JP2009025381A - Digital camera - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To prevent false focusing by eliminating influence of a photographer's unintended subject which is included in a focus detection area. <P>SOLUTION: An image by a photographic lens is picked up by an imaging device, and image information is output, and then a specified subject (person's face) is recognized based on the image information, so as to detect the area (face detection area 26) of the specified subject in the picked-up image. The area (face detection area 26) of the specified subject is divided into a plurality of areas (face areas 26a to 26d), and the focusing position of the photographic lens is detected in each of the plurality of divided areas (face areas 26a to 26d) and also the final focusing position is decided based on a plurality of focusing positions detected in the plurality of divided areas (face areas 26a to 26d), then focus adjustment of the photographic lens is performed with the final focusing position as a target drive position. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a digital camera.

  The focus evaluation value is calculated while moving the focusing lens little by little at a predetermined interval (hereinafter referred to as AF (Auto focus) search), and the focus adjustment is performed with the position where the focus evaluation value is maximized as the in-focus position. In a contrast detection type automatic focus adjustment device (hereinafter simply referred to as contrast type AF), a captured image is analyzed to recognize a human face, and a shooting distance to the face region is estimated based on the size of the face region. There has been known a digital camera that determines a lens position as a starting point of an AF search based on the estimated shooting distance (see, for example, Patent Document 1).

Prior art documents related to the invention of this application include the following.
JP 2006-201282 A

  In general, a human face has a low contrast, and a landscape has a high contrast because it contains many high-frequency components. Since contrast AF greatly depends on the contrast level in the focus detection area, when taking a commemorative photo at a scenic spot or the like, if the high-frequency part of the background is included in the focus detection area, the person's face will be focused. Even when captured in the detection area, the focus may be pulled toward the background.

(1) The invention of claim 1 is an image pickup means for picking up an image by a photographing optical system and outputting image information, and a specific subject is recognized based on the image information, and a region of the specific subject in the image is detected. Subject detection means for dividing the area of the specific subject into a plurality of areas, and a focus position detection means for detecting the focus position of the photographing optical system in each of the plurality of divided areas, and detected by the plurality of divided areas A focus position determining unit that determines a final focus position based on a plurality of focus positions; and a focus adjustment unit that performs focus adjustment of the photographing optical system with the final focus position as a target drive position.
(2) In the digital camera according to claim 2, the difference between the infinitely far focus position and the closest focus position among the plurality of focus positions is less than a predetermined value by the focus position determining means. In such a case, the closest focusing position among the plurality of focusing positions is set as the final focusing position, or the average position of the plurality of focusing positions is set as the final focusing position. It is a thing.
(3) In the digital camera according to claim 3, the difference between the focus position on the infinity side and the focus position on the closest side among the plurality of focus positions is greater than or equal to a predetermined value by the focus position determination unit. In this case, the final focusing is performed based on the focusing positions of the remaining divided areas excluding the divided areas in which the difference between the focused position of each divided area and the median value of the focused positions of the plurality of divided areas is a predetermined value or more. The focal position is determined.
(4) In the digital camera according to claim 4, the in-focus position determining unit excludes the divided areas where the difference between the in-focus position of each divided area and the median value of the in-focus positions of the plurality of divided areas is a predetermined value or more. The focus position closest to the focus position of the divided areas is set as the final focus position, or the average position of the focus positions of the remaining divided areas is set as the final focus position. It is a thing.

  According to the present invention, it is possible to prevent false focusing by eliminating the influence of a subject unintended by the photographer included in the focus detection area, and to accurately focus on a specific subject.

  FIG. 1 is a diagram illustrating a configuration of a digital camera according to an embodiment. The photographing lens 1 includes a focusing lens, and forms a subject image on the imaging surface of the image sensor 2. The focusing lens of the photographic lens 1 is driven by a focus motor 10. A stepping motor is normally used as the focus motor 10, and the open loop control pulse drive is performed by the CPU 9.

  The image sensor 2 outputs a subject image signal corresponding to the light intensity distribution of the subject image formed on the imaging surface. The analog signal processing unit 3 includes a CDS circuit, an AGC circuit, a color separation circuit, and the like (not shown), and processes a subject image signal output from the image sensor 2. The subject image signal is subjected to correlated double sampling (CDS) processing in the CDS circuit, and the level of the subject image signal is adjusted in the AGC circuit. The A / D converter 4 converts the processed subject image signal output from the analog signal processing unit 3 into a digital signal. The digital signal processing unit 5 includes signal processing circuits such as a γ correction circuit, a luminance signal, and a color difference signal generation circuit, and performs various processes on the subject image signal.

  The CPU 9 includes an AE calculation unit, an AF calculation unit, an AWB calculation unit, and the like, and performs sequence control, exposure calculation control, focus detection and focus adjustment control, white balance control, and the like for the entire camera. In the conventional AF, as shown in FIG. 2, focus detection areas 21 to 25 are set at the center of the shooting screen and at five locations on the top, bottom, left, and right. Based on the calculation result, the focus motor 10 drives the focusing lens of the photographing lens 1 to adjust the focus.

  The buffer memory 12 is a frame memory that can store data for a plurality of frames imaged by the imaging device 2, and stores image data that has been subjected to a series of processing by the digital signal processing unit 5. When the face recognition function is selected by the user, the face recognition calculation unit 11 performs face recognition calculation processing on the still image display still image held in the buffer memory 12, and the presence of a face in the still image. If detected, the CPU 9 outputs the coordinates of the position and size of the recognized face detection area. FIG. 3 shows an example of the face detection area 26.

  The CPU 9 inputs the position and size of the face detection area, and sets the face detection area 26 as the focus evaluation value calculation area of the contrast AF, that is, the focus detection area as shown in FIG. When the position and size information of the face detection area 26 is updated, or when the user performs a half-press operation on the release button, the focus evaluation value calculation sequence of the contrast AF is started. In contrast AF, there is a correlation between the degree of image blur and the contrast, and focusing is performed using the fact that the contrast of the image is maximized when the image is focused. The level of contrast can be evaluated based on the magnitude of the AF evaluation value obtained from an evaluation function that extracts a high frequency component of the subject image signal.

  When searching for the peak of the focus evaluation value, usually, the focusing lens is shifted by a predetermined amount toward the infinity side or the close side, and the focus evaluation value calculated at that time is compared with the focus evaluation value before the movement. When the focus evaluation value after the movement is large, it is considered that the resolution (in-focus level) tends to increase, and the focusing lens is further moved in the same direction, and the same calculation and comparison are performed. On the other hand, when the focus evaluation value after the movement is small, it is considered that the degree of focusing tends to be low, and the focusing lens is moved in the reverse direction to perform the same calculation and comparison. The position at which the focus evaluation value is maximized, that is, the in-focus position is searched by so-called “mountain climbing AF” in which such processing is repeated.

  Although different from the AF search described above, the focus evaluation value is acquired at predetermined intervals while moving the focusing lens from a predetermined position, for example, from the infinite end to the closest end, and the maximum value of the focus evaluation value is searched from the entire range. A so-called “entire search” may be performed.

  By the way, when contrast type AF is performed using the detected face detection area as the focus detection area 26 (see FIG. 3), a characteristic curve of the focus evaluation value with respect to the position of the focusing lens as shown by a solid line in FIG. 4 may be obtained. is there. As described above, since the background subject includes a lot of high-frequency components, the characteristic curve of the focus evaluation value of only the background is a characteristic curve indicated by a broken line in FIG. On the other hand, since the human face has a small contrast change (spatial frequency change), the characteristic curve of the focus evaluation value of only the person becomes a characteristic curve as shown by a one-dot chain line in FIG. However, as shown in FIG. 3, when a low-contrast human face and a high-contrast background overlap in the focus detection area 26, that is, when there is a perspective conflict, a solid line in FIG. As shown, the background characteristic curve (broken line) and the person characteristic curve (one-dot chain line) may be combined into a focus evaluation value characteristic curve.

  When the focus search is performed based on the characteristic curve of the focus evaluation value, the lens driving is performed with the lens position L3 as a target driving position instead of the position of the person who is the original main subject (lens position L2 shown in FIG. 4). Will go. In this case, although the focus detection area is set by the face detection function, the out-of-focus photo is not correctly focused on the person.

  In addition, the focus position may not be detected correctly for other reasons. Depending on the configuration of the optical system, the focal length variation within the focus adjustment movable range may be large. For example, if the photographic optical system has a large difference in actual focal length between the focused state at the infinity end and the focused state at the close end, that is, the photographic optical system is moved in the direction of the subject to move the focal point to the close side. This is a case where a lens barrel having a relatively large fluctuation amount is used in an optical system in which the focal length becomes shorter (closer to the wide) as the lens is moved forward.

  The situation in this case is shown in FIG. There may be a focus evaluation value characteristic curve as indicated by a broken line in the figure, but a focus evaluation value characteristic curve as indicated by a solid line in the figure. This is especially because the human face has low contrast, the focusing lens moves closer to the person during the search, the actual focal length becomes shorter (wider), and the background (high-frequency subject) near the human face enters. In this case, such a phenomenon occurs, and in this case, not the correct focus position L1, but the position L2 that is blurred toward the closest side is mistaken as the focus position.

  In order to cope with these problems, in one embodiment, the detected face detection area is divided into a plurality of face areas, and a characteristic curve of a focus evaluation value in each divided face area is obtained. This is based on the premise that the difference in focus position between the divided face areas is not large. Here, a plurality of calculated in-focus positions are statistically observed, and a value greatly deviated is regarded as a state where a correct in-focus position cannot be detected due to the phenomenon described above, and is excluded from in-focus position determination candidates.

  FIG. 6 shows an example in which the face detection area 26 (see FIG. 3) is divided into four face regions 26a to 26d in the horizontal direction. In this example, five focus evaluation values are obtained from the face detection area 26. That is, the focus evaluation value of each of the face areas 26a to 26d obtained by dividing the face detection area into four parts and the area obtained by adding these four face areas 26a to 26d, that is, the focus evaluation value of the face detection area 26 itself.

  The determination processing shown in FIG. 7 is performed on the in-focus positions obtained from the focus evaluation values in the face areas 26a to 26d and the face detection area 26. First, an in-focus position between the face areas 26a to 26d and the face detection area 26, that is, a position having the maximum contrast is obtained. In step 1 of FIG. 7, the difference between the in-focus positions of the face area on the most infinite side and the face area on the closest side is obtained. This may be converted from the distance from the image plane, but the focusing lens of this embodiment is driven and controlled by the stepping motor, so the difference between the relative pulse amounts from the reference position may be used. It is determined whether or not the difference between the in-focus face area and the closest face area is larger than a preset threshold th1.

  If the difference in focus position between the face area on the infinity side and the face area on the nearest side is greater than or equal to the threshold th1 in step 1, it is considered that the above-described malfunction phenomenon has occurred, and the process proceeds to step 2. If it is less than the threshold th1, the process proceeds to Step 3. First, if the difference in focus position is greater than or equal to the threshold value th1, the face of a person who is a single subject is divided and a focus position search is performed in each face area, but each face area is searched. This is a case where the variation in focus position is large. For this reason, it is necessary to eliminate the influence of a face area greatly deviating from other in-focus positions by performing statistical processing or the like.

In step 2, the difference δ _n between the in-focus positions of the face areas 26a to 26d and the face detection area 26 and the median value (median) of the in-focus positions is obtained by the following equation.
δ _n = | AF _n −median [AF] | (1)
(1) In the equation, AF _n is the focus position of the face region 26a~26d and the face detection area 26, median [AF] is the median of the focus position of the face region 26a~26d and the face detection area 26.

If the difference δ _n is greater than the predetermined threshold th2, the focus position is excluded from the focus position selection candidates as being a face area greatly deviating from the focus position of another face area, and the remaining face areas The closest focus position among the focus positions is determined as the target drive position. Note that the average position of the focus positions of the remaining face regions may be set as the target drive position. The threshold values th1 and th2 are preferably set to about the depth of field, but are set to table values so that they can be changed according to the focal length (zoom position) of the actual photographing optical system and various photographing conditions. Keep it.

  On the other hand, when the difference in focus position between the face area closest to the infinity side and the face area closest to the closest side is less than the threshold th1, in step 3, there is a variation in the focus position of each face area. Since it is small, the focus position search between the face areas is performed as before, and the closest focus position is determined as the target drive position. Note that an average position of a plurality of in-focus positions may be set as the target drive position.

  In step 4, the focusing lens is moved to the target drive position determined by the focus motor 10 to adjust the focus.

  In the above-described embodiment, an example in which the number of divisions of the face detection area is four is shown, but the number of divisions of the face detection area is not limited to four. Further, in the above-described embodiment, a face area whose difference between the focus position of each face area and the median value (median) of the focus positions is a predetermined threshold or more is excluded from the focus position selection candidates. However, the method for determining the face area to be excluded is not limited to the method based on the median value, and the same effect can be obtained by using an average value or other statistical methods.

  In the above-described embodiment, an example is shown in which a human face is detected as a specific subject and a region of the specific subject in the captured image, that is, a human face region is recognized. However, the specific subject is limited to a human face. First, by storing the pattern of the main subject to be photographed, various subjects can be specified subjects.

  As described above, according to one embodiment, an image taken by the photographing lens 1 is picked up by the image pickup device 2 and image information is output, and a specific subject (a person's face) is recognized based on the image information. An area of the specific subject (face detection area 26) in the image is detected. Then, the specific subject area (face detection area 26) is divided into a plurality of areas (face areas 26a to 26d), and the focus position of the photographing lens 1 is detected in each of the plurality of divided areas (face areas 26a to 26d). In addition, the final focus position is determined based on a plurality of focus positions detected in the plurality of divided areas (face areas 26a to 26d), and the photographing lens 1 is set with the final focus position as the target drive position. The focus was adjusted. As a result, even if a specific subject such as a person is captured in the focus detection area, if a high-contrast subject such as a background is included in the focus detection area, the subject is pulled by the high-contrast subject and becomes a specific subject such as a person. The inconvenience of being unable to focus accurately can be prevented, and the specific subject can be accurately focused.

  In addition, according to the embodiment, the infinitely far-end focus position and the closest focus position among the plurality of focus positions detected in the plurality of divided areas (face areas 26a to 26d). Is equal to or greater than the predetermined value th1, the difference between the focus position of each divided area (face areas 26a to 26d) and the median value of the focus positions of the plurality of divided areas (face areas 26a to 26d) is Since the final in-focus position (target drive position) is determined based on the in-focus positions of the remaining divided areas excluding the divided area of the predetermined value th2 or more, the photographer included in the focus detection area does not intend. It is possible to accurately focus on a specific subject by excluding the subject.

The figure which shows the structure of one embodiment The figure which shows the example of arrangement | positioning of the conventional focus detection area The figure which shows the example of a setting of the focus detection area using the face recognition function of one Embodiment The figure which shows the characteristic curve of the focus evaluation value with respect to a focusing lens position when a low contrast person's face and a high contrast background exist in a focus detection area The figure which shows the characteristic curve of the focus evaluation value with respect to a focusing lens position when the focal distance of an imaging optical system changes The figure which shows the example of a division | segmentation of the focus detection area by the face recognition of one Embodiment The flowchart which shows the target drive position determination process of the focusing lens of one embodiment

Explanation of symbols

1 Shooting Lens 2 Image Sensor 9 CPU
10 Focus motor 11 Face recognition calculation unit

Claims (4)

An imaging means for capturing an image by the imaging optical system and outputting image information;
Subject detection means for recognizing a specific subject based on the image information and detecting a region of the specific subject in the image;
An in-focus position detecting means for dividing the area of the specific subject into a plurality of areas, and detecting the in-focus position of the imaging optical system in each of the plurality of divided areas;
Focusing position determination means for determining a final focusing position based on a plurality of focusing positions detected in the plurality of divided areas;
A digital camera comprising: a focus adjustment unit that performs focus adjustment of the photographing optical system with the final focus position as a target drive position. The digital camera according to claim 1, wherein
The in-focus position determining means, when the difference between the infinitely far side in-focus position and the closest in-focus position among the plurality of in-focus positions is less than a predetermined value, the plurality of in-focus positions. A digital camera characterized in that a closest focusing position among the positions is set as the final focusing position, or an average position of the plurality of focusing positions is set as the final focusing position. . The digital camera according to claim 1 or 2,
The in-focus position determining means, when the difference between the infinitely far side in-focus position and the closest in-focus position among the plurality of in-focus positions is equal to or greater than a predetermined value, The final focus position is determined based on the focus positions of the remaining divided areas excluding the divided areas where the difference between the focus position and the median of the focus positions of the plurality of divided areas is a predetermined value or more. A digital camera characterized by The digital camera according to claim 3, wherein
The in-focus position determining unit is configured to focus in the remaining divided areas excluding the divided areas in which the difference between the in-focus position of each of the divided areas and the median value of the in-focus positions of the plurality of divided areas is a predetermined value or more. The closest focusing position among the positions is set as the final focusing position, or the average focusing position of the remaining divided areas is set as the final focusing position. A digital camera.

Images