after a predetermined time by an internal correlation operation circuit 248 based on the stored element outputs. Further, the CPU 201 controls the photoelectric conversion operation of the AFIC 240. In the EEPROM 237, non-uniformity correction data of a photoelectric conversion element output, which will be described later, and various adjustment data such as an interval between two images at the time of focusing are written at the time of shipment from a factory, for example. Data that needs to be stored is written even in the OFF state.

The motor driver 241 is a driver for controlling a large current of the mirror shutter motor 218, the winding motor 220, the zoom motor 223, the AF motor 225 and the like. The auxiliary light LED 242 is an LED for illuminating the subject when the luminance is low. The auxiliary light LED 242 is turned on when the AFIC 240 does not complete the photoelectric conversion within a predetermined time and the image interval between the two images cannot be detected, so that the AFIC 240 can photoelectrically convert the subject image by the illumination light. is there.

The key switch (SW) group (2) 243 includes:
Switches for controlling the operation of the camera, including a first stroke signal (1R) and a second stroke signal (2R) of a release switch, a switch for driving a zoom lens to a long focal length side, a switch for driving a short focal length side, and a spot. A switch and the like for storing the photometric value are included.

The state of each of these switches is read into the CPU 201 via the interface IC 202, and the operation of the camera is controlled. Sequence switch (S
The W) group 244 detects the state of the camera.
This includes a switch for detecting a mirror raised position, a switch for detecting completion of shutter charge, a switch for detecting completion of running of a shutter front curtain, a power switch, a switch for detecting a strobe pop-up state, and the like. In addition, the buzzer 245 displays a sound when the AF is in focus, when the AF is out of focus, when the power is turned on, or when a camera shake warning is issued.

Next, FIG. 4 shows a detailed configuration of the AFIC 240 and will be described.

Referring to FIG.
The charge generated according to the incident light on 50L and 250R is
Each photo sensor is stored in a storage capacitor inside the amplifier circuit AP and amplified. And this amplifier circuit AP
Each of the pixel signals amplified inside is converted by the monitor output circuit MO into a signal corresponding to the peak (maximum value) of all the pixels, and is output from the terminal MDATA to the AD converter inside the CPU 201. The signal output to this MDTA is
The signal indicates the accumulation level of the electric product stored in the storage capacitor.

On the other hand, each pixel signal amplified by the amplifier circuit AP is sent to the CPU 201 by the shift register SR.
The signal is sequentially output from the terminal SDATA according to the clock signal of the terminal CLK output from the terminal SDATA. And the terminal SDAT
The output of A is input to an AD converter inside the CPU 201. The control circuit CNT receives control signals RESET, END, and CLK from the CPU 201 and
The operation of each block in 40 is controlled. The reference voltage VREF is input from the IFIC 202 to the amplifier circuit AP,
This is the reference potential inside the amplifier circuit AP.

Here, FIG. 5 shows and describes the state of the signal at each terminal in FIG.

In FIG. 10, when an “L” level RESET signal is output from the CPU 201, the AFIC 240
Each internal block is initialized and “H” level RESE
The accumulation operation starts in synchronization with the T signal. As aforementioned,
A pixel output corresponding to the peak of the accumulation level of each pixel is output to the terminal MDATA.

Symbols a and b shown in FIG. 5 indicate the magnitude of the amount of incident light, where a is a relatively large amount of incident light and b is a relatively small amount of incident light. After the start of accumulation, the CPU 201 performs AD conversion of the MDATA signal by an AD converter and monitors the accumulation level.

As can be seen from FIG. 5, the smaller the value of MDATA, the larger the accumulation amount of each pixel. Then, during the time when the amount of storage becomes appropriate, the CPU 201 stops the storage operation by changing the END signal from the “H” level to the “L” level. Further, the CPU 201 changes the accumulated time, that is, the RESET signal from “L” to “H”.
Level, and the time from when the END signal changes from “H” to “L” level is counted by an internal counter and stored as an integration time.

Next, the CPU 201 outputs a clock for reading a pixel signal to the terminal CLK. The pixel signal is sequentially supplied to the terminal SDA in accordance with the clock of the terminal CLK.
The signal is output to the TA, and the CPU 201 AD-converts the SDATA signal and stores it as pixel data in an internal RAM.

Hereinafter, the correlation calculation of the subject image signal will be described in detail.

In the apparatus according to the first embodiment, two types of correlation calculations are performed. On the other hand, similarly to the conventional focus detection device, the correlation between the first object image formed on the photosensor array 250L divided by the separator lens 239 and the second object image formed on the photosensor array 250R is obtained. The calculation is performed to obtain the defocus amount from the shift amount between the two images. On the other hand, a correlation operation is performed between the subject image at time t0 and the subject image at time t1 to obtain the moving amount of the subject image.

First, the calculation of the correlation between the first subject image and the second subject image will be described. In the following description, for the sake of convenience, the first subject image is referred to as image L, the first subject image signal is referred to as L (I), the second subject image is referred to as image R, and the second subject image signal is referred to as R (I ). .., 64 in this embodiment in order from the left. That is, each element row has 64 elements.

Hereinafter, referring to the flowchart of FIG.
The above-described correlation calculation will be described.

First, 1, 20, and 8 are set to variables SL, SR, and J as initial values (steps A1, A2). Here, SL is a variable that stores the head number of the small block element row for which the correlation is detected from the subject image signal L (I), and similarly, SR is the variable for which the correlation is detected from the subject image signal R (I). A variable for storing the head number of the block element row, and J is a variable for counting the number of movements of the small block in the subject image signal L (I).

Next, the correlation output F (s) is calculated by the following equation (step A3).

[0057]

(Equation 1)

In this case, the number of elements in the small block is 44. The number of elements in the small block is determined by the size of the distance measurement frame displayed on the finder and the magnification of the detection optical system.

Next, the minimum value of the correlation output F (s) is detected (step A4). That is, F (s) is compared with FMIN, and if F (s) is smaller than FMIN, F (s) is substituted for FMIN, and the SL and SR at that time are stored in SLM and SRM (step A5). ), And proceed to step A6. On the other hand, in step A4, if F (s) is larger than FMIN,
Proceed directly to step A6.

In step A6, 1 is subtracted from SR, and 1 is subtracted from J. If J is not 0 (step A7), the correlation calculation of the above equation (1) is repeated.
That is, the position of the small block in the image L is fixed, and the correlation is obtained while shifting the position of the small block in the image R by one element. When J becomes 0, 4 is added to SL, and 3 is added to SR to continue the correlation (step A8). That is, the correlation is repeated while shifting the small block position in the image L by four elements.
Thus, when the value of SL becomes 17, the correlation calculation ends (step A9).

As described above, it is possible to efficiently perform the correlation calculation and detect the minimum value of the correlation output. The position of the small block indicating the minimum value of the correlation output indicates the positional relationship of the image signal having the highest correlation.

Next, the correlation of the detected image signal of the block having the highest correlation is determined. First, at step A10, the values of FM and FP are calculated as shown by the following equations.

[0063]

(Equation 2)

That is, with respect to the subject image R, the correlation output when the block position showing the minimum correlation output is shifted by ± 1 element is calculated. At this time, FM, FMIN, FP
Has a relationship as shown in FIG. The image interval detected here is
If the correlation is high, the correlation output F (s) becomes 0 at the point S0 as shown in FIG. on the other hand,
If the correlation is low, it does not become 0 as shown in FIG.

Here, a correlation index Sk as shown in the following equation is obtained (step A11).

When FM ≧ FP, Sk = (FP + FMIN) / (FM−FMIN) (4) When FM <FP Sk = (FM + FMIN) / (FP−FMIN) (5) The correlation index Sk is As can be seen from the figure, when the correlation is high, Sk = 1, and when the correlation is low, Sk> 1. Therefore, it can be determined from the value of the correlation index Sk whether or not the detected image shift amount is reliable (step A12). Actually, a non-coincidence component between the image L and the image R is generated due to variations in the optical system, noise of the photoelectric conversion element, conversion error, and the like.
It does not become. Therefore, when Sk ≦ α, it is determined that there is a correlation, and the image shift amount is obtained (steps A13 and A15). S
If k> α, it is determined that there is no correlation and it is determined that AF detection is not possible (step A14). The value of the determination value α is about 2
~ 3.

Further, when the auxiliary light is turned on, since the relativity deteriorates due to the influence of the color, aberration and the like of the auxiliary light, the judgment value is increased to make it difficult for AF detection to become impossible. If there is a correlation, a two-image interval S0 between the image L and the image R is obtained from the relationship shown in FIG.

When FM ≧ FP, S0 = SRM−SLM + (1/2) · {(FM−FP) / (FM−FMIN)} (6) When FM <FP, S0 = SRM−SLM + (1/2) ) · {(FP−FM) / (FP−FMIN)} (7) The image shift amount ΔZd from the in-focus state is obtained by the following equation.

ΔZd = S0−ΔZ0 (8) Here, ΔZ0 is the image shift amount at the time of focusing, which is measured for each product and stored in the storage device. Note that the first S0 at time t0 is ΔZ1 and the second S0 at time t1 is ΔZ2.
, The future predicted S0 at time t2 is denoted by ΔZ '.

The defocus amount ΔD on the optical axis from the image shift amount ΔZd can be obtained by the following equation.

ΔD = B / (A−ΔZd) −C (9) (A, B, and C are constants determined by the optical system.) The method of obtaining the lens drive amount from the defocus amount ΔD on the optical axis is as follows. Since many proposals have been made, no detailed description is given here. For example, JP-A-64-544
In the method disclosed in Japanese Patent Publication No. 09-0909, it can be obtained as in the following equation.

ΔL = b− (a × b) / (a + ΔD) + c × ΔD (10) (where a, b, and c are constants obtained for each focal length) By driving only the photographing lens ΔL, a focused state can be obtained. In this embodiment, the movement of the subject image is obtained by the method disclosed in Japanese Patent Laid-Open No. 5-93850.

Here, the correlation calculation for obtaining the movement of the subject image will be described.

Subject images L '(I) and R' (I) at time t0
And the correlation block positions SLM 'and SRM', the correlation coefficient Sk ', and the image shift .DELTA.Z obtained by the above-described correlation operation between the two images are temporarily stored in a storage area in the CPU.

Next, at time t1, the subject image signals L (I),
R (I) is detected. First, for the signal of the image L, the time t
A correlation operation is performed on the subject image signal L '(I) at 0 and the subject image signal L (I) at time t1.

The manner in which the correlation is obtained will be described below with reference to FIGS. Here, only the method of calculating the movement amount of the image L will be described.

First, SLSTR-10 is substituted for a variable SL (step B1). This SLSTR is an element number for starting the correlation operation, and details thereof will be described later. Variable J
Is a variable for counting the number of correlations, and an initial value 20 is substituted here (step B2).

Subsequently, a correlation output F (s) is calculated based on the following correlation equation (step B3).

[0079]

(Equation 3)

Next, similarly to the above-described correlation calculation, F (s)
And FMIN are compared (step B4), and if smaller than F (s), F (s) is substituted for FMIN and the SL at that time is stored in the SLM (step B5). In this case, the number of elements of the block to be correlated Is 12, which is smaller than the number of block elements 44 when the above-described image shift amount is obtained.

Next, 1 is added to SL and 1 is subtracted from J (step B6). Until this J becomes a negative number, the correlation equation F
(s) is repeated (step B7). In this case, the correlation was obtained by changing to ± 10 elements, but this correlation range is determined by the movement amount range to be detected.

Next, the correlation is determined. Time t described above
In the same manner as when the image interval of 0 is obtained, it is obtained by the following equation (step B8).

[0083]

(Equation 4)

The correlation coefficient Sk is obtained by the above equations (4) and (5) (step B9). And
If Sk≤β, it is determined that there is a correlation, and the movement amount is obtained (step B10). This determination value β is set to a value larger than the determination value α used for obtaining the image interval at time t0 (β becomes about 7). This is because when the subject is moving, the waveform often changes, so that there is a high possibility that the correlation will deteriorate.

Next, the moving amount ΔXL of the image is obtained (step B11). In the same manner as when the image interval at the time t0 is obtained, it is obtained by the following equation.

When FM ≧ FP, ΔXL = SLM−SLSTR + (1/2) · {(FM−FP) / (FM−FMIN)} (14) When FM <FP, ΔXL = SLM−SLSTR + (1/2) ) ・ {(FM−FP) / (FP−FMIN)} (15) Then, the undetectable flag is cleared (step B1).
3) Return.

Similarly, a correlation operation is performed on the image R to obtain a correlation block position SRM and a movement amount ΔXR.
When the movement amounts .DELTA.XR and .DELTA.XL of the object images of the images L and R are obtained, the two image interval .DELTA.Z2 at time t1 is obtained from the two image interval .DELTA.Z1 at time t0 as follows.

ΔZ2 = ΔZ1 + ΔXR−ΔXL (16) In order to further reduce the calculation error, the image L at time t1
Alternatively, the correlation calculation shown in FIG. 6 may be performed again on the basis of the signal of the image R to obtain the interval between the two images and calculate ΔZ2. The image movement amount ΔZ01 between the times t0 and t1 is obtained by the following equation.

.DELTA.Z01 = | .DELTA.XR-.DELTA.XL | (17) The interval .DELTA.Z 'between the two images at the time t2 is predicted by the following equation as described above.

ΔZ ′ = ΔZ1 + {(t2−t1) / (t1−t2)} · (ΔXR−ΔXL) (18) The lens is moved at the time t2 by driving the lens by an amount based on ΔZ ′. You can focus on the subject you are doing.

On the other hand, in step B10, Sk ≦ β
If not, the process proceeds to step B12, where the undetectable flag is set.

If the moving amount ΔXR or ΔXL of the subject image is too large, it is determined that focusing is impossible, and no image shift amount is predicted. On the other hand, when the movement amount of the subject image is small and is regarded as a detection error, the movement amount is set to zero. This determination value is increased in accordance with the focal length, the subject distance, and the subject brightness when the movement amount of the subject image is predicted to be larger than the movement amount of the subject.

FIG. 10 shows the subject image signals L '(I) and R' (I) at time t0 for a moving subject.
And examples of the subject image signals L (I) and R (I) at time t1 will be described. As shown in FIG.
M 'is the head number of the block element row (44 elements) that has the smallest FMIN when detecting the image shift amount between the subject images L' (I) and R '(I) as described above.

As described above with reference to FIG. 8, when the correlation between the subject image signals at time t0 and time t1 is calculated to calculate the amount of movement of the image L and the image R, 44
An image moving amount is calculated by dividing a block row composed of elements into, for example, three.

Here, as shown in FIG.
To the third block, each having 20 elements. The first element numbers of the small blocks are SLM'2 (= SLM ') for the first block, SLM'2 (= SLM'1 + 12) for the second block, and SLM'3 (= SLM') for the third block. 1 + 24).

That is, when calculating the image movement amount of each block, first, SLSTR = SLM'1 in FIG.
To determine the image movement amount of the first block, and then SLS
The image movement amount of the second block is obtained by setting TR = SLM'2, and finally, the image movement amount of the third block is obtained by setting SLSTR = SLM'3.

The same applies to the first to third images R in the same manner.
And the image movement amount ΔZ01 between the time t1 and the time t0 is obtained by Expression (17).

The operation of the entire camera to which the first embodiment of the present invention is applied will be described below with reference to the flowchart of FIG.

The CPU 201 is a microcomputer that performs sequence control of the entire camera and various operations. When the main switch of the camera is turned on by the photographer, the CP
U201 is power-on reset and starts operation. First, initialization of the I / O port, initialization of the RAM, and the like are performed (step C1).

Then, the output of the photometric element 233 is computed by the photometric circuit in the interface IC 202, and the computation of the shutter speed and the aperture value, that is, the apex computation is performed (step C2). Subsequently, the output of the AFIC 240 is calculated as described above, and an AF calculation including a moving object prediction function is performed (step C3). This step C3 will be described later. The release button of the camera according to the present invention has two stages. It performs photometry and AF with a half-pressed first stroke (hereinafter referred to as 1R), and performs exposure with a fully-pressed second stroke (hereinafter referred to as 2R). Has been reached.

Subsequently, it is determined whether or not 1R is on (step C4). If 1R is off, step C2 is performed.
Return to On the other hand, if 1R is turned on in step C4, the lens is driven by the lens drive amount calculated in step C3 (step C5). The lens drive is known and has no direct relation to the present invention, and a description thereof will be omitted. I do. Then, it is determined whether the lens is in focus (step C).
6). This determines a focus flag described later.

If it is determined that the camera is not in focus, the process returns to step C2. If it is determined that focus is achieved, 2R
Is turned on (step C7), 2R
Is off, the process returns to step C3. If 2R is on, the diaphragm is driven to the value calculated in step C2 (step C8), and the mirror 102 is raised (step C9).

Then, control is performed so that the shutter 118 opens at the shutter speed calculated in step C2 (step C10). Next, when the shutter 118 is opened for a predetermined time, the mirror 102 is lowered (step C11), and the aperture is set to open (step C12).
8 is charged to the initial position (step C13), one frame is wound up (step C14), and the process returns to step C2 to repeat the above operation.

Next, referring to the flowchart of FIG.
The operation of the AF subroutine executed in step C3 of FIG. 11 will be described in detail.

First, in step D1, an AF detection subroutine described later is executed. This subroutine is a subroutine from the start of integration to the calculation of the defocus amount ΔZ, and includes a moving object prediction calculation.

Then, it is determined whether or not the detection is impossible by a detection impossible flag (step D2). Here, if it is determined that detection is impossible, the focus flag is cleared (step D3), and the routine returns. On the other hand, if it is determined that detection is possible, it is next determined whether or not the camera is in the continuous AF mode using the continuous AF flag (step D4). Here, if it is determined that the AF is not the continuous AF, the process proceeds to step D6 because it is not necessary to determine whether or not the first distance measurement is performed, but if it is determined that the AF is the continuous AF, First, it is determined whether or not the first distance measurement has been performed using the first calculation completion flag (step D5).

If it is determined that the distance measurement is the first distance measurement, the process proceeds to step D3. If it is determined that the distance measurement is the second time, the process proceeds to step D6 to calculate the defocus amount. I do.

In step D6, a defocus amount is calculated from the defocus amount calculated in step D1 based on equations (8) and (9). Subsequently, the calculated defocus amount is compared with the focus determination value (step D7). This determination value is a value obtained based on the permissible circle of confusion. If it is within the determination value, it is already in focus.

If it is determined in step D8 that the defocus amount is within the allowable focus range, it is not necessary to drive the lens, so that the focus flag is set (step D9), and the process returns. If it is determined that the in-focus range is not within the allowable range, the in-focus flag is cleared (step D10).
), Calculate the amount of lens drive necessary for focusing (step D11), and return.

Next, the operation of the subroutine for AF detection will be described in detail with reference to the flowcharts of FIGS.

First, the integration operation is reset (step E1). Then, it waits until the integration of the AFIC 240 is completed (step E2). Specifically, the process waits until the value of MDATA becomes a proper value at which the integration is completed. Next, data of all elements (pixels) is read out for each pixel (step E).
3). The output of the AFIC 240 is an analog value, and is converted into a digital signal by an A / D converter in the CPU 201 every time one pixel is read, and stored in a predetermined storage area.

Next, non-uniformity correction is performed on the obtained subject image signal (step E4). This is for correcting subtle variations in sensitivity for each pixel that occur during manufacturing, and unevenness in illuminance of the re-imaging optical system in the AF unit 210.
Correction is made so that the output of another pixel is matched with the pixel with the lowest sensitivity among all the pixels. The correction coefficient is adjusted for each product and stored in the EEPROM 237. The details are described in Japanese Patent Application Laid-Open No. 5-93850, and are omitted here.

Next, MDATA which is a value indicating the integration level
Is stored in a predetermined storage area (step E41).
This is for use in step E42 described later.

Subsequently, it is determined whether the moving object mode (mode for performing moving object prediction) is selected (step E5), and whether the self-timer shooting mode is selected (step E6), and the remote control shooting mode is shot. Is determined (step E7), whether landscape photography mode is selected (step E8), whether night landscape photography mode is selected (step E9), and portrait photography mode is selected. Is determined (step E10),
It is determined whether the camera shake prevention mode has been selected (step E11), and it is determined whether the auxiliary light LED 242 has been turned on during the current integration operation (step E12).

In the above eight kinds of judgment items, the continuous AF flag is set only when it is judged that the moving object mode is selected, all the other photographing modes are not selected, and the auxiliary light is turned off. (Step E1)
3). If this flag is set, moving object prediction AF is performed hereinafter. On the other hand, if the determination result is other than that, the continuous AF flag is cleared (step E1).
4) The process moves to step E16, where the moving object prediction AF is not performed.

The reason why the auxiliary light is determined in step E12 is that when the auxiliary light LED 242 is on, the subject is darker, so that the AF detection accuracy is lower than when the subject is brighter, and the error in the moving object prediction calculation becomes larger. Because. Basically, it is unsuitable for photographing a moving object in a dark situation because the shutter speed becomes slow.

Subsequently, it is determined whether the first image shift calculation has been completed (step E15). This is to determine the first calculation completion flag which is set and cleared in steps E18 and E10 described later. This flag indicates whether or not the first image shift amount has been calculated, and the initial value has been cleared in advance in step C1 in FIG. 1
If the first image shift calculation is not completed, the correlation calculation described with reference to FIG. 6 is performed to calculate the image shift amount ΔZ1 (step E16).

Subsequently, it is determined whether the image shift amount ΔZ1 has been calculated (step E17). That is, a detection impossible flag that is set and cleared in steps A14 and A15 in FIG. 6 is determined. If it is determined in step E17 that detection is impossible, the first calculation completion flag is cleared (step E18), the detection impossible flag is set (step E19), and the routine returns. On the other hand, if it is determined in step E17 that detection is possible, the first calculation completion flag is set (step E20), and the routine returns. If it is determined that the lens cannot be detected, the process shifts to lens scanning in the lens driving subroutine to search for a position of a lens that can be detected.

On the other hand, if it is determined in step E15 that the first image shift amount calculation has been completed, a second image shift amount calculation is performed. That is, first, the first calculation completion flag is cleared for the next time (step E21). Next, the first and second integration levels are corrected to make the integration amounts substantially the same (step E42). This will be described later. Then, the same correlation calculation is performed as in the first time to calculate the image shift amount ΔZ2 (step E22).

Subsequently, it is determined whether or not the image shift amount ΔZ2 has been calculated, as in the first case of step E17 (step E23).

If the calculation has not been performed, the process proceeds to step E40, and the calculated ΔZ1 is changed to the time t2.
Is the image shift amount ΔZ ′. If you can calculate,
The amount of movement of the image L of the first block is calculated according to the flowchart of FIG. 8 by performing the correlation calculation of the image L of the first block described in FIG. 10 (step E24).

Subsequently, a correlation calculation of the images L of the second and third blocks is performed to calculate the moving amounts of the images L of the second and third blocks, respectively (steps E25 and E26). continue,
It is determined whether the calculated amount of movement of the image L of the three blocks is larger than a predetermined first determination value (step E27).

This first determination value is a relatively large value. Step E27 is performed when the object deviates from the distance measurement area in the viewfinder and the distance cannot be measured, or when the moving speed of the object is too high and the moving object is detected. It is provided to detect a case where focusing cannot be performed even if it is predicted. If the calculated amount of movement of the image L is larger than the predetermined first determination value, it is determined that the moving object is unpredictable, and the process shifts to step E38 to be described later.

Similarly, the movement amount of the image R is calculated (steps E28, E29, E30) and the calculated movement amount is determined (step E31). If the calculated amount of movement of the image R is larger than the first determination value, the process proceeds to step E38 as a moving object cannot be predicted. The first calculated above
Based on the reliability index of the third block, a block having the highest correlation is selected. That is, a block having the smallest reliability index Sk is selected (step E32).

Next, in the selected correlation block, a detection impossible flag is determined (step E33). If the selected block cannot be detected, the process shifts to step E38 to perform a non-detection process. If the selected block can be detected, the image movement amount ΔZ01 moved between the first integration and the second integration is obtained based on the equation (17) (step E).
34). Then, it is determined whether or not the subject is moving (step E35). The moving object flag output from step E35 is determined (step E36). If it is determined that the subject is moving, a future image shift amount ΔZ ′ is predicted based on equation (18) (step E36). E3
7). Then, the undetectable flag is cleared (step E39), and the routine returns.

On the other hand, if it is determined that the subject is stationary, it is not necessary to predict the moving object, so that ΔZ 'is set to the image shift amount ΔZ2 calculated in step E22 (step E38), and the routine returns.

The operation of the subroutine integral reset executed in step E1 in FIG. 13 will be described below with reference to the flowchart in FIG.

First, the value of the integration time timer is read as the current integration time (step F1). This timer may be configured to stop counting of the timer, for example, in synchronization with the integration end signal of the AFIC 240. Next, the value of the integration interval timer is read as the previous and current integration intervals (step F2).

Next, the integration time timer and the integration interval timer are reset (steps F3 and F4). Finally, AFI
At the same time as starting the next integration of C240, the integration time timer and the integration interval timer are started (step F5), and the process returns.

FIG. 16 is a diagram showing the concept of integration level correction according to the first embodiment. FIG.
1A and 1B schematically show examples of the sensor data of the line chart already described with reference to FIG.
6 (a) and 6 (b) correspond to the uncorrected views of FIGS. 2 (b) and (c), and the lower views of FIGS. 16 (a) and (b) correspond to FIG. d) and (e) correspond to the corrected figures. Here, the value of the first MDATA is M1, and the second MDATA is
The value of ATA is M2, and an example of M1 <M2 is shown. As can be seen from the figure, since the integration amount of the first sensor data is larger than the integration amount of the second sensor data, the second sensor data is left unchanged.
The first sensor data is corrected on software to make the apparent integration level the same.

Here, FIG. 16 (c) shows and explains the correction method. It should be noted that the vertical axis in the figure indicates the correction amount, and the horizontal axis indicates the output of each pixel. That is, the larger the pixel output is, the larger the correction is, and the smaller the pixel output is, the smaller the correction is.

Specifically, when the integration amount of the first sensor data is larger than the integration amount of the second sensor data, the obtained correction amount is subtracted from the first sensor data. to correct. Conversely, when the integration amount of the first sensor data is smaller than the integration amount of the second sensor data, the obtained correction amount is added to the first sensor data to perform the correction. Here, the relationship between the correction amount and the output of the pixel is linear, and the coefficient A is shown in FIG.
(D).

FIG. 16D shows the first and second MDAs.
6 is a table showing a relationship between an absolute value B of a difference between TA values and a correction coefficient A, which is obtained experimentally. This table is CPU
201 is stored as a table in the ROM in
After calculating the absolute value B of the DATA difference, the correction coefficient A is obtained by referring to the absolute value B.

FIG. 17 is a flowchart showing a method of correcting the integration level described in FIG. 16, and is a subroutine executed in step E42 in FIG.

First, MDATA indicating the integration level of the first sensor data stored in step E41 of FIG.
Is read from a predetermined storage area (step G).
1). Then, the MDATA value M2 indicating the integration level of the second sensor data is also read from the predetermined storage area (step G2).

Next, the absolute value B of the MDATA difference is calculated (step G3).

B = | M1−M2 | (19) Then, by referring to the table described in FIG. 16D from the ROM table, a correction coefficient A is obtained (step G4), and the correction coefficient A is first stored in a predetermined storage area. The value of the first pixel data of the stored first sensor data is read into X (step G).
5).

Next, the correction amount Y is calculated as described with reference to FIG. 16C (step G6).

Y = A × X (20) Subsequently, the magnitude relationship between M1 and M2 is determined (step G).
7) In the case of M1 ≧ M2, since the integration amount of the first sensor data is smaller than the integration amount of the second sensor data, the obtained correction amount Y is added to the first sensor data. To correct the first sensor data (step G8).

Conversely, when M1 <M2, the integration amount of the first sensor data is greater than the integration amount of the second sensor data. , The first sensor data is corrected (step G9), the sensor data after each correction is set to Z, and the process proceeds to step G10.

Next, the corrected sensor data Z is stored in the original predetermined storage area (step G10). Then, it is determined whether or not the correction has been completed for all pixels (here, 64 × 2 = 128 pixels) (step G11). If the correction has been completed, the process returns. If not, the output of the next pixel is set to X. Read (step G12), return to step G6,
The above processing is performed for all pixels.

By configuring the first embodiment as described above, even if the integration level is not the same at the end of the integration, the sensor data is corrected so as to be the same. Performance can be improved. It is needless to say that modifications can be made without departing from the spirit of the first embodiment of the present invention. For example,
The current sensor data may be corrected without changing the previous sensor data.

Next, a second embodiment of the present invention will be described.

In the second embodiment, the integration time when integrating the current sensor data is set to be the same as the integration time when integrating the previous sensor data. The same. This is because a change in the distance by which the subject moves between the previous and current integration intervals has a small effect on the change in the subject brightness on the AFIC 240. It is the same.

The second embodiment differs from the first embodiment only in the subroutine portion for AF detection, and among them, only the portion corresponding to FIG. 13 is different, so only that portion will be described, and other description will be omitted. I do.

FIG. 18 is a flowchart showing a subroutine for AF detection according to the second embodiment. FIG.
Steps that perform the same processing as in step 3 are assigned the same reference numerals as in FIG. 13 and will not be described.

First, it is determined whether the first calculation has been completed after the integration reset, that is, whether the current processing is the first calculation (step E43). If it is determined that the calculation is the first one, after waiting until the integration is completed, the integration time at that time is stored in a predetermined storage area (step E44).
Move to step E3.

On the other hand, if it is determined in step E43 that the calculation is the second one, the process waits until the current integration time reaches the first integration time stored in step E44 (step E43).
45) When the integration time becomes the same as the first integration time, the integration is forcibly terminated (step E46), and step E is performed.
Move to 3.

In the second embodiment, the integration level correction subroutine of step E42 described in the first embodiment is executed. Can be corrected in step E42, but step E42 may be omitted in the second embodiment.

By configuring the second embodiment as described above, it is possible to control the integration levels to be the same.

The embodiments of the present invention have been described above, but it goes without saying that modifications can be made without departing from the gist of the present invention. For example, Japanese Unexamined Patent Application Publication No.
Although the technique of Japanese Patent Application No. 93850 was used, it is a matter of course that the moving object can be detected without being limited to this.

The above embodiment of the present invention includes the following inventions.

(1) Focus detection means comprising a photoelectric conversion element for outputting a focus detection signal for detecting the focus state of the photographing lens, and predictive calculation based on the output of the focus detection means, and Means for performing focus adjustment so as to focus on a subject moving in the optical axis direction, and a camera having a moving object predicting function, the charge of the photoelectric conversion element when the latest focus detection signal is detected. A focus detection device for a camera, comprising: control means for controlling an accumulation level to be substantially the same as a charge accumulation level of the photoelectric conversion element when the past focus detection signal is detected.

(2) The control means controls the charge accumulation level of the photoelectric conversion element when the latest focus detection signal is detected and the charge accumulation level of the photoelectric conversion element when the past focus detection signal is detected. The focus detection device for a camera according to (1), further including a correction unit that corrects at least one of the focus detection signals so that the level becomes the same as the level.

(3) The control means controls the charge storage time of the photoelectric conversion element when the latest focus detection signal is detected, and the charge storage time of the photoelectric conversion element when the past focus detection signal is detected. The automatic focusing device for a camera according to (1), wherein the control is performed so that the time is substantially the same as the time.

(4) The control means controls the charge accumulation level of the photoelectric conversion element when the latest focus detection signal is detected, and the charge accumulation level of the photoelectric conversion element when the past focus detection signal is detected. The level and the level are the same, the correction means for correcting at least one of the focus detection signals, the charge accumulation time of the photoelectric conversion element when the latest focus detection signal is detected, and the past The focus detection device for a camera according to (1), wherein the control is performed such that a charge accumulation time of the photoelectric conversion element when a focus detection signal is detected is substantially the same.

(5) The photoelectric conversion element includes monitor means for outputting a charge accumulation level, and the correction means makes the correction based on the output of the monitor means. Camera focus detection device.

[0158]

As described in detail above, according to the present invention,
It is possible to provide a focus detection device for a camera in which the charge accumulation level of the sensor data relating to the moving object prediction calculation is set to be the same and the error of the image movement amount detection is reduced.

[Brief description of the drawings]

FIG. 1 is a conceptual diagram showing a configuration of a camera focus detection device according to a first embodiment of the present invention.

FIG. 2 is a diagram conceptually showing a difference between an integration level and a difference between sensor outputs.

FIG. 3 is a block diagram showing in detail a control system of a camera system to which the camera focus detection device according to the first embodiment is applied.

FIG. 4 is a diagram showing a detailed configuration of an AFIC 240.

FIG. 5 is a timing chart showing in detail a signal state of each terminal in FIG. 4;

FIG. 6 is a flowchart illustrating a correlation operation according to the first embodiment.

FIG. 7 shows FM, when the subject image R is shifted by ± 1 element from the block position showing the minimum correlation output,
It is a figure which shows the relationship of FMIN and FP.

FIG. 8 is a flowchart illustrating a sequence of calculating a moving amount of the image L;

FIG. 9 is a diagram for explaining how to obtain a correlation.

FIG. 10 shows examples of subject image signals L '(I) and R' (I) at time t0 and subject image signals L (I) and R (I) at time t1 for a moving subject. FIG.

FIG. 11 is a flowchart for explaining the operation of the entire camera to which the first embodiment of the present invention is applied;

FIG. 12 is a flowchart showing in detail an operation of an AF subroutine executed in step C3 of FIG. 11;

FIG. 13 is a flowchart showing an operation of a subroutine for AF detection in detail.

FIG. 14 is a flowchart showing an operation of a subroutine for AF detection in detail.

FIG. 15 is a flowchart showing an operation of a subroutine integral reset executed in step E1 of FIG. 13;

FIG. 16 is a diagram showing a concept of integration level correction according to the first embodiment.

FIG. 17 is a flowchart illustrating a method of correcting an integration level described with reference to FIG.

FIG. 18 is a flowchart illustrating a subroutine for AF detection according to the second embodiment.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 focus detection unit 2 integration level control unit 3 image movement amount detection unit 4 moving object prediction unit

Claims (3)

[Claims] 1. A detecting means including a charge storage type photoelectric conversion element, which repeatedly detects a focus adjustment state of a photographing lens and outputs a focus detection signal in time series, a charge storage level of a latest focus detection signal, Control means for controlling the charge accumulation level of the next focus detection signal to be substantially the same; and moving in the optical axis direction of the photographing lens in response to a time-series signal output from the detection means. A focus detection device for a camera, comprising: focus adjustment means for repeatedly performing a focus adjustment prediction operation for focusing on a subject and adjusting the focus of the photographing lens based on the result. 2. A charge storage type photoelectric conversion means for repeatedly photoelectrically converting an input subject image and outputting detection signals in time series, and storing a plurality of recent detection signals including a latest detection signal. Storage means, control means for controlling the charge accumulation levels of the plurality of detection signals output in time series to be substantially constant, and an optical axis of a photographing lens based on the plurality of stored detection signals. And a calculation means for performing calculation for focus adjustment so that the subject moving in the direction is also focused. 3. The control means controls the charge accumulation time of the photoelectric conversion means to be substantially constant, or corrects at least one of the detection signals so that the level of the detection signal is substantially constant. The focus detection device for a camera according to at least one of claims 1 and 2, wherein the correction is performed so that