JP2004004923A - Automatic focus controller - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an automatic focus controller by which high speed and highly accurate focusing operation is performed. <P>SOLUTION: In the automatic focus controller by which a focal position is decided by outputting an image data by performing image pickup at an initial focus lens position, and calculating each AF evaluated value for an image data group restored for each restoring filter by restoring the image data by using several restoring filters stored in a storing means and corresponding to the radius of a dot image for the image data and collating each AF evaluated value calculated, a corresponding restoring filter is stored in the storing means with the narrower interval of the radius of the dot image at a short distance as compared with a long distance. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an automatic focus control device, and more particularly, to an automatic focus control device applied to an image input device using an image pickup device such as a video camera and a still video camera.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, as an in-focus position determining method in an automatic focus control device, there are an AF method using infrared rays or ultrasonic waves, an external light passive method, and a passive AF method such as a hill-climbing servo. In particular, in a digital still video camera (hereinafter abbreviated as "DSVC") and the like, a passive AF method that does not require a special distance measuring component is often used. In the passive AF method, there is an "in-focus position detecting device for an electronic camera" disclosed in Patent Document 1 as a device that detects a focus position in one shot in recent years. Such a focus position detecting device uses one-shot AF, that is, detects a focus position in one shot using a restoration filter.
[0003]
More specifically, the in-focus position detection apparatus obtains a point spread function of an optical imaging system or a function obtained by converting the point spread function at a focal position and a plurality of lens positions before and after the focus position and stores and stores the characteristic values. Storage means, image restoration means for restoring image data of one screen or a part thereof for each lens position of the plurality of points using the characteristic values stored in the characteristic value storage means, and image restoration image data And a focus position estimating means for estimating a focus position by obtaining an evaluation value of a focus position for each lens position and comparing the respective evaluation values, thereby detecting the focus position faster and more accurately. It is possible to do.
[0004]
[Patent Document 1]
JP-A-6-181532
[0005]
[Problems to be solved by the invention]
However, in the related art, the one-shot AF using the restoration filter has problems such as a pseudo peak, a luminance dependence, and a large amount of calculation when specifying a high-speed and high-precision focal point.
[0006]
In the conventional one-shot AF, the minimum value of the negative integrated value of the restored image data (which does not originally exist) is used as the AF evaluation value. However, such a method can be applied only when there is no pseudo peak of the negative integrated value, and in fact, there is a problem that a pseudo peak often exists and the focusing index is caught other than the original focusing point. .
[0007]
Further, when the focus index is calculated by the conventional one-shot AF, there is a problem that the focus index follows the brightness of the captured AF area and indicates a position shifted from the original focus.
[0008]
Further, when performing the one-shot AF process, the normal Fourier transform is performed. However, when this Fourier transform (FFT) process is performed on the AF area, the vicinity of the side of the AF area becomes an edge, and unnecessary high-frequency components may be generated. is there. In the one-shot AF, an integrated value of a high-frequency component is sometimes used to specify a focal point, and there is a problem in that if the unnecessary high-frequency component is mixed, the accuracy of the focal point specification is reduced.
[0009]
Also, in the one-shot AF, when specifying the focal point, the focal point actually has a certain range of width in consideration of not only one point but also the depth of field, that is, It is determined in a range where the lens is apparently in focus. In particular, the farther the subject is, the wider the focusing allowable range becomes, and it is meaningless to arrange a plurality of corresponding restoration filters in this range. Conversely, the closer the distance is, the narrower the allowable focusing range becomes. Therefore, there is a problem that more corresponding restoration filters must be arranged.
[0010]
Also, when there are a plurality of subject candidates and the photographer is selecting which subject to target, and when there is only one AF area, the AF area is sequentially adjusted to the subject candidate, and One-shot AF focusing operation must be performed, and there is a problem that a redundant operation is performed.
[0011]
In general, as the number of restoration filters increases, the more accurate and versatile the lens can be, the more practically it is necessary to prepare a large number of restoration filters due to the limitation of the mounted ROM (memory). Is difficult.
[0012]
Further, the one-shot AF involves an image restoration process using a restoration filter, so that the amount of calculation becomes enormous. If all of these processes are performed by software, the flexibility of changing and improving the algorithm is excellent, but the processing speed of the CPU is improving year by year, but it is still not enough in terms of speed. There is a problem that cannot be said.
[0013]
The present invention has been made in view of the above, and an object of the present invention is to provide an automatic focus control device capable of performing a high-speed and high-accuracy focusing operation.
[0014]
[Means for Solving the Problems]
The automatic focus control device according to claim 1 captures an image at an initial focus lens position, outputs image data, and applies a plurality of restoration filters corresponding to a point image radius stored in a storage unit to the image data. An automatic focus control device that calculates an AF evaluation value for each image data group restored for each restoration filter by performing image data restoration using the image data, and compares the calculated AF evaluation values to determine a focus position. In the storage means, a corresponding restoration filter is stored at an interval of a small point image radius at a short distance as compared with a long distance.
[0015]
BEST MODE FOR CARRYING OUT THE INVENTION
A preferred embodiment in which an automatic focus control device according to the present invention is applied to a digital still camera will be described below in detail with reference to the accompanying drawings.
[0016]
(Common parts of the embodiment)
FIG. 1 is a configuration diagram of a digital still camera according to the present embodiment. The digital still camera shown in FIG. 1 is roughly divided into a lens system 1 for forming a subject image on a CCD, a diaphragm 2, a front end unit for outputting image data corresponding to the subject image, and various data processing of the image data. An image pre-processor (hereinafter referred to as “IPP”) unit 4, a CPU I / F 5, a CPU 6 for controlling the operation of each unit of the digital still camera, and a focus lens control unit 7 for driving and controlling the lens system 1 And an aperture control unit 8 for controlling an aperture value (f-value) of the aperture 2.
[0017]
The lens system 1 includes an imaging lens and a focus lens. The front end unit 3 converts a subject image formed by the lens system 1 into an electric signal (analog image data) and outputs the same, and performs noise removal and gain adjustment of the electric signal input from the CCD 31. A signal processing unit 32 and an A / D converter 33 that converts analog image data input from the CCD 31 via the signal processing unit 32 into digital image data and outputs the digital image data to the IPP unit 4 are provided.
[0018]
The IPP unit 4 separates the digital image data input from the front end unit 3 into R, G, and B components (RGB digital signals), and an RGB separation unit 41, and separates each color component of the separated RGB digital signals. An RGB gain adjustment unit 42 that adjusts the gains and outputs the luminance signal to a luminance value generation unit 43, and a luminance value generation unit 43 that converts an input RGB digital signal into a luminance signal and outputs the luminance signal to an FFT (IFFT) calculation unit 44. Then, the spatial component of the luminance signal input from the luminance value generating unit 43 is converted into a frequency component and output to the filtering unit 45, and the frequency component passed through the filtering unit 45 is subjected to IFFT and inversely transformed into the spatial component. And an FFT (IFFT) calculation unit 44 that outputs the result to the negative value integration unit 48.
[0019]
Further, the IPP unit 4 includes a filtering unit 45 that performs a restoration process of the signal converted into the frequency component by the FFT (IFFT) operation unit 44 based on the restoration filter stored in the restoration filter data ROM 46, and a plurality of point images. A restoration filter data ROM 46 in which data of a plurality of restoration filters respectively corresponding to radii are stored, and a high frequency obtained by extracting and accumulating a specific high frequency component after filtering a signal converted into a frequency component. A high-frequency component integrator 47 that outputs the component integrated value to the CPU 6; and a negative value integration that outputs to the CPU 6 a negative value integrated value obtained by integrating the negative value of the spatial frequency component input from the FFT (IFFT) calculation unit 44. A part 48. Although each function of the IPP unit 4 can be realized by software, it is preferable to configure the function by hardware as shown in the figure, since high-speed processing can be performed.
[0020]
As described above, the CPU 6 controls the operation of each unit of the digital still camera. Specifically, for example, the CPU 6 performs the AF based on the input integrated value of the high-frequency component and / or the integrated value of the negative value. An evaluation value is calculated (a detailed calculation method of the AF evaluation value will be described later), and AF control and the like are performed.
[0021]
Next, an outline of the AF (one-shot AF) operation of the digital still camera having the above configuration will be described. First, the RGB digital color signals obtained from the CCD 31 through the A / D converter 33 are input to the IPP unit 4 that performs image signal processing. In the IPP unit 4, first, the luminance value is converted into a luminance signal by the luminance value generation unit 43, and then the spatial component is converted into a frequency component by the FFT (IFFT) calculation unit 44. The filtering unit 45 performs a restoration process on the signal converted into the frequency component based on the data (1 restoration filter) read from the restoration filter data ROM 46. The filtered signal is output to the high-frequency component integrator 47 and also to the FFT (IFFT) calculation unit 44.
[0022]
The high-frequency component integrator 47 extracts a specific high-frequency component of the input signal, and outputs a high-frequency component integrated value obtained by accumulating the specific high-frequency component to the CPU 6 via the CPU I / F 5.
[0023]
On the other hand, the data subjected to the filtering process is again input to the FFT (IFFT) calculation unit 44, converted into a spatial component again, and output to the negative value integration unit 48. Here, the re-converted data (image data) may include a negative value. In the negative value integrating section 48, a negative value integrated value obtained by integrating negative values of the reconverted data is output to the CPU 6 via the CPU I / F5. That is, the high frequency component integrated value and the negative value integrated value corresponding to one restoration filter are input to the CPU 6.
[0024]
These series of processes of filtering processing → high frequency component integration → IFFT → negative value integration are repeated by the number of restoration filters, and the high frequency component integration values and the negative value integration values by the number of restoration filters are output to the CPU 6.
[0025]
The CPU 6 calculates an AF evaluation value based on the input high-frequency component integrated value and / or negative value integrated value, and searches, for example, the minimum value. Then, the corresponding focus lens position (target position) is specified using the restoration filter number corresponding to the searched minimum value as the focus index. Then, the CPU 6 outputs control data for instructing the focus lens control unit 7 to move the focus lens to the target position. In response, the focus lens control unit 7 drives the focus lens to the target position. With the above operation, the AF operation ends.
[0026]
(Embodiment 2 (Reference Example))
The second embodiment (the first embodiment is omitted) will be described with reference to FIG. In the second embodiment, a first method for specifying a focal point in a digital still camera having the configuration shown in the common part of the embodiment will be described.
[0027]
FIG. 2 is an explanatory diagram for explaining a first method of specifying the focal point, and particularly shows an example of a relationship between each restoration filter (No. 1 to No. 12) and a negative value integrated value. I have. In the figure, the horizontal axis represents the restoration filter No. and the vertical axis represents the negative integrated value of the image data restored by each restoration filter. When specifying the focal point, first, a negative value integrated value of the AF area is calculated for each of the restored image data, and is used as an AF evaluation value.
[0028]
Normally, in specifying the focal point, since the negative value integrated value becomes smallest at the focal point, that point is often set as the focal point. In the second embodiment, a search is made for a minimum value from the calculated negative value integrated value group, and when there is one minimum value, that focus point is set. On the other hand, as shown in FIG. In this case, the focal point having the largest point image radius (filter number) is set as the focal point.
[0029]
In the second embodiment, the AF evaluation value is calculated from the negative value integrated value in the AF area, the position of the minimum value of the negative value integrated value is set as the focal point, and further, if there are a plurality of the minimum values, Since the focal point is determined using the largest point image radius as the focal point, the focal point can be determined with high accuracy even when a pseudo peak exists in the negative integrated value.
[0030]
(Embodiment 3 (Reference Example))
Embodiment 3 will be described with reference to FIG. In the third embodiment, a second method for specifying a focal point in a digital still camera having the configuration shown in the common part of the embodiment will be described.
[0031]
FIG. 3 is an explanatory diagram for explaining a second method of specifying the focal point. In particular, FIG. 3 shows the relationship between each restoration filter (No. 1 to No. 12), the negative value integrated value, and the high frequency component integrated value. 9 shows an example of a ratio relationship. In the figure, the horizontal axis represents the restoration filter No., and the vertical axis represents the ratio between the negative value integrated value and the high frequency component integrated value of the image data restored by each restoration filter.
[0032]
In the figure, when specifying the focal point, first, in each of the restored images, the negative value integrated value of the AF area and the high frequency component of the specific area are calculated. Next, a ratio between the calculated negative value integrated value group and the calculated high frequency component integrated value is calculated and used as an AF evaluation value, and a minimum value thereof is searched for and set as a focal point. The reason why the above-mentioned focus is set is that the negative value integrated value is the same as the reason described above, and the high-frequency component integrated value has the largest high-frequency component at the focal point. When viewed from these ratios, at the focal point, the negative value integrated value becomes the minimum and the high frequency component integrated value becomes the maximum, so that the ratio becomes the minimum.
[0033]
Here, the ratio between the negative value integrated value and the high frequency component integrated value can be calculated, for example, by the following equation (1).
[0034]
(Ratio between negative value integrated value and high frequency component integrated value)
= | 1000- (negative value integrated value) | / (high frequency component integrated value) ... (1)
Here, || indicates an absolute value.
[0035]
As described above, if the method of searching for the minimum value of (the ratio between the negative value integrated value and the high frequency component integrated value) is used, the number of minimum values is smaller than in the first method, and the accuracy is reduced. Is improved. On the other hand, since the amount of calculation is smaller in the first method, the time required for specifying the in-focus position is better in the first method.
[0036]
In the second method, when there are a plurality of minimum values of the ratio between the negative value integrated value and the high frequency component integrated value, the one with the largest point image radius (filter number) is set as the focal point. Expression (1) is an example of the definition of the ratio between the negative value integrated value and the high frequency component integrated value, and the present invention is not limited to this.
[0037]
As described above, in the third embodiment, the AF evaluation value is calculated by the ratio between the negative value integrated value and the high frequency component integrated value in the AF area, and the minimum value of the ratio between the negative value integrated value and the high frequency component integrated value is calculated. The focal point is defined as the value, and when there are a plurality of minimum values, the focal point is determined by determining the focal point with the largest point image radius.Therefore, there is a pseudo peak in the negative integrated value. However, it is possible to calculate the focal point with high accuracy.
[0038]
(Embodiment 4 (Reference Example))
Embodiment 4 will be described with reference to FIG. In the fourth embodiment, a third method for specifying the focal point in the digital still camera having the configuration shown in the common part of the embodiment will be described.
[0039]
FIG. 4 is an explanatory view for explaining a fourth method for specifying the focal point. In particular, FIG. 4 shows the relationship between each restoration filter (No. 1 to No. 12), the negative value integrated value, and the high frequency component integrated value. 9 shows an example of a ratio relationship. In the figure, the horizontal axis represents the restoration filter No., and the vertical axis represents the ratio between the negative value integrated value and the high frequency component integrated value of the image data restored by each restoration filter.
[0040]
In the figure, to specify the focal point, first, in each of the restored image data, the negative value integrated value of the AF area and the integrated value of the high frequency component in the specific area are calculated. Next, the ratio of the calculated negative value integrated value to the high frequency component integrated value is calculated, and the calculated value is used as the AF evaluation value to search for the minimum value, and the position corresponding to the minimum value is set as the temporary focus point (FIG. Middle Fa). Also in the third method, the above equation (1) is used to define the ratio between the negative value integrated value and the high frequency component integrated value.
[0041]
Next, scanning is performed again from the provisional focal point toward the larger point image radius (filter number). At this time, if the ratio between the negative value integrated value and the high frequency component integrated value is equal to or less than the specific threshold value (Rth) (Fb, Fc, Fd in the figure), the temporary focus point and the threshold value (Rth) or less are set. Based on the filter number obtained, the focal point is calculated by, for example, the following equation (2).
[0042]
(Focus)
= Fa + (c1 * Fb + c2 * Fc + c3 * Fd) / Num (2)
However, c1, c2, c3: coefficient c1 = c2 = c3 = 1.0
Fa: Temporary focusing point (filter number)
Fb, Fc, Fd: filter numbers below the threshold Rth
Num: the number of filter numbers when the value is equal to or less than the threshold value Rth
The above equation (2) is obtained by adding the weighted average value of the filter numbers that have become equal to or less than the threshold Rth to the temporary focus when calculating the focus. Note that the components to be considered are not limited to the above equation (2), and the coefficients are set to c1 = c2 = c3 = 1.0, but are not limited to this.
[0043]
As described above, in the fourth embodiment, the AF evaluation value is calculated by the ratio between the negative value integrated value and the high frequency component integrated value in the AF area, and the ratio between the negative value integrated value and the high frequency component integrated value is calculated. The minimum value of is set as a temporary focal point, and the ratio between the negative value integrated value and the high-frequency component integrated value is re-scanned in the direction in which the point image radius increases from the temporary focal point, and the ratio falls below a specific threshold value Since the true focal point is determined by fitting the point image radius at that time to the temporary focal point as a weight, the focal point is calculated with high precision even if there is a pseudo peak in the negative integrated value. It becomes possible.
[0044]
(Embodiment 5 (Reference Example))
Embodiment 5 will be described with reference to FIG. In the fifth embodiment, a fourth method for specifying the focal point in the digital still camera having the configuration shown in the common part of the embodiment will be described.
[0045]
FIG. 5 is an explanatory diagram for explaining a fourth method for specifying the focal point. In particular, FIG. 5 shows the relationship between each restoration filter (No. 1 to No. 12), the negative value integrated value, and the high frequency component integrated value. 9 shows an example of a ratio relationship. In the figure, the horizontal axis represents the restoration filter No., and the vertical axis represents the ratio between the negative value integrated value and the high frequency component integrated value of the image data restored by each restoration filter.
[0046]
In the figure, to specify the focal point, first, in each of the restored image data, the negative value integrated value of the AF area and the integrated value of the high frequency component in the specific area are calculated. Next, the ratio between the calculated negative value integrated value and high frequency integrated value is calculated and used as an AF evaluation value, and the minimum value is searched for and set as a temporary focal point (Fa in the figure). In the fifth embodiment, the above equation (1) is used to define the ratio of the ratio between the negative integrated value and the high-frequency integrated value.
[0047]
Next, scanning is performed again from the temporary focal point toward the point image having a larger radius (filter number). At this time, the ratio between the negative integrated value and the high-frequency integrated value is compared with a specific threshold (Rth), and the filter No (Fb, Fc, Fc, The focal point is calculated in consideration of Fd). More specifically, when calculating the focal point, the difference between the filter number that is equal to or less than the threshold value Rth and the ratio value and the threshold value for each filter number, that is, the depth component (Db, Dc, Dd in the figure) ) As a factor, and the focal point is calculated by, for example, the following equation (3).
[0048]
(Focus)
= Fa + (c1 * Fb * Db + c2 * Fc * Dc + c3 * Fd * Dd) / Num (3)
Where c1, c2 and c3 are coefficients
c1 = c2 = c3 = 1.0
Fa: Temporary focusing point (filter number)
Fb, Fc, Fd: filter numbers below the threshold Rth
Db: Depth factor from threshold Rth Db = 1−Rb / Rth
Dc: Depth factor from threshold Rth Dc = 1−Rc / Rth
Dd: depth factor from threshold Rth Dd = 1−Rd / Rth
Num: the number of filter numbers when the value is equal to or less than the threshold value Rth
The above equation (3) is obtained by adding a weighted average value of the filter number that has become equal to or smaller than the threshold Rth and a factor of the depth from the threshold Rth to the temporary focus when calculating the focal point. Note that the components to be considered are not limited to the above equation (3), and the coefficients are set to c1 = c2 = c3 = 1.0, but are not limited thereto.
[0049]
As described above, in the fifth embodiment, the AF evaluation value is calculated by the ratio between the negative value integrated value and the high frequency component integrated value in the AF area, and the minimum value of the ratio between the negative value integrated value and the high frequency component integrated value is calculated. The value is set as a temporary focus position, and the ratio between the negative value integrated value and the high frequency component integrated value is re-scanned in a direction in which the point image radius increases from the temporary focus point, and the ratio becomes equal to or less than a specific threshold value. In this case, the true focus point is specified by adding the weight of the point image radius at that time and the ratio of the ratio to the threshold value or less to the temporary focus point. It is possible to calculate the focal point with high accuracy even if there is a pseudo peak.
[0050]
(Embodiment 6 (Reference Example))
Embodiment 6 will be described with reference to FIG. In the fifth embodiment, a fifth method for specifying the focal point in the digital still camera having the configuration shown in the common part of the embodiment will be described. FIG. 6 is a flowchart for explaining the fifth method for specifying the focal point, and shows an outline of the one-shot AF operation according to the sixth embodiment.
[0051]
In FIG. 6, first, at the initial focus lens position, an image is captured and image data (for example, image capture 8 bits / pixel) is captured (step S10), and an AF area (for example, an area of 256 × 256 pixels) is cut out (step S11). ). Next, in this AF area, an average luminance (Yave) is calculated (step S12), and it is determined whether or not this average luminance (Yave) is within a specific luminance range (appropriate luminance range). Specifically, it is determined whether or not Yellow (for example, 100) ≦ Yave ≦ Yhigh (for example, 130) (step S13). If the result of this determination is that Yellow ≦ Yave ≦ Yhigh, the flow shifts to step S14 to start the substantial processing of the one-shot AF.
[0052]
On the other hand, if Ylow ≦ Yave ≦ Yhigh, the process proceeds to step S15, and adjustment is performed so that the average brightness value (Yave) falls within the specific brightness range (Ylow ≦ Yave ≦ Yhigh). Specifically, this adjustment is performed arithmetically for all pixels in the AF area. That is, when the target average luminance value is 100 and the average luminance is 95, the difference of 5 is added (subtracted) as an offset. In this embodiment, the thresholds of the appropriate luminance range are set to Ylow = 100 and Yhigh = 130, and the target average luminance to be adjusted when the threshold is exceeded is set to 100. However, A / D conversion is performed. The value is not limited to these values as long as it is 40% to 50% with respect to the maximum luminance setting standard. The average luminance value is not limited to the method of adjusting the average luminance value by arithmetic operation, and the photographing (image capturing) may be performed again by changing photographing conditions and the like, and the average luminance value may be calculated. When the adjustment of the average luminance value is completed, the process proceeds to step S14, and the substantial processing of the one-shot AF is started.
[0053]
Hereinafter, the substantial processing of the one-shot AF will be described. First, in step S14, the luminance signal corresponding to the image data in the AF area is converted from a spatial component to a frequency component by the FFT (IFFT) calculation unit 44. Then, a search for a focal point is performed for each restoration filter (for example, FiltNo. 0 to 120) (step S16). Specifically, first, the filtering unit 45 accesses the data (1 restoration filter) read from the restoration filter data ROM 46 for the signal converted to the frequency component (step S24), and performs restoration processing. (Step S17). Then, the high-frequency component integrator 47 extracts a specific high-frequency component of the input signal and accumulates the specific high-frequency component to calculate a high-frequency component integrated value (step S18). Further, the FFT (IFFT) calculation unit 44 converts the data subjected to the filtering process into a spatial component (step S19). The negative value integrating unit 48 integrates the negative values of the reconverted data to calculate a negative value integrated value (step S20). The calculation of the high frequency component integrated value and the negative value integrated value is performed for all restoration filters (Filt Nos. 0 to 120) (step S21).
[0054]
Finally, based on the calculated high-frequency component integrated value group and negative value integrated value group, the in-focus position is calculated (step S22), and the focusing lens is driven to the in-focus position (step S23).
[0055]
As described above, in the sixth embodiment, prior to performing a series of one-shot AF operations, the average luminance in the AF area is calculated, and when the average luminance is out of the specific luminance range, the average luminance is calculated. Since the focal point is specified by adjusting the luminance so that the AF area average luminance falls within the specific luminance range, and then entering the one-shot AF operation, it is taken in when calculating the focal point index by the one-shot AF. It is possible to prevent deviation from the original focal point caused by the brightness of the AF area, and it is possible to determine the focal point with high accuracy.
[0056]
(Embodiment 7 (Reference Example))
Embodiment 7 will be described with reference to FIG. 7 and FIG. In the seventh embodiment, in the digital still camera having the configuration shown in the common part of the embodiment, an example in which window function processing is performed on an AF area prior to performing one-shot AF to reduce the influence of the vicinity of the AF area. Will be described.
[0057]
FIG. 7 shows an example of the relationship between the pixel position in the AF area and the relative intensity. In the figure, the horizontal axis indicates the pixel position in the AF area, and the vertical axis indicates the relative intensity. Although the drawing is shown in one dimension for simplicity, it is actually expanded to two dimensions.
[0058]
First, photographing (image capture) is performed at the initial focus lens position, and an AF area is cut out. Here, the size is 256 × 256 pixels. In this AF area, window function processing as shown in the following equation (4) is performed two-dimensionally.
[0059]
V [i]
= Data [i] * {0.54-0.46 * cos (2.0 * 3.14 * i / N)} (4)
Here, V: pixel data at pixel position i after window function processing
i: Pixel position 0 to 255
data: pixel data at pixel position i before window function processing
N: window size N = 256 here
This window function process is a so-called weighting based on the pixel position, and indicates that the weight as the image information decreases as the edge of the AF area approaches the edge as shown in FIG. Then, after the window function process is completed, the subroutine enters the substantial process of the one-shot AF. In this embodiment, a Hamming window is used as a window function, but the present invention is not limited to this.
[0060]
FIG. 8 shows a configuration example of a window function processing unit for performing the window function processing. The window function processing unit shown in FIG. 8 includes a window function processing unit 20 that performs window function processing based on the window function selected by the selector 21, a window function data ROM 22 storing a plurality of types of window functions 1 to M, A selector 21 for selecting an optimal window function from a plurality of window functions.
[0061]
Next, the operation of the window function processing unit will be briefly described. First, the image data taken in from the CCD is A / D converted and then converted into luminance data. Then, image data (luminance data) of the AF area is extracted and input to the window function processing unit 20. The window function processing unit 20 calculates the window function data and the image luminance data read from the window function data ROM 22, and outputs the image data subjected to the window function processing. At this time, a plurality of window functions are stored in the window function data ROM 22, and the selector 21 selects which window function process is to be performed. The selector 21 may select a window function based on information from an interchangeable lens or information from a restoration filter, for example. In order to implement the seventh embodiment, the window function processing unit shown in FIG. 8 may be added to the digital still camera having the configuration shown in the common part of the embodiment.
[0062]
As described above, in the seventh embodiment, the window function processing is performed on the AF area before performing a series of one-shot AF, so that the influence near the AF area can be reduced. Become. In addition, generally, when Fourier transform (FFT) processing, which is one of the processings when performing the one-shot AF processing, is performed on the AF area, an edge near the side of the AF area becomes an edge, and unnecessary high-frequency components may be generated. In addition, in the one-shot AF, the high-frequency component integrated value may be used to specify the focal point. When the unnecessary high-frequency component is mixed, the precision of the focal point specification is reduced. Since the window function processing is performed on the area, the influence can be prevented.
[0063]
(Embodiment 8 (Embodiment of the present invention))
Embodiment 8 will be described with reference to FIG. In the eighth embodiment, an example of efficient arrangement of restoration filters in a digital still camera having the configuration shown in the common part of the embodiment will be described.
[0064]
FIGS. 9A and 9B are diagrams for explaining a restoration filter according to the eighth embodiment. FIG. 9A illustrates an example of the arrangement of a conventional restoration filter, and FIG. 9B illustrates the arrangement of the restoration filter according to the eighth embodiment. Here is an example. FIG. 9 shows the case where the initial focus lens position is set to infinity. However, FIG. 3 shows the structure of the restoration filter in an easy-to-understand manner, and this filter does not necessarily exist at an actual distance.
[0065]
In the configuration of the conventional restoration filter shown in FIG. 9A, 120 filters are used, and in the configuration of the restoration filter of the present embodiment shown in FIG. 9B, 60 restoration filters are used. The restoration filter is assigned to a distance of 30 [cm] to 300 [cm].
[0066]
By the way, when specifying the focal point, actually, the focal point is not only one point but has a certain range of width in consideration of the depth of field and the like, that is, it is considered that the focal point is apparently focused. Is determined within the range of In particular, the farther the subject is, the wider the allowable focus range becomes, and it is meaningless to arrange a plurality of corresponding restoration filters in this range. On the other hand, the closer the distance is, the narrower the permissible focusing range becomes. Therefore, it is necessary to arrange more corresponding restoration filters.
[0067]
In the eighth embodiment, as shown in FIG. 9B, the restoration filters are arranged so as to be dense on the near side and coarse on the far side. Thereby, highly accurate one-shot AF can be realized. In this embodiment, from the viewpoints of speed, cost, and mounting, the number of restoration filters is reduced as much as possible, and the number of restoration filters is set to 120 to 60 and allocated to each distance. However, the present invention is not limited to this. Also, the number of restoration filters is not limited to those listed here.
[0068]
(Embodiment 9 (Reference Example))
Embodiment 9 will be described with reference to FIG. In the tenth embodiment, an operation example in which a plurality of AF areas are provided for focusing in a digital still camera having the configuration shown in the common part of the embodiment will be described.
[0069]
FIG. 10 is an explanatory diagram for explaining an AF area according to the ninth embodiment. FIG. 10A shows a case where AF areas are set at several places near the center of an image frame. ) Indicates an AF area obtained by equally dividing the entire image frame.
[0070]
FIG. 10A shows an image frame of 1280 × 1024 pixels and three AF areas (256 × 256 pixels) set in parallel near the center of the image frame. Area, center AF area, and right AF area. FIG. 10B shows a case where an image frame is set to 1280 × 1024 pixels and an AF area in which the image frame is equally divided into 5 × 4 is set. For convenience, from the upper left to the lower right, , AF area 21, AF area 22,... AF area 45.
[0071]
In FIG. 10A, it is assumed that there are a plurality of subject candidates and the photographer selects which subject is to be a target. First, in a digital still video camera, areas corresponding to a left AF area, a center AF area, and a right AF area are cut out from a captured image frame by one-shot AF operation, and a focusing index is set for each area. A value is calculated.
[0072]
The photographer operates an operation button (not shown) to select an AF area including the target subject. In response to this, the digital still video camera takes out the focus index value of the selected AF area and controls the focus lens to move to a predetermined position based on the value. Accordingly, when there is only one AF area, a redundant operation in which the AF area is sequentially adjusted to the subject candidate and the one-shot AF operation is repeated each time is avoided, and a high-speed one-shot AF operation can be obtained. .
[0073]
In FIG. 10B, the same operation as that described in FIG. 10A is performed. However, in this case, when calculating the focus index value, the focus index value need not be calculated for all AF areas, but may be calculated only for some AF areas picked up by the photographer. When the focus index value is calculated for all AF areas, if the subject distance information is stored together with the image information, appropriate image processing means typified by a personal computer and application software can be used as post processing. Accordingly, it is possible to generate an in-focus image from near to far.
[0074]
In the present invention, the size, shape, and arrangement of the AF areas shown in FIG. 10 are not limited, and each AF area may have an overlapping portion.
[0075]
(Embodiment 10 (Reference Example))
Embodiment 10 will be described with reference to FIG. 11 and FIG. In the tenth embodiment, an operation example will be described in which a one-shot AF is used for coarse adjustment and a hill-climbing servo method is used for fine adjustment in a digital still camera having the configuration shown in the common part of the embodiment.
[0076]
In the tenth embodiment, one-shot AF is used for coarse adjustment to determine an approximate subject position, the focus lens is moved to the corresponding position, and then fine adjustment is performed by a hill-climbing servo method. The focus lens is moved to an in-focus position.
[0077]
FIG. 11 is an explanatory diagram for explaining the conventional one-shot AF and hill-climbing servo method, and FIG. 12 is an explanatory diagram for explaining the one-shot AF and the hill-climbing servo method according to the tenth embodiment. FIGS. 11 and 12 correspond to a diagram for explaining the one-shot AF operation (a diagram shown in the upper part of the figure) and a diagram for explaining the hill-climbing servo method (a figure shown in the lower part of the figure). Shown.
[0078]
In FIG. 12, a diagram for explaining the one-shot AF operation shows a relationship among a lens position, a subject position, a restoration filter, and a one-shot AF evaluation value. As shown in the figure, when the focal position is between ∞ and Near (closest), seven restoration filters No. 7 each corresponding to each distance. 1 to No. 7 (filter group). Each restoration filter No. 1 to No. 7, the image is restored, and the AF evaluation value is obtained. Here, the focal point is a filter number indicating the minimum value of the AF evaluation value. In the case of the example shown in FIG. In step 6, the AF evaluation value is output to indicate that it is the in-focus position. Based on this signal, the focus lens is set to the filter No. Move to the position corresponding to 5. Conventionally, as shown in FIG. Had moved to 6.
[0079]
After that, the control is switched to the control of the hill-climbing servo method to perform fine adjustment. Conventionally, as shown in FIG. 11, in order to detect the direction of the focal point, it is necessary to move at least two points to acquire data and to know whether the AF evaluation value is in the inclination direction (a hollow slope or a rising slope). . On the other hand, in the present embodiment, as shown in FIG. 12, the focus lens is moved to the index value immediately before the in-focus index value of the coarse adjustment, and then the No. 1 lens is moved by the hill-climbing servo method. 5 to No. 5 6 to move to the final focus position. In the hill-climbing servo method, the position at which the AF evaluation value becomes maximum is detected as the focus position. That is, the scanning direction by the hill-climbing servo method is determined to be one direction (in this embodiment, the direction in which the filter number increases). Therefore, a redundant operation for detecting a new focusing direction as in the related art is unnecessary.
[0080]
Although seven restoration filters are provided for the one-shot AF coarse adjustment, the invention is not limited to this.
[0081]
As described above, in the tenth embodiment, a temporary focus position is determined by one-shot AF, and then scanning is performed by a hill-climbing servo method from a position immediately before the temporary focus position to finally determine the focus position. Since the in-focus position is determined, the number of restoration filters can be reduced, the memory cost can be reduced, and a high-precision and high-speed one-shot AF can be realized with an inexpensive configuration. .
[0082]
In the tenth embodiment, the high-frequency component used in the hill-climbing servo method can use the FFT engine (FFT (IFFT) calculation unit 44 and high-frequency component integrator 47, etc.) used in the one-shot AF. No new device is required for the servo method.
[0083]
It should be noted that the present invention is not limited to the above-described embodiment, and can be appropriately changed without changing the gist of the invention.
[0084]
Further, the automatic focus control device of the present invention is widely applicable to an image input device using an image sensor such as a video camera and a still video camera.
[0085]
【The invention's effect】
The automatic focus control device according to claim 1 captures an image at an initial focus lens position and outputs image data, and a plurality of restorations corresponding to the point image radius stored in the storage unit for the image data. An autofocus that calculates an AF evaluation value for each of the image data groups restored for each restoration filter by performing image data restoration using a filter, and compares the calculated AF evaluation values to determine a focus position. In the control device, since the corresponding restoration filter is stored in the storage means at a small point image radius interval at a short distance compared to a long distance, the restoration filter can be efficiently arranged without waste. A highly accurate one-shot AF can be realized.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating a configuration of a digital still camera to which an automatic focus control device according to an embodiment is applied.
FIG. 2 is an explanatory diagram according to a second embodiment (reference example), for explaining a first method of specifying a focal point;
FIG. 3 is an explanatory diagram according to a third embodiment (reference example) for explaining a second method of specifying a focal point;
FIG. 4 is an explanatory diagram according to a fourth embodiment (reference example), for explaining a third method of specifying a focal point;
FIG. 5 is an explanatory diagram according to a fifth embodiment (reference example), for explaining a fourth method of specifying a focal point;
FIG. 6 is an explanatory diagram according to a sixth embodiment (reference example), for explaining a fifth method of specifying a focal point;
FIG. 7 is for describing an operation example in which a window function process is performed on an AF area to reduce the influence near the AF area before performing one-shot AF according to the seventh embodiment (reference example); FIG.
FIG. 8 is a block diagram showing a configuration of a window function processing unit according to the seventh embodiment (reference example).
FIG. 9 is an explanatory diagram related to Embodiment 8 (Embodiment of the present invention) for explaining an example of arrangement of restoration filters.
FIG. 10 is an explanatory diagram for describing an AF area according to the ninth embodiment (reference example).
FIG. 11 is an explanatory diagram for explaining a one-shot AF and a hill-climbing servo method according to the related art.
FIG. 12 is an explanatory diagram for explaining a one-shot AF and a hill-climbing servo method according to the tenth embodiment (reference example);
[Explanation of symbols]
1 lens system (initial focus lens)

Claims (1)

The image data is output at the initial focus lens position, and the image data is restored by using a plurality of restoration filters corresponding to the point image radius stored in the storage means. In an automatic focus control device that calculates an AF evaluation value for each of the image data groups restored for each filter and compares the calculated AF evaluation values to determine an in-focus position, An automatic focus control device wherein corresponding restoration filters are stored at intervals of a point image radius finer than at a long distance.

Images