JP5137808B2 - Imaging device, control method thereof, and program - Google Patents

Description

  The present invention relates to an imaging apparatus such as a video camera, and more particularly to a technique for detecting a specific subject such as a face.

  In recent years, imaging devices such as a still camera, a video camera, and a surveillance camera that detect a specific subject such as a face from a photographed image and perform exposure control, focus control, and the like suitable for the specific subject have been realized.

  As a face detection method, the face inclination, orientation, size, position on the photographed image, etc. included in the photographed image are usually indeterminate, so detection conditions such as the face tilt, orientation, size, position, etc. A method of detecting a face while changing the position is often used.

  Patent Document 1 discloses a technique for detecting a face by scanning a sub window on a digital image and sequentially discriminating whether or not the sub window image is a face image including a face. ing.

In Patent Document 2, face detection processing is performed while changing the detection region. At that time, when the face inclination is small, the detection area movement width is reduced to increase the face detection accuracy, and when the face inclination is large, the detection area movement width is increased to detect the face. Instead of lower accuracy, the detection processing time is shortened.
US2002 / 0102024 A1 (US Patent Application Publication) JP 2007-128127 A

  However, in the method of detecting a specific subject while changing the detection condition, normally, if it is attempted to increase the detection accuracy of the specific subject, it is necessary to set the detection condition in detail, and therefore the processing time becomes long. On the other hand, if it is attempted to speed up the detection process, the detection conditions need to be set roughly, so that the detection accuracy of the specific subject is lowered.

  That is, the detection accuracy of a specific subject and the detection processing time are in a so-called trade-off relationship, and it is difficult to achieve both high detection accuracy and a short detection processing time. In this case, the processing time for detecting a desired specific subject such as a face from the captured image becomes long, and if the specific subject is not detected within the specified time, appropriate focusing control and exposure control for the specific subject can be performed. This is a big problem in terms of image quality.

  In particular, in a video camera that shoots a moving image, a specific subject often moves, so that it is desired to detect the specific subject continuously within a specified time for each frame image.

  In the face detection system according to Patent Literature 1, the face detection processing time is not taken into consideration, and proper focus control and exposure control on the face may not be performed. On the other hand, in the face detection system according to Patent Document 2, consideration is given to shortening the processing time when the inclination of the face to be detected is large.

  However, in Patent Document 2, when the inclination of the face is small, only the processing is performed to increase the detection accuracy, and the face detection processing time is not considered. For this reason, in Patent Document 2, when the inclination of the face is small, it may be impossible to perform proper focus control and exposure control on the face, as in Patent Document 1 and the like.

  The present invention has been made under such a technical background, and an object thereof is to continuously detect a specific subject within a specified time and improve the detection accuracy of the specific subject. is there.

  In order to achieve the above object, an imaging apparatus according to the present invention sets a detection unit for detecting a specific subject from a captured image shot by a shooting unit, and a detection condition for the detection unit to detect the specific subject. Setting means for switching between the first detection mode and the second detection mode within a range in which the detection means can detect the specific subject within a predetermined time, The first detection mode is a detection mode in which priority is given to detecting the specific subject having a smaller size compared to the second detection mode, and the second detection mode is compared with the first detection mode. It is a detection mode in which priority is given to detecting the specific subject in an inclined state.

  According to the present invention, it is possible to continuously detect a specific subject within a specified time and improve the detection accuracy of the specific subject.

  Hereinafter, the best mode for carrying out the invention will be described with reference to the drawings.

  FIG. 1 is a configuration diagram of an imaging apparatus according to an embodiment of the present invention. The imaging apparatus is configured as a digital video camera (hereinafter referred to as a video camera) that records a moving image on a disk-type recording medium.

  In FIG. 1, a video camera 100 includes a CPU 101, a RAM 102, a ROM 103, and an operation unit 104 as a control system. The CPU 101 is connected to the RAM 102, the ROM 103, and other blocks via the internal bus 105, and controls each block based on a control program stored in the ROM 103.

  At this time, the CPU 101 uses the RAM 102 as a work area. The ROM 103 also stores programs according to the flowcharts of FIGS. When the user operates various operation levers or buttons, the operation unit 104 outputs an operation signal related to the operation to the CPU 101, and the CPU 101 executes various processes according to the operation signal.

  The video camera 100 has camera system blocks denoted by reference numerals 110 to 115 in FIG. The optical image that has passed through the zoom lens 110 and the focus lens 111 is formed on the imaging surface of the CCD image sensor 114 and is photoelectrically converted. The driver 113 drives the variable power lens 110, the focus lens 111, and the aperture 112 under the control of the CPU 101, and adjusts the optical zoom magnification, focus, and aperture amount.

  The AFE 115 drives the CCD image sensor 114 based on the shutter speed instructed by the CPU 101, reads an image signal for one frame every 1/60 seconds, samples and holds and amplifies the image signal, and converts it into a digital signal. I do.

  Further, the video camera 100 has image processing system blocks denoted by reference numerals 120 to 125 in FIG. Under the control of the CPU 101, the image input unit 120 inputs an image signal from the AFE 115 during shooting, and inputs an image signal from the CODEC 130 during reproduction. The image processing unit 121 uses the buffer memory 122 to perform image signal processing such as color conversion, resolution conversion, and image clipping. The focus evaluation unit 123 and the exposure evaluation unit 124 evaluate the focus and exposure states for a plurality of regions of the image (image signal) output from the image processing unit 121, respectively.

  The feature detection unit 125 reads an image from the buffer memory 122 in units of one frame, and detects a specific subject such as a face from the image. As this detection method itself, for example, a known detection method such as the detection method according to Patent Document 1 can be used. In this case, it is assumed that the feature detection unit 125 detects the feature of a specific subject such as a face for the entire region of the image while shifting the detection region in units of 1920 × 1080 pixels.

  The feature detection unit 125 according to the present embodiment further detects the number of specific subjects and the size, position, and inclination as the feature information of the specific subjects. Then, the CPU 101 stores the size, position, and tilt characteristic information of the specific subject detected in each frame in the RAM 102 as a subject list. Note that detection conditions used when the feature detection unit 125 detects feature information of a specific subject are stored in the RAM 102.

  Further, the video camera 100 has recorder blocks shown by reference numerals 130 to 133 in FIG. When recording a captured image, the CODEC 130 encodes the image signal output from the image processing unit 121 and temporarily stores it in the buffer memory 131 under the control of the CPU 101. The disk controller 132 reads an image signal related to the encoding from the buffer memory 131 and records it on a disk 133, that is, a recording medium such as a DVD.

  When playing back a captured image, the disk controller 132 reads an image signal related to encoding from the disk 133 under the control of the CPU 101 and temporarily stores it in the buffer memory 131. The CODEC 130 decodes an image signal related to encoding temporarily stored in the buffer memory 131.

  Further, the video camera 100 has display system blocks indicated by reference numerals 140 to 146 in FIG. Under the control of the CPU 101, the drawing processing unit 140 draws characters and graphics superimposed on the image signal output from the image processing unit 121, and temporarily stores them in the buffer memory 141. The LCD controller 142 reads an image signal from the buffer memory 141 and displays it on the LCD 143. Further, the external output controller 145 reads an image signal from the buffer memory 141 and outputs the image signal to the external output device 151 via the cable 150 connected to the external terminal 146.

  Next, the subject list stored in the RAM 102 will be described. FIG. 2 is a conceptual diagram illustrating a subject list stored in the RAM 102. The CPU 101 stores the position (X, Y), the width W, the inclination A, and the number of lost frames F as feature information of a specific subject such as a face detected by the feature detection unit 125 in the RAM 102 as a subject list. The feature detection unit 125 detects the width W of the specific subject as the size of the specific subject.

  As the specific subject, not only a human face but also various subjects such as human eyes and hands, pets such as cats and dogs are conceivable. For convenience, the following description will be made assuming that the specific subject is a face.

  When the CPU 101 updates the subject list, it is based on the correlation between the size (width W), the position (X, Y), and the inclination A related to the face detected this time and the face already listed. It is determined whether the faces are the same face.

  If the CPU 101 determines that the faces are the same, the CPU 101 updates the size (width W), position (X, Y), and inclination A relating to the faces to those detected this time. On the other hand, if it is determined that they are not the same face, the CPU 101 newly registers the face detected this time in the subject list.

  The number of lost frames F is information indicating how many consecutive frames the face has not been detected. When the face information is not updated, the CPU 101 sets the number of lost frames F to “1”. Increment. Further, when the number of lost frames F becomes “3”, that is, when the face is not detected for three consecutive frames, the CPU 101 deletes the feature information related to the face from the subject list.

  Next, detection mode information when the feature detection unit 125 detects a face will be described. This detection mode information is stored in the RAM 102, and is roughly classified into priority information (priority mode information) and image information (image mode information) (see FIG. 6). The priority mode information includes a size priority mode (first detection mode) and a tilt priority mode (second detection mode). The size priority mode as the first detection mode is a mode that prioritizes the face size when detecting the face, and the tilt priority mode as the second detection mode prioritizes the face inclination. In this mode, the face is detected.

  Image information (image mode information) includes a normal image mode and an enlarged image mode. In the normal image mode, face detection processing is performed using a normal captured image that has not been subjected to enlargement processing or reduction processing. In the magnified image mode, face detection processing is performed using a magnified image extracted from a normal captured image and enlarged to the same size as the normal captured image.

  The CPU 101 determines whether or not the priority mode information should be updated every time the feature detection unit 125 completes the face detection process for the frame.

  Next, priority mode information selection processing will be described based on the flowchart of FIG. This priority mode information selection process is performed for each frame.

  The CPU 101 determines whether the currently set priority mode information is the size priority mode or the tilt priority mode (S401). If the currently set priority mode information is the size priority mode, the CPU 101 determines whether or not a specific subject (face) has been detected by the feature detection unit 125 in the current frame (S402).

  Here, the CPU 101 functions as a counter using the variable Count. This variable Count is a variable indicating how many frames the face is not detected continuously. That is, if no face is detected in the current frame, the CPU 101 increments (increases) the variable Count by “1” (S403). On the other hand, if a face is detected, the CPU 101 clears the variable Count to “0” (S404). In other words, this variable Count can be used to count frames for which no face could be detected continuously.

  When the currently set priority mode information is the tilt priority mode, the CPU 101 determines whether or not a specific subject (face) has been detected by the feature detection unit 125 in the current frame (S405). When the face is not detected, the CPU 101 increments the variable Count by “1” (S407).

  On the other hand, when a face is detected, the CPU 101 determines whether or not the degree of inclination (inclination angle) of the face is within 45 degrees (S406). As a result, if the inclination of the face is within 45 degrees, the CPU 101 increments the variable Count by “1” (S407). When the inclination of the face is not within 45 degrees and the face is greatly inclined, the CPU 101 clears the variable Count to “0” (S408).

  After performing the update processing of the variable Count in S403 or S404, the CPU 101 determines whether or not the variable Count is “30” or more (S409). If the variable Count is less than “30”, the CPU 101 sets the size priority mode, which is the same priority mode information as that set for the current frame, for the next frame (S411).

  On the other hand, if the variable Count is “30” or more, the CPU 101 sets an inclination priority mode, which is priority mode information different from that set for the current frame, for the next frame (S412).

  Further, after performing the update processing of the variable Count in S407 or S408, the CPU 101 determines whether or not the variable Count is “30” or more (a specified number or more) (S410). If the variable Count is less than the prescribed number “30”, the CPU 101 sets the tilt priority mode, which is the same priority mode information as that set for the current frame, for the next frame (S412). .

  On the other hand, if the variable Count is equal to or greater than the prescribed number “30”, the CPU 101 sets a size priority mode, which is priority mode information different from that set for the current frame, for the next frame ( S411).

  As described above, in this embodiment, even if a face is detected in the tilt priority mode, if the face tilt is within 45 degrees (S406, S406), the same as when no face is detected. The variable Count is incremented (S407).

  This is because a face inclination of 45 degrees or less is likely to be able to detect a face even in the size priority mode, so that it is possible to make an early transition to the size priority mode in which a smaller size face can be detected.

  Next, the image mode information used by the feature detection unit 125 for face detection will be described in detail. When the normal image mode is set, the feature detection unit 125 detects a face using the captured image itself, that is, a captured image (normal image) that has not been subjected to enlargement / reduction processing.

  When the enlarged image mode is set, the feature detecting unit 125 cuts out the central 3/4 region of the captured image by the image processing unit 121 and enlarges the same size as the normal image to generate the enlarged image. The face is detected using. Both the normal image and the enlarged image are stored in the buffer memory 122 for each frame by the image processing unit 121.

  The CPU 101 sets the image mode information alternately between the normal image mode and the enlarged image mode for each frame, unlike the priority mode information for switching according to the face detection result. Then, the feature detection unit 125 reads the normal image or the enlarged image related to the set normal image mode or the enlarged image mode from the buffer memory 122, and detects the face from the read image.

  Next, the detection conditions shown in FIG. 6 will be described. This detection condition is set by the CPU 101 according to the combination result of the priority mode information and the image mode information. The CPU 101 can set a minimum face size (minimum size) and a tilt range as detection conditions when the feature detection unit 125 detects a specific subject (face).

  The minimum size can be set in units of 200 pixels, 250 pixels, and 300 pixels, and the inclination range can be set to 45 degrees and 135 degrees centering on the upward vertical direction (see FIG. 3). The smaller the minimum size and the wider the tilt range, the longer the face detection process. Further, the minimum size and the range of inclination can be changed for each frame.

  Here, in this embodiment, since it is assumed that processing for one frame is performed every 1/60 seconds, the detection condition for detecting a face in each frame is that the detection processing is 1 / The condition must be completed within 60 seconds (within the specified time).

  FIG. 3 shows the relationship between the set values of the minimum size and the range of inclination and the time required for the face detection process. When it is desired to detect the size preferentially, the detection condition is set to a minimum size of 200 pixels and an inclination range of 45 degrees (minimum size condition: first detection condition). When it is desired to detect the inclination with priority, the detection condition is set to a minimum size of 300 pixels and the inclination range is set to 135 degrees (maximum inclination condition: second detection condition). The detailed meaning and contents shown in FIG. 3 will be described later with a specific example.

  Next, face detection processing will be described based on the flowchart of FIG. The feature detection unit 125 sets the initial value of the detection mode information (the priority mode information is the size priority mode, the image mode information is the normal image mode, and the detection condition is the minimum size condition) for the first frame of the captured image ( S501). Then, the feature detection unit 125 performs face detection processing under the detection condition (see FIG. 6) corresponding to the initial value of each information set in S501 (S502). In this case, the CPU 101 updates the subject list based on the detection result by the feature detection unit 125.

  Next, the CPU 101 performs priority mode information selection processing in FIG. 4 for face detection processing in the next frame (S503). Similarly, the CPU 101 selects image mode information so that the normal image and the enlarged image are alternately switched for the face detection process in the next frame (S504). Further, the CPU 101 selects a face detection condition for the next frame based on the priority mode information and the image mode information selected in S503 and S504 (S505).

  In this case, as shown in FIG. 6, the CPU 101 sets the detection condition to the minimum size when the image information (image mode information) is a normal image even if the priority information (priority mode information) is the tilt priority mode. The condition (that is, the detection condition with size priority). Only when the priority information is the tilt priority mode and the image information is an enlarged image, the detection condition is set to the maximum tilt condition (that is, the tilt priority detection condition).

  This is because the subject is tilted sideways with respect to the video camera 100 or the tilt of the subject compared to the probability that a subject with a small size will enter the screen during movie shooting, or the subject will become smaller due to the zoom operation. This is because the probability of changing is considered to be low. Therefore, even in the tilt priority mode, the maximum tilt condition is set only for one of the two frames, and the other frame is set to the minimum size condition so that a small subject appears as much as possible. It is configured so that it can be detected.

  Next, the operation when a subject is photographed with the video camera 100 will be described using a specific example. Here, as the initial value of the face detection mode information, the size priority mode is set as the priority mode information, the normal image mode is set as the image mode information, and the minimum size condition (minimum size 200 pixels, inclination 45 degrees) is set as the detection condition. Shall.

  It is assumed that a variable Count indicating the number of consecutive frames in which a face (specific subject) cannot be detected is “0”. Further, it is assumed that the face is stationary and the same image is captured for 1 second (period of 60 frames) from the start of capturing.

  7 and 8 respectively show the normal image and the enlarged image stored in the buffer memory 122 for the first frame F0 of the captured image. A person 710 is shown in the normal image 700 of FIG. 7 and the enlarged image 800 of FIG. The image size of the normal image 700 and the enlarged image 800 is 1920 × 1080 pixels, the width of the face portion of the person 710 is 450 pixels and 600 pixels, and the inclination is 80 degrees.

  The feature detection unit 125 performs face detection processing from the normal image 700 based on the previously set detection mode information, and the CPU 101 updates the subject list based on the face detection result by the feature detection unit 125. To do.

  In this case, since the size priority mode is set, the feature detection unit 125 can detect the face only when the inclination of the face is within 45 degrees and cannot detect the face of the person 710. Therefore, the CPU 101 sets “1” to the variable Count.

  Next, the CPU 101 updates the detection mode information. In this case, since the variable Count is “1”, the CPU 101 maintains the size priority mode for the priority mode information, and changes the image mode information to the enlarged image mode because the image mode information was the normal image mode before the update. Since the CPU 101 maintains the size priority mode as the priority mode information, the minimum size condition is set as the detection condition.

  The image stored in the buffer memory 122 30 frames after the frame F0 remains the same as the frame F0 shown in FIGS. 7 and 8 because the person 710 is still. Therefore, the feature detection unit 125 updates the subject list and increments the variable Count in a state where the face of the person 710 cannot be detected in the frames after 30 frames.

  In this case, as a result of the CPU 101 updating the subject list and incrementing the variable Count for each frame, the variable Count becomes “30”. Therefore, the CPU 101 changes the priority mode information (size priority mode) set so far to the tilt priority mode (see FIG. 4). Further, the CPU 101 changes the image mode information to an enlarged image because the image mode information was a normal image before the update.

  Furthermore, the CPU 101 changes the priority mode information to the tilt priority mode, and changes the image mode information to the enlarged image mode, so that the detection condition is the maximum tilt condition (minimum size 300 pixels, tilt 135 degrees). Set.

  This is because the face can not be detected even though the minimum detection size is set to the smallest 200 pixels, and the face can be detected by the control when the minimum inclination size is 200 pixels. This is because it may be larger than 45 degrees.

  However, if the detectable face angle is simply increased, the time required for the face detection process exceeds 1/60 seconds as shown in FIG. In this case, particularly when shooting a moving image, even if a face is detected, the face to be detected moves to another position at the time of actual shooting with focus control and exposure control suitable for the face. Therefore, the face detection process is wasted. For this reason, the time required for face detection needs to be within 1/60 seconds (within a specified time).

  Therefore, the CPU 101 sets a maximum inclination condition (135 degrees) so that the face can be detected, and sets the minimum detection size from 200 pixels to 300 in order to detect the face within a specified time within 1/60 seconds. Make a pixel.

  The image stored in the buffer memory 122 after 31 frames from the frame F0 remains the same as the frame F0 shown in FIGS. 7 and 8 because the person 710 is still. However, since the maximum inclination condition (135 degrees) is set, the feature detection unit 125 can detect the face of the person 710 from the enlarged image 800 related to the frame after 31 frames. Therefore, the CPU 101 clears the variable Count to “0”.

  Next, the CPU 101 updates the detection mode information. In this case, since the variable Count is “0” in the priority mode information, the CPU 101 maintains the inclination priority mode in the priority mode information, and the image mode information is updated to a normal image because the image was updated before the update. Further, the CPU 101 changes the priority mode information to the tilt priority mode, but the image mode information is the normal image mode, so the detection condition is changed to the minimum size condition (minimum size 200 pixels, tilt 45 degrees). As a result, the face of the person 710 cannot be detected in this frame, and the variable Count is incremented. However, in the next frame, the mode is changed to the enlarged image mode again and the maximum inclination condition is set, so the variable Count is cleared. “0”.

  After that, it is assumed that the person 710 rises from the states of FIGS. 7 and 8 and then remains stationary for 1 second (60 frames), and the same image is captured. FIGS. 9 and 10 show the normal image 900 and the enlarged image 1000 stored in the buffer memory 122 in the frame F1 of the captured image at the moment when the person 710 lying in FIGS. 7 and 8 wakes up. .

  The image size of the normal image 900 and the enlarged image 1000 in FIGS. 9 and 10 is 1920 × 1080 pixels as in FIGS. 7 and 8, and the width of the face portion of the person 710 is as in FIGS. 7 and 8. Similarly, it is assumed that the pixels are 450 pixels and 600 pixels. However, the inclination of the face is 10 degrees in all cases, and changes from 80 degrees in FIGS. 7 and 8.

  Further, the variable Count is “0”, and as the priority mode information, the tilt priority mode is subsequently set, and the image mode information is an enlarged image. Further, it is assumed that the maximum inclination condition is set as the detection condition.

  The feature detection unit 125 performs face detection processing from the enlarged image 1000 based on the detection mode information, and the CPU 101 updates the subject list based on the face detection result by the feature detection unit 125. In this case, the person 710 is detected, but since the inclination is within 45 degrees, the CPU 101 increments the variable Count.

  If the variable Count reaches “30” in the state where the inclination priority mode is set, the CPU 101 changes from the inclination priority mode to the size priority mode.

  Next, the CPU 101 updates the detection mode information. In the priority mode information, since the variable Count is “1”, the inclination priority mode is maintained, and the image mode information is switched to the normal image because the image before update is an enlarged image. The CPU 101 sets the minimum size condition as the detection condition because the priority mode information is the tilt priority mode but the image mode information is a normal image.

  The image stored in the buffer memory 122 30 frames after the frame F1 is the same image as the frame F1 shown in FIGS. 9 and 10 because the person 710 is still. The feature detection unit 125 detects a face from the normal image 900 based on the image mode information. The CPU 101 updates the subject list based on the face detection result by the feature detection unit 125.

  The face of the person 710 is also detected in the detection process in the frame F1, but the CPU 101 increments the variable Count because the inclination of the face is within 45 degrees. As a result, the variable Count is “30”. Therefore, the CPU 101 switches the priority mode information from the tilt priority mode to the size priority mode. For the image mode information, the CPU 101 switches to the enlarged image because the image before the update was a normal image. Further, since the CPU 101 switches the priority mode information to the size priority mode, the detection condition maintains the minimum size condition.

  As described above, the CPU 101 sets so that switching can be performed for each frame in which priority is given to detecting a small face within a range in which the feature detection unit 125 can detect a face within 1/60 seconds.

  By such detection condition switching processing, it is possible to detect a specific subject (face) continuously within a specified time and to detect with high accuracy even when the specific subject is small or inclined. . Furthermore, it is possible to properly focus on a desired specific subject and take a high-quality image with an appropriate exposure state.

  The present invention is not limited to the above-described embodiment. For example, the timing for alternately switching between the size priority mode and the tilt priority mode is not limited to the case where 30 frames in which a specific subject is not detected are consecutive, The number of frames may be changed.

  The object of the present invention can also be achieved by supplying a storage medium storing a program code of software programmed to realize the functions of the above-described embodiments to a system or apparatus. In this case, the computer (or CPU, MPU, etc.) 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 program code and the storage medium storing the program code constitute the present invention.

  Further, as a storage medium for supplying the program code, for example, a floppy (registered trademark) disk, a hard disk, a magneto-optical disk, a CD-ROM, a CD-R, a CD-RW, a magnetic tape, a nonvolatile memory card, A ROM or the like can be used. Furthermore, optical disks such as DVD-ROM, DVD-RAM, DVD-RW, DVD + RW, etc. can also be used as a storage medium for supplying program codes. Alternatively, the program code may be downloaded via a network.

  Further, based on the instruction of the program code read by the computer, an OS (operating system) 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. This is also included.

  Furthermore, the case where the program code read from the storage medium is written in a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer is also included. In addition, based on the instruction of the program code, a CPU or the like provided with the extension function in the extension board or the extension unit may perform part or all of the actual processing, and the function of the above-described embodiment may be realized by the processing. included.

It is a block diagram of the imaging device which concerns on the Example of this invention. It is the figure which illustrated the contents of the subject list. It is a figure which shows the relationship between detection conditions (minimum size, the range of inclination) of specific subjects (such as a face) and detection processing time. It is a flowchart which shows the selection switching process of the size priority mode and inclination priority mode as face detection conditions. It is a flowchart which shows a face detection process. It is a figure which shows the relationship between each detection mode information of a face. It is the figure which illustrated the picked-up image (normal image) which has not been enlarged / reduced. It is a figure which shows the enlarged image which extracted and expanded a part of normal image of FIG. It is a figure which shows the normal image at the time of image | photographing, after the specific subject of FIG. 7 moves. It is a figure which shows the enlarged image which extracted a part of normal image of FIG. 9, and expanded.

Explanation of symbols

100 ... Imaging device 101 ... CPU
102 ... RAM
103 ... ROM
114: CCD image sensor 121 ... Image processing unit 122 ... Buffer memory 125 ... Feature detection unit

Claims (8)

Detecting means for detecting a specific subject from a photographed image photographed by the photographing means;
Setting means for setting detection conditions for the detection means to detect the specific subject;
Have
The setting means switches between a first detection mode and a second detection mode within a range in which the detection means can detect the specific subject within a specified time. The first detection mode is the second detection mode. The second detection mode is a detection mode in which priority is given to detecting the specific subject having a smaller size compared to the first detection mode, and the second detection mode is for the specific subject in an inclined state as compared to the first detection mode. An imaging apparatus characterized by being in a detection mode in which priority is given to detection.   The setting means is unable to detect the specific subject even when the detection means performs the detection processing of the specific subject over a predetermined number of frames in a state where the first detection mode is set. In addition, even if the second detection mode is switched and the detection means performs the detection process of the specific subject over a predetermined number of frames in the state where the second detection mode is set, the specific subject The imaging apparatus according to claim 1, wherein the image pickup device is switched to the first detection mode when the image cannot be detected.   The setting unit detects the specific subject by alternately using the photographed image and an enlarged image obtained by extracting a part from the photographed image and enlarging it to the same size as the photographed image. The imaging device according to claim 1 or 2.   In the state in which the first detection mode is set, the setting unit sets a first detection condition that prioritizes detection of the specific subject having a small size, and the second detection mode is set. The imaging apparatus according to claim 3, wherein the imaging apparatus is configured to switch between the first detection condition and the second detection condition that prioritizes the detection of the specific subject in an inclined state. The setting means sets the second detection condition only when the detection means detects the specific subject using the enlarged image in a state where the second detection mode is set. The imaging apparatus according to claim 4, wherein the imaging apparatus is characterized. Counter means for increasing the variable when the detection means cannot detect the specific subject;
Even when the specific object is detected by the detection means in a state where the second detection mode is set, the counter means detects the inclination of the specific object related to the detection in the first detection mode. The imaging apparatus according to claim 2, wherein when the inclination is such that the specific subject can be detected even under the mode, the variable is increased according to a frame in which the specific subject is detected. A detection step of detecting a specific subject from a photographed image photographed by the photographing means;
A setting step for setting a detection condition for detecting the specific subject by the detection step;
Have
The setting step switches between a first detection mode and a second detection mode within a range in which the specific subject can be detected within a specified time in the detection step. The first detection mode is the second detection mode. The second detection mode is a detection mode in which priority is given to detecting the specific subject having a smaller size compared to the first detection mode, and the second detection mode is for the specific subject in an inclined state as compared to the first detection mode. A control method for an imaging apparatus, characterized in that the detection mode prioritizes detection.   A program that is programmed so that a computer can execute the control method of the imaging apparatus according to claim 7.

Images