JP2002006203A - Focusing device and image pickup device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain an image which is most suitable for portrait-photographing by surely setting a focusing peak on eyes of a person as an object. SOLUTION: A digital camera used for portrait-photographing is provided with a CCD imaging device for picking up the image of the object, a focusing information generating part 201 for generating the focusing information based on the video signal outputted from the imaging device, a focusing control part 202 for controlling the focusing state of the optical system to be controlled based on the focusing information, an eye area detecting part 203 for detecting the eye area of the person when the object is the person and a focusing area setting part 204 for setting the focusing area as the object to be focused by the focusing control part 202, and in the case the eye area is detected by the eye area detecting part 203, the eye area is set as the focusing area by the focusing area setting part 204.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a focusing device for focusing on an eye area of a person as a subject, and an imaging device using the same.

[0002]

2. Description of the Related Art Conventionally, as an automatic focusing device (focusing device) of an electronic camera, image data of an area of a predetermined portion on an image pickup element on which an image of a subject incident through a photographing lens is formed is obtained by using a lens. A so-called hill-climbing method (one of the contrast methods) in which reading is performed for each predetermined drive unit and the lens is driven to a position where the contrast is maximized.
Has been adopted.

There is also known a technique for changing a focus area in accordance with a focusing operation. For example, in a hill-climbing method, a large area is required to avoid a problem of focusing on a background and so-called rear focus. Automatic switching of a small area is also performed.

[0004]

In this type of method, the effect of the background can be prevented. However, in portrait photography in which a person is up, a strong contrast component at the outline of the face or at the boundary with the hair causes a problem. There is a problem that a focus peak does not come to an ideal eye. In other words, although the face was in focus, the best focus was at the contour of the face (hence, from the cheek to the ear) or at the hairline of the hair, and the eyes tended to slightly deviate from the peaks. .

[0005] On the other hand, portrait photography is characterized by using a medium-telephoto large-aperture lens and using a shallow depth of field to obtain a peripheral omission effect, as compared with normal photography, and in general, subject photography. It is the best case that the eye has a focus peak from the request of the psychological description of the person. Therefore, the above-mentioned slight deviation is easily noticeable, and a method for solving this is desired.

The present invention has been made in consideration of the above circumstances, and an object thereof is to enable a focus peak to be surely set to the eyes of a subject person, and to be suitable for portrait photography or the like. It is to provide a focusing device and an imaging device.

[0007]

(Structure) In order to solve the above problem, the present invention employs the following structure.

That is, the present invention provides an image pickup device for picking up an image of a subject, focus information generating means for generating focus information based on an output video signal of the image pickup device, and focusing adjustment of an optical system to be adjusted based on the focus information. Focusing control means for controlling the state, eye area detection means for detecting an eye area of the person when the subject is a person, and setting a focus area to be used when controlling the focusing adjustment state by the focusing control means Focusing area setting means for performing
The focus area setting means sets the eye area as the focus area when the eye area is detected by the eye area detection means.

According to another aspect of the present invention, there is provided a focusing device for focusing on the eyes of a person as a subject, wherein an image sensor for capturing an image of the subject and focus information for generating focus information based on an output video signal of the image sensor. Generating means, focusing control means for controlling a focusing adjustment state of the optical system to be adjusted based on the focus information, eye area detecting means for detecting an eye area of a person when the subject is a person, and the focus information When the focus information is generated by the generation means, a focus information weighting means for varying the contribution of the information on the eye area detected by the eye area detection means is provided.

Here, preferred embodiments of the present invention include the following.

(1) The focus area setting means switches the focus area from a predetermined area set independently of the eye area to the eye area in the course of the focusing adjustment operation controlled by the focusing control means.

(2) The focus information weighting means sets the contribution of the information on the eye area larger than the contribution of the other areas when the focusing control operation controlled by the focusing control means comes near the focal point. thing.

(3) The eye region detecting means captures the output video signal of the image pickup device as a binary image, sequentially cuts out a predetermined region from the image, detects candidate points of the eye region, and determines whether the candidate point is two or more. And when these distances are within a predetermined range, it is to be detected as an eye area.

(4) In order to detect candidate points of the eye region, a plurality of donut-shaped regions having the same center and different diameters are sequentially cut out from the binarized image, and the black-and-white ratio of the cut-out portion is set to a set value. To compare.

(5) As a plurality of donut-shaped regions, a first donut-shaped region having an outer diameter smaller than the iris and a second donut-shaped region having an inner diameter larger than the iris are cut out to form a first donut-shaped region. Detecting a case where the black ratio is almost 100% and the white ratio of the second donut-shaped region is 80% or more as a candidate point.

According to the present invention, there is also provided an imaging apparatus having the above-mentioned focusing apparatus, wherein the imaging element is also used as an imaging element for photographing a subject, and the optical system to be adjusted is an imaging optical system of the imaging element. And wherein the detection of the eye region performed by the eye region detection means is performed based on an output video signal of the imaging device.

(Operation) According to the present invention, when the subject is a person, the eye area detecting means detects the eye area of the person, and the focus area setting means sets the focus area of the focusing control means to the eye area. By doing
The focus can be adjusted to the eyes of the subject person. Similarly, when the subject is a person, the eye area detecting means detects the eye area of the person, and the focus information weighting means increases the degree of contribution of the information on the eye area when generating the focus information in the focus information generating means. It is possible to focus on the eyes of the subject person. That is, by setting the eye area as the focus area or increasing the focus weight of the eye area, it is possible to reliably set the focus peak on the eyes of the subject person, which is suitable for portrait photography or the like.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The details of the present invention will be described below with reference to the illustrated embodiments.

(Embodiment) FIG. 1 is a block diagram showing a schematic configuration of a digital camera according to an embodiment of the present invention.

In FIG. 1, reference numeral 101 denotes a lens system including various lenses (focusing lens, zooming lens, and others);
02, a lens driving mechanism for driving the lens system 101; 103, an exposure control mechanism for controlling the aperture of the lens system 101; 104, a filter system; 105, a CCD color image sensor; Driver 107, a pre-processing circuit including an A / D converter, etc., 108, a digital processing circuit for performing color signal generation processing, matrix conversion processing, and other various digital processing, 109, a card interface,
Reference numeral 10 denotes a memory card such as a CF, and 111 denotes an LCD image display system.

In the figure, reference numeral 112 denotes a system controller (CPU) for comprehensively controlling each unit, 113 denotes an operation switch system including various switches, and 114 denotes an operation display for displaying an operation state, a mode state, and the like. System, 115 is a lens driver for controlling the lens driving mechanism 102, 116 is a strobe as a light emitting means, 117 is an exposure control driver for controlling the strobe 116, 118
Is a nonvolatile memory (E) for storing various setting information and the like.
EPROM).

In the digital camera of the present embodiment,
A system controller 112 performs overall control, and includes a shutter device included in the exposure control mechanism 103 and a CCD image sensor 105 by a CCD driver 106.
The exposure (charge accumulation) and signal reading are performed by controlling the driving of the memory, and the signals are taken into the digital processing circuit 108 via the pre-processing circuit 107, and subjected to various signal processings. The information is recorded on the card 110. In addition, CCD
The image sensor 105 is, for example, an interline type having a vertical overflow drain structure.

The basic configuration up to this point is the same as that of the conventional apparatus, but in the present embodiment, the digital process circuit 108 has the following functional circuits in order to focus on the eye region.

FIG. 2 is a functional block diagram for performing focus adjustment in the present embodiment. A focus information generating unit 201 that generates focus information based on an output video signal of the image sensor 105; a focusing control unit 202 that controls a focusing adjustment state of an optical system to be adjusted based on focus information (contrast information); , An eye area detecting means 203 for detecting an eye area of the person, and a focus area setting means 204 for setting a focus area of the focusing control means 202
The focus area setting means 204 includes:
The eye area detected by the eye area detection means 203 is set as a focus area.

The generation of the contrast information signal (contrast evaluation value) may be performed as follows, for example. First, an absolute value of a high-frequency component obtained by applying a predetermined high-pass filter to a known A / D-processed video signal is set as a contrast value for each image. Then, the obtained contrast value is added to all pixels in the designated focus area.

Next, the focus adjusting operation in the present embodiment will be described with reference to the flowchart shown in FIG.

First, the focus area is set near the center of the screen, and the same hill-climbing AF operation as in the prior art is performed. That is,
While driving the lens, image data of a predetermined area on the subject is taken in at each position of the lens, and each contrast value is detected. (Step S1). The focus area is
As shown in FIG. 4B with respect to the screen shown in FIG.
Near the center of the screen, it is 3/4 of the image frame both vertically and horizontally. 4 in the figure
01 indicates an image frame, and 402 indicates a focus area.

Next, the lens is temporarily stopped when the peak is detected by the hill-climbing AF operation (step S2). The lens position at this time is as shown in FIG.
The position is slightly above the peak point P1, for example, P2.
FIG. 5 shows a change in the contrast value when the lens is moved. A shows the case where the center area of the screen is the focus area, and B shows the case where the eye area is the focus area.

Next, the image signal is taken in, binarized, and analysis is started (step S3). Although the binarization level is separately optimally controlled, as an example, histogram analysis is performed and “intermediate value ((maximum + minimum) / 2)” “average value”
The “median value” is the minimum value among the three values. That is, the level of the iris is fairly low, so a lower level setting ensures separation from the surrounding cheeks.

Next, a pupil point corresponding to the first half of the detection of the eye area is detected under the following conditions. The pupil point referred to here is the so-called center point of the black eye. The black eye is present in the white eye (the white eye is present around the black eye). This detection is performed based on the fact that it is almost constant.

First, "φ2 or more and φ7 around the point of interest"
The white ratio (the number of white pixels with respect to the number of pixels) of the following donut-shaped first region is 0, and the white ratio of the second donut-shaped region of φ16 to φ24 is 80% or more. "Points satisfying these conditions are detected as candidate points (step S4).

Here, when the point of interest is a pupil point, FIG.
As shown in the figure, the entire first region 601 is
, The white ratio is zero. Second area 60
2 overlaps the white-eye region surrounding the black-eye region 603 and the upper and lower eyelid regions, so that the white ratio increases to 80% or more. That is, the above condition is satisfied, and it is detected as a candidate point.

In the above description, the first region is formed in a donut shape and φ2 at the center is excluded because the pupil point is a luminescent point due to the influence of illumination (if it is a photograph, this is called a catch light. The reason why the white ratio is set to 80% or more in the determination of the second region is that a certain amount of black is often included due to the outline of the eyelids and the frame when wearing glasses. Because it will be. In the above description, "the pupil point is the center point of the black eye". However, the meaning is not necessarily limited to the geometric distribution center of the black eye. When a catchlight is generated in the eyes of a person, it is rather caught as the position of the bright spot.

On the other hand, when the point of interest deviates from the pupil point, FIG.
As shown in (b), a part or the entirety of the first region 601 is located outside the black eye region 603 of the eye, so that the white ratio is not 0. In addition, since the second region 602 partially overlaps with the black-eye region 603, the white ratio is significantly reduced. Therefore, in this case, it is not detected as a candidate point. Also, when the point of interest deviates greatly from the eye area, for example, near the hair, the white ratio of the first area 601 becomes 0,
Since the white ratio of the second area 602 is also 0, no candidate point is erroneously detected.

When the above detection is simply performed, no problem occurs when the set resolution of the point of interest is appropriate and only one candidate point is generated for one eye, but when the resolution is high, Since many points satisfying the above condition (substantially the same point in a group adjacent to each other) are detected for one eye, the next "number determination of candidate points" is performed as it is. Failure will occur. In order to avoid this, the candidate point "If the detected candidate points are consecutively adjacent (accordingly, they constitute one area), this is regarded as the only candidate point" May be added. As the position coordinates of the candidate point at this time, for example, the coordinates of the center of gravity of the area can be adopted. In the following, it is assumed that the “candidate points” are those after performing this narrowing-down processing as necessary.

Next, the number of candidate points is determined (step S5). If there are no candidate points, or if there are any, the number is not two, there is no pupil point. Then, the lens is driven in the reverse direction to return to the previous peak point P1 (step S6). In this case, the focusing operation is the same as the conventional one.

When the number of candidate points is 2, the distance between the two points is 40 or more and 75 or less (pupil width condition: also consider a case where the child is held diagonally), and the inclination is 0 ± 25 degrees (condition H) or 90 degrees. ± 2
In the case of 5 degrees (condition V), these two points are detected as pupil points. Here, the condition V considers the vertical position shooting, and when the vertical position detection by a gravity (acceleration) sensor or the like is separately used, only one of the two conditions is detected by matching the detection result. Can be targeted. In this example, when both the vertical position and the horizontal position are tilted by 25 degrees or more, they are excluded from the detection target because they are regarded as deviating from the category of general portrait photography.
The numerical value of the angle condition is one example value having a remarkable effect, and can be arbitrarily changed according to circumstances. Further, the configuration may be such that the angle condition is not questioned at all.

Specifically, it is determined whether or not the distance between the two points is 40 to 75 (step S7). If the distance is not 40 to 75, the process proceeds to S6 with no pupil point. In the case of 40-75,
It is determined whether the inclination is 0 ± 25 degrees (step S8),
If it is not 0 ± 25 degrees, it is further determined whether the inclination is 90 ± 25 degrees (step S9). And 90 ± 2
If it is not 5 degrees, the process proceeds to S6 with no pupil point. If it is determined in S8 that the inclination is 0 ± 25 degrees, or if it is determined in S9 that the inclination is 90 ± 25 degrees, two points are detected as pupil points.

In the above (and below), the unit of the length numerical value is mm.
The processing is performed by calculating the imaging magnification based on the focal length of the lens (using the zoom information) and the subject distance (using the AF information at this time = the peak position information) and converting it into pixel coordinates.

Next, an eye area corresponding to the latter half of the eye area detection is set. In the case of the condition H, the shape of the area is a rectangle of 15 × 30 in the left and right sides, and the position is the left area: the center of the rectangle is shifted to the left by 7 with respect to the pupil point / the right area: the center of the rectangle is It is assumed that the pupil point is shifted right by 7 (step S10). On the other hand, in the case of condition V,
The shape of the area is a rectangle of 30 × 15 in the vertical direction, and the position is the upper area: the center of the rectangle is shifted upward by 7 from the pupil point / the lower area: the center of the rectangle is the pupil point 7 below
It is assumed that the image is shifted (step S11). FIG. 4B shows an example in the case of the condition H. Although FIG. 4B also shows a straight line 404 indicating the inclination of both pupil points, it is more preferable that the rectangle 403 in the above-mentioned region is inclined according to this inclination.

As described above, when an eye area is detected, the focus area is switched to the eye area set in S10 or S11 (step S12), the lens is driven in reverse (step S13), and the peak again occurs. At the time of detection, this is determined as a true (corresponding to the eye) peak, and the lens is finally driven to the peak corresponding point (step S).
13). As a result, as shown in FIG.
From 2 ', the peak position P1' for the eye region is set.

Note that S13 is reversed after a predetermined overrun. For example, the lens is overrun by about the maximum value of the shift amount between the peak position when the eye area is set as the focus area and the peak position when the entire eye area is set as the focus area. As a result, the peak for the eye region can be reliably detected only by the reverse driving of the lens.

More specifically, the predetermined overrun amount may be a lens extension amount (or more) that corresponds to an assumed maximum value of the depth deviation between the contour of the face and the eyes. In other words, when the lens is reversed at the lens position at the time of peak crossing detection, the focus position of the eye (the peak position by the eye area) is shifted to the overrun side with respect to the focus position of the face outline (the peak position by the original focus area). In such a case, a reverse re-rotation drive is required for final focusing. On the other hand, if the reverse rotation is performed after the predetermined amount of overrun, the re-reversal drive only needs to be performed once, and such a modified form is extremely preferable.

In this case, as in the above embodiment, the image signal for the eye detection should be taken in the vicinity of the peak position by the initial focus area. This can be said to be more desirable because the degree and probability can be reduced overall. As described above, according to the present embodiment, when the subject is a person, S4
By detecting the eye area of the person in the steps S9 to S9 and setting the final focus area to the eye area, the focus can be surely focused on the eyes of the subject person, which is suitable for portrait photography and the like. Becomes Further, in the present embodiment, since focusing is performed by the contrast method from the imaging output of the imaging CCD 13, it is not necessary to newly provide an imaging device, and the configuration can be simplified.

(Modification) The present invention is not limited to the above embodiment. In the embodiment, the area is switched, but the weighting of the information of the eye area and the information of the other focus area may be switched. Since a certain amount of information on the surrounding area is included, even when there is a motion of the subject or an erroneous detection of the eye area, focusing cannot be performed, and stable focusing can be obtained at the same level as in the related art.

Specifically, after a focusing operation is first performed based on the total contrast value of all the focus areas, (the total contrast value of the eye area × ｛(the total number of pixels in the focus area) / (the number of pixels in the eye area) The focusing operation is performed based on {) + (total contrast value of the focus area other than the eye area) × {(the total number of focus area pixels) / (the number of pixels of the focus area other than the eye area)}. Since the contribution degree of the region is the same as that of the region whose area is several times, the eye region is substantially weighted.

If the eye region is compared with the rest of the eye region, it goes without saying that the weight is the same in S1, but in S13, a weight corresponding to ｛｝ is added. Since the number of pixels in the eye region is smaller than that in the other regions, the weight of the eye region is increased, and the same effect as in the above embodiment is obtained. In addition, since a certain amount of information on the surrounding area is included, there is a change in the subject (for example, when the subject has closed his eyes, the contrast component of the eye region may be extremely small) or an erroneous detection of the eye region. In this case, focusing is not impossible, and stable focusing can be obtained at the same level as in the related art.

If the image of each point driven by the lens is stored in the memory and the contrast information is obtained for each of the subdivided areas, when the eye area is detected, the contrast information for the area is referred to. Can be focused on the eye region.

Specifically, instead of finding one contrast value for each lens position, the focus area is subdivided into a number of partial areas (pixel units may be used), All the contrast values of the partial areas are stored. Then, in the conventional hill-climbing control in S1, the same control as described above is performed using the sum of the respective contrast values (this is equal to the conventional contrast value of all focus areas) as the contrast evaluation value.

If an eye area is detected in S4 to S9, the process proceeds to S1.
In No. 3, only the eye area is selected and added from the “contrast values of each of the plurality of partial areas for each lens position” which have already been obtained (this is the eye area focus area in S13). (Corresponding to a contrast value) is used as a contrast evaluation value, and a peak point of a “lens position-contrast evaluation value” curve is found by calculation, and the lens is driven to that point.

At this time, since the data already used at the time of the operation corresponding to S3 is used, it is not necessary to newly drive for peak detection. That is, although a large amount of buffer memory is required, the reverse rotation only needs to be performed once as in the conventional hill-climbing AF. Also in this case, it is extremely preferable to perform a predetermined amount of overrun.

Throughout the above description, various numerical values and the like are exemplifications, and such values themselves have special significance closely related to the remarkable effects of the present invention, but are not limited thereto. Not at all. That is, for example, φ2
φ7, φ16, φ24, white ratio 0,80% or more, ± 25
Numerical values such as degrees, 15 × 30 (30 × 15), and 7 shifts, and examples of setting binarization levels can be appropriately changed as needed.

The eye region is not limited to a rectangle, but may be a circle,
Any shape such as an ellipse and a polygon can be used. Furthermore, the boundary between the white eye and the eyelid may be determined by an analysis method such as contour detection, and the eyelid area (preferably set slightly larger than the real one) may be used as the eye area. Still,
It goes without saying that the present invention can be applied to any type of movie. In addition, various modifications can be made without departing from the scope of the present invention.

[0054]

As described above in detail, according to the present invention, focus information generating means for generating focus information based on an output video signal of an image sensor, and a focusing adjustment state of an optical system to be adjusted based on the focus information. Focusing control means for controlling the focus area, an eye area detecting means for detecting an eye area of the person when the subject is a person, and a focus for setting a focus area to be controlled when the focusing control state is controlled by the focusing control means. A region setting unit, and when the eye region is detected by the eye region detection unit, the focus region is set as the focus region by the focus region setting unit, so that the focus peak is reliably set in the eyes of the subject person. Is possible, and an ideal image in portrait photography can be obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a schematic configuration of a digital camera according to an embodiment of the present invention.

FIG. 2 is a functional block diagram for performing focus adjustment in the embodiment.

FIG. 3 is a flowchart for explaining a focus adjustment operation in the embodiment.

FIG. 4 is a diagram showing an example of setting a focus area.

FIG. 5 is a characteristic diagram showing a change in contrast value when a lens is moved.

FIG. 6 is a schematic diagram showing a relationship between a donut-shaped first region and a second region and a point of interest.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 101 ... Lens system 102 ... Lens drive mechanism 103 ... Exposure control mechanism 104 ... Filter system 105 ... CCD image sensor 106 ... CCD driver 107 ... Pre-process part 108 ... Digital process part 109 ... Card interface 110 ... Memory card 111 ... LCD image display System 112 ... System controller (CPU) 113 ... Operation switch system 114 ... Operation display system 115 ... Lens driver 116 ... Strobe 117 ... Exposure control driver 118 ... Non-volatile memory (EEPROM) 201 ... Focus information generation means 202 ... Focusing control means 203 ... Eye area detecting means 204 Focus area setting means 401 Image frame 402 Focus area 601 First area 602 Second area 603 Iris

Claims (6)

[Claims] An image pickup device for picking up an image of a subject; focus information generating means for generating focus information based on an output video signal of the image pickup device; and controlling a focusing adjustment state of an optical system to be adjusted based on the focus information. Focusing control means, an eye area detecting means for detecting an eye area of a person when the subject is a person, and a focus area for setting a focus area to be controlled when the focusing control means controls the focusing adjustment state. A focusing device, wherein the focus region setting device sets the eye region as the focus region when the eye region is detected by the eye region detection device. 2. The focusing area setting means switches the focus area from a predetermined area set independently of the eye area to the eye area in a focusing adjustment operation controlled by the focusing control means. The focusing device according to claim 1, wherein: 3. An image sensor for imaging a subject, focus information generating means for generating focus information based on an output video signal of the image sensor, and controlling a focusing adjustment state of an optical system to be adjusted based on the focus information. Focusing control means, an eye area detecting means for detecting an eye area of the person when the subject is a person, and an eye area detected by the eye area detecting means when generating focus information in the focus information generating means. A focusing device comprising: focus information weighting means for varying the degree of contribution of information. 4. The focus information weighting means, when the focusing adjustment operation controlled by the focusing control means is near the focal point, makes the contribution of the information on the eye area larger than the contribution of other areas. The focusing device according to claim 3, wherein the focusing device is set. 5. The eye area detecting means captures an output video signal of the image sensor as a binary image, and calculates a black and white ratio of a portion cut out from the image into a plurality of donut-shaped areas having the same center and different diameters. 4. The method according to claim 1, wherein a candidate point of the eye region is detected based on the detected region, and the candidate point is detected as an eye region when there are two candidate points and their distance is within a predetermined range. Focusing device. 6. An image pickup apparatus having the focusing device according to claim 1, wherein said image pickup element is also used as an image pickup element for photographing a subject, and said optical system to be adjusted is: An image pickup apparatus, comprising: an image pickup optical system of the image pickup device, wherein an eye region detection performed by the eye region detection means is performed based on an output video signal of the image pickup device.