JP2012128287A - Focus detection device and focus detection method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve such problems that, in a low illuminance or when a subject is a low contrast, an S/N ratio of an AF evaluation value is deteriorated to lower the focusing accuracy, or, when a subject looks away or the like and the subject and a background are mixed in an AF frame, the background may be omitted.SOLUTION: When the size of the detected face is smaller than a prescribed size, a focus detection device makes the AF frame set to a body region larger than the AF frame set to the position of the face. The AF frame set to the body region is horizontally divided with respect to the body according to the condition of the face, hence the plurality of AF frames are set.

Description

  The present invention relates to a focus detection apparatus and a focus detection method applied to a digital camera or the like having a focus adjustment function.

  Conventionally, when auto-focusing (hereinafter referred to as AF) is performed in an electronic still camera, a video camera, or the like, a lens that maximizes a high frequency component of a luminance signal obtained from an image sensor using a CCD (charge coupled device) or the like. A system in which the position is the in-focus position is used.

  As one of these methods, the following scan method is known. Each time an evaluation value (hereinafter referred to as an AF evaluation value) based on a high frequency component of a luminance signal obtained from a focus detection area (hereinafter referred to as an AF frame) in the image sensor while moving the lens in the focus detection range. I will remember it. When the contrast in the AF frame is high and the size of the AF frame is large, the AF evaluation value becomes large. In the stored method, a lens position corresponding to the maximum value (hereinafter referred to as a peak position) is used as a focus position. Here, the in-focus position is a lens position where the subject is assumed to be in focus.

  However, since the S / N ratio deteriorates when the illumination is low or when the contrast of the subject is low, the AF evaluation value fluctuates randomly due to the influence of noise, and accordingly the calculation accuracy of the peak position of the AF evaluation value is increased. There is also the problem of getting worse. In addition, when a subject to be focused (hereinafter referred to as a main subject) and the background are mixed in the AF frame, there is a problem that the peak position becomes closer to the far side due to the influence of the background. This situation is called background loss.

  Therefore, the method of Patent Document 1 is disclosed as a countermeasure for suppressing the influence due to the deterioration of the S / N ratio. In Patent Document 1, first, an AF frame is set in an area including a face area of a subject, an area including a partial area inside the face, an area including a body area, or the like. When the level of the AF evaluation value of the face area or an area including a partial area inside the face is low, the in-focus position is determined from the peak position in the AF frame including the body area that may have a higher contrast than the face. decide.

  In addition, there are the following methods as measures for suppressing the influence of background loss. Japanese Patent Application Laid-Open No. 2004-228867 discloses that the main subject and the background are prevented from being mixed in the AF frame by changing the inclination and shape of the AF frame set in the face area in accordance with the face angle and size of the subject.

JP 2008-077686 A JP 2007-034261 A

  However, in Patent Document 1, there is a problem in that background missing occurs when the subject and the background are mixed in the AF frame set in the body region because the subject faces sideways.

  Further, in Patent Document 2, since the face often has a low contrast, there is a problem that the S / N is deteriorated when the AF evaluation value is acquired only from the position of the face. Also, when the size of the face is small, such as when the subject is small or far away, the size of the AF frame set at that position decreases, and the AF evaluation value decreases accordingly, and the AF focusing performance deteriorates. There is a problem.

  The present invention has been made in view of the above-described problems, and an object of the present invention is to reduce the influence of the deterioration of the S / N ratio when the illuminance is low or when the subject is low contrast and the influence of background omission and the focus position. It is to improve accuracy.

  In order to achieve the above object, the technical features of the present invention include a face detection step for detecting the position and size of a person's face from a captured image, and a focus for detecting the in-focus state of the imaging optical system. A setting step of setting a first focus detection area where the person's face is present and a second focus detection area where the person's body is predicted to be located when viewed from the position of the person's face as detection areas And a focus adjustment step of adjusting the focus by moving the imaging optical system based on a signal output in the focus detection region, and the setting step determines whether the face size detected by the face detection step is a predetermined size. When the size is smaller than the size, the size of the second focus detection region is set to be larger than the size of the first focus detection region.

  Another technical feature of the present invention is that the person is used as a face detection step for detecting a position where a person's face is present from a captured image, and as a focus detection region when detecting the in-focus state of the imaging optical system. A setting step for setting a first focus detection area where the face of the person exists and a second focus detection area where the person's body is predicted to be located when viewed from the position of the face of the person; and the focus detection area A focus adjustment step of adjusting the focus by moving the imaging optical system based on the signal output in the step, and in the setting step, the second focus detection region is divided horizontally with respect to the body A plurality of focus detection areas are provided.

  According to the present invention, it is possible to improve the accuracy of the in-focus position by reducing the influence of deterioration due to the S / N ratio at the time of low illuminance and low contrast and the influence of background loss.

It is a block diagram which shows the structure of the focus adjustment apparatus to which this invention is applied. It is a flowchart figure showing AF operation. FIG. 3 is a flowchart illustrating a frame setting subroutine in FIG. 2. It is a figure explaining AF frame setting at the time of face non-detection. It is a figure explaining AF frame setting at the time of camera side position photography at the time of face detection. It is a figure explaining AF frame setting at the time of camera side position photography at the time of face detection. It is a figure explaining AF frame setting at the time of camera side position photography at the time of face detection. It is a figure explaining AF frame setting at the time of camera side position photography at the time of face detection. It is a figure explaining AF frame setting at the time of camera side position photography at the time of face detection. It is a figure explaining AF frame setting at the time of camera vertical position photography at the time of face detection. It is a figure explaining AF frame setting at the time of camera vertical position photography at the time of face detection. It is a figure explaining AF frame setting at the time of camera vertical position photography at the time of face detection. It is a figure explaining AF frame setting at the time of camera vertical position photography at the time of face detection. It is a figure explaining AF frame setting at the time of camera vertical position photography at the time of face detection. It is a figure explaining AF frame setting at the time of camera vertical position photography at the time of face detection. It is a figure explaining AF frame setting at the time of camera vertical position photography at the time of face detection. It is a figure explaining the subroutine of a focus determination in FIG. FIG. 3 is a flowchart for explaining a subroutine for determining an in-focus position in FIG. 2. FIG. 9 is a flowchart illustrating a reference selection subroutine in FIG. 8. It is a figure explaining the reliability determination at the time of big deviation in FIG. 9, and reliability determination at the time of middle deviation. FIG. 9 is a flowchart for explaining a frame selection subroutine in FIG. 8. FIG. 9 is a flowchart for explaining a frame addition subroutine in FIG. 8.

  Embodiments will be described below with reference to the drawings. FIG. 1 is a block diagram showing a configuration of a focus adjusting apparatus to which the present invention is applied.

<Device configuration>
Reference numeral 101 denotes a photographing lens including a zoom mechanism. Reference numeral 102 denotes a focus lens for focusing on an image sensor described later. Reference numerals 101 and 102 are also referred to as imaging optical systems. Reference numeral 103 denotes an AF processing unit. An image sensor 104 converts reflected light from the subject into an electrical signal and outputs a captured image. Reference numeral 105 denotes an A / D converter including a CDS circuit that removes output noise from the image sensor 104 and a non-linear amplifier circuit that is executed before A / D conversion. Reference numeral 106 denotes an image processing unit. Using this image processing unit, it is possible to acquire color information and the like in a predetermined area of the image. Reference numeral 107 denotes a system control unit (hereinafter referred to as CPU) that controls the system such as a photographing sequence. Reference numeral 108 denotes a shooting mode switch for performing settings such as switching the face detection mode to ON or OFF. Reference numeral 109 denotes a main switch for turning on power to the system, and reference numeral 110 denotes a switch (hereinafter referred to as SW1) for performing a shooting standby operation such as AF or AE. Reference numeral 111 denotes a photographing switch (hereinafter referred to as SW2) that performs photographing after the operation of SW1. A face detection module 112 performs face detection using the image signal processed by the image processing unit 106 and sends one or more pieces of detected face information (position / size / reliability) to the CPU 107. Since the face detection method is not the main point of the present invention, detailed description thereof is omitted.

  A subject detection unit 113 sends one or more pieces of detected subject information (position / size / reliability) to the CPU 107 based on color information and luminance information of the image signal processed by the image processing unit 106. is there. It should be noted that the subject detection method is not the main point of the present invention and will not be described in detail.

  An angular velocity sensor unit 114 detects the angular velocity of the camera itself and sends camera motion information to the CPU 107. It is also possible to detect whether the camera is held in a vertical position or a horizontal position by using the angular velocity sensor unit. An acceleration sensor unit 115 detects the acceleration of the camera itself and sends camera movement information to the CPU 107. Using this acceleration sensor unit, the camera can detect whether the taking lens 101 is held upward or whether the taking lens 101 is held downward. Reference numeral 116 denotes a high-speed built-in memory (for example, a random access memory, hereinafter referred to as a DRAM). Reference numeral 117 denotes an operation display unit that displays not only an image display but also a display for assisting operation and a display of a camera state, and a shooting screen and an AF frame at the time of shooting.

<AF operation>
Next, the AF operation at the time of face detection by the electronic camera of the present invention will be described using the flowchart of FIG.

  In step S201, an AF frame is set based on the face information (position, size, number of detected faces) obtained from the face detection module 112. The procedure of step S201 will be described later with reference to FIGS.

  In step S202, AF scanning is performed with each AF frame set in step S201, and the process proceeds to step S203. In the AF scan, the AF evaluation value at each focus lens position is stored in the CPU 107 while moving the focus lens 102 by a predetermined amount from the scan start position to the scan end position. For example, the scan start position is set to infinity, and the scan end position is set to the closest end of the range where AF scanning is possible.

  In step S203, the peak position of the AF evaluation value obtained in step S202 is obtained by calculation, and the process proceeds to step S204.

  In step S204, focus determination of the peak position of each AF frame is performed, and the process proceeds to step S205. The procedure of step S204 will be described later with reference to FIG.

  In step S205, if a face is detected, a reference position is determined from the estimated distance calculated based on the peak position of the AF frame (hereinafter referred to as a face frame) set in the face area and the detected face size. select. Then, the focus position is determined by comparing the reference position with the peak position of each AF frame, and the process proceeds to step S206. The procedure of step S205 will be described later with reference to FIG.

  If no face is detected, an AF frame is set in a predetermined area in the screen, as will be described later in the procedure of step S201. The focus determination at that time, determination of the focus position, and the like are not the main points of the present invention and will not be described.

  In step S206, it is checked whether or not the focus position has been determined in step S205. If the focus position has been determined, the process proceeds to step S207, and if the focus position has not been determined, the process proceeds to step S209.

  In step S207, the focus lens 102 is moved to the in-focus position determined in step S205, and the process proceeds to step S208.

  In step S208, the focusing frame is displayed on the image display unit of the operation display unit 117, and the AF operation is terminated. Here, the in-focus frame is a frame for indicating which region in the image region is in focus. For example, when the face is in focus, a frame is displayed in the face area. In addition, the frame is displayed with an in-focus color (for example, green) so that the state of being in focus can be easily identified.

  In step S209, the focus lens 102 is moved to the reference position determined in step S205, and the process proceeds to step S210.

  In step S210, an in-focus frame is displayed on the image display unit of the operation display unit 117, and the AF operation is terminated. Here, the out-of-focus frame is a frame that is displayed in a region where a subject is present or a predetermined region in the image region at the time of out-of-focus, so that it can be easily discriminated that it is out of focus. A frame of a color different from (for example, yellow) is set.

<Frame setting>
Next, the AF frame setting in step S201 of FIG. 2 will be described with reference to the flowchart of FIG. 3 and FIGS.

In step S301, it is checked whether or not a face has been detected from the face information (position, size, and number of detected faces) obtained from the face detection module 112. If no face has been detected, the process proceeds to step S302 to detect the face. If so, the process proceeds to step S303.
In step S302, a predetermined number of AF frames are set in a predetermined range of the screen.

For example, as shown in FIG. 4, three AF frames may be set at the center of the screen in the vertical direction and in the horizontal direction with respect to the screen, or other settings may be used.
Reference numeral 401 in FIG. 4 denotes an AF frame.

  In step S303, a face frame is first set, and the process proceeds to step S304. The face frame is a predetermined magnification of the size of the face obtained from the face detection module 112 at the position of the face information obtained from the face detection module 112 regardless of whether the camera is held in the horizontal position or the camera is held in the vertical position. Set with.

  In step S304, AF frames (hereinafter referred to as face assist frames) are set on the left and right sides of the face, and the process proceeds to step S305. The AF frame to be set to the left and right can be used when the camera is held horizontally (FIGS. 5-a to 5-e) or the camera is held vertically (FIGS. 6a to 6-g). The left and right sides of the face position are set at a predetermined magnification of the face size.

  This is because the position of the hair that is easily contrasted with the face is placed in the AF frame by arranging it on the left and right sides of the face.

  In step S305, it is checked whether the positions of the face and the body region are aligned in the vertical direction with respect to the screen (hereinafter referred to as a case where the face is vertical). If the face is vertical, the process proceeds to step S306. If the position of the face and body region is aligned in the horizontal direction with respect to the screen (hereinafter referred to as a case where the face is horizontal), the process proceeds to step S309.

  FIGS. 5A to 5E show an AF frame (hereinafter referred to as a body frame) set in a body region where the human body is predicted to be located when viewed from the position of the human face when the face is vertical. It is a figure for demonstrating the setting method.

  Reference numeral 501 in FIG. 5A denotes a shooting screen when shooting with the camera held in the horizontal position. Reference numeral 502 denotes a face frame. Reference numeral 503 denotes a face assist frame. Reference numeral 504 denotes a body frame. 505 is the horizontal size of the body frame. Reference numeral 506 denotes the vertical size of the body frame. As shown in the figure, the body frame is set as a plurality of AF frames by dividing the body frame in the horizontal direction with respect to the position where the human body is predicted to be located from the position of the human face. As will be described later, the number of divisions is changed depending on the position, size, angle, and the like of the face.

  In step S306, the horizontal size of the body frame is determined from the face size, and the process proceeds to step S307. The horizontal size is set at a predetermined magnification with respect to the face size so that the AF frame is not set to an area outside the body area. This magnification may be determined from the number of body frames in the horizontal direction with respect to the screen to be set and the magnification of the body region size with respect to the face size, or from the position / size of the body region obtained from the subject detection unit 113, The size may be changed according to the number of body frames in the horizontal direction with respect to the screen to be set.

  However, when the subject is small or far away, the body region size becomes small, so that the focus set as shown in FIG. 5-c with respect to FIG. Reduce the number of inspection areas.

  In step S307, a body frame creation start position is set, and the process proceeds to step S308. The body frame start point is set at a position below the face position obtained from the face detection module 112 by a predetermined amount. This is because if the body frame is set immediately below the face, that is, at the neck position, the background may enter the body frame and the background may be lost.

  This predetermined amount may be set in advance as several times the face size, or may be determined from the position of the body region obtained from the subject detection unit 113.

  In step S308, the vertical size of the body frame is set, and the body frame setting is completed. If the face position is near the top of the screen and the face size is larger than a predetermined size, the vertical size is set up to the top of the body area (FIG. 5-a). This is because if the setting is made to the lower part, there is a high possibility that subjects other than the main subject will enter the body frame.

  When the face position is at the center or the lower part of the screen and the face size is larger than a predetermined size, the vertical size of the body frame is set to the end of the screen (FIG. 5-d). If the face size is smaller than the predetermined size, the position up to the position corresponding to the waist is set as the vertical size (FIG. 5-b).

  This is to ensure the output of the AF evaluation value regardless of the face size. By freely changing the aspect ratio of the AF frame in this way, the S / N of the AF evaluation value is improved, and an AF frame that does not cause background omission even when the subject faces sideways is provided. It is possible to set.

  Also, as shown in FIG. 5E, when there is another face in front of the main subject and the face is on the body area of the main subject, the other body frame is not overlapped with the other face. The body frame is the vertical size of the body frame. Here, the face of a person is shown as another subject, but the present invention is not limited to this. For example, when another subject is recognized by subject recognition such as pattern matching, the vertical size of the body frame is adjusted.

  FIG. 6A to FIG. 6G are diagrams for explaining a body frame setting method in the body region when the face is vertical.

  Reference numeral 601 in FIG. 6A denotes a shooting screen when shooting with the camera held in the vertical position. Reference numeral 602 denotes a face frame. Reference numeral 603 denotes a face assist frame. Reference numerals 604 to 606 denote body frames. Reference numeral 607 denotes the horizontal size of the body frame. Reference numeral 608 denotes the vertical size of the body frame.

  In step S309, the size of the body region is calculated from the face size, and it is determined whether or not a body frame larger than a predetermined size can be set vertically on the screen as indicated by 604 and 605 in FIG. To do.

  If it can be set, the setting is made as shown in FIG. 6A, and the vertical size of the body frame is calculated according to the size of the body region.

  If it is determined that they cannot be set side by side, the vertical size of the body frame is set to a predetermined size and set side by side with respect to the screen as shown in FIG. 6B.

  The size of the body region may be obtained from the size of the body region obtained from the subject detection unit 113.

  Also, as shown in FIG. 6G, when there is another face in front of the main subject and the face covers the body area of the main subject, the body frame of the body area must overlap the other face. In this way, the vertical size of the body frame is set up to other faces. Here, the face of a person is shown as another subject, but the present invention is not limited to this. For example, when another subject is recognized by subject recognition such as pattern matching, the vertical size of the body frame is adjusted.

  In step S310, the creation start position of the body frame is set, and the process proceeds to step S311. The body frame is set to a position below the face position by a predetermined amount, as in the case where the face is vertical. In step S309, if a body frame of a predetermined size or more can be set in the body region in the vertical direction with respect to the screen, it is also set in the horizontal direction with respect to the screen. This is because even if a subject other than the main subject enters the body frame 604 or 605 in FIG. 6A, it is possible to focus using the information on the body frame 606. It is.

  Further, when the subject is small or far away, the body region size is small, so that the body frame is set as shown in FIG. 6-c with respect to FIG. 6-b so that the body frame can secure a predetermined size. Increase the amount of frame overlap.

  In step S311, the horizontal size of the body frame is set, and the body frame setting is completed. When the face position is near the center in the horizontal direction with respect to the screen, that is, as shown in FIG. 6A, the horizontal size is set so that two subject detection areas can be set near the position corresponding to the abdomen of the subject. To do. Although the number of subject detection areas is two in this embodiment, it may be plural. The horizontal size may be calculated in advance from the face size, or may be set from the position / size of the body region detected by the subject detection unit 113.

  When the face position is near the lower part in the horizontal direction with respect to the screen, that is, in the case of FIG. 6D or FIG. 6E, the horizontal size of the subject detection area is reduced to a predetermined size. . When the horizontal size is smaller than the predetermined size, two body frames are set as shown in FIG.

  If the subject is small or far away and the subject is at the lower side in the horizontal direction with respect to the screen, one body frame is set as shown in FIG.

  As described above, by changing the size and position of the body frame in accordance with the position, size, and angle of the face in the screen, the size of the body frame is secured, the S / N is improved, and the subject is It is possible to set a body frame that does not cause background loss even when facing sideways.

<Focus determination>
Next, focusing determination in step S204 in FIG. 2 will be described with reference to FIG.
When there is only one subject within the subject distance range corresponding to the moving position of the focus lens at the time of AF scanning, the AF evaluation value is the focus lens position on the horizontal axis and the AF evaluation value on the vertical axis. If it takes, the shape will become a mountain shape as shown in FIG. Therefore, by determining the shape of the mountain from the difference between the maximum value and the minimum value of the AF evaluation value, the length of the inclined portion with a slope equal to or greater than a certain value (SlopeThr), and the slope of the inclined portion, In-focus determination can be performed.

  The length L of the inclined part and the slope SL / L of the inclined part for judging the shape of the mountain will be described with reference to FIG.

  Points that are recognized to be inclined from the top of the mountain (point A) are defined as points D and E, and the width between points D and E is defined as the width L of the mountain. The range in which it is recognized that the inclination continues is a range in which a scan point whose AF evaluation value has decreased by a predetermined amount (SlopeThr) or more continues from point A. The scan point is a point at which the AF evaluation value is acquired while moving the focus lens continuously and moving from the scan start point to the scan end point. The sum SL1 + SL2 of the difference SL1 between the AF evaluation values at the points A and D and the difference SL2 between the AF evaluation values at the points A and E is defined as SL.

To determine the focus from the shape of the mountain, check the following conditions.
[1] Whether the length L of the inclined portion is equal to or larger than a predetermined value (L0) or more. [2] The average value step SL / L of the inclined portion is predetermined value (SL0). / L0) or more [3] Whether the difference between the maximum value (Max) and the minimum value (Min) of the AF evaluation value is a predetermined value (defmaxmin) or more The determination result in is output in a ○ determination and a × determination as shown below.
○ Determination: The subject exists at a distance within the subject distance range corresponding to the moving position of the focus lens when the subject has sufficient contrast and is scanned.
X determination: The subject is insufficient in contrast, or the subject is located at a distance outside the subject distance range corresponding to the moving position of the focus lens when scanned.
Here, the predetermined values L0, SL0, and defmaxmin are referred to as in-focus conditions, and the focus determination can be tightened or loosened by setting these values.

<Determining the in-focus position>
Next, the operation for determining the in-focus position in step S205 in FIG. 2 will be described using the flowchart in FIG.
In step S801, a reference position is selected from the estimated distance calculated from the peak position in the face frame and the face size, and the process proceeds to step S802. The procedure of step S801 will be described later with reference to FIGS.

  In step S802, the AF frame used for determining the in-focus position is selected from the face frame / auxiliary frame / body frame based on the reference position selected in step S801, and the process proceeds to step S803. The procedure of step S802 will be described later with reference to FIG.

  In step S803, it is checked whether there is an AF frame selected in step S802. If there is no AF frame selected, the process proceeds to step S804.

  In step S804, AF evaluation value addition processing is performed based on the reference position determined in step S801, and the operation for determining the in-focus position ends. The procedure of step S804 will be described later with reference to FIG.

<Selection of reference position>
Next, the reference selection operation in step S801 of FIG. 8 will be described using the flowchart of FIG.

  In step S901, the size of the face detected by the face detection module 112 is compared with predetermined upper and lower limit values of the face size. If the face size is within the upper and lower limits, the process proceeds to step S902. If the face size is outside the upper and lower limits, the process proceeds to step S911. Here, the range of the upper and lower limits of the face size is determined in consideration of the accuracy of the estimated distance.If the face detection size is larger than the upper limit value, the error of the estimated distance accompanying the error of the face detection result is If it is large and the face detection size is smaller than the lower limit value, the face detection accuracy is low. For this reason, when the face size outside the upper and lower limits is detected, the estimated distance is not reliable, so the estimated distance is excluded from the reference position candidates.

  In step S902, the result of the focus determination in step S204 in FIG. 2 of the AF evaluation value in the face frame is checked. If YES, the process proceeds to step S903, and if NO, the process proceeds to step S910.

In step S903, an outlier determination is performed to check whether the peak position of the face frame and the estimated distance are within a predetermined depth (hereinafter, referred to as a large outlier depth). If it is out of the large out-of-out depth, it is determined that it is out-of-out, and the process proceeds to step S905. Here, the “large deviation depth” is a value determined in consideration of an error amount of an estimated distance such as an error due to an individual difference of devices / an error due to variations in face detection accuracy / an individual difference of face size. Even when the in-focus position corresponding to the face frame can be acquired correctly, a difference in the outlier depth may occur between the estimated distance and the peak position of the face frame. However, if the difference between the peak position of the face frame and the estimated distance is greater than or equal to the depth of deviation, it is considered that one of the following situations has occurred, and which of the face peak position and the estimated distance is reliable It is necessary to make a judgment.
[1] The situation where the peak position of the face frame deviates significantly from the actual in-focus position [2] The subject is a doll or a poster, and the estimated distance is calculated based on the face size deviating greatly from the human face size. The situation where the error of the estimated distance is large

  In step S904, an out-of-phase determination is performed to check whether the peak position of the face frame and the estimated distance are within a predetermined depth (hereinafter referred to as an out-of-center depth) smaller than the out-of-range depth. If it is outside the mid-out depth, it is determined that it is out-of-center, and the process proceeds to step S908. If it is within the mid-out depth, it is determined that it is not out-of-center, and the process proceeds to step S911. Here, the “out-of-center depth” is set to a value smaller than the maximum out-of-field depth, and if the difference between the peak position of the face frame and the estimated distance is within the out-of-center depth, the peak position of the face frame and the estimated distance Both are considered reliable. On the other hand, if the difference between the peak position of the face frame and the estimated distance is out of the middle depth, it is necessary to determine which of the face frame peak position and the estimated distance is more reliable.

  In step S905, reliability determination at the time of large deviation is performed to determine which of the face frame peak position and the estimated distance is reliable, and the process proceeds to step S906. Here, the content of the reliability determination at the time of large deviation will be described with reference to FIG. FIG. 10A shows the relationship between the peak position of the AF evaluation value in each AF frame and the estimated distance when the estimated distance is unreliable, and FIG. 10B shows the case where the peak position of the face frame is unreliable. It is the figure shown simplified. 1101a to 1101f and 1102a to 1102f indicate the positions of the AF frames and the peak positions of the AF evaluation values, and 1101g and 1102g indicate the positions of the estimated distances, respectively.

  When the estimated distance is unreliable at the time of large deviation, the case where a face extremely larger than the face of an actual person is detected by an object or poster representing a person as shown in FIG. This is considered to be a case where a face extremely smaller than the face of a person is detected in a photograph or a photograph. In this case, it is assumed that a correct in-focus position can be detected in the face frame. In such a case, when the correct focus position can be detected in the face auxiliary frame or the body frame, the peak positions 1001b to 1001d are distributed in the vicinity of the peak position 1001a of the face frame. Further, even if there are face assist frames and body frames for which a correct in-focus position cannot be detected, there is a low possibility that the peak positions 1001e to f are close to the estimated distance 1001g. Therefore, if there is no peak position of the face assist frame or the body frame within a predetermined depth from the estimated distance (hereinafter referred to as an estimated reliability depth), it is determined that the estimated distance is not reliable.

  Next, the case where the peak position of the face frame is unreliable at the time of large deviation means that a false peak has occurred in the AF evaluation value when the background is turned sideways as shown in FIG. 10B. This is considered to be the case. In this case, it is assumed that the estimated distance is a correct value. In such a case, the peak positions 1002c and 1002f of the AF frame in which the correct focus position can be detected without missing the background are distributed in the vicinity of the estimated distance 1002g. Therefore, when the peak position of the face assist frame or the body frame exists within the estimated reliability depth from the estimated distance, it is determined that the estimated distance is reliable.

  As described above, in the out-of-bounds reliability determination, when the peak position of the auxiliary frame or the body frame exists within the estimated reliability depth from the estimated distance, it is determined that the estimated distance is reliable. Further, when the peak positions of the auxiliary frame and the body frame do not exist within the estimated reliability depth from the estimated distance, it is determined that the estimated distance is not reliable.

  In step S906, if it is determined that the estimated distance is more reliable in the out-of-shift reliability determination in step S905, the process proceeds to step S907. If it is determined that the estimated distance is lower, the process proceeds to step S911.

  In step S907, setting is made so that the face frame is not used during the frame selection process in step S802 and the frame addition process in step S804 in FIG. 8, and the process proceeds to S910.

  In step S908, reliability determination for out-of-center is performed to determine which of the face frame peak position and the estimated distance is reliable, and the process proceeds to step S909. Here, although the determination similar to the reliability determination at the time of large deviation is performed, the value of the estimated reliability depth may be different from that at the time of large deviation.

  In step S909, if it is determined that the estimated distance is more reliable in the determination of reliability when the estimated distance is out of step S908, the process proceeds to step S910, and if it is determined that the estimated distance is less reliable, step S909 is performed. The process proceeds to S911.

  In step S910, the estimated distance is set as the reference position. In step S911, the peak position of the face frame is set as the reference position, and the reference selection operation is terminated.

  As described above, the more reliable face focus position is selected as the reference position from the peak position of the face frame and the estimated distance. In particular, even when the difference between the peak position of the face frame and the estimated distance is large and the focus position in the face frame is wrong or the estimated distance error is large, the more reliable one can be selected as the reference position. it can.

<Frame selection>
Next, the frame selection operation in step S802 in FIG. 8 will be described using the flowchart in FIG.

  In step S1101, whether or not the face frame can be used is checked. If the face frame cannot be used in step S807 in FIG. 8, the process proceeds to step S1104. If the face frame is usable, the process proceeds to step S1102.

  In step S1102, the result of the focus determination in step S204 in FIG. 2 in the face frame is checked. If yes, the process proceeds to step S1103, and if x, the process proceeds to step S1104.

  In step S1103, the peak position of the face frame is set as the in-focus position, and the frame selection operation ends.

  In step S1104, it is checked in step S201 in FIG. 2 whether a face auxiliary frame or body frame is set. If it is set, the process proceeds to step S1105. If it is not set, the frame selection operation is terminated.

In step S1105, it is checked whether or not the following two conditions are satisfied for the face assist frame and the body frame.
Condition 1: In-focus determination at S204 in FIG. 2 is ◯. Determination condition 2: When the determination is made for all AF frames within the predetermined depth of the reference position and peak position set in S801 in FIG. 8, the process proceeds to step S1106.

  In step S1106, it is checked whether there is an AF frame set as a frame selection candidate frame in step S1105. If there is, the process proceeds to step S1107, and if not, the frame selection operation is terminated.

  In step S1107, an AF frame whose peak position is closest to the reference position is selected from the AF frames set as frame selection candidates, and the process proceeds to step S1108.

  In step S1108, the frame selection operation is terminated with the peak position of the AF frame selected in step S1107 as the in-focus position.

Regarding steps S1107 and S1108,
・ Select all candidate frames for frame selection and use the average position of the peak positions of the selected AF frames as the in-focus position. ・ Select and select the AF frame with the closest peak position from the candidate frames for frame selection. A method may be used in which the peak position of the AF frame is used as the in-focus position.

  As described above, when the face frame is usable and the focus determination is “good”, the face frame is selected, and when the face frame cannot be selected, the face auxiliary frame or body frame close to the reference position is selected. Thereby, a more reliable AF frame can be selected as the focus position of the face, and the focus accuracy can be improved.

<Frame addition unit>
Next, the frame addition operation in step S804 in FIG. 8 will be described using the flowchart in FIG.
In step S1201, focus determination is performed for all AF frames, and the process proceeds to step S1202. Here, the values of the focusing conditions L0, SL0, and defmaxmin are changed from the values set in step S204 in FIG. The case where the focus is judged to be less ◯ is when the AF evaluation value has a peak, but the output is low and the reliability is low when used as it is as the in-focus position.

In step S1202, it is checked whether or not the following two conditions are satisfied for all AF frames, and if they are satisfied, it is determined as a frame addition target frame (hereinafter referred to as an addition target frame).
Condition 1: In-focus determination in step S1201 is ◯. Determination condition 2: When determination is made for all AF frames whose peak positions are within the predetermined depth from the reference position, the process proceeds to step S1203. In step S1203, the object to be added in step S1202 It is checked whether or not there are two or more AF frames determined as frames. If there are two or more AF frames, the process proceeds to step S1204. If it is less than two, the process proceeds to step S1212.

In step S1204, a variable i for counting the number of AF frames is initialized to 1, and the process proceeds to step S1205.
In step S1205, the variable i is incremented and the process proceeds to step S1206.

  In step S1206, the AF evaluation value of the i-th addition target frame is added to the AF evaluation value of the addition target frame closest to the reference position, and the process proceeds to step S1207. Here, the size of each AF frame may be different, and the output of the AF evaluation value varies depending on the size. Therefore, the AF evaluation value is normalized in accordance with a predetermined AF frame size for normalization and then added.

  In step S1207, focus determination is performed with respect to the values of the focusing conditions L0, SL0, and defmaxmin according to the size of the AF frame for normalization with respect to the added AF evaluation value, and the process proceeds to step S1208.

  In step S1208, the in-focus determination result in step S1207 is checked with respect to the added AF evaluation value. If YES, the process proceeds to step S1210, and if X, the process proceeds to step S1209.

  In step S1209, it is checked whether or not addition has been performed for all addition target frames. If addition has been performed for all addition target frames, the process proceeds to step S1212. If there is an addition target frame that has not yet been added, the process returns to step S1205, and the AF evaluation value of the addition target frame is further added.

  In step S1210, the peak position of the added AF evaluation value is obtained by calculation, and the process proceeds to step S1211.

  In step S1211, it is checked whether the peak position of the AF evaluation value after addition is within a predetermined depth from the reference position. If it is within the predetermined depth, the process proceeds to step S1222, and if not within the predetermined depth, the process proceeds to step S1212.

  In step S1212, focusing determination is not taken into consideration, and if all AF frames are targeted, and the peak position is within a predetermined depth from the reference position, it is determined as an addition target frame, and the process proceeds to step S1213.

  In step S1213, it is checked whether or not there are two or more AF frames determined as addition target frames in step S1212. If there are two or more AF frames, the process proceeds to step S1214. If there are less than two, the frame addition operation is terminated.

In step S1214, a variable i for counting the number of AF frames is initialized to 1, and the process proceeds to step S1215.
In step S1215, the variable i is incremented and the process proceeds to step S1216.

  In step S1216, the AF evaluation value of the i-th addition target frame is normalized and added to the AF evaluation value of the addition target frame closest to the reference position in the same manner as in step S1206, and the process proceeds to step S1217. .

  In step S1217, in the same manner as in step S1207, a focus determination based on a focus condition (predetermined values L0, SL0, defmaxmin) that matches the size of the AF frame for normalization is performed on the added AF evaluation value. The process proceeds to step S1218.

  In step S1218, the determination result in the focus determination in step S1217 is checked with respect to the added AF evaluation value, and if it is ◯, the process proceeds to step S1220, and if it is x, the process proceeds to step S1219.

  In step S1219, it is checked whether or not addition has been performed for all the addition target frames. If addition has been performed for all the addition target frames, the frame addition operation is terminated. If there is an addition target frame that has not yet been added, the process returns to step S1215 to further add the AF evaluation value of the addition target frame.

  In step S1220, the peak position of the added AF evaluation value is obtained by calculation, and the process proceeds to step S1221.

In step S1221, it is checked whether or not the peak position of the AF evaluation value after addition is within a predetermined depth from the reference position. If it is within the predetermined depth, the process proceeds to step S1222, and if it is not within the predetermined depth, a frame addition operation is performed. finish.
In step S1222, the peak position of the added AF evaluation value is determined as the in-focus position, and the frame addition operation ends.

  As described above, the accuracy of the peak position of the AF evaluation value after the addition is assured by performing the addition in order from the AF frame close to the reference position with respect to the AF frame whose peak position is within the predetermined depth from the reference position. be able to.

  As described above, when performing AF on the face area at the time of face detection, it is possible to reduce the influence of the deterioration of the S / N ratio due to noise and the influence of background omission and focus on the face with higher accuracy.

  In the above-described embodiments, the face of the subject has been described as a person's face. However, the present invention is not limited to this. For example, an animal's face and body are extracted by subject recognition such as pattern matching, and the face region and body region are extracted. An AF frame may be set.

  In the above-described embodiment, the so-called scan-type AF operation in which the focus lens is moved within a predetermined movement range to perform the scan operation has been described. However, the focus lens is moved in the direction in which the AF evaluation value increases, and the AF frame is set as described above in the hill-climbing method (also referred to as continuous AF) in which the position where the AF evaluation value is maximum is the in-focus position. May be.

(Other embodiments)
The object of each embodiment is also achieved by the following method. That is, a storage medium (or recording medium) in which a program code of software that realizes the functions of the above-described embodiments is recorded is supplied to the system or apparatus. Then, the computer (or CPU or MPU) of the system or apparatus reads and executes the program code stored in the storage medium. In this case, the program code itself read from the storage medium realizes the functions of the above-described embodiments, and the storage medium storing the program code constitutes the present invention. Further, by executing the program code read by the computer, not only the functions of the above-described embodiments are realized, but the present invention includes the following cases. That is, based on the instruction of the program code, an operating system (OS) running on the computer performs part or all of the actual processing, and the functions of the above-described embodiments are realized by the processing.

  Furthermore, the following cases are also included in the present invention. That is, the program code read from the storage medium is written in a memory provided in a function expansion card inserted into the computer or a function expansion unit connected to the computer. Thereafter, based on the instruction of the program code, the CPU or the like provided in the function expansion card or function expansion unit performs part or all of the actual processing, and the functions of the above-described embodiments are realized by the processing.

  When the present invention is applied to the above storage medium, the storage medium stores program codes corresponding to the procedure described above.

DESCRIPTION OF SYMBOLS 101 Shooting lens 102 Focus lens 103 AF process part 104 Image pick-up element 105 A / D conversion part 106 Image processing part 107 System control part 108 Shooting mode switch 109 Main switch 110 Shooting standby switch 111 Shooting switch 112 Face detection module 113 Moving body detection part 114 Angular velocity sensor unit 115 Acceleration sensor unit 116 DRAM

Claims (9)

Face detection means for detecting the position and size of a human face from a captured image;
As a focus detection area for detecting the in-focus state of the imaging optical system, the first focus detection area where the person's face exists and the person's body are predicted to be located as seen from the position of the person's face. Setting means for setting a second focus detection area;
Focus adjusting means for adjusting the focus by moving the imaging optical system based on the signal output in the focus detection region;
The setting means is configured to make the size of the second focus detection area larger than the size of the first focus detection area when the size of the face detected by the face detection means is smaller than a predetermined size. A focus detection device characterized by setting.   The focus detection apparatus according to claim 1, wherein the setting unit divides the second focus detection region in a horizontal direction with respect to the body to form a plurality of focus detection regions. Face detection means for detecting a position where a human face is present from a captured image;
As a focus detection area for detecting the in-focus state of the imaging optical system, the first focus detection area where the person's face exists and the person's body are predicted to be located as seen from the position of the person's face. Setting means for setting a second focus detection area;
Focus adjusting means for adjusting the focus by moving the imaging optical system based on the signal output in the focus detection region;
The focus detection apparatus, wherein the setting means divides the second focus detection region in a horizontal direction with respect to the body to form a plurality of focus detection regions. The face detection means further detects the angle of the face,
4. The focus detection apparatus according to claim 2, wherein the setting unit changes a direction in which the second focus detection area is divided in accordance with an angle of the face. 5.   The said setting means changes the division | segmentation number of a said 2nd focus detection area according to the size of the face detected by the said face detection means, The any one of Claim 2 thru | or 4 characterized by the above-mentioned. The focus detection apparatus described.   The said setting means changes the aspect ratio of the said 2nd focus detection area according to the position of the face detected by the said face detection means, The any one of Claim 2 thru | or 5 characterized by the above-mentioned. Focus detection device.   The setting means, when another face is detected by the face detection means at a position where the body of the person is predicted to be seen from the position of the face of the person, the second focus detection area The focus detection apparatus according to claim 2, wherein the setting is changed. A face detection step of detecting the position and size of a human face from a captured image;
As a focus detection area for detecting the in-focus state of the imaging optical system, the first focus detection area where the person's face exists and the person's body are predicted to be located as seen from the position of the person's face. A setting step for setting the second focus detection area,
A focus adjustment step of performing focus adjustment by moving the imaging optical system based on a signal output in the focus detection region,
In the setting step, when the size of the face detected in the face detection step is smaller than a predetermined size, the size of the second focus detection region is made larger than the size of the first focus detection region. A focus detection method comprising: setting. A face detection step of detecting a position where a human face exists from the captured image;
As a focus detection area for detecting the in-focus state of the imaging optical system, the first focus detection area where the person's face exists and the person's body are predicted to be located as seen from the position of the person's face. A setting step for setting the second focus detection area,
A focus adjustment step of performing focus adjustment by moving the imaging optical system based on a signal output in the focus detection region,
In the setting step, the second focus detection region is divided in a horizontal direction with respect to the body to form a plurality of focus detection regions.

Images