JP2004029105A - Autofocus device and af evaluated value acquiring method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an autofocus device capable of realizing accurate focusing operation, and an AF evaluated value acquiring method. <P>SOLUTION: When an AF evaluated value is calculated in a specified cycle in the autofocus control of a contrast detecting system, a contrast value is calculated for respective pixels constituting a specified evaluation range on a subject image. On the basis of the calculated value, contrast vector (arrow in figure) showing the relative contrast relation of the respective pixels to their peripheral pixels in eight directions is calculated. Thereafter, the respective pixels are grouped according to the difference of the direction of the contrast vector, and the AF evaluated value is calculated by the elements (pixels at a painted-out spot in figure) of the vector group selected under a specified condition based on the contrast values of the pixels except the pixels (5 and 12) isolated in the group. The noise component of the subject image is prevented from being mixed with the AF evaluated value, and a probability that the noise component is mixed is restrained to be extremely low. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an autofocus device suitable for use in a camera device such as a digital camera and an AF evaluation value acquisition method.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, a contrast detection method is often used for autofocus control (AF control) of a digital camera that captures an image of a subject using a CCD or MOS type solid-state imaging device, converts the image into image data, and records the image data. . Such a method is also employed in digital video cameras, silver halide cameras using a CCD as a sensor, and the like.
[0003]
In the AF control of the contrast detection method, the focus lens is moved to some points in the optical axis direction, and a contrast value (AF evaluation value) in a predetermined range of a subject image captured at each position is calculated. For example, for each pixel in a predetermined range, a contrast value that is a luminance difference from surrounding pixels is obtained, and a total value thereof is calculated. Then, based on the comparison of the AF evaluation values calculated at a plurality of points and the gradient of the changing contrast value, the peak position where the contrast value becomes maximum is determined, and the focus lens is moved with the peak position as the focus position. I have. In such AF control, the evaluation range in the subject image, that is, the range for which the AF evaluation value is to be calculated may be selectable from preset sizes such as spots and wides.
[0004]
[Problems to be solved by the invention]
However, in the conventional autofocus method, even if the size and position of the evaluation range can be selected, all the contrast values within the evaluation range are related to the evaluation value. Therefore, for example, when there is much noise in the subject image as in the case of low illuminance, the noise component tends to be mixed into the AF evaluation value, and the detection accuracy of the peak position is reduced.
[0005]
In addition, during the detection of the peak position, the evaluation range needs to be fixed to some extent in the subject image (to such an extent that a high-contrast portion of the evaluation range remains in the evaluation range). For this reason, when the subject moves or the apparatus body moves due to camera shake or the like during the detection period of the peak position, there is a problem that a deviation occurs between the peak position where the AF evaluation value is maximum and the focusing position. Was.
[0006]
The present invention has been made in view of such conventional problems, and has as its object to provide an autofocus apparatus capable of performing a highly accurate focus operation, an AF evaluation value acquisition method, and a program used for realizing them. And
[0007]
[Means for Solving the Problems]
In order to solve the above problem, according to the first aspect of the present invention, an image of a subject is taken while moving an optical system in an optical axis direction, and an AF evaluation value is obtained based on a contrast value for each pixel of the taken subject image. In an autofocus apparatus that determines a focus position of an optical system based on an acquired AF evaluation value, a direction of a height relationship between contrast values of each pixel and peripheral pixels is determined based on a contrast value of each pixel included in the subject image. Acquisition means for acquiring the vector data indicated by, a decision means for deciding a pixel for which the AF evaluation value is to be calculated based on the vector data of a plurality of pixels acquired by the acquisition means, and a pixel decided by the decision means And a calculating means for calculating the AF evaluation value based on the contrast value.
[0008]
In such a configuration, the AF evaluation value sequentially acquired while the optical system is moved in the optical axis direction indicates, in the direction, the level relationship of the contrast value of each pixel among the pixels constituting the subject image with the surrounding pixels. It is calculated based on the contrast value of only a predetermined pixel determined according to the vector data. Therefore, it is possible to prevent the noise component of the subject image from being mixed into the AF evaluation value used for determining the focus position of the optical system, or to extremely reduce the probability that the noise component is mixed.
[0009]
Further, in the invention according to claim 2, when determining a pixel for which an AF evaluation value is to be calculated, the determining unit determines, among pixels having the same direction indicated by the vector data, pixels which are adjacent to each other. Where the number of pixels is equal to or less than a predetermined number, that is, pixels having the same direction are not adjacent at all (isolated pixels) and pixels having the same direction are adjacent to each other by a predetermined number or less (connected pixels) And were excluded.
[0010]
In such a configuration, the AF evaluation value is calculated based on the contrast value of only pixels that exclude pixels that are likely to be affected by noise among the pixels that form the subject image. Therefore, it is possible to prevent the noise component of the subject image from being mixed in the AF evaluation value.
[0011]
Further, in the invention according to claim 3, the determining means divides the pixels constituting the subject image into vector groups for each direction of the vector data, and includes pixels included in the vector group selected based on a predetermined condition. Is determined as a pixel for which the AF evaluation value is to be calculated.
[0012]
In such a configuration, the AF evaluation value is calculated based on the contrast values of only the pixels that contribute to the determination of the focus position of the optical system among the pixels that form the subject image. Therefore, the probability that the noise component of the subject image is mixed in the AF evaluation value can be extremely reduced.
[0013]
In the invention according to claim 4, the predetermined condition includes a size of an aggregate composed of a plurality of pixels in which the vector data adjacent to each other are the same.
[0014]
Even in such a configuration, the probability that the noise component of the subject image is mixed in the AF evaluation value can be extremely reduced.
Further, in the invention according to claim 5, the predetermined condition includes the number of pixels whose directions indicated by the vector data are the same.
[0015]
Even in such a configuration, the probability that the noise component of the subject image is mixed in the AF evaluation value can be extremely reduced.
[0016]
Further, in the invention of claim 6, the predetermined condition includes a magnitude of a contrast value for each vector group.
[0017]
Even in such a configuration, the probability that the noise component of the subject image is mixed in the AF evaluation value can be extremely reduced.
[0018]
Further, in the invention according to claim 7, the deciding means decides, from among the pixels included in the vector group selected based on a predetermined condition, a pixel constituting an aggregate part in which pixels adjacent to each other are the largest. , Are determined as pixels for which AF evaluation values are calculated.
[0019]
Even in such a configuration, the probability that the noise component of the subject image is mixed in the AF evaluation value can be extremely reduced.
[0020]
Further, in the invention according to claim 8, the acquisition unit is configured to determine, based on a contrast value for each pixel included in the AF evaluation area in the subject image, a contrast value between each pixel in the AF evaluation area and a peripheral pixel. Obtaining vector data indicating the height relationship in the direction, wherein the determining unit determines a pixel to be calculated for the AF evaluation value in the AF evaluation area based on the vector data of a plurality of pixels obtained by the obtaining unit. It was decided.
[0021]
In such a configuration, the AF evaluation values sequentially acquired while the optical system is moved in the optical axis direction are different from the pixels constituting the AF evaluation area in the subject image in the level of the contrast value between each pixel and the surrounding pixels. It is calculated based on a contrast value of only a predetermined pixel determined according to vector data indicating a relationship in a direction. Therefore, it is possible to prevent the noise component in the AF evaluation area from being mixed into the AF evaluation value used for determining the focus position of the optical system, or to extremely reduce the probability that the noise component is mixed.
[0022]
In the ninth aspect of the invention, the contrast value of each pixel of the subject image is a value indicating a difference between the luminance value of the pixel and the luminance values of peripheral pixels.
[0023]
According to the tenth aspect of the present invention, an image of a subject is taken while moving the optical system in the optical axis direction, and an AF evaluation value is obtained based on a contrast value for each pixel of the taken subject image. In an autofocus device that determines a focus position of an optical system based on a value, based on a contrast value of each pixel forming the subject image, vector data indicating a direction of a height relationship of a contrast value between each pixel and a peripheral pixel is obtained. Acquiring means for acquiring, setting means for setting an AF evaluation range of the subject image for which the AF evaluation value is to be calculated based on the vector data of a plurality of pixels acquired by the acquiring means, and setting by the setting means Calculating means for calculating the AF evaluation value based on the contrast value of the pixel in the AF evaluation range.
[0024]
In such a configuration, when the subject relatively moves while the optical system is moved in the optical axis direction, the position of the AF evaluation range in the subject image is set or updated following the movement of the subject and updated. An AF evaluation value is calculated based on the contrast values of the pixels in the AF evaluation range.
[0025]
Further, in the invention according to claim 11, the setting unit divides the pixels constituting the subject image into vector groups for each direction of the vector data, and sets the pixels included in the vector group selected based on a predetermined condition. Of these, the set portion having the largest number of pixels adjacent to each other is set as the AF evaluation range.
[0026]
In such a configuration, the position of the AF evaluation range in the subject image can more reliably follow the movement of the subject.
[0027]
According to the twelfth aspect of the present invention, an image of a subject is taken while moving the optical system in the optical axis direction, and an AF evaluation value is obtained based on a contrast value for each pixel of the taken subject image. A method of acquiring the AF evaluation value in autofocus control for determining a focus position of an optical system based on a contrast value, wherein a contrast between each pixel and peripheral pixels is determined based on a contrast value of each pixel included in the subject image. A step of obtaining vector data indicating the direction of the height of the value, a step of determining a pixel for which the AF evaluation value is to be calculated based on the obtained vector data of a plurality of pixels, and a step of determining the AF based on the determined contrast value of the pixel. And a step of calculating an evaluation value.
[0028]
According to such a method, the AF evaluation value sequentially acquired while the optical system is moved in the optical axis direction is determined by the direction of the contrast value of each pixel among the pixels constituting the subject image with respect to the surrounding pixels. Is calculated based on the contrast value of only predetermined pixels determined in accordance with the vector data indicated by. Therefore, it is possible to prevent the noise component of the subject image from being mixed into the AF evaluation value used for determining the focus position of the optical system, or to extremely reduce the probability that the noise component is mixed.
[0029]
Further, according to the invention of claim 13, an image of a subject is taken while moving the optical system in the optical axis direction, an AF evaluation value is obtained based on a contrast value for each pixel of the taken subject image, and the obtained AF evaluation A computer having an autofocus device that determines an in-focus position of an optical system based on a value indicates a direction of a height relationship between contrast values of each pixel and peripheral pixels based on a contrast value of each pixel constituting the subject image. Obtaining means for obtaining vector data, determining means for determining a pixel for which the AF evaluation value is to be calculated based on vector data of a plurality of pixels obtained by the obtaining means, and contrast of the pixel determined by the determining means This is a program for functioning as a calculating means for calculating the AF evaluation value based on the value.
[0030]
BEST MODE FOR CARRYING OUT THE INVENTION
An embodiment of the present invention will be described below with reference to the drawings.
(First Embodiment)
FIG. 1 is a block diagram of a digital camera 1 showing an embodiment of the present invention. The digital camera 1 has an AF function, and includes a focus lens 2, a zoom lens 3, a CCD 4, a CDS / AD block 5, a TG (Timing Generator) 6, a CCD data pre-processing block 7, a color processing (CP). A block 8, a JPEG encoder 9, a DRAM 10, a ROM 11, a RAM 12, a CPU 13, an image display unit 14, a key block 15, a card interface 16, and a motor drive block 17 are provided. A memory card 18 detachably mounted in a card slot of the main body is connected.
[0031]
Each of the focus lens 2 and the zoom lens 3 is composed of a lens group (not shown). The motor drive block 17 includes a focus motor 170a for driving the focus lens 2 in the optical axis direction, a zoom motor 170b for driving the zoom lens 3 in the optical axis direction, and a focus motor 170a and a zoom motor 170b in accordance with a control signal sent from the CPU 13. And motor drivers 171a and 171b for driving the respective components.
[0032]
The CCD 4 photoelectrically converts the subject image projected via the focus lens 2 and the zoom lens 3 and outputs the image as an imaging signal. The TG 6 drives the CCD 4 by generating a timing signal of a predetermined frequency. The CDS / AD block 5 samples the output signal from the CCD 4, removes noise, and converts it into a digital signal. The CCD data pre-processing block 7 performs data processing such as luminance signal processing on the imaging signal converted into a digital signal. The color processing (CP) block 8 performs color processing such as color separation on the image signal on which the luminance signal processing and the like have been performed, and generates Y, Cb, and Cr image data. The DRAM 10 sequentially stores the Y, Cb, and Cr image data after the color processing.
[0033]
The image display unit 14 includes a color LCD, a driving circuit for driving the color LCD, and the like. When the shutter key is not pressed while the shooting mode is set (shooting standby state), the image is taken in from the CCD 4 and stored in the DRAM 10. A through image based on the stored image data for one frame (screen) is displayed, and in a state where the reproduction mode is set, a recorded image read from the memory card 18 and based on the expanded image data is displayed. When recording an image, the JPEG encoder 9 compresses the image data input from the color processing (CP) block 8 by JPEG. The memory card 18 records the compressed image data sent via the card interface 16. The recorded image data is read out when the recorded image is reproduced, decompressed by the JPEG encoder 9, and displayed on the image display unit 14. The key block 15 includes a switching key used for switching operations between a shooting mode and a reproduction mode, and various operation keys such as a shutter key, and sends an operation signal corresponding to a key operation to the CPU 13.
[0034]
The ROM 11 stores programs and various data necessary for control of the digital camera 1 by the CPU 13, that is, AF control, AE control, AWB control, and the like. The CPU 13 stores an operation signal from the key block 15 and the program. Operates as a working memory, thereby functioning as an acquisition unit, a determination unit, a calculation unit, and a setting unit of the present invention.
[0035]
Note that the program need not always be stored in the ROM 11, but may be stored in a predetermined area of the memory card 18, for example. In addition, as long as the program includes a rewritable memory such as an EEPROM in addition to the memory card 18, the program may be supplied to the memory from another device by an arbitrary method such as communication.
[0036]
Next, the operation of the digital camera 1 having the above configuration will be described. FIG. 2 is a flowchart showing an outline of the AF processing in the digital camera 1. The AF processing in the present embodiment is based on the contrast detection method described in the related art. In the AF processing, the moving range of the focus lens is set to the entire area from the end point in the close direction to the end point in the infinity direction. , A series of processes such as imaging (exposure) by a CCD or the like, transfer of an image signal, AD conversion, and calculation of an AF evaluation value (contrast value) are repeatedly performed (steps SA1 to SA4). FIG. 3 is a timing chart showing the operation timing during this period, and the above-described series of processing is performed in the illustrated AF evaluation cycle. When the calculation of the AF evaluation value for each point is completed (YES in step SA5), a peak position at which the AF evaluation value becomes maximum is calculated (step SA6), and the focus lens is moved to the peak position (step SA6). SA7).
[0037]
On the other hand, FIG. 4 is a flowchart showing details of the AF evaluation value calculation process (step SA4) performed for each AF evaluation cycle. Hereinafter, the description will be made assuming that the image (hereinafter, the original image) 100 as shown in FIG. In this process, first, pixel data in an (AF area) evaluation range (center of the screen in the figure) of a predetermined position and size is cut out from the original image 100 (step SA101). The evaluation range (search area) is an area that is fixedly set or an area that is selected (or automatically) by a user from a plurality of areas prepared in advance.
[0038]
Next, a contrast value of each pixel is calculated based on the luminance component data of each pixel of the image 101 (hereinafter referred to as an evaluation image) in the cut-out evaluation range as shown in FIG. 6A (step SA102). Specifically, after the evaluation image 101 is once converted to black and white (only the luminance signal is extracted), a contrast value, which is a luminance difference between each pixel and surrounding pixels, is calculated. To obtain a contrast image 102 in which the magnitude of the contrast value is expressed by the brightness. FIG. 6B is a partially enlarged view of the contrast image 102 corresponding to a part (A part) of the evaluation image 101. It should be noted that the image processing such as obtaining the contrast image 102 based on the contrast ratio may not be performed, and the contrast value of each pixel may be simply calculated. Instead of calculating the contrast value from the black and white image, the contrast value may be obtained by another method.
[0039]
Subsequently, a contrast vector is calculated based on the contrast value (contrast image 102) of each pixel in the evaluation image 101 calculated in step SA102 (step SA103). When calculating a contrast vector based on the contrast image 102, the contrast vector is data indicating a relative luminance relationship between each pixel and a peripheral pixel in a direction (luminance direction). In the embodiment, the calculation is performed based on the luminance relationship between the target pixel and its eight neighboring pixels. That is, FIG. 7A shows the luminance direction when the luminance of (1, 1) is higher than the luminance of the other eight neighboring elements in a pixel area of 3 × 3 [(0, 0) to (2, 2)]. This represents the display rule of (vector). In the present embodiment, the possible luminance directions of each pixel are the eight directions shown in the drawing. FIG. 8 is a schematic diagram showing the contrast vectors of the respective pixels in the region corresponding to FIG. 6B. In the pixels adjacent to each other near the edge of the subject, the contrast vectors are the same. Tends to be the direction.
[0040]
Next, the calculated contrast vectors of the respective pixels are grouped according to the vector (luminance) direction (step SA104), and, among the eight vector groups, the range is wide (the number of pixels having vectors in the same direction is large). A vector group having a large number and a high contrast (sum of the contrast values of all the pixels in the group) is selected (step SA105). Subsequently, it is determined whether or not there is an isolated point among the elements (each pixel having the same luminance direction) of the selected vector group, that is, an isolated point around which no pixel having the same luminance direction exists. Step SA106).
[0041]
Note that FIG. 8 shown above is a diagram in which, when the luminance direction of the vector group that matches the selection condition in step SA105 is “lower left”, the elements (pixels) of the vector group are painted out. In the area, pixels (5, 12) among the pixels (0, 0) to (18, 18) are isolated points, and the isolated points are different from the edge portions of the subject and have noise. Will appear in the part.
[0042]
If there is an isolated point as described above (YES in step SA106), the contrast value of the pixel serving as the isolated point is excluded from the data used for calculating the AF evaluation value (step SA107), and a predetermined value is determined in step SA105. The AF evaluation value is calculated by adding the contrast values of the pixels (excluding the isolated points) of the vector group selected based on the selection condition using a predetermined conditional expression (step SA108). If there is no isolated point (NO in step SA106), the AF evaluation value is calculated by adding the contrast values of all the pixels of the selected vector group using a predetermined conditional expression (step SA108).
[0043]
As described above, in the present embodiment, the calculation of the AF evaluation value uses the contrast of the remaining pixels excluding the elements of the vector group selected under the above-described conditions and excluding the pixels that are isolated points in the group. Use only values. Here, as described above, the isolated point appears in a portion different from the edge portion of the subject and in a portion where noise exists in the evaluation range of the original image 100, and thus, for example, as in the case of low illuminance, Even if 100 contains a lot of noise, it becomes difficult for that noise component to be mixed into the calculated AF evaluation value. Therefore, the peak position of the AF evaluation value can be detected with high accuracy, and as a result, a highly accurate focus operation can be performed.
[0044]
In the present embodiment, when selecting a vector group for which an AF evaluation value is to be calculated from the eight types of vector groups in the AF evaluation value calculation process shown in FIG. Although a vector group having a wide range and a high contrast is selected, for example, a vector group having the widest range and a vector group having the highest contrast may be simply selected. In addition, each group may be ranked based on one or both of “extent of the range” and “height of the contrast”, and a plurality of vector groups having higher ranks may be selected. In ranking based on both criteria, one of the criteria may be weighted. Further, one or a plurality of vector groups for which an AF evaluation value is to be calculated may be selected based on one or more criteria other than those described above. However, when a plurality of vector groups are to be calculated for the AF evaluation value, it is necessary to perform the isolated pixel exclusion processing described in step SA107 for each vector group. Further, in each vector group, a set portion in which a plurality of pixels (pixels having a contrast vector in the same direction) are continuous is specified to be the largest, and the size of the set portion of each specified vector group is compared. The vector group having the largest aggregate part may be selected.
[0045]
Also, when calculating the final AF evaluation value, the contrast value of a pixel that is an isolated point in which no pixel having the same luminance direction exists around the selected vector group is excluded. A pixel in which the number of pixels having the same luminance direction existing therearound (eight peripheral pixels) is equal to or less than a predetermined number, for example, two adjacent pixels having the same luminance direction (two connected pixels) or three pixels Contrast values of adjacent pixels (three connected pixels) may be excluded.
[0046]
Further, in the calculation process of the AF evaluation value shown in FIG. 4, the processes of steps SA106 and SA107 may not be performed, and the contrast value of the pixel serving as an isolated point may not be excluded.
[0047]
Separately from this, the final AF evaluation value is calculated by confirming a set portion where a plurality of pixels are continuous in the vector group selected under the predetermined condition in step SA106 (note that the direction in which each pixel contacts each other). May be any of eight directions), and specifies the largest aggregate portion having the widest range (most pixels) among them, and uses only the contrast values of the pixels constituting the largest aggregate portion to determine the final AF evaluation The value may be calculated. That is, assuming that the pixel area shown in FIG. 8 is the evaluation range of the original image 100 in FIG. 5, the pixel area is filled in a rectangular area having (4, 0) and (12, 8) as diagonal points. The portion indicated by the dotted line is the above-described aggregated portion, and the final AF evaluation value may be calculated using only the contrast value of the filled pixel. In this case, as described above, in the case where a plurality of vector groups are selected as the calculation target of the AF evaluation value, the final value is determined using the contrast value of the pixel constituting each of the aggregate portions having the widest range in each vector group. A typical AF evaluation value is calculated.
[0048]
On the other hand, the eight types of contrast vectors (data in the luminance direction) calculated in step SA103 are shown in FIG. 7A, but may be four types shown in FIG. 7B. That is, FIG. 11B shows a luminance direction when the luminance of (1, 1) is higher than the luminance of the other four neighboring elements in a pixel area of 3 × 3 [(0, 0) to (2, 2)]. This is a display rule of (vector), and thus, the luminance direction that each pixel can take may be four directions. In this case, the load on the CPU 13 during the AF control can be reduced by reducing the amount of data processing related to the AF evaluation value calculation process. However, in this case, after the contrast vector of each pixel is divided into four vector groups in step SA104 described above, a predetermined vector group is selected from the four vector groups in step SA105. In this case, a selection condition (or selection criterion) different from the above-described embodiment may be applied.
[0049]
Further, in the present embodiment, during the AF process, when the calculation process of the AF evaluation value is executed for each AF evaluation cycle (see FIG. 3), the process of selecting the vector group for which the AF evaluation value is to be calculated (see FIG. The fourth step SA105) is executed every time, but the selection process may be omitted for the second and subsequent times. In other words, for the second and subsequent times, the vector group selected (used) for the first time (however, the element pixels vary) is used to calculate the vector group for which the AF evaluation value is to be calculated for the second and subsequent times. May be locked. In this case, in step SA104 in FIG. 4, it is not necessary to perform the process of dividing the contrast vector of each pixel into eight vector groups, and only the process of extracting the pixels of the first selected vector group may be performed. .
[0050]
(Second embodiment)
Next, a second embodiment will be described. In the present embodiment, a program for causing the digital camera 1 to perform an AF evaluation value calculation process different from that of the first embodiment during the AF process of FIG. It is.
[0051]
Hereinafter, the calculation process of the AF evaluation value in the present embodiment will be described with reference to the flowchart shown in FIG. 9. The evaluation image 101 is cut out from the original image 100 captured in step SA2 in FIG. 2 (step SA201), and the cut out evaluation image After calculating the contrast value of each pixel in 101 (step SA202), the contrast vector of each pixel is calculated (step SA203), and the calculated contrast vectors of each pixel are grouped by vector direction (step SA204). Up to this point, the processing is the same as the processing of steps SA101 to SA104 in FIG. 4 described in the first embodiment.
[0052]
Next, for each of the eight vector groups grouped in step SA204, it is determined whether or not there is an isolated point in the element where no pixel having the same luminance direction exists in the periphery (step SA205). ). If there is an isolated point (YES in step SA206), the contrast value of the pixel that becomes an isolated point in each vector group is excluded from the data used for calculating the AF evaluation value (step SA207), and then all of the vector groups are deleted. An AF evaluation value is calculated using a contrast value of a pixel (excluding an isolated point), that is, a contrast value of a pixel obtained by removing a pixel that becomes an isolated point in each vector group from all pixels of the evaluation image 101 (step SA208). ). If there are no isolated points in all the vector groups (NO in step SA206), the AF evaluation value is calculated using the contrast values of all the pixels of all the vector groups, that is, the contrast values of all the pixels of the evaluation image 101. (Step SA208).
[0053]
As described above, in the present embodiment, the calculation of the AF evaluation value is performed by subtracting pixels that are isolated points in each vector group from all pixels (elements of all vector groups) of the evaluation image 101. Only the contrast value of is used. For this reason, for the same reason as in the first embodiment, even if the original image 100 contains a lot of noise, for example, at low illuminance, the noise component is mixed in the calculated AF evaluation value. It becomes difficult to do. Therefore, the peak position of the AF evaluation value can be detected with high accuracy, and as a result, a highly accurate focus operation can be performed.
[0054]
Also in the present embodiment, the isolated points (pixels) to be excluded when calculating the final AF evaluation value are such that no pixel having the same luminance direction exists around the same point in the same vector group. However, as in the first embodiment, the number of pixels having the same vector direction existing therearound may be a predetermined number (a small number such as 0 to 3) or less.
[0055]
Separately from this, the contrast values of the pixels (excluding the isolated points) of all the vector groups are used in the calculation of the final AF evaluation value. After confirming the set parts that are in contact with each other (the directions in which they contact each other may be any of eight directions), and specifying the largest set part having the largest range (most pixels) among them, The final AF evaluation value may be calculated using only the contrast values of the pixels constituting the maximum aggregate part.
[0056]
Also in this embodiment, as described in the first embodiment, eight types of contrast vectors (data in the luminance direction) calculated in step SA203 are used, and in step SA204, eight contrast vectors are calculated for each pixel. Although the contrast vector is divided into four groups, the contrast vector of each pixel may be divided into four vector groups, and the processing after step SA205 may be performed based on the divided contrast vectors. In this case, the load on the CPU 13 during the AF control can be reduced by reducing the amount of data processing related to the AF evaluation value calculation process.
[0057]
(Third embodiment)
Next, a third embodiment will be described. In the present embodiment, a program for causing the digital camera 1 to perform an AF evaluation value calculation process different from the first and second embodiments during the above-described AF process of FIG. Is what it is.
[0058]
Hereinafter, the process of calculating the AF evaluation value in the present embodiment will be described with reference to the flowchart shown in FIG. That is, after calculating a contrast value for each pixel for the entire region of the original image 100 captured in step SA2 in FIG. 2 (step SA301), a contrast vector for each pixel is calculated (step SA302), and the calculated contrast vector for each pixel is calculated. Are grouped by vector direction (step SA303).
[0059]
Next, in the entire region of the original image 100, an aggregate portion in which the range in which pixels having the same direction of the contrast vector are continuous is wide and the contrast (the sum of the contrast values of all the pixels in the range) is high is selected. A pixel region corresponding to the set portion is determined as an evaluation range (search area) (step SA304). Thereafter, the AF evaluation value is calculated by adding the contrast values of all the pixels in the determined evaluation range using a predetermined conditional expression (step SA305).
[0060]
As described above, in the present embodiment, the evaluation range to be calculated for the AF evaluation value is not set in advance, but the evaluation range is set for each AF evaluation cycle (see FIG. 3), and each pixel in the evaluation range is set. The AF evaluation value is calculated based on the contrast value of. Therefore, even when the subject moves or the apparatus body moves due to camera shake during the execution of the AF process, the evaluation range follows the subject. Therefore, even in such a case, a highly accurate focus operation can be performed.
[0061]
In the present embodiment, when determining an evaluation range for which an AF evaluation value is to be calculated for each AF evaluation cycle (step SA304), the eight vector groups (the entire area of the original image 100) that have been previously grouped are used. , The vector group (set part) according to the above-mentioned predetermined condition is selected each time, but it may be omitted for the second and subsequent times. In other words, for the second and subsequent times, the evaluation range for the second and subsequent times is determined by using the vector group having the set portion selected for the first time (however, the pixel area formed by the pixels as its elements fluctuates). May be locked. For example, if the luminance direction of the vector group selected at the time of the first calculation of the AF evaluation value is “lower left”, the second and subsequent times include a plurality of aggregate portions composed only of pixels having a “lower left” contrast vector. Is determined as the evaluation range (search area). In this case, the subject captured at the beginning of the AF control can more reliably follow the evaluation range for which the AF evaluation value is calculated, and the accuracy of the focus operation can be further improved.
[0062]
Also in the present embodiment, as described in the first and second embodiments, eight types of contrast vectors (data in the luminance direction) calculated in step SA302 are used, and in step SA303, the contrast vector of each pixel is calculated. Is divided into eight vector groups, but four types of contrast vectors may be used, and the contrast vector of each pixel may be divided into four vector groups. In this case, the load on the CPU 13 during the AF control can be reduced by reducing the amount of data processing related to the AF evaluation value calculation process.
[0063]
In the first to third embodiments described above, the CPU 13 moves the focus lens in the entire range from the approach direction to the infinity direction during the AF control, and searches for the peak position (focus position) of the AF evaluation value. The case where the entire area search is performed has been described. In addition to such control, for example, for the purpose of shortening the AF time, the focus lens is moved from the approach direction to the infinity direction at a large movement interval, and the AF evaluation value is temporarily set. The present invention is also effective in a case where after roughly determining the vicinity of the peak, a peak position is detected in a narrow range around the peak, or in a peak search performed when the search range is narrow.
[0064]
That is, in the peak search, as shown in FIG. 11, a series of processes such as imaging by a CCD or the like, transfer of an image signal, AD conversion, and calculation of an AF evaluation value are repeatedly performed while moving the focus lens at a small interval (step S1). SB1 to SB5), when a peak of the AF evaluation value can be detected during that time (YES in step SB4), the focus lens is moved to the detected peak position (step SB6), and the focus operation is performed. The present invention may be applied to the calculation of the AF evaluation value.
[0065]
Further, the “pixel” used in the first to third embodiments described above does not necessarily match the pixel forming the CCD 4 or the pixel forming the image data recorded on the memory card 18. For example, one contrast value and one contrast vector may be calculated using a plurality of pixels constituting the CCD 4 as one pixel (one unit).
[0066]
Further, in the above description, the case where the present invention is mainly applied to a digital camera for photographing a still image has been described. However, the present invention uses a CCD and another image sensor as a sensor to detect a contrast detection AF. The present invention may be applied to another camera device such as a movie camera that performs control, and in such a case, the same effect as in the above-described embodiment can be obtained.
[0067]
【The invention's effect】
As described above, according to the first, eighth, ninth, and twelfth aspects of the present invention, it is possible to prevent the noise component of the subject image from being mixed into the AF evaluation value used for determining the focus position of the optical system. The probability that components are mixed can be extremely reduced. Therefore, a highly accurate focus operation can be performed.
[0068]
According to the second aspect of the present invention, it is possible to prevent the noise component of the subject image from being mixed in the AF evaluation value. Therefore, a highly accurate focus operation can be performed.
[0069]
Further, in the third, fourth, fifth, sixth and seventh aspects of the present invention, the probability that the noise component of the subject image is mixed in the AF evaluation value can be made extremely low. Therefore, a highly accurate focus operation can be performed.
[0070]
According to the tenth aspect, when the subject relatively moves while moving the optical system in the optical axis direction, the position of the AF evaluation range in the subject image is set, that is, updated, following the movement of the subject. Then, the AF evaluation value is calculated based on the updated contrast value of the pixel in the AF evaluation range. Therefore, a highly accurate focusing operation is possible even when the subject relatively moves.
[0071]
In the invention of claim 11, the position of the AF evaluation range in the subject image can be made to more reliably follow the movement of the subject. Therefore, the accuracy of the focus operation when the subject relatively moves can be further improved.
[0072]
According to the invention of claim 13, by using the same, the apparatus and method of the present invention can be implemented in an autofocus apparatus having a computer. This is also possible with existing devices.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a main part of a digital camera common to each embodiment of the present invention.
FIG. 2 is a flowchart showing an AF processing procedure by a CPU common to the embodiments.
FIG. 3 is a timing chart showing an operation related to autofocus control common to the embodiments.
FIG. 4 is a flowchart illustrating a procedure for calculating an AF evaluation value by a CPU according to the first embodiment;
FIG. 5 is a diagram showing an original image used for calculating an AF evaluation value and an evaluation range thereof.
6A is a diagram illustrating an evaluation image corresponding to an evaluation range of an original image, and FIG. 6B is a partially enlarged view illustrating a part of a contrast image obtained from the evaluation image.
7A is a diagram illustrating a display rule of a contrast vector in a luminance direction, and FIG. 7B is a diagram illustrating another display rule.
FIG. 8 is a schematic diagram corresponding to FIG. 6B showing a contrast vector of each pixel in a part of the evaluation image.
FIG. 9 is a flowchart illustrating a procedure for calculating an AF evaluation value by a CPU according to the second embodiment;
FIG. 10 is a flowchart illustrating a procedure for calculating an AF evaluation value by a CPU according to the third embodiment;
FIG. 11 is a flowchart showing another AF processing procedure by the CPU.
[Explanation of symbols]
1 Digital camera
2 Focus lens
4 CCD
11 ROM
12 RAM
13 CPU
18 Memory card
100 original image
101 Evaluation image
102 Contrast image

Claims (13)

An image of a subject is taken while moving the optical system in the optical axis direction, an AF evaluation value is obtained based on a contrast value for each pixel of the taken subject image, and a focus position of the optical system is determined based on the obtained AF evaluation value. In an autofocus device,
An acquisition unit configured to acquire vector data indicating a direction of a height relationship of a contrast value between each pixel and peripheral pixels based on a contrast value of each pixel included in the subject image;
Determining means for determining a pixel for which the AF evaluation value is to be calculated based on the vector data of a plurality of pixels acquired by the acquiring means;
Calculating means for calculating the AF evaluation value based on the contrast value of the pixel determined by the determining means. The determining means excludes, when deciding a pixel for which an AF evaluation value is to be calculated, pixels in which the number of pixels adjacent to each other is equal to or less than a predetermined number among pixels having the same direction indicated by the vector data. The autofocus device according to claim 1, wherein: The deciding means divides the pixels constituting the subject image into vector groups for each direction of the vector data, and regards the pixels included in the vector group selected based on a predetermined condition as pixels for which an AF evaluation value is to be calculated. 3. The autofocus device according to claim 1, wherein the autofocus device is determined. 4. The autofocus apparatus according to claim 3, wherein the predetermined condition includes a size of an aggregate including a plurality of pixels in which the vector data adjacent to each other are the same. The autofocus apparatus according to claim 3, wherein the predetermined condition includes a number of pixels whose directions indicated by the vector data are the same. 6. The autofocus apparatus according to claim 3, wherein the predetermined condition includes a magnitude of a contrast value for each vector group. The determining means determines, as pixels to be calculated for an AF evaluation value, a pixel included in a vector group selected based on a predetermined condition and which constitutes an aggregate portion having the largest number of pixels adjacent to each other. 7. The auto-focusing device according to claim 3, wherein The obtaining means obtains vector data indicating a direction of a height relationship of a contrast value between each pixel in the AF evaluation area and peripheral pixels based on a contrast value of each pixel included in the AF evaluation area in the subject image. ,
8. The method according to claim 1, wherein the determining unit determines a pixel for which the AF evaluation value is to be calculated in the AF evaluation area based on vector data of a plurality of pixels acquired by the acquiring unit. 9. The autofocus device according to any one of the above. 9. The autofocus apparatus according to claim 1, wherein the contrast value of each pixel of the subject image is a value indicating a difference between a luminance value of the pixel and luminance values of peripheral pixels. . An image of a subject is taken while moving the optical system in the optical axis direction, an AF evaluation value is obtained based on a contrast value for each pixel of the taken subject image, and a focus position of the optical system is determined based on the obtained AF evaluation value. In an autofocus device,
An acquisition unit configured to acquire vector data indicating a direction of a height relationship of a contrast value between each pixel and peripheral pixels based on a contrast value of each pixel included in the subject image;
Setting means for setting an AF evaluation range of the subject image for which the AF evaluation value is to be calculated based on the vector data of a plurality of pixels obtained by the obtaining means;
An auto-focusing device comprising: a calculating unit configured to calculate the AF evaluation value based on a contrast value of a pixel in the AF evaluation range set by the setting unit. The setting unit divides the pixels constituting the subject image into vector groups for each direction of the vector data, and sets a pixel having the largest number of adjacent pixels among the pixels included in the vector group selected based on a predetermined condition. The autofocus apparatus according to claim 10, wherein the portion is set as an AF evaluation range. An image of a subject is taken while moving the optical system in the optical axis direction, an AF evaluation value is obtained based on a contrast value for each pixel of the taken subject image, and a focus position of the optical system is determined based on the obtained AF evaluation value. A method of acquiring the AF evaluation value in auto focus control,
Based on a contrast value for each pixel that constitutes the subject image, a step of obtaining vector data indicating a direction of a height relationship between contrast values of each pixel and peripheral pixels,
Determining a pixel to be a target for calculating the AF evaluation value based on the acquired vector data of a plurality of pixels;
Calculating the AF evaluation value based on the contrast value of the determined pixel. An image of a subject is taken while moving the optical system in the optical axis direction, an AF evaluation value is obtained based on a contrast value for each pixel of the taken subject image, and a focus position of the optical system is determined based on the obtained AF evaluation value. A computer included in the autofocus device,
An acquisition unit configured to acquire vector data indicating a direction of a height relationship of a contrast value between each pixel and peripheral pixels based on a contrast value of each pixel included in the subject image;
Determining means for determining a pixel for which the AF evaluation value is to be calculated based on the vector data of a plurality of pixels acquired by the acquiring means;
A program for functioning as calculating means for calculating the AF evaluation value based on the contrast value of the pixel determined by the determining means.

Images