JP2009021893A - Imaging device and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To prevent erroneous operation in the case where a photograph including a particular item is selected as an imaging object in the imaging device in which the focus is placed on a particular object by detecting the particular object. <P>SOLUTION: In an image obtained with an imaging device for detecting a face of person as the particular object and setting the focus on the particular object by moving the focus distance corresponding to a size of the face, detecting regions are set to the upper and lower portions and right and left portions of the face of detected person. Characteristics such as profiles of changes when focus distance of an evaluation value used for focusing is changed are compared with that in the region including the particular object and the detecting regions. When the characteristics are not similar to that in the regions including the particular object in the predetermined number of detecting regions, for example, two or more detecting regions, an actual person is judged as an object locating in front of the background. When the characteristics are similar to that in the region including the particular object under two detecting regions, a photograph or a picture locating in front of the person is judged as an imaging object. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an imaging apparatus and an imaging method capable of automatically focusing, and more particularly to an imaging apparatus capable of accurately focusing on a subject when the target subject has a specific part.

  Currently, digital imaging devices such as digital cameras and digital video cameras are widely used. In such an image pickup apparatus, the focus of the lens is generally an autofocus (AF) that is automatically controlled.

  As a focusing control method, an AF evaluation value based on a high-frequency component of a luminance signal in an image is detected while changing a focusing distance from the nearest point to infinity or vice versa, and the AF evaluation value is a maximum value. There is a way to do it by focusing on the position to take.

  When capturing an image having a composition in which an actual person is placed in front of the background as shown in FIG. 14, the relationship between the AF evaluation value and the focal distance of the lens is as shown in the graph of FIG. In the graph of FIG. 15, the horizontal axis is the in-focus distance (lens position), the vertical axis is the AF evaluation value, and the horizontal axis is the left end at infinity and the right end is the closest distance. In the graph of FIG. 15, maximum values appear at two locations. The local maximum on the infinity side is that of the background mountain, and the local maximum on the nearest side is that of the person in front. In this case, the local maximum closest to the imaging device is selected.

In an imaging device having an AF function, it is necessary to reduce the time required for focusing and improve the focusing accuracy so that the user can take a focused image immediately when he / she wants to take a picture. is there. Therefore, in Patent Document 1, in a digital camera having the focusing control method as described above, the face of a person in an image is detected, and the approximate distance to the person is determined based on the size of the face in the image. A method is disclosed in which the distance is calculated, the focus position is moved in the closest direction from the distance, and focus control is started. For example, the distance to the person is estimated to be around 1.6 m from the digital camera based on the size of the person's face detected from the image of FIG. 14, and when detecting the AF evaluation value, the distance (“Start” shown in FIG. The focus distance is changed from the position “)”. According to this method, the time required to focus on the person can be shortened compared to the case where the focus distance is changed from infinity.
JP 2006-201282 A

  However, the digital camera disclosed in Patent Document 1 does not distinguish whether a person's face is an actual person or a person on a screen of a television screen, a photograph, or a picture. Therefore, if there is a photograph that includes a person nearby as a subject and the person is small, it may be mistakenly assumed that the person is far away, and conversely, the time required for focusing will become longer. Become.

  For example, when a photograph including a person hung on the wall as shown in FIG. 16 is used as a subject, the relationship between the AF evaluation value and the in-focus distance is as shown in the graph of FIG. The vertical and horizontal axes are the same as in FIG. In the graph of FIG. 17, the maximum value appears only in one position corresponding to the photograph as the subject. In the digital camera disclosed in Patent Document 1, when the approximate distance to a person is calculated based on the size of the face of the person included in the photograph in the image, the distance to the actual photograph (for example, 1.6 m) is greater. A far-inferred position (for example, 20 m, “start position” in FIG. 17) is erroneously estimated.

  Therefore, an object of the present invention is to provide an imaging apparatus that can determine whether or not a subject having a specific portion on a screen is being photographed.

  In order to achieve the above object, an imaging apparatus according to the present invention includes a lens unit that can change a focusing distance, an imaging unit that acquires a captured image including a subject through the lens unit, and focusing from the captured image. An evaluation value calculation unit that calculates the evaluation value of the lens unit, and the focus of the lens unit is at a position where the evaluation value of the captured image acquired at a focus distance within a variable range in the lens unit is a maximum value. An imaging apparatus for setting a position includes a specific part detection unit that detects a specific part of the subject from the captured image, and sets a plurality of detection regions around the specific part detected by the specific part detection unit. When the specific part of the subject is detected, the characteristic of the evaluation value in the specific part detection unit is compared with the characteristic of the evaluation value in the detection region, and the detection region similar to the characteristic of the specific part is compared. The number of is smaller than the predetermined number, characterized by continuing the setting operation of the focal length from by moving the lens unit in focusing distance corresponding to the size of the particular portion. As a characteristic of the evaluation value, a direction (sign) of a change in the evaluation value accompanying a change in the evaluation value or the focus distance can be used.

  In addition, after the lens unit is adjusted to the in-focus distance calculated from the size of the specific portion, the in-focus distance setting operation may be continued based on the evaluation value in the specific portion.

  In addition, the image processing apparatus includes an area dividing unit that divides the captured image into a plurality of divided areas, wherein the detection area includes the previous divided area and has the same shape as the divided area including the specific part, However, it may be set at four locations.

  The specific portion may be a human face.

  In the focusing distance setting operation, the direction in which the focusing distance of the lens unit is changed may be from the infinity side toward the closest distance side.

  In the gradation correction method of the present invention, a first step of acquiring a captured image including a subject through the lens unit at a predetermined focusing distance, and a second step of calculating a focus evaluation value from the captured image. A third step of detecting a specific part of the subject from the captured image, characteristics of the evaluation value of the specific part, and characteristics of the evaluation value of a plurality of detection regions set around the specific part A fifth step of calculating the number of the detection regions whose characteristics of the evaluation value compared in the fourth step are similar to the specific part, and the fifth step In the case where the number of the detection areas similar to the specific portion calculated in step S is less than a predetermined number, the focus corresponding to the size of the specific portion detected in the third step Said to the distance Characterized in that it comprises a sixth step of combining the lens unit.

  According to the present invention, for example, in an imaging apparatus that detects a person's face as a specific part and focuses based on the size of the face, the evaluation value of the person's face is compared with the evaluation value of the detection area around the face To determine whether it is a photograph or an actual person. Therefore, when a photograph including a person is nearby as a subject and the person is small, it is possible to prevent the focusing operation from being delayed due to an erroneous estimation that the actual person is far away.

<First Embodiment>
A first embodiment of the present invention will be described with reference to the drawings. In the following description, an imaging apparatus such as a digital camera or a digital video that performs the shooting method of the present invention will be described as an example. The imaging device may be capable of capturing a moving image as long as it can capture a still image.

(Configuration of imaging device)
First, the internal configuration of the imaging apparatus will be described with reference to the drawings. FIG. 1 is a block diagram illustrating an internal configuration of the imaging apparatus according to the first embodiment.

  The imaging apparatus of FIG. 1 includes a solid-state imaging device (image sensor) 1 such as a CCD or CMOS sensor that converts incident light into an electrical signal, and a zoom lens and a zoom lens that form an optical image of a subject on the image sensor 1. A lens unit 2 having a motor for changing the focal length of the zoom lens, that is, the optical zoom magnification, and a motor for focusing the zoom lens on the subject, and an image signal which is an analog signal output from the image sensor 1 is converted into a digital signal. An AFE (Analog Front End) 3, a microphone 4 that converts externally input sound into an electrical signal, and various image processing including blurring processing and gradation correction on an image signal that is a digital signal from the AFE 3 An AF evaluation value used for focusing control is calculated from the image processing unit 5 to be applied and the image signal input from the image processing unit 5 An evaluation value calculation unit 23, an audio processing unit 6 that converts an audio signal that is an analog signal from the microphone 4 into a digital signal, and a JPEG (Joint) for the image signal from the image processing unit 5 when a still image is taken. When a moving image is shot, such as a Photographic Experts Group (compression) compression method, compression coding processing such as an MPEG (Moving Picture Experts Group) compression method is performed on the image signal from the image processing unit 5 and the audio signal from the audio processing unit 6. A compression processing unit 7 that performs the processing, a driver unit 8 that records the compressed encoded signal compressed by the compression processing unit 7 in an external memory 22 such as an SD card, and a compressed code read from the external memory 22 by the driver unit 8 A decompression processing unit 9 that decompresses and decodes the converted signal, a video output circuit unit 10 that converts an image signal obtained by decoding by the decompression processing unit 9 into an analog signal, and a video output circuit unit 1 The video output terminal 11 for outputting the signal converted in the above, the display unit 12 having an LCD or the like for displaying an image based on the signal from the video output circuit unit 10, and the audio signal from the expansion processing unit 9 as an analog signal An audio output circuit unit 13 for conversion, an audio output terminal 14 for outputting a signal converted by the audio output circuit unit 13, a speaker unit 15 for reproducing and outputting audio based on an audio signal from the audio output circuit unit 13; A timing generator (TG) 16 that outputs a timing control signal for matching the operation timing of each block, a CPU (Central Processing Unit) 17 that controls the drive operation of the entire imaging apparatus, and each program for each operation And a memory 18 for temporarily storing data during program execution, and a shutter button for still image shooting An operation unit 19 to which an instruction from the user is input, a bus line 20 for exchanging data between the CPU 17 and each block, and a bus for exchanging data between the memory 18 and each block. And a line 21. In the lens unit 2, the CPU 17 controls the focus and the diaphragm by driving the motor in accordance with the image signal detected by the image processing unit 5.

(Basic operation of the imaging device)
Next, the basic operation of the image pickup apparatus at the time of still image shooting will be described with reference to the flowchart of FIG. First, when the user turns on the power of the imaging apparatus (STEP 201), the imaging mode of the imaging apparatus, that is, the driving mode of the image sensor 1 is set to the preview mode (STEP 202). The preview mode is a mode in which an image to be photographed is displayed on the display unit 12, and can be used to determine the photographing object and determine the composition. Subsequently, the camera enters a shooting mode input waiting state, and a mode according to the function of the imaging device, such as a mode suitable for human photographing, a mode suitable for photographing a moving object, a mode suitable for photographing with backlight, and the like are selected. If the shooting mode is not input, the preview mode is selected (STEP 203). In the preview mode, an image signal that is an analog signal obtained by the photoelectric conversion operation of the image sensor 1 is converted into a digital signal by the AFE 3, subjected to image processing by the image processing unit 5, and compressed by the compression processing unit 7. An image signal for the current image is temporarily recorded in the external memory 22. The compressed signal is expanded by the expansion processing unit 9 via the driver unit 8, and an image with an angle of view at the zoom magnification of the lens unit 2 set at the present time is displayed on the display unit 12.

  Subsequently, the user sets the zoom magnification with the optical zoom so that the desired angle of view is obtained with respect to the subject to be imaged (STEP 204). At this time, the lens unit 2 is controlled by the CPU 17 based on the image signal input to the image processing unit 5, and optimum exposure control (Automatic Exposure; AE) / focusing control (Auto Focus; AF) is performed. It is carried out (STEP 205). When the user determines the shooting angle of view and composition and presses the shutter button of the operation unit 19 halfway (STEP 206), the AE is adjusted (STEP 207) and the AF optimization process is performed (STEP 208). The AF optimization process will be described later.

  Thereafter, when the shutter button is fully pressed (STEP 209), the TG 16 gives timing control signals to the image sensor 1, the AFE 3, the image processing unit 5 and the compression processing unit 7 to synchronize the operation timing of each unit. Then, the drive mode of the image sensor 1 is set to the still image shooting mode (STEP 210), and the image signal (raw data) that is an analog signal output from the image sensor 1 is converted into a digital signal by the AFE 3, and the image processing unit 5 is temporarily set. Is written in the frame memory (STEP 211). This digital signal is read from the frame memory, and various image processing such as signal conversion processing for generating a luminance signal and a color difference signal is performed in the image processing unit 5, and the signal subjected to the image processing is processed in the compression processing unit 7. After being compressed to JPEG (Joint Photographic Experts Group) format (STEP 212), the compressed image is written in the external memory 22 (STEP 213), and the photographing is completed. Thereafter, the process returns to the preview mode (STEP 202).

(Basic operation of the imaging device)
The operation during movie shooting will be described. In this imaging apparatus, when the operation unit 19 instructs to perform an imaging operation, an image signal that is an analog signal obtained by the photoelectric conversion operation of the image sensor 1 is output to the AFE 3. At this time, the image sensor 1 receives the timing control signal from the TG 16 to perform horizontal scanning and vertical scanning, and output an image signal as data for each pixel. In the AFE 3, when an image signal (raw data) that is an analog signal is converted into a digital signal and input to the image processing unit 5, various image processing such as signal conversion processing that generates a luminance signal and a color difference signal. Is given.

  Then, the image signal subjected to the image processing by the image processing unit 5 is given to the compression processing unit 7. At this time, an audio signal which is an analog signal obtained by inputting the sound into the microphone 4 is converted into a digital signal by the audio processing unit 6 and given to the compression processing unit 7. As a result, the compression processing unit 7 compresses and encodes the image signal and the audio signal, which are digital signals, based on the MPEG compression encoding method, gives the image signal and the audio signal to the driver unit 8, and records them in the external memory 22. At this time, the compressed signal recorded in the external memory 22 is read by the driver unit 8 and applied to the decompression processing unit 9, and decompressed to obtain an image signal. This image signal is given to the display unit 12 to display a subject image currently photographed through the image sensor 1.

  When the imaging operation is performed in this way, a timing control signal is given to the AFE 3, the image processing unit 5, the audio processing unit 6, the compression processing unit 7, and the expansion processing unit 9 by the TG 16, and one frame by the image sensor 1. An operation synchronized with each imaging operation is performed.

  When the operation unit 19 instructs to play back the moving image recorded in the external memory 22, the compressed signal recorded in the external memory 22 is read out by the driver unit 8 and given to the decompression processing unit 9. It is done. Then, the decompression processing unit 9 decompresses and decodes the image signal and the audio signal based on the MPEG compression encoding method. Then, the image signal is given to the display unit 12 to reproduce the image, and the audio signal is given to the speaker unit 15 via the audio output circuit unit 13 to reproduce the audio. Thereby, a moving image based on the compressed signal recorded in the external memory 22 is reproduced together with the sound.

  When it is instructed to reproduce an image, the decompression processing unit 9 decompresses and decodes the compressed signal recorded in the external memory 22 based on the JPEG compression encoding method, and acquires an image signal. Then, the image signal is given to the display unit 12 to reproduce the image.

(AF optimization process)
Next, AF optimization processing will be described. FIG. 3 is an example of an image obtained by dividing a photographed image area into a plurality of regions, and FIG. 4 is a flowchart of AF optimization processing according to the present embodiment.

  In the flowchart of FIG. 4, when the optimization process is started, first, the focusing distance of the lens unit 2 is set to infinity, and the distance information and the AF evaluation value stored in the distance information storage unit 17 a provided in the CPU 17 are set. Erasing and initializing the evaluation value calculation unit 23, for example, initializes a portion related to focusing and distance detection (STEP 401). Subsequently, an image at the in-focus distance (in this case, infinity because it is immediately after initialization) is read, and the image is read into a plurality of areas, for example, 256 areas of 16 columns and 16 rows as shown in FIG. The AF evaluation values of all the divided areas are acquired (STEP 402). Here, the intensity of the high-frequency component of the luminance signal can be used as the AF evaluation value.

  Next, it is detected whether or not there is a human face in the image read in STEP 401 (STEP 403). The human face detection will be described later. If no face is detected (No in STEP 403), the evaluation value calculation unit 23 compares the sum of the AF evaluation values of all the divided areas of the image at the current focus distance with that at the previous focus distance (STEP 411). . If the focus distance of the current image is infinity, there is no image of the previous focus distance to be compared because it is the first image, so STEP 411 is omitted and the focus distance is shortened. A predetermined distance is moved (STEP 413). If the result of comparing the sum of the AF evaluation values of all the divided areas has decreased (No in STEP 412), the lens unit is assumed to have a maximum AF evaluation value and focus on the subject at the previous focusing distance. 2 is returned to the previous focusing distance (STEP 414), and the AF optimization process is terminated. If the result of comparing the AF evaluation values has increased (Yes in STEP 412), the lens is moved by a predetermined distance in the direction in which the AF evaluation value is predicted to increase again, that is, in the direction in which the focusing distance of the lens unit is shortened ( (STEP413). Thereafter, when the CPU 17 receives a vertical synchronization (hereinafter referred to as V sync) signal from the TG 16, the process returns to STEP 402, an image is read at a short focal distance by a predetermined distance, and an AF evaluation value in the divided area is acquired (STEP 402).

(Face detection process)
Here, the face detection process of the imaging apparatus will be described. The image processing unit 5 includes a face detection device 40 and can detect a person's face from the input image signal. The configuration and operation of the face detection device 40 will be described below.

  FIG. 5 is a block diagram showing a configuration of the face detection device 40. The face detection device 40 is stored in the memory 18 and reduced image generation means 42 for generating one or a plurality of reduced images based on the image data obtained by the AFE 3, each hierarchical image composed of the input image and the reduced image, and the memory 18. A face determination unit 45 that determines whether or not a face exists in the input image using a face detection weight table, and a detection result output unit 46 that outputs a detection result of the face determination unit 45 are provided. When a face is detected, the detection result output means 46 outputs a distance to the face estimated from the size and position of the detected face input image and the size of the face.

  The weight table stored in the memory 18 is obtained from a large number of teacher samples (face and non-face sample images). Such a weight table can be created using, for example, a known learning method called Adaboost (Yoav Freund, Robert E. Schapire, “A decision-theoretic generalization of on-line learning and an application to boosting” ", European Conference on Computational Learning Theory, September 20, 1995.).

Adaboost is an adaptive boosting learning method. Based on a large number of teacher samples, Adaboost selects multiple weak classifiers that are effective for identification from among a plurality of weak classifier candidates. This is a learning method for realizing a highly accurate classifier by weighting and integrating. Here, a weak classifier refers to a classifier that has a higher discrimination ability than a coincidence but is not high enough to satisfy sufficient accuracy. When a weak classifier is selected, if there is a weak classifier that has already been selected, the learning is focused on the teacher sample that is misrecognized by the selected weak classifier. To select the most effective weak classifier.

  FIG. 6 shows an example of a hierarchical image obtained by the reduced image generating means 42. This example shows a plurality of hierarchical images generated when the reduction ratio is set to 0.8. In FIG. 6, 50 indicates an input image, and 51-55 indicate reduced images. Reference numeral 61 denotes a determination area. In this example, the determination area is set to a size of 24 pixels vertically and 24 pixels horizontally. The size of the determination area is the same for the input image and each reduced image. In this example, as indicated by an arrow, a face image matching the determination area is detected by moving the determination area from left to right on the hierarchical image and performing horizontal scanning from the top to the bottom. I do. However, the scanning order is not limited to this. The reason why the plurality of reduced images 51 to 55 are generated in addition to the input image 50 is to detect faces of different sizes using one kind of weight table.

  FIG. 7 is a diagram for explaining the face detection process. The face detection processing by the face determination unit 45 is performed for each hierarchical image, but since the processing method is the same, only the face detection processing performed on the input image 50 will be described here. FIG. 7 shows an input image 50 and a determination area 61 set in the input image.

  The face detection process performed for each hierarchical image is performed using an image corresponding to the determination region set in the image and a weight table. The face detection process includes a plurality of determination steps that sequentially shift from a rough determination to a fine determination. When a face is not detected in a certain determination step, the process does not proceed to the next determination step, and the determination area includes It is determined that no face exists. In all the determination steps, only when a face is detected, it is determined that a face exists in the determination area, and the determination area is scanned to shift to determination in the next determination area. When a face is detected, the size of the face and the distance to the face can be estimated based on the input image depending on which layer image is used. Thus, the detection result output means 46 outputs the position and size of the detected face and the distance to the person having the face. Such face detection processing is described in detail in Japanese Patent Application No. 2006-053304, which is a patent application filed by the present applicant.

  If a human face is detected in STEP 403 (YES in STEP 403), it is checked whether or not a face has been detected in the image at the previous focusing distance (STEP 404). When a face is not detected in the image at the previous focusing distance (No in STEP 404), a face area is set, and detection of the same size as the face area is performed above, below, right, and left of the face area. Set the area. Then, AF evaluation values of the face area and the detection area are acquired. The AF evaluation value in each area in the image at the previous focusing distance is compared. The number of detection areas in which the direction of change (positive, negative, or no change) between the previous and current ones is different from the face area is counted (STEP 406). In the present embodiment, the number of detection areas whose direction of change is different from the face area is referred to as reliability R.

  When the reliability R is less than 2, that is, 0 or 1 (No in STEP 407), the process proceeds to the above-described STEP 411. In this case, there are three or four detection areas in which the direction of change of the AF evaluation value is the same as that of the face area, and the face and the background are at the same distance, that is, a TV screen including a person as a subject. It is assumed that a normal focusing operation is performed for a photograph or a picture. Examples of such a subject include a photograph including a person hung on a wall surface shown in FIG. FIG. 8 shows the relationship between the AF evaluation value and the focus distance of the face area and the four detection areas set in the area obtained by dividing the image of FIG. 16 into 256 areas of 16 columns and 16 rows. Is shown in FIG. In FIG. 9, although the AF evaluation value of the face area and the AF evaluation value of the detection area are different in size, the focal distance of the maximum value is almost the same.

  When the reliability R is 2 or more, that is, 2, 3 or 4 (Yes in STEP 407), a face detection frame is displayed in the face area on the display unit 12 (STEP 408). By displaying the face detection frame, the subject to be focused can be made clear to the user. When the reliability R is 2 or more, there are 2 to 4 detection areas in which the direction of change of the AF evaluation value differs from the face area, and the face and the background are at different distances, that is, the subject is actually It is assumed that the focusing operation is performed on the assumption that the person is a background and a background away from the person. Therefore, the distance to the person is estimated based on the size of the face area, and the focusing distance of the lens unit 2 is moved to a position farther than the distance by a predetermined ratio (STEP 409). Estimating the distance to the person based on the size of the face area can be performed from the size of the hierarchical image used when the face is detected as described above.

  As such a subject, for example, a background mountain as shown in FIG. 3 and an actual person arranged in front of the mountain can be cited. In the image of FIG. 3, the face area and the four detection areas are each composed of four areas in two columns and two rows. FIG. 10 shows the relationship between the AF evaluation value and the in-focus distance for the face area and the four detection areas set in the image of FIG. In FIG. 10, although the AF evaluation value of the face area and the AF evaluation value of the detection area at the position corresponding to the human body are different in size, the position of the maximum value is almost the same. However, the manner in which the AF evaluation value of the detection area at the position corresponding to the background changes with respect to the in-focus distance is different from the AF evaluation value of the face area, and the position of the maximum value is also different.

  Thereafter, when the CPU 17 receives the V sync signal from the TG 16, the process returns to STEP 402, an image is read at a short focal distance by a predetermined distance, and an AF evaluation value in the divided area is acquired.

  If a face of a person is detected in STEP 403 and a face is also detected in an image at the previous focusing distance (YES in STEP 404), the process proceeds to the above-described STEP 411. This is because the in-focus distance has already been moved to a position based on the distance estimated from the face size (STEP 409), and it is not necessary to move the in-focus distance to that distance again. Then, the sum of the AF evaluation values of all the divided areas is compared with the previous one (STEP 411). If the result is decreased (No in STEP 412), the in-focus distance of the lens unit 2 is set to the previous focus. The distance is returned (STEP 414), and the AF optimization process is terminated. If the result of comparing the AF evaluation values has increased, the focusing distance of the lens unit 2 is moved by a predetermined distance in the direction of shortening (STEP 413), and the CPU 17 receives a V sync signal from the TG 16 (STEP 410). Return to.

  Here, the AF optimization processing of the present embodiment will be described with reference to the graph of FIG. 11 that shows the sum of the AF evaluation values of all the divided regions and the relationship between the AF evaluation values of the face regions and the in-focus distance. The graph of FIG. 11 is an enlarged view of the vicinity of the total maximum value of the AF evaluation values of all the divided areas, and one scale on the horizontal axis is a predetermined distance for one step in which the focal distance of the lens unit 2 is moved. . From the first focusing distance on the infinity side to the fourth focusing distance in FIG. 11 is a case where no face is detected in the image at each focusing distance (No in STEP 403), and AF in all the divided areas is performed. This is the case where the result of comparing the sum of the evaluation values with that of the image at the previous focusing distance has increased (No in STEP 412). At each focusing distance, the direction in which the AF evaluation value increases, that is, the lens portion A predetermined distance is moved in a direction to shorten the focusing distance (STEP 413). At the fifth in-focus distance, a face is detected in the image at this in-focus distance, and the reliability R is 2 or more (Yes in STEP 407). Therefore, the in-focus distance based on the distance estimated from the size of the face area ( In the case of FIG. 11, the lens is moved to a focus distance shorter by two steps of a predetermined distance. Therefore, the reading of the image at the focusing distance corresponding to the sixth is omitted. Thereafter, the face is detected from the images at the respective in-focus distances from the seventh in-focus distance to the tenth in-focus distance, but is continuously detected. Therefore, the sum of the AF evaluation values of all the divided areas is compared with the image of the previous focus distance (the fifth focus distance acquired last time for the seventh focus distance) at each focus distance. Since the result increases from the seventh in-focus distance to the ninth in-focus distance (No in STEP 412), the in-focus distance of the lens unit is moved by a predetermined distance (STEP 413). At the 10th in-focus distance, the result of comparing the sum of the AF evaluation values of all the divided areas with that of the image at the 9th in-focus distance is decreased (Yes in STEP 412). Return to the in-focus distance (STEP 414). The graph of FIG. 11 is a schematic diagram for simply showing the operation of the AF optimization process. Therefore, the predetermined distance for one step of moving the focusing distance of the lens unit 2 may actually be finer than that of FIG. 11 in order to increase the focusing accuracy, and the AF may be optimized. It may be rough to make the process faster.

  In this way, by moving the focusing distance of the lens unit 2 to the position based on the distance estimated from the size of the face area, the focusing distance can be changed more quickly than when changing the focusing distance in steps of a predetermined distance from infinity. Focus on a person. Furthermore, since the reliability is evaluated that the subject is an actual person, if the photograph including the person is nearby as the subject and the person is small, the person is mistakenly assumed to be far away. In this way, it is possible to avoid an increase in the time required for focusing. The in-focus distance is moved to a position that is a predetermined percentage away from the estimated distance. If the estimated distance is closer than the actual distance, the in-focus distance is moved to the estimated distance. This is because it becomes impossible to focus on a person.

  In the present embodiment, in the AF optimization process, the high-frequency component of the luminance signal may be used as the AF evaluation value, and the AF evaluation values of the entire area may be added together or used in a specific area, for example, FIG. The 64 regions of 8 columns in the center and 8 rows in the center may be used by adding more weight than other regions. Further, the AF evaluation values for only the specific area may be added together.

(Second Embodiment)
A second embodiment of the present invention will be described with reference to the drawings. FIG. 12 is a flowchart for AF optimization processing of the imaging apparatus according to the second embodiment. The present embodiment is the same as the first embodiment except that the operation when a face is detected in the AF optimization process is different, and substantially the same parts are denoted by the same reference numerals.

  In the flowchart of FIG. 12, when the optimization process is started, first, the focusing distance of the lens unit 2 is set to infinity, and the AF evaluation value stored in the distance information storage unit 17a provided in the CPU 17 is deleted. Then, the part related to distance detection is initialized (STEP 401). Subsequently, an image at the in-focus distance (in this case, infinity because it is just after initialization) is read, the image is divided into a plurality of areas, and AF evaluation values of all the divided areas are acquired (STEP 402).

  Next, it is detected whether or not there is a human face in the image read in STEP 401 (STEP 403). If no face is detected (No in STEP 403), the sum of the AF evaluation values of all of the current divided areas is compared with the previous one (STEP 1214). When the focus distance of the current image is infinite, since it is the first image, there is no previous focus distance image to be compared, so STEP 1214 is omitted and the focus distance is shortened. A predetermined distance is moved (STEP 413). If the result of comparing the sum of the AF evaluation values of all the divided areas has decreased (No in STEP 412), the lens unit is assumed to have a maximum AF evaluation value and focus on the subject at the previous focusing distance. 2 is returned to the previous focusing distance (STEP 1215), and the AF optimization process is terminated. If the result of comparing the AF evaluation values has increased (Yes in STEP 412), the lens is moved by a predetermined distance in the direction in which the AF evaluation value is predicted to increase again, that is, in the direction in which the focusing distance of the lens unit is shortened ( (STEP413). Thereafter, when the CPU 17 receives the V sync signal from the TG 16, the process returns to STEP 402, an image at a predetermined short focal distance is read, and an AF evaluation value in the divided area is acquired (STEP 402).

  When the face of a person is detected in STEP 403, it is checked whether or not a face has been detected in the image at the previous focusing distance (STEP 404). When a face is not detected in the image at the previous focusing distance (No in STEP 404), a face area is set, and detection of the same size as the face area is performed above, below, right, and left of the face area. Set the area. Then, the AF evaluation values of the face area and the detection area are acquired and compared with those of the previous in-focus distance (STEP 405), and the reliability R is calculated (STEP 406).

  When the reliability R is less than 2 (No in STEP 407), the process proceeds to the above STEP 1214. When the reliability R is 2 or more (YES in STEP 407), a face detection frame is displayed in the face area on the display unit 12 (STEP 408). Then, the distance to the person is estimated from the size of the face area, and the focusing distance of the lens unit 2 is moved to a position farther than the distance by a predetermined ratio (STEP 409).

  Thereafter, when the CPU 17 receives the V sync signal from the TG 16, the process returns to STEP 402, an image is read at a short focal distance by a predetermined distance, and an AF evaluation value in the divided area is acquired.

  If a human face is detected in STEP 403 (Yes in STEP 403) and a face is also detected in the image at the previous focusing distance (Yes in STEP 404), the AF evaluation value of the current face area is the previous one. (STEP411). If the result of comparing the total AF evaluation values of the face area has decreased (No in STEP 412), the focusing distance of the lens unit 2 is returned to the previous focusing distance (STEP 1215), and AF optimization processing is performed. finish. If the result of comparing the AF evaluation values has increased (Yes in STEP 412), the lens is moved by a predetermined distance in the direction in which the AF evaluation value is predicted to increase again, that is, in the direction in which the focusing distance of the lens unit is shortened ( (STEP413). Thereafter, when the CPU 17 receives the V sync signal from the TG 16, the process returns to STEP 402, an image at a predetermined short focal distance is read, and an AF evaluation value in the divided area is acquired (STEP 402).

  The AF optimization process when a human face is detected in STEP 403 will be described using the graph of FIG. 13 that shows the sum of the AF evaluation values of all the divided areas and the relationship between the AF evaluation values of the face areas and the in-focus distance. To do. The graph of FIG. 13 is the same graph as FIG. From the first focusing distance on the infinity side to the seventh focusing distance in FIG. 13 is a case where no face is detected in the image at each focusing distance (No in STEP 403), and AF in all the divided areas is performed. This is the case where the result of comparing the sum of the evaluation values with that of the image at the previous focusing distance has increased (No in STEP 412). At each focusing distance, the direction in which the AF evaluation value increases, that is, the lens portion A predetermined distance is moved in a direction to shorten the focusing distance (STEP 413). At the fifth in-focus distance, the face is detected for the first time in the image at this in-focus distance (Yes in STEP 403, No in STEP 404), and the reliability R is 2 or more (Yes in STEP 407). Is moved to a focusing distance based on the distance estimated from (a focusing distance shorter by two steps of a predetermined distance in the case of FIG. 13). After that, although the face is detected from the images at the respective in-focus distances from the seventh in-focus distance to the ninth in-focus distance, they are continuously detected (STEP 404). Therefore, the sum of the AF evaluation values of the face area is compared with that of the image at the previous focus distance (the fifth focus distance acquired at the seventh focus distance) at each focus distance (STEP 411). Since the results of the seventh and eighth in-focus distances increase (No in STEP 412), the in-focus distance of the lens unit is moved by a predetermined distance (STEP 413). The result is decreased at the ninth focus distance (Yes in STEP 412), and the focus distance is returned to the previous focus, that is, the eighth focus distance (STEP 1215).

  In this way, by bringing the focusing distance of the lens unit 2 to a position based on the distance estimated from the size of the face area, it is possible to quickly focus on the person and trust that the subject is an actual person. Since the evaluation is based on the degree R, it is possible to avoid an increase in the time required for focusing when a photograph including a person is nearby as a subject. Further, by performing the focusing operation using the AF evaluation value of the face area, it is possible to accurately focus on the person who is the subject.

  In the present embodiment, the AF optimization process when a face is not detected may use the high frequency component of the luminance signal as the AF evaluation value and add the AF evaluation values of the entire area, or use a specific area. For example, in FIG. 3, 64 regions of 8 columns in the center and 8 rows in the center may be used by adding more weight than other regions. Further, the AF evaluation values for only the specific area may be added together.

  In the AF optimization process of the first embodiment and the second embodiment, when two or more faces are detected as in the case where there are two or more persons including the face, focus on any person. The face detection frame displayed on the display unit 12 may be selected by the operation unit 19 or may be automatically selected based on a predetermined criterion such as nearer or farther. Good.

  In addition, the focusing operation is performed by paying attention to a person including a face as a subject, but the subject to be focused is not limited to a person including a face as long as the subject has a specific part.

  The present invention can be applied to an imaging apparatus having a function of detecting a specific part of a subject like a face detection function.

These are block diagrams which show the internal structure of the imaging device which concerns on 1st Embodiment. These are the flowcharts for demonstrating the basic operation | movement of the imaging device which concerns on 1st Embodiment. Is an example of a captured image composed of a person and a background, which is divided into regions and detects a face region. These are the flowcharts for demonstrating the optimization process of AF of the imaging device which concerns on 1st Embodiment. These are block diagrams which show the structure of a face detection apparatus. Is an example of a hierarchical image obtained by the reduced image generating means. FIG. 10 is a diagram for explaining face detection processing; Is an example of a captured image made up of a photograph including a person hung on a wall surface, which is divided into regions and detects a face region. These are the graphs of the relationship between the AF evaluation value and the focus distance of the face area and detection area of the image of FIG. These are the graphs of the relationship between the AF evaluation value and the focus distance of the face area and detection area of the image of FIG. These are graphs of the relationship between the AF evaluation values and the focus distances of all face areas and divided areas, for explaining the AF optimization processing according to the first embodiment. These are the flowcharts for demonstrating the optimization process of AF of the imaging device which concerns on 2nd Embodiment. These are graphs of the relationship between AF evaluation values and focus distances of all face areas and divided areas, for explaining AF optimization processing according to the second embodiment. Is an example of a captured image composed of a person and a background. These are graphs of the relationship between the AF evaluation value of the image of FIG. 14 and the in-focus distance. Is an example of a captured image composed of a photograph including a person hung on a wall surface. FIG. 17 is a graph showing the relationship between the AF evaluation value of the image of FIG. 16 and the in-focus distance.

Explanation of symbols

1
DESCRIPTION OF SYMBOLS 1 Image sensor 2 Lens part 3 AFE
4 microphone 5 image processing unit 6 audio processing unit 7 compression processing unit 8 driver unit 9 expansion processing unit 10 video output circuit unit 11 video output terminal 12 display unit 13 audio output circuit unit 14 audio output terminal 15 speaker unit 16 timing generator 17 CPU
17a Distance information storage unit 18 Memory 19 Operation unit 20 Bus line 21 Bus line 22 External memory 23 Evaluation value calculation unit 40 Face detection device 42 Reduced image generation means 45 Face determination means 46 Detection result output means 50 Input image 51 Reduced image 52 Reduction Image 53 Reduced image 54 Reduced image 55 Reduced image 61 Determination area

Claims (6)

A lens unit whose focus distance can be changed;
An imaging unit that acquires a captured image including a subject through the lens unit;
An evaluation value calculation unit for calculating an evaluation value for focusing from the captured image;
With
In the imaging apparatus for setting the focusing position of the lens unit at a position where the evaluation value of the captured image acquired at a focusing distance in a range that can be changed in the lens unit is a maximum value,
A specific part detector for detecting a specific part of the subject from the captured image;
Set a plurality of detection areas around the specific part detected by the specific part detection unit,
When the specific part of the subject is detected, the characteristic of the evaluation value in the specific part detection unit is compared with the characteristic of the evaluation value in the detection area, and the number of the detection areas similar to the characteristic of the specific part is predetermined. The imaging apparatus is characterized by continuing the focusing distance setting operation after moving the lens unit to a focusing distance corresponding to the size of the specific portion.   The focus distance setting operation is continued based on the evaluation value in the specific portion after the lens unit is adjusted to the focus distance calculated from the size of the specific portion. The imaging device described. An area dividing unit that divides the captured image into a plurality of divided areas;
2. The detection area according to claim 1, wherein the detection area is composed of a divided area in the previous period and has the same shape as the divided area including the specific part, and is set at four positions on the specific part. 2. The imaging device according to 2.   The imaging device according to claim 1, wherein the specific portion is a human face.   5. The direction of changing the focusing distance of the lens unit in the focusing distance setting operation is from the infinity side toward the closest distance side. 6. Imaging device. A first step of acquiring a captured image including a subject through the lens unit at a predetermined focusing distance;
A second step of calculating an evaluation value for focusing from the captured image;
A third step of detecting a specific part of the subject from the captured image;
A fourth step of comparing the characteristic of the evaluation value of the specific part with the characteristic of the evaluation value of a plurality of detection regions set around the specific part;
A fifth step of calculating the number of the detection regions whose characteristics of the evaluation value compared in the fourth step are similar to the specific part;
The size of the specific portion detected in the third step when the number of the detection areas whose characteristics of the evaluation value calculated in the fifth step are similar to the specific portion is equal to or less than a predetermined number. A sixth step of adjusting the lens unit to a focusing distance corresponding to
An imaging method comprising:

Images