JPH01284183A - Automatic focusing video camera - Google Patents

Abstract

PURPOSE:To promptly correct the dislocation between a lens position and a focused position by gradually increasing the quantity of the fine movement of the lens position when the lens position is finely moved to a side approaching the focused position. CONSTITUTION:The title device is equipped with an evaluated value detecting means 26 to detect the levels of the high frequency components of an image picked up video signal obtained from an image pickup element passed through a first HPF9 and a second HPF10 as focus evaluated values and a focus control means 27 to change the distance between a focus lens and the image pickup element, once fix the distance between both the focus lens and the image pickup element and complete focusing action when either of the focus evaluated values becomes the maximum evaluated value, and execute fine adjusting action to finely move the focus lens in the direction of an optical axis by dividing the fine movement of the focus lens into plural fine movements and make the lens position follow the focused position when the focus evaluated value is changed. Further, the quantity of the fine movement of the lens position is gradually increased when the lens position is finely moved to the side approaching the focused position. Namely, the evaluated value is processed by a microcomputer 26 according to a software way, and the microcomputer 26 sends a command to the focus motor control circuit 27 based on the processed result. Thus, the dislocation between the lens position and the focused position can be promptly corrected when the distance between the device and an object is slightly changed.

Description

DETAILED DESCRIPTION OF THE INVENTION (A) Field of Industrial Application The present invention is based on a captured video signal obtained from an image sensor.
This invention relates to a video camera that automatically adjusts focus.

(b) Conventional technology In the autofocus device of a video camera, the method of using the video signal itself from the image sensor to evaluate the focus control state has no physical variation and the depth of field is limited. It has many advantages, such as being able to accurately focus on shallow or distant objects, and is mechanically extremely simple, requiring no special sensor for autofocus. Japanese Unexamined Patent Publication No. 105978/1983 (H04N
5/232) discloses an example of an autofocus device as described above.

In the conventional technology, the high frequency component level of the captured video signal is controlled by A/D within the range of the focus area set at the center of the screen.
This converted data is integrated for each field in the adjustment circuit, and the digital data for one field is held as a focus evaluation value.The focus evaluation value is always determined by comparing it with the evaluation value one field before. The focus motor is driven and controlled in the direction where the maximum value is achieved.

In this type of autofocus device, the lens is usually held at the in-focus position and the evaluation value changes over time, and if the amount of change exceeds a certain level, it indicates that the subject has changed. The camera is configured to resume autofocus operation as soon as possible.

Here, if the subject moves or changes significantly, it is effective to restart all focusing operations from the beginning.
If the change in the subject is extremely small, use Focus L.
It is faster to maintain the in-focus state by correcting slight deviations by slightly changing the lens in the front and rear directions of the optical axis to the extent that no change in the angle of view occurs, and it is possible to perform focusing operations that suppress changes in the angle of view. In Japanese Patent Application No. 61-273212, the present applicant has previously proposed an autofocus device that corrects deviations due to minute movements as described above.

The technology of the prior application includes a coarse adjustment means that largely displaces the focus lens in the optical axis direction and uses mountain climbing control to reach a focus position where the focus evaluation value reaches the maximum evaluation value at the top of the mountain; After the focusing operation, the focus lens is slightly moved in the front and back direction of the optical axis, for example, in the front direction.
Move the focus tep and compare the focus evaluation values before and after the fine movement. If the focus evaluation value tends to decrease, move the focus 25 steps in the opposite direction and compare the focus evaluation values again. If the value before the fine movement is larger, move forward. l Move the lens 5 steps, return it to the light position, confirm this position as the in-focus position, hold the focus evaluation value here as the maximum evaluation value, and confirm that there has been no change in the subject. have the means. During the confirmation operation using this fine adjustment means, it was observed that the subject moved slightly, that is, when the optical axis was slightly moved forward or backward, the focus evaluation value after the fine movement was larger than before the fine movement. In this case, the maximum evaluation value is updated with the focus evaluation value after this slight movement, this position is set as the in-focus position, and a similar confirmation operation is executed from this position. Therefore, slight deviations can be corrected by repeating the above-described slight movement in one direction.

(c) Problems to be Solved by the Invention In the prior art, the deviation correction by the fine adjustment means is
The maximum evaluation value is updated with the focus evaluation value after fine movement, and the deviation is slightly large because it is finely moved by 15 steps for each field, but the coarse adjustment means is restarted and the focusing operation is restarted from the beginning. If the deviation is not large enough to cause a change in the angle of view, it will take an extremely long time.

(2) Means for solving the problem: An evaluation value detection means for detecting the high-frequency component of the imaged video signal obtained from the image sensor as a focus evaluation value, and a distance between the focus lens and the image sensor, which can be changed. When the focus evaluation value reaches the maximum evaluation value, the distance between the two is temporarily fixed and the fine focusing operation is finished, and when the focus evaluation value changes, the focus lens is moved slightly in the optical axis direction multiple times to adjust the lens position. The present invention is characterized by comprising a focus control means for performing a fine adjustment operation to follow the in-focus position, and increasing the amount of fine movement in stages when the lens position is slightly moved toward the in-focus position!.

(E) Effect Since the present invention is constructed as described above, even if the distance to the subject changes slightly, it is possible to quickly correct the deviation from the focal position.

(f) Examples An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a circuit block diagram of this embodiment.

(1) is the video camera section, and the focus lens (2
) and a focus motor (4) that drives a focus ring (3) that supports the focus ring (3) and moves it forward and backward in the optical axis direction, and an end point switch (5) that detects the drive limit of the focus ring (3).
), an aperture mechanism (6) that controls exposure, and this aperture mechanism (
6), and an imaging circuit (8) having a solid-state image sensor (CCD) that converts subject light into a captured video signal.

The brightness in the imaged video signal obtained by the imaging circuit (8) is determined by a first bypass filter (HPF) (9) with a cutoff frequency, a second HPF (10), and a low-pass filter (LPF). (11) and synchronous separation circuit (12)
) will be sent to.

The vertical synchronization signal (VD) and horizontal synchronization signal (HD) separated from the luminance signal by the synchronization separation circuit (12) are supplied to the switching control circuit (13) in order to set a sampling area. The switching control circuit (13) uses a fixed oscillator output as a clock to drive the vertical/horizontal synchronizing signals (VD) and (HD) as well as the COD to drive a rectangular first switch in the center of the screen as shown in Figure 2. Sampling area (A1) and
A second sampling area (A2) which includes this area (A1) and whose area is four times that of area (A1) and this area (A
2) The third to sixth sampling areas (A 3
) (A 4) (A 5) (A 6) can be set, the selection signal (Sl) is output to the subsequent selection circuit (15),
Also, the output of the first HPF (9) and the second HPF (10) is 1.
Switches for each field, and then turns L once every 32 fields.
Switching signal (Sl) for selecting the output of P F (11)
is output to the switching circuit (14).

The switching circuit (14) receives the switching signal (Sl) and selects the output of the first HPF (9) and the output of the second HPF (lO) for each field.
The output is selected and output to the subsequent selection circuit (15), and the L P F (11) output is selected once every 32 fields and output to the selection circuit (15).

Based on the selection signal (Sl), the selection circuit (15) selects the output selected by the switching circuit (14) to the integration circuits (16), (17)...(21) according to the sampling area. Output. That is, the first sampling area (A1)
The filter outputs related to the second sampling area (A2) are sent to the integration circuit (16), and the filter outputs related to the second sampling area (A2) are sent to the integration circuit (17). ), the filter outputs are sent to integration circuits (18), (19), (20), and (21), respectively.
) is output.

The integration circuit (16) includes an A/D converter (22) and an adder (2
3), it is composed of a memory circuit (24), and the A/D converter (22) sequentially A/D converts each filter output that passes through the selection circuit (15), and sends it to the adder (23). Output. The adder (23) constitutes a digital integrator together with the A/D converter (22) at the front stage and the memory circuit (24) at the rear stage, and the output of the memory circuit (24) and the A/D converter (22)
The outputs are added and the addition result is sent to the memory circuit (24) again.
). The memory circuit (24) is reset for each field and holds one field of the output of the adder (23), that is, the digitally converted value of the level of the luminance signal that has passed through the filter, for the first sampling area (A1). become.

Regarding the integration circuit (17) (1g)...(21),
It has exactly the same configuration as the integrating circuit (16), and the memory circuit built into each integrating circuit stores the level 1 of the luminance signal that has passed through the filter selected in the current field for each sampling area. The integral value for the field can now be retained. The integral value of each of these memory circuits is further sent to the subsequent memory circuit (25).
It is stored collectively.

1st HPF (9), 2nd HPF (10) and LPF
The permissible passing range according to each of (11) is 600kHz or more,
It is set to 200kHz or more and 2.4MHz or less, but in reality it is 600kHz to 2.4MHz, 200k.
It can be set with a BPF having a passband of Hz to 2.4 MHz and O to 2.4 MHz. 2.4MH at this time
z is an extremely high frequency that has little to do with the luminance signal, so even if L P F (11) is omitted, it will be referred to as yaw. The high-frequency or low-frequency components of the luminance signal that have passed through either the VGI, the second HPF (9), (10), or the LPF (11) are digitally integrated for one field, and each sampling This will be stored in the memory circuit (25) as the evaluation value of the current field for each area. Here, among the integral values stored in the memory circuit (25), the integral value of the low frequency component when the LPF (11) is selected is used as the exposure evaluation value for exposure control and also as the first HPF (
9) Alternatively, when the second HP F (10) is selected, the integral value of the high frequency component is processed by the microcomputer (26) at the subsequent stage as a focus evaluation value for focus control.

These evaluation values are processed by software by the microcomputer 26), and based on the processing results, a command is issued to the focus morph control circuit (27) to control the focus morph (
4) to advance and retreat the focus lens (2),
The autofocus operation is executed so that the focus evaluation value is maximized, and a command is issued to the iris and motor control circuit (28) to drive the iris motor (7) and control the aperture mechanism (6).
), and automatic exposure adjustment becomes possible so that the exposure evaluation value becomes a predetermined value.

Next, referring to the flowchart in Figure 3, the microcontroller (2
The main routine of 6) autofocus operation and auto iris operation will be explained.

When the video camera enters the operating state, the microcomputer (26) executes the main routine shown in FIG.

First, at 5TEP (30), the integral value for one field in each sampling area in the current field is read into the microcomputer (26) from the memory circuit (25), and S T
The rotational direction of the zoom motor (101) is detected at EP (31). Here, the zoom motor (101) is a zoom ring (102) rotatably arranged in the radial direction on the lens barrel projecting forward of the video camera body (1).
), and this zoom ring (10
2) supports a zoom lens (not shown) which is a variable magnification lens, and this lens moves forward and backward in the optical axis direction according to the rotation of the zoom ring (102), and changes from the telephoto (τeke) flI range to the intermediate (middle) range. Wide angle via area (ν1de)
1. Normally, the photographer presses the zoom drive switch (not shown) located on the camera body.
Move the zoom motor (1e) in the 1e direction and Wide direction.
01) in either direction, a desired zoom position can be obtained.

Next, the counter (AECNT) provided to perform the autofocus operation and the auto iris operation in a time-sharing manner is decremented, that is, 5 TEP (32)) is used to determine whether the counter value is 0 or not. (3
3)) If the counter value is not 0, autofocus operation is executed, and only when the counter value is 0, auto iris operation is executed. Furthermore, Zoono by 5TEP (31),
The 5THP (34) confirms that the detection result of the rotational direction of the motor (101), that is, the zoom direction, is in the wide-angle direction,
Furthermore, in S T E P (100), the operating mode hood (MODE) for AF, which will be described later, is 11.
If the normal autofocus operation has been completed and it can be determined that the peak of the focus evaluation value has been reached, the AP routine for executing the basic autofocus operation is skipped without being executed. This is because when the zoom mechanism moves toward the wide-angle side, the depth of field gradually deepens, so if the focus is reached before the zoom operation, autofocus will continue while zooming toward the wide-angle direction. There is no need to perform the operation again; on the other hand, if unnecessary autofocus operations are repeated due to fluctuations in the focus evaluation value due to changes in the angle of view, the screen will become unsightly, so this is to prevent this. If the zoom direction is not the wide-angle direction, or if the camera is not in focus just before zooming even if it is in the wide-angle direction,
The AF scene (35) is executed.

When the AF period ends, the counter (A E CNT
) is subtracted by 1 and it is determined whether the counter value becomes 0 or not (STEP (36)), and if it becomes 0,
A control signal is issued from the microcomputer (26) to the switching control circuit (13), and in response to this, the switching circuit (14)
A switching signal (Sl) is issued to select (11),
The selection of L P F (11) is made (STEP (
37)) In this way, when P F (11) is selected, it waits for the evaluation value obtained by this selection to be read.

On the other hand, when auto iris operation is selected at 5TEP (33), the AE routine (38) which is the basis of auto iris operation is executed, and then the counter (AECNT
) to the initial state (STEP (39)), select the filter as the first HPF (9) (STEP (40>
), waits for the evaluation value of the next field to be integrated.

Here, the initial state of the counter (AECNT) refers to a state where the initial value "32" is set in order to calculate the exposure evaluation value based on the luminance signal that passes through L P F (11) every 32 fields.

Next, the autofocus operation according to the present invention will be explained according to the flowchart shown in FIG.

Autofocus operation is selected in STEP (33) of the main routine, and if the zoom drive switch is not moved to the wide-angle side, an AF sequence (35) is executed. This AF screen is shown in a flowchart as shown in FIG.

First, in the 5TEP (41), the zero-order STEP is used to calculate the focus evaluation value and its relative ratio from the integral value in each sampling area stored in the memory circuit (25).
P (42) After the end point switch of the focus ring and the zoom position, which is the zoom range of the zoom mechanism, are confirmed at (42) and (43>), mountain climbing control is entered.

The mountain climbing control includes an evaluation value stability confirmation routine (45), a direction determination routine (46), a mountain climbing routine (47), an apex return routine (48), and an evaluation value fluctuation routine (49).
), these mode selections are made by specifying the operating mode code (MODE) from 0 to 4 in 5TEP (44) according to each condition, and usually the evaluation value stability confirmation routine The routine is executed in the following order: (45) → direction determination routine (46) → mountain climbing routine (47) → apex return routine (48) → evaluation value fluctuation monitoring routine (49).

After each routine is completed, the HPF is switched at 5TEP (50). That is, if the switching circuit (14) has executed the AF sequence for the current field using the first HPF (9), the second HPF (
10), and vice versa, the second HP F(10)
A control signal is supplied to the switching control circuit (13) to execute switching from the first HPF (9) to the first HPF (9). Therefore, while the AF screen (35) is selected by 5TEP (33) (34) in the main routine, the filter selects the first HPF (9) and the 28th HPF (10) for each field.
) will be switched alternately.

Next, the structure, operation, and advantages of each operation that can be executed in the AF screen shown in FIG. 4 will be explained individually. First 5TE
The calculation operation of the focus evaluation value of P(41) and its relative ratio will be explained with reference to the flowchart of FIG. 5.

First, among the integral values stored in the memory circuit (25) in the current field, the integral values integrated by the integrating circuits (16) and (17), that is, the first and second sampling areas (A1) ( A2) Integral value DATA(ri,
DATA (2) is the first HPF (9) and second HPF (1
0) to determine which HPF was used for extraction.
It was done at T E P (51), and 5 TEP (52)
53>, if it is based on the 1st HPF (9), DAT
A(1) and DATA(2) are memories A(1) and A, respectively.
(2), and if it is based on the second HP F(10), it is assigned to memories B(1) and B(2). However, as described above, the second sampling area (A2) has an area four times that of the first sampling area (AI), and also includes the first sampling area (A1).

Next, at 5TEP (54), the focus evaluation values of the first and second sampling areas <AI) (A2) of the previous field are transferred to the memories Y(1) and Y(2).

In S T E P (55), memories A(1) and A(2
), B(1), and B(2), the focus evaluation values X(1) and X(2) of the first and second sampling areas (A1) and (A2) of the current field are calculated. Here, the focus evaluation value X(1) of the first sampling area (AI) is the sum of the value of memory A(1) and the value of memory B(1), that is, the focus evaluation value X(1) of the first HPF( The change cloth is the sum of the latest integrated value when using 9) and the latest integrated value when using 2nd HP F(10), and similarly, the focus evaluation value x(2) is the memory A(2) For example, as shown in FIG. 6, DATA(1) is 1111.0 for each field as shown in FIG. bl.

a2. b2. ...and the data is imported, D
c as ATA(2) 1. d1. c2. d
2. ...and the data is imported (however, a L
m L a L ” is the integral value of the first sampling area (A1) by the output of the first HPF (9), b 1. b
2. b3. ... is the integral value of the first sampling area (A1) by the second HP F (10) output, c
Lc2°c3. ... is the integral value of the second sampling area (A2) based on the output of the first HPF (9), d1. d
2. d3. ... is the integral value of the second sampling area (A2) by the second HP F(10) output), and the focus evaluation value X(1) is calculated by a for each field.
l+b1. bl+a2. a2+ b2. b2+ A3.
・”, and the focus evaluation value
．． ...will change. Therefore, the focus evaluation value for each sampling area is the sum of the integral value of one of the HPF outputs in the current field and the integral value of the other HPF output in the previous field, and the integral value of the odd field and even field is 1. As a result, the effects of noise and variations in the evaluation value for each field due to ink race etc. can be alleviated, and the evaluation value becomes stable.

In 5TEP (56), relative ratios R(1) and R(2) in each sampling area are calculated. Relative ratio R〈1
) is the ratio of the value of memory A(1) to the value of memory B(1) B(1)/
A(1), that is, the latest integral value when using the first HPF (9) within the first sampling area (A1), and the second H
It is the ratio of the latest integral value when using P
The ratio of the values of 2) is B(2)/A(2).

The purpose of the end point switch processing routine (42) is to detect whether the focus ring (3) has reached the periapsis or ω point, which is the rotation limit, and when it reaches the periapsis or ψ point, A near-point side reflector and a ψ side reflector are respectively fixed to the focus ring (3) so as to face the end point switch (5) fixed to the cabinet of the video camera.
The light from the light emitting element in the end point switch (5) is reflected by each reflecting plate, and the light is received by the light receiving element in the end point switch (5), so that each reflecting plate faces the end point switch (5). It detects the position, that is, reaching the near point or the ω point, and furthermore, the near point and the ψ point can be distinguished by the rotation direction of the focus motor. As soon as the near point or ω point is reached, the focus motor (4) is controlled to rotate in reverse.

Next, there are five routines that form the core of autofocus operation: evaluation value stability confirmation routine, direction determination routine,
The mountain climbing routine, the apex return routine, and the evaluation value fluctuation observation routine will be explained in order.

First, the evaluation value stability confirmation routine is executed according to the flowchart shown in FIG. Note that this evaluation value stability confirmation operation can be executed when the power is turned on when the operation mode hood (MODE) becomes ``0'', or when the subject changes and the autofocus operation is restarted again.

First, the first threshold value (THR
I) has been defined, and in 5TEP (62), the focus evaluation values X(1) and Y(1) in the first sampling area (A1) in the current field and the previous field are compared, and the difference is defined first. If the evaluation value is larger than the 115th value (THRI), it is assumed that the evaluation value is not stable and the evaluation value is 5TEP (63
) to reset the counter (INN) described later, and if the AF sequence can be executed again in the next field, set the operation mode hood (MODE) to MO to execute this evaluation value stability confirmation routine again. Leave this as is and exit this routine.

The difference between the focus evaluation values X(1) and Y(1) is the vg1 threshold (TH
Rl) or less, the counter (INN> is incremented at STEP (64), that is, 1 is added to the count value, and the count value of the counter (INN) becomes '5' at 5TEP (65). ”, that is, whether this state has continued for 5 fields. If the count value of the counter (INN) has not reached “5”, the evaluation value is The stability confirmation routine is ended, but since the operation mode hood (MODE) remains at 10'', this evaluation value stability confirmation routine is executed again in the next AF scene.

Focus evaluation value X (1), Y4 at S T E P (65)
1) is smaller than the 11th Il value (THR1) for 5 fields, 5TEP (6
6), reset the counter (INN) and 5TEP (
In step 67), initial settings for autofocus operation are made. In other words, AF in the next field at 5TEP (67)
In the screen, in order to execute the direction determination routine,
Change the operation mode code (MODE) to 1", initialize the rotation direction of the focus motor (4) to the ω point direction,
The focus evaluation value X(1) in the first sampling area (A1) of the current field is stored as the reference evaluation value X5(1) and the first maximum evaluation value Xm(1), and the focus evaluation value )'s focus evaluation value X(2) as the second
It is stored as the maximum evaluation value XM(2) and the evaluation value safety confirmation routine is ended.

Next, a subroutine for executing the direction determination operation will be explained using FIG. The purpose of the direction determination routine is to determine whether the vertex of the focus evaluation value is in the direction of the near point or the ω point with respect to the current lens position.

Therefore, first, the second threshold value (T
HR2) is defined in advance, and as determined in 5TEP (72), the focus evaluation value X(1) of the current field is the reference evaluation value Xa(1) defined by the focus evaluation value of the final field of the evaluation value stability confirmation routine. If it is larger, it is determined that there is a peak that is the maximum value of the focus evaluation value on the rotation direction side of the focus motor (4), and conversely, if it is smaller, it is determined that there is a peak on the opposite direction side. 51°, the 21st iIl value (THR2) is set in advance to be the difference between the focus evaluation value X(1) of the current field and the reference evaluation value X5(1), taking into account the variation in focus evaluation values due to noise etc. 5 TEP when exceeding
The above judgment is made only after the confirmation in (73) is completed.

By the way, as shown in FIG. 9, the focus evaluation value at point G, where the focus lens position is far away from the in-focus position and is in a state of great blur, is set as the reference evaluation value X5(1), and a one-way determination routine is executed to determine the focus. If the evaluation value is small and the slope is not clear, focus lens (2) in the direction where the focus evaluation value decreases;
That is, by moving the focus ring (4), the end point (ψ
The focus lens (2) moves up to the point), giving an unsightly impression. Therefore, the second threshold (T
HR2) is applied only when the current field's focus evaluation value X (1 success) is larger than the standard evaluation value
If it is 3% smaller, jump over S T E P (73) and get 5
When TEP (74) is reached, the counter (CHK) is incremented by 1 for each field along with 5TEP (75), and the count value of the counter (CHK) becomes 3, that is, the focus evaluation value X. (1) is the standard evaluation value x
8 (1) or greater than the standard evaluation value
In (76), when the focus evaluation values among the three fields are both larger than the standard evaluation value When it is smaller than the standard evaluation value X5(1), it is determined that the evaluation value is on a decreasing trend.

When the focus evaluation value tends to increase, the focus motor (4) maintains the current rotation direction, and when it tends to decrease, the focus motor (4) reverses at 5TEP (77). Moshite S T E P (78) to S T E
Focus evaluation value in the current field at P (80)
is held as the maximum evaluation value XM(1), and the counter (CH
K), set the operation mode hood to a2'', and execute the mountain climbing routine with the AF screen in the next field.

When it is not yet determined that the above-mentioned condition has continued for 3 fields at 5TEP (75), the focus motor (
The direction determination routine is temporarily terminated with the rotational direction and operation mode code (MODE) of 4) left as they are, and the direction determination routine is executed again at the AF scene in the next field.

On the other hand, at 5TEP (73), the focus evaluation value
If the threshold (THR2) is not exceeded, 5TEP (81
), reset the counter (CHK) at 5TEP (8
2), the first sampling area (A1
)'s focus evaluation value x(1) is the maximum evaluation value up to that point
If it is judged to be larger than (1), 5TEP (
83), the maximum evaluation value
M(1) is updated, and the first ring position counter P(1) is reset to maintain the mechanical position of the focus ring (3) at this time. In 5TEP (84), the focus evaluation value X (2) in the second sampling area (A2) of the current field is the maximum evaluation value X up to that point in the same area.
It is determined whether or not it is larger than M (2), and if it is larger, the maximum evaluation value XM (2) is updated with this focus evaluation value x (2) in 5TEP (85), and the focus ring at this time is The second ring position counter P(2) is reset to maintain the mechanical position of (3). Here, both ring position counters P(1) and P(2) have a constant rotation amount of the focus motor (4) in a predetermined rotation direction as a first step, and are counted up every time this one step is driven. When it is inverted, it is a counter that counts down every step of driving.

Then, the first and second sampling areas (Al) (A
2) The focus ring (3) moves between the two end points without the focus evaluation value That is, when it can be determined that while repeating the direction determination routine using the end point switch (5), both the flag that is set due to reaching the periapsis point and the flag that is set due to reaching point ■ are set. , 5TEP (87
), the maximum evaluation value in both areas XM (1), XM (2)
Compare with. However, when making this comparison, the first and second
Area ratio of sampling area (Al) (A2) (1:4
), the maximum evaluation value X is normalized by multiplying <1) by 4 and compared with the maximum evaluation value XM(2). In the above comparison result, if it is determined that the maximum evaluation value XM (2) is larger, a flag (M A
The count value of the ring position counter P(1) is updated by the count value of the second ring position counter P(2), and the following autofocus operation is performed for the second sampling area (A2). Also, the maximum evaluation value XM (1
) is larger, the autofocus operation is performed for the first sampling area (A1). Then, in STE P (89), the operation mode code <MODE) is changed to 3'', and in the AF scene in the next field, the transition is made to the return routine.
If it is not determined at TEP (86) that the focus ring (3) has moved completely between both end points, this direction determination routine is executed again in the AF scene in the next field.

Next, the operation of the mountain climbing routine will be explained using chart 70 in FIG. The objective of the mountain climbing routine is to find the vertex where the focus evaluation value is maximum.

First, S T E P (91)! ,= te 3rd w
The 4-value (THR3) is defined as a fixed value, and the focus motor (4) is moved in the direction determined by the direction determination routine described above.
That is, the focus ring (3) continues to rotate at the first speed (Sl), and as a result of the determination at 5TEP (92), the focus evaluation value x(1) of the current field becomes the maximum evaluation value XM(1).
), the maximum evaluation value xm at 5TEP (93)
(x) is updated with the focus evaluation value x(1), and the first ring position counter P(1) is also reset.

If the focus evaluation value x(1) of the current field is not larger than the previous maximum evaluation value XM(1), 5TEf described below
' (94) to 5TEP (98), 5TEP (9
9) t, knee r focus evaluation value X(1) is the third w4 value (TH
When it is determined that the value is higher than R3) or lower than the maximum evaluation value XM (1), -CTIL is sent to S T E P (110).
Immediately stop the focus motor (4) and
If it is determined in (111) that this state continues regarding the focus evaluation value between three fields, S T E
Set the relative ratio OK flag (OK) at P (112), and turn the focus motor (4) at 5TEP (113).
is reversed, the operation mode code (MODE) is set to '3', and the apex return routine is executed in the AF scene in the next field. Note that (CHK) in S T E P (111) is 1) -XM (January> This is a counter for detecting whether or not the THR3 status is within 3 fields. Also, the check between 3 fields, the 311th! value (THR3), The relationship between focus evaluation values is as shown in FIG.

Here, if the amount of overshoot from the vertex (ΔF(1)) is too large, the amount of out-of-focus will be large and the impression in the actual photograph will be poor. Therefore, in order to make ΔP(1) as small as possible, the rotational speed of the focus motor (4) is controlled as follows. First, during the mountain climbing motion, the relative ratio R(1) is set to 1 for each field. S T E P (114))
, as shown in FIG. 12, when it exceeds the first reference relative ratio (rl), it is considered that the peak is approaching, and S T E
Primary focus evaluation value x (1) that reduces the rotational speed (SPEED) of the focus motor (4) from the standard speed (81) to a slower intermediate distance (s2) at P (115). When exceeds the maximum evaluation value YM (1), the speed can be reduced to slow speed (S3), which is even slower than the intermediate speed (82) at 5TEP (95), and the following 5TEP (99) to (10G) The third R value (
T) When IR3) is exceeded, the focus motor (4) is stopped and the focus evaluation value
Confirm that it is below the threshold value (THR3).

However, after reducing the rotational speed of the focus motor (4) to the slow speed (A3), if the above mountain climbing operation is not completed even after 20 fields have passed, STE P (96) (9
When it is determined in step 7), the rotational speed of the focus motor (4) is returned to the maximum standard speed (sl) at 5TEP (98), the hill climbing operation is continued, and the focus motor (4) is slowed down. Prevent you from moving for long periods of time. In 5TEP (96) and (97), (summer NN) is a counter for detecting that the above-mentioned state has continued for 20 fields, and is counted up every field.

In addition, in 5TEP (94), when the focus evaluation value X (1) exceeds the maximum evaluation value X M (1), the relative ratio R (1)
If this is too small for the relative ratio of the vertices and less than the second reference relative ratio (r2), then S T
E P (114), and the mountain climbing routine is forcibly executed in the next field, only when the relative ratio R(1) at the top is greater than or equal to the second reference relative ratio (r2). , S T E P (95) to (113), and then the process moves to the next apex return routine. This reduces the frequency of malfunctions in which the apex is incorrectly recognized due to noise or the like and the focus lens 4) stops in an out-of-focus state. Here, the second reference relative ratio (r2) is
In FIG. 14, the relative ratio (for example, 0.1) is set in advance when the degree of blur is large enough to be recognized as being out of focus.

By the way, the reason why the relationship between the relative ratio R(1) and the focus ring position changes linearly as shown in FIG. 12 will be explained in detail below. As mentioned above, this relative ratio R(1) is determined by the first HPF (
9), the integral value for one field and the second H
This is the ratio to the integral value for one field when using P F (10), and the relationship between both horizontal values and the focus ring position when the subject is the same is as shown in Figure 13. . In other words, the integral value at the first HPF (9) with a high cutoff frequency becomes a steep mountain, and the integral value at the second HPF (9) with a low cutoff frequency becomes a steep mountain.
The integral value in (10) becomes a gentle peak. Therefore, when the relationship between this relative ratio and the degree of blurring of the object (the amount of movement or deviation from the lens position at the time of focusing) is plotted in a graph, it becomes a monotonically decreasing characteristic curve as shown in FIG.

This is because the relative ratio state quantity is a function value that can express the focus state (bow/degree) of the subject in the same way as the focus evaluation value, and because it is expressed as a ratio, it is a kind of normalized state. For example, if the illuminance of the subject changes, the absolute value of the focus evaluation value will change, but
There is no significant change in the relative ratio (0). Since the above properties are usually independent of the type of subject, it is possible to use this relative ratio as a parameter for the degree of blur.

When this monotonically decreasing characteristic curve in Fig. 14 is made to correspond to the lens position, that is, the focus lens position, it changes approximately linearly toward the near point and the ψ point with the in-focus position as the apex, as shown by the dashed line in Fig. 12. A characteristic diagram is obtained.

Also, in this mountain climbing routine, as in the direction determination routine, the difference between the focus evaluation value x(1) and the maximum evaluation value and S T
If it is determined by E P (116), the maximum evaluation value of Nasi Pulling Area (At) (A2) X5s (1)
, Xm(2) per unit area, that is, XM(1)
Comparing X4 and Xm(2) at 5TEP(17),
If M(1)
(11g) Set the flag (MAX2) at (119), update the value of the first ring position counter (Pl) with the count value of the second ring value counter (P2), and S T
E P (113), and the second sampling area (A2) is selected as the focus area in the next vertex return routine.

Note that S T E P (120) and (121) are The maximum evaluation value XM (2) is updated with the focus evaluation value X (2) at the time, and the second ring position counter (P2
It plays the role of resetting the .

Next, the operation of the apex return routine for returning the focus ring (3), that is, the focus lens (2) to the apex position again after the apex is recognized by the mountain-climbing operation, will be explained using FIG. 15. The first ring position counter P(1)% is incremented as the amount of overshoot from the apex in the mountain climbing routine, or when the count is incremented when the focus ring (3) moves between the two end points in the direction determination routine or the mountain climbing routine. In this routine, the first and second ring position counters P(1) and P(2) are decremented (according to each step of rotation of the focus motor (4)) by rotating the focus ring (3) in the opposite direction. (count down by 1), and when the count value reaches zero, it can be determined that the position has returned to the apex position.

Specifically, in S T E P (131), the flag (M A
2) is in the set state or not, and the flag (M
If AX2) is set, S T E P (1
32), the second sampling area (A2) is specified as the subsequent focus area, and if the flag (MAX2) is not set, the first sampling area (A2) is specified as the subsequent focus area in S T E P (133). Specify Al).

Then, when it is determined in 5TEP (134) that the count value of the ring position counter on the side designated as the focus area has become zero, STE P (135
), the focus evaluation value X(J) and the maximum evaluation value XM(J) of the sampling area on the focus area side (Here, if the first sampling area (A1) has been specified as the focus area, 2 sampling areas (A
2) is specified and it is confirmed that the difference from 1, % lever J -2) is less than the preset 4th YR value (THR4(J)), the focus is set at 5TEP (36>). In order to stop the motor (4) and execute the evaluation value fluctuation monitoring routine in the AF scene in the next field, STE P (137) 4: Set the operation mode hood (MODE) to 4''. Furthermore, S T E P
(138) Vertex confirmation permission flag (TL) at (139)
・ is set, the apex confirmation counter (MC) is reset, and the series of focusing operations is completed.

In addition, when the difference between the focus evaluation value X (J) and the maximum evaluation value Assuming that, in 5TEP (140), set the operation mode code (MOD E > to 'O',
The evaluation value stability confirmation routine is restarted at the AF scene in the next field. Note that the fourth threshold value (THR4(J)) is individually set to an optimal value in advance according to each sampling area.

After the focusing operation is completed, a focus evaluation value variation microscopic routine is executed to monitor changes in the subject and to determine whether or not it is necessary to restart the focusing operation when there is a change. The operation of this routine will be explained below using FIG. 16.

First, in the field immediately after entering this routine, in order to check whether there is an error in the apex detected in the series of focusing operations described above, the apex is set at 5TEP (151) immediately before the end of the apex return routine. The set state of the vertex confirmation permission flag (TL) is determined, and if it is the first field after transitioning from the vertex return routine, it is in the set state, so the vertex confirmation routine is requested to be executed.

This vertex confirmation routine is shown in the flowchart of FIG. To explain this flowchart, first, the focus motor (
4) in any direction, for example, towards the near point, by a minute amount (one step of the focus motor (4), which is a stepping motor) to the extent that the angle of view fluctuation is not perceivable.
When it is determined at T E P (181) that the slight movement has finished, the focus motor (4) is immediately stopped at S T E P (183), and the focus evaluation of the focus area in the field at that time is performed. J ) and the maximum evaluation value X (J) that is determined to be the vertex.
It is done in E P (184). Here, the focus area refers to STEP (8g) in Figure 8 and S in Figure 10.
The flag (MA
Depending on the set state of X2), S T E in Fig. 15
P This is the sampling area that has been fully specified in (131) to (133), and when it is J-1, the first sampling area (A1) corresponds to it, and when it is J-2, it is the second sampling area (A2). . As a result of the comparison in STEP (1g4), the focus evaluation value X (J) is the maximum evaluation value XM (
J), 5TEP (185)
Set the flag (Fl) at (188), and set the flag (Fl) at 5TEP (
At 18, the focus motor (4) is reversed and the lens (2) is slightly moved by 2 steps in the opposite direction, that is, in the vertical direction.
When it is confirmed in TEP (184) that the focus evaluation value X (J) is smaller than the maximum evaluation value X M (J >
S T E P (185) and then S T E P (
1g? ), the flag (F2) is set, and S T E
Reverse the 1 focus point (4) again at P (188) and set the flag (Fl) (F2) at 5TEP (189).
When it is determined that both are set, the lens is slightly moved in both directions as shown by the arrows in Fig. 18, and it is recognized that the obtained focus evaluation value is smaller than the maximum evaluation value, and the apex is determined. It is confirmed that there is no error in the detected position. Then, in S T E P (195) and (196), the focus motor (4) is operated in the periapsis direction by the amount of extra operation in the ψ vertical direction beyond the apex, and returned to the apex again, and the focus is returned to the apex. Stop the motor (4) and
The vertex confirmation flag (TL) is reset at P (190), and the vertex confirmation routine is ended.

Furthermore, as shown in FIG. 19, if the apex detection in the focusing operation is incorrect, the focus evaluation value x(1) will change to the maximum evaluation value XM(1) when the lens is slightly moved in any direction. , so without reversing the focus motor (4), continue to move the lens slightly in the same direction until the 5TEP (
191), the number of movements is incremented by a fine movement number counter (MC).

By the way, this apex confirmation routine can be used in two ways: immediately after the apex return routine ends, to confirm that the subject has returned to the apex, and to confirm that the subject has changed after the apex return has been confirmed. As will be described later, the vertex confirmation flag (TN) is set only in the latter case to distinguish between the two. S T E
In P (192), the above-mentioned two cases are divided depending on whether the apex confirmation flag (TN) is set or not, and it is determined whether the lens position has definitely returned to the apex immediately after the apex return routine is completed. When you want to check the status, skip S T E P (193), which will be described later, and maintain the minute fluctuation amount of the focus motor (4) in one step, and set the flag (Fl) at S T E P (i94). (F
2) is cleared, and the maximum evaluation value XM(J) is updated with the new focus evaluation value X(J).

Therefore, by repeating the vertex confirmation routine,
Immediately after the apex return routine is completed, as shown in FIG. 18, when the apex return is reliably performed, the focus evaluation value changes as shown by the arrow, confirming that the apex has been reliably returned.
As shown in FIG. 19, when the apex return is not reliably performed, the focus evaluation value changes as shown by the arrow, and each time, the maximum evaluation value XM (J) is updated and the lens position gradually adjusts. Approach the focal point. In addition, in FIGS. 18 and 19, ■ to ■ indicate the order of variation.

In the evaluation value fluctuation monitoring routine, the vertex confirmation routine (15
Every time step 2) is completed, the count value of the movement counter (MC) is checked in STE It is recognized that the vertex has moved due to a change in . That is, in FIG. 19, by setting the allowable number of times to "3", it is recognized that the vertex was incorrect after the ■th operation. When recognized in this way, S T E P
The vertex confirmation permission flag (TL) is reset at (154), the rotation speed of the focus motor (4) is switched to the standard speed (sl) at S T E P (155), and the operation mode hood (MODE) is set. "2" and restarts the focusing operation from the above-mentioned mountain climbing routine in order to detect the correct apex again at the next AF scene. In addition, if the apex confirmation routine is executed immediately after the apex return routine ends, once the apex is confirmed, this operation cannot be executed unless the focus evaluation value changes. The operation of monitoring evaluation value fluctuations will be explained.

The focus evaluation value X (J) of the side of the first or second sampling area (AI) (A2) that has been designated as the focus area in the above-mentioned apex return routine, and the maximum evaluation value XM detected so far. (J) defines the amount of variation in the focus evaluation value in the current field, and S T E
The major company with the 5th g4 value (T HR5 (J)) in the focus area defined in P (156), that is, 1/8 of the maximum evaluation value XM (J) in the focus area, is S T E P (157). When the SW4 value (THR5(J)) is exceeded, it is recognized that the subject has changed.

Here, the aforementioned fifth threshold value (T HR5 (J)) is the maximum evaluation value XM (
The setting of 178 for J) is determined experimentally in advance, for example, the maximum evaluation value XMH) of 174
If the setting is large, such as, it is difficult to detect even if the subject changes slightly, and the maximum evaluation value
If it is set to a small value such as , the false detection of camphor will easily occur due to the influence of noise etc. even if there is no change in the subject.

When the change in the subject is recognized, the 5TEP (158) identifies whether the first or second sampling area is designated as the focus area at that point, and selects the first sampling area (A1) as the focus area. If the specified subject change confirmation as described in L has been completed, that is, if there is a change in the subject when viewed in the first sampling area (AI), the focus evaluation value of the second sampling area (A2) In other words, the magnitude relationship between the difference between the focus evaluation value x(2) and the maximum evaluation value
(159), 1 discrimination mill, So (7) result, No. 5811
If the value (T HR (2)) is also small, the subject is
As shown in Figure 0 (b), it escapes from the first sampling area (A1), but it is assumed that it remains in the second sampling area (A2), and from then on at S T E P (160). The focus area, which is the focus evaluation value variation focused viewing area, is switched to the second sampling area (A2) and set to J-2, and the single viewing operation is continued. In this way, by observing the movement of the subject in two large and small sampling areas, even if the subject moves away from the center of the screen after focusing on the center of the screen, it will maintain focus over a wider area of the screen. This makes it possible to stabilize autofocus operation. In Fig. 20 (a) to (d), the dotted line indicates the subject immediately before movement, and the solid line indicates the subject immediately after movement. In Fig. 20 (a), the subject is in the first sampling area (
This is the case when moving within Al).

Focus evaluation value x (2) at next S T E P (159>)
The difference between the maximum evaluation value XM(2) and the fifth threshold value (THR5(
2)) If it is determined that the subject will be photographed as shown in Figure 20 (
As shown in C), it is considered that the object has strayed from the second sampling area (A2), and it is necessary to restart the focusing operation on a new object.

However, if the fluctuations in the focus evaluation value in this case are caused by vanning or tilting for subjects at the same distance, or changes in the brightness of the entire screen, the focus evaluation value will change as shown in Figure 21. In such a case, restarting a new focusing operation will result in an unsightly movement. Therefore, in order to distinguish between the case shown in Fig. 21 and the case where the distance to the subject actually changes as shown in Fig. 22, a change in the focus evaluation value in the second sampling area (A2) is detected. The STE
At P (ift2) (163), focus evaluation value X (1) in the current field, maximum evaluation value XM ( at X (2)
1), updates XM(2), and also sets the vertex confirmation permission flag (TL) to execute the vertex confirmation routine in the next field.
) and vertex confirmation flag (TN) as S T E P (1
64), then S T E in the next field.
At P (152), the vertex confirmation routine is restarted, and it is determined whether the change in the focus evaluation value is due to a change in the shape of the subject pattern or a change in distance.

By the way, when the distance to the subject actually changes, it is desirable to restart the focusing operation immediately. Therefore, in S T E P (192) of the focus confirmation routine, the apex confirmation flag (TN ) is set, the focus evaluation value X(J) in the current field is set to the maximum evaluation value X&1(J) in 5TEP (184).
It can be determined that the above is the case, and the focus motor (4
) is rotated by a small amount of movement, the focus motor ( 4) The amount of minute rotation, that is, the amount of minute movement of the focus lens (2), is
1 step every time you pass TEP (193) → 3
Increase step by step → 5 steps →...

As a result, as shown in Fig. 22, the lens position is changed from the lens position 1t (L'), which is the apex of the heavy liquid object, to the apex of the new object from ■-■→■-■-■-■. In case of ■, ■, ■, the displacement is one step at a time.

In (2), the lens moves 3 steps at a time, and in (2), it moves 5 steps at a time to speed up the approach to the in-focus position.

At the same time, the allowable number of times is set smaller than when the vertex confirmation flag (TN) is in the reset state. Furthermore, in the case of Fig. 21, even if the characteristics of the focus evaluation value change from the heavy liquid subject to the new liquid subject, S T E P (184) in Fig. 17
As a result of this determination, the operations following S T E P (185) can be executed, so that the lens position immediately maintains its original position.

The above vertex confirmation operation is performed as shown in Figs. 21 and 22.
It is necessary that the slope of the focus evaluation value peak is clear and that differences in focus evaluation values due to slight lens movement can be detected. Therefore, as shown in FIG. 23, there is a risk of malfunction in the part where the slope is unclear at the base of the peak of the focus evaluation value characteristic of the new liquid photographic object. Therefore, in STEP (161), the relative ratio R(J) (J 1 or 2) in the sampling area on the side that can be specified as the current focus area is experimentally set in advance as a limit value that will not be recognized as out of focus. When it is smaller than the fourth reference relative ratio (r4), it is judged that the slope is unclear as shown in Fig. 23, and the operation is immediately performed at S T E P (165) without checking the apex. Set the mode control F (MODE) to 0'' and start over from the evaluation value stability confirmation routine at the AF scene in the next field.

Next, as a result of the determination in S T E P (157), the difference between the focus evaluation value X (J) and the maximum evaluation value XM (J) is
The monitoring operation when the threshold value THR5(J) or less, that is, when the subject has not changed, will be described.

Even if there is no change in the focus evaluation value, there may be a case where the subject changes as shown in FIG. 24. Therefore, even if no change in focus evaluation value is detected, STE P (166)
, the relative ratio R(J) in the current field is set to the third standard relative ratio (r
3) If it is determined that S T
The process moves to E P (161) and the same processing as when a change in the focus evaluation value is recognized is performed. However, this relative ratio judgment is based on the vertex of the subject currently in focus from 5TEP (167). If the relative ratio is determined to be equal to or higher than the second reference relative ratio (r2) in the mountain climbing routine,
That is, only when the relative ratio OK flag (OK) is set.

Next, in S T E P (168), the current focus area is set to the first and second sampling areas (AI>(A2
) is specified, and further STE
The flag (MAX2) is not set in P (169), and the focus evaluation value per unit area is
The sampling area (A1) is larger, and the difference between the focus evaluation value x (1) and the maximum evaluation value XM (1) at 5TEP (170) is less than or equal to the fifth threshold THR5 (1). If it is below this threshold, it is assumed that the subject has returned to the first sampling area (AI) from the state shown in Figure 20 (b) to the first sampling area (AI) as shown in Figure 20 (d), and focus evaluation is performed. The focus area, which is the area where you can see the value fluctuations, is S
At T E P (171), switching is made to the first sampling area (A1) and the tube viewing operation is continued. In this way, once the subject returns to the first sampling area (A1), the area is immediately reduced in size, thereby reducing the possibility that the focus will be drawn to the background of the large area.

In each drawing, the focus lens position on the horizontal axis is
Among the movable strokes of the focus lens in the optical axis direction, the front end position furthest from the image sensor is taken as a reference, and corresponds to the distance from this position to the focus lens in the optical axis direction.

As described above, the focusing operation in the AF scene, the confirmation of changes in the subject after focusing, and the operation of correcting deviation are completed.

Further, as shown in the main routine, the AF screen is stopped every 32 fields, and the auto iris operation is completed at the AP screen. This auto iris operation is
The LPF (11) output is divided into each sampling area (AI>(
A2)...In (A6), integrate for one field to obtain the exposure evaluation value for each area, and if the normalized exposure evaluation value includes an extremely high value, the corresponding Assuming that there is an abnormal brightness area such as a light source in the area where
Excluding this area, the average value of the exposure evaluation value per unit area for the entire remaining area is maintained at a predetermined level.
The iris motor (7) is driven via the iris motor control circuit (28) to adjust the aperture amount of the aperture mechanism (6) and to control the gain of the AGC circuit (not shown) through which the captured video signal passes. This allows optimal exposure adjustment to be made.

(G) Effects of the Invention As described above, according to the present invention, when the distance to the subject changes slightly after focusing, the shift correction is performed with the top priority being to prevent the angle of view from changing if the amount of change is small. Furthermore, even if the amount of change is slightly large, it is possible to minimize the change in the angle of view and quickly correct the deviation without having to restart the focusing operation from the beginning.

[Brief explanation of the drawing]

The drawings all relate to one embodiment of the present invention, and 1st rM is an overall circuit block diagram, 2nd figure is an explanatory diagram of dividing the sampling area, 3rd figure is a flowchart of the main routine, and 4th figure is an illustration of the division of the sampling area.
Figure 5 is a flowchart of the AF cylinder, Figure 5 is a flowchart of the focus evaluation value and relative ratio calculation routine, Figure 6 is an explanatory diagram of sunspot evaluation value calculation, and Figure 7 is a flowchart of the evaluation value stability confirmation routine (coach yard). , Fig. 8 is a flowchart of the direction determination routine, Fig. 9 is a diagram of the relationship between the focus evaluation value and the lens position, Fig. 10 is the broach route of the Rei Yamato routine, and Fig. 11 is the focus evaluation value when returning to the apex. Diagram showing changes, 1st
Figure 2 shows the relationship between the focus evaluation value, focus morph rotation speed, relative ratio, and lens position. Figure 13 shows the relationship between the integral value of each HPF output and lens position. Figure 14 shows the relative ratio and degree of blur. FIG. 15 is a flowchart of the vertex return routine, FIG. 16 is a flowchart of the evaluation value fluctuation monitoring routine, FIG. 17 is a flowchart of the vertex confirmation routine, and FIGS. 18, 19, and 20 indicate the lens Diagram showing the fluctuation of the focus evaluation value due to the change in position, 2nd
1@, FIG. 22, FIG. 23, and FIG. 24 are diagrams showing changes in the focus evaluation value due to changes in the subject. (26)...Microcomputer (evaluation value detection means), (27)...Focus morph control circuit (focus control means)

Claims (1)

[Claims] (1) evaluation value detection means for detecting a high-frequency component level of a captured video signal obtained from an image sensor as a focus evaluation value; and changing a distance between a focus lens and the image sensor;
When the focus evaluation value reaches the maximum evaluation value, the distance is temporarily fixed and the focusing operation is finished, and when the focus evaluation value changes, the focus lens is slightly moved in the optical axis direction several times to adjust the lens position. and a focus control means that executes a fine adjustment operation to make the lens follow the in-focus position, and when the lens position is moved slightly toward the in-focus position,
An autofocus video camera characterized by increasing the amount of micromotion in stages.