JPH07154666A - Video camera - Google Patents

Abstract

PURPOSE:To attain tracking image pickup for an object by applying rotation control of the video camera in tracking with a motion of the object. CONSTITUTION:The video camera consists of a thresholding means 5 for a video signal of an object as to object areas 5C, 6A, 7C, 7E and other areas, a fill processing means 7 applying fill processing to a binary picture signal relating to the object areas 5C, 6A, 7C, 7E, a data calculation means 8 outputting respectively an area information signal and a gravity center position information signal of the object areas 5C, 6A, 7C, 7E based on an output result of the fill processing means 7, a drive control means 9 outputting a control signal based on the calculation result, and a drive means 10 implementing panning/tilt control so that a picture of the object is always displayed on an image pickup screen based on a control signal.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a camera-integrated VTR (so-called video camera) which performs pan / tilt control according to the movement of a subject so that an image of the subject is always displayed on an image pickup screen.

[0002]

2. Description of the Related Art FIG. 7 is a block diagram of a conventional video camera. Hereinafter, with reference to the drawings, the related art (1),
(2) will be described. (1) When a subject is imaged by a video camera, for example,
There was a demand for a video camera that tracks and shoots a child on a carousel in an amusement park. As such a video camera, the applicant first proposed a video camera for automatically tracking an object (Japanese Patent Application No. 4-
103802).

Such a video camera divides a captured image of a subject into a plurality of regions on a screen, obtains a region in which a specific color signal (color information representative of the subject) is most included in this divided region, and moves in the direction of this region. The subject was automatically tracked by rotating the video camera.

That is, as shown in FIG. 7, a line-sequential video signal obtained by a CCD 1 for photoelectrically converting an optical image of a subject (not shown) is supplied to a signal processing section 2, and this signal is generated by a known conversion means. It was supplied to a recording system (not shown) as a composite video signal. Further, the composite video signal is supplied to the signal detection unit 12 and the color setting unit 13, respectively, and the color setting unit 13 stores a specific color signal by a color signal related to the video signal according to an instruction signal from an operation unit (not shown). Then
The signal detection unit 12 compares the color signal related to the video signal with the specific color signal stored by the color setting unit 13 and supplies the comparison result to the division / comparison unit 14. Then, the dividing / comparing unit 14 divides the video signal input from the signal detecting unit 12 into left and right regions on the screen (dividing region), and the specific color signal detected by the signal detecting unit 12 is in this divided region. Compare which one contains more,
The comparison result is supplied to the drive control unit 15. The drive control unit 15 outputs a control signal from the input comparison result to the drive unit 1.
Based on this control signal, the driving unit 16 supplies the video camera 6 with the video camera so that the optical axis of the optical lens of the video camera is directed in the direction of the divided area containing more specific color signals. An instruction signal for rotationally driving the platform is supplied to the drive motor 11.

In this video camera, for example, a child riding on a merry-go-round in an amusement park is used as a subject, and a region in which the specific color signal of this subject is most included in the above-mentioned divided regions is obtained, and the direction of the video camera By rotating the pan / tilt head with the video camera so that the optical axis of the optical lens is directed, automatic tracking was always performed so that this subject was always photographed. (2) Further, the following has been proposed as a method for automatically tracking a subject. That is, 1) When a binarized image obtained by picking up a moving object as a subject and binarizing the same is small, enlargement processing is performed to cope with a background change, and a luminance histogram of the subject image is generated. By this, the subject is distinguished from the background, and an automatic tracking device for a moving object that prevents disturbance due to noise in the background of the subject (Japanese Patent Laid-Open No. 2-1.
No. 23877), 2) Obtain a representative color from the color density level of the subject within the search range of the subject, find a position where the number of pixels of the representative color is maximum, and set this as the subject.
An automatic tracking device for automatically tracking this (Japanese Patent Laid-Open No. 4-2829).
89), 3) For example, a person wearing a helmet of either R: red, G: green, B: blue is designated as a subject,
There has been proposed an automatic tracking video photographing device (Japanese Patent Laid-Open No. 4-354490) that automatically tracks a subject by determining in which area on the image pickup screen the color exists.

[0006]

However, in the above-described conventional technique, for example, when a child riding on a merry-go-round is used as a subject for tracking shooting, another child wearing clothes of a color similar to the subject's child takes an image. When it appears on the screen,
There was a risk that the child would be tracked and photographed by mistake. Further, even if a color (subject color information) specified by the user to identify the subject is present on the same screen as the subject, a position of the subject is detected if a plurality of colors have similar colors to the subject color information. However, there are drawbacks such as being unable to identify the subject, and performing an erroneous operation of identifying an object other than the subject having a similar color as the subject. Furthermore, this object is approximately the same size as the subject and moves so as to overlap the subject,
There is a defect that the object and this object cannot be distinguished and the object is not tracked. Furthermore, for example, in which area on the image pickup screen the color of the hat is present by covering the child as a subject with a hat of R; red, G; green, B; blue, etc. However, there is a drawback that it is troublesome because it is necessary to specify in advance the color of the subject clearly distinguished from the color of the object other than the subject.

[0007]

In order to solve the above problems, the present invention provides the following configurations.

Binarization processing means for binarizing a video signal of a subject into a subject area and an area other than the subject area, a fill processing means for performing a fill process on a binary image signal relating to the subject area, and the fill Data calculation means for respectively outputting the area information signal and the barycentric position information signal of the subject area in the fill processed image signal output from the processing means, the area information signal and the barycentric position information signal, and the color information signal relating to the subject area. The color information signal and the center-of-gravity position information signal that are held are output to the binarization processing unit and the fill processing unit, respectively, and are updated and held based on the initially-set center-of-gravity position information signal. A drive control unit that outputs a control signal based on a difference from the center-of-gravity position information signal, and an image of the subject is constantly captured on the screen based on the control signal. Video camera and having a driving means for performing pan and tilt control so reflect.

[0009]

1 is a block diagram for explaining an embodiment of a video camera according to the present invention, FIG. 2 is a diagram for explaining the operation of a binarization processing means, and FIG. 3 is an operation for a logical filter. FIGS. 4 and 5 are views for explaining the operation of the fill processing means, respectively, and FIG. 6 is a view for explaining the operation of the fill processing means after the initial setting state.
An embodiment of a video camera according to the present invention will be described below with reference to the drawings. As shown in FIG. 1, the video signal processing system of the video camera according to the present invention includes a CCD 1, a signal processing unit 2, an A / D conversion unit 3, a preprocessing unit 4, a binarization processing unit 5, and a logical filter 6. , A fill processing unit 7, a data calculation unit 8, a drive control unit 9, a drive unit 10, a drive motor 11, and a video signal recording system (not shown). Although not described in detail here, it goes without saying that there is an audio signal processing system in parallel with this video signal processing system.

The drive motor 11 transmits the power for driving the pan / tilt head (not shown) on which the video camera is installed to rotate in the up / down / left / right directions. By this rotary drive, the video camera installed on the pan / head is constructed. The direction of the optical axis that is incident on an optical system (focus lens, zoom lens, etc.) not shown in the figure is controlled to be vertically and horizontally (so-called pan-tilt) control, and the incident optical axis of this optical system is adjusted to the direction in which the subject is present. is there.

The CCD 1 outputs a line-sequential video signal obtained by photoelectrically converting the optical image incident through the optical system to the signal processing unit 2 in the next stage.

The signal processing unit 2 described above processes the line-sequential video signal output from the CCD 1 by subjecting the composite video signal to predetermined processing, and the A / D conversion unit 3 at the next stage and a well-known V signal.
The signals are output to video signal recording systems that are substantially the same as the TR recording system.

The A / D converter 3 described above outputs a digital composite video signal (hereinafter referred to as a digital video signal) obtained by A / D converting the analog composite video signal output from the signal processor 2. Output to the pre-processing unit 4.

The above-mentioned pre-processing unit 4 performs the binarized image processing to be described later on the digital video signal output from the A / D conversion unit 3. It outputs to the digitization processing unit 5. That is, the pre-processing unit 4 has a low-pass filter and a sub-sampling circuit (neither of which is shown), and an unnecessary high level of the digital video signal output from the A / D conversion unit 3 is generated by the low-pass filter. Pre-processing of removing the band component and compressing the digital video signal output from the low-pass filter in the preceding stage into a predetermined amount of information in this sub-sampling circuit is performed.

The binarization processing unit 5 described above is an image region to be a subject (hereinafter referred to as a subject region. Here, the subject region does not refer to an entire subject to be imaged, but to a point specified by a user to be described later. For example, a region is recognized as a subject itself for convenience.For example, when the subject is a person wearing a red shirt and wearing jeans, when the user specifies the region of the red shirt, binarization corresponding to this red shirt is performed. The image region of the signal is the subject region, and the point (pixel) designated by the user in the region 5C shown in FIG.
The digital video signal output from the above-mentioned preprocessing unit 4 is converted into the digital image signal based on the color information (color information including hue, saturation, lightness, etc.) constituting point 5A in FIG. The binarized signal obtained by dividing the image area including the color information and the image area not including the color information into two (binarization) is output to the logical filter 6 in the next stage.

The logical filter 6 described above is equivalent to the above-mentioned 2
A large number of small regions (a large number of regions 5E such as isolated points and thin lines in FIG. 2) existing in the vicinity of the subject region 5C in the binarized signal output from the binarization processing unit 5 are removed, and the subject region 5 is removed.
Image processing for shaping (filtering) C is performed, and filter processing for clarifying an image area (subject candidate area) having specified color information including the subject area 5C is performed (filtering the image illustrated in FIG. 2). To obtain the two regions 6A shown in FIG. 3). In this way, the shaped binarized signal subjected to this filter processing is supplied to the fill processing unit 7 in the next stage.

The above-mentioned fill processing section 7 has an area corresponding to the object area 5C in the shaped binarized signal outputted from the above-mentioned logical filter 6 (area 6 on the left side in FIG. 3).
As a result of performing image processing (hereinafter referred to as fill processing) for filling the whole of A) using the color information specified by the user, the area 7D (illustrated in FIG. 5) that has been subjected to the fill processing with the specified color information is performed. The binarized signal including the data is supplied to the data calculation unit 8 in the next stage.

The above-mentioned data calculating section 8 calculates the total number of pixels of the partial area having the above-mentioned color information within the area 7C for the area 7C which has been subjected to the fill processing in the binarized signal output from the fill processing section 7. Area information is obtained by integrating (counting) the coordinates, and the coordinates (X, Y) including the entire area 7C are virtually set, and the X value and the Y value at which the above-described number of pixels becomes maximum are respectively calculated. Then, the center of gravity information is obtained, which is the center of gravity of the total number of pixels (that is, the center of gravity of the area information), and the obtained area information signal regarding the area information, the center of gravity position information signal regarding the center of gravity information, and the color information are obtained. The color information signal is supplied to the drive control unit 9 in the next stage. The value of the center-of-gravity position information signal supplied to the drive control unit 9 in the "initial setting state" described later is stored as a reference value, and the center-of-gravity position information sequentially supplied for each frame period after the initial setting. As the signal value, a signal corresponding to the difference value obtained by comparing with the reference value is supplied to the drive unit 10 of the next stage as a control signal for controlling the rotary drive of the platform. As a result, when this difference value becomes zero, this control signal is not output. In addition, a “gravity state (after initial setting)”, which will be described later, and a barycentric position information signal and a color information signal that are sequentially supplied for each frame period are stored (updated) in memory. Then, the center-of-gravity position information signal and the color information signal of the immediately preceding one frame period (previous frame screen) are sequentially supplied to the fill processing unit 7 and the binarization processing unit 5, respectively. As a result, the fill processing unit 7 and the binarization processing unit 5 output the barycentric position information signal and the color information signal in the previous frame screen and the barycentric position information signal and the color information signal in the current one frame period (current frame screen). If the comparison result is within a range having a correlation with that of the previous frame screen, the current frame screen is determined to be substantially the same as the previous frame screen, and the subject image is continuously recognized.

The drive unit 10 supplies an instruction signal to the drive motor 11 based on the control signal output from the drive control unit 9, and the drive motor 11 rotates at the rotation angle and the rotation based on the instruction signal. Rotate to get the number.
As a result, the power based on the control signal can be transmitted to the platform with the video camera installed.
It is possible to rotate the platform in accordance with this. That is, it becomes possible to pan and tilt the video camera installed on this platform.

Now, regarding the operation of the video camera having the above-mentioned configuration, the "initial setting state" in which the position where the subject image is photographed is initially initialized, and then the subject image moves and the position where this image appears The operation in the "imaging state" in which the tracking imaging (automatic tracking control) is performed so that the object is always imaged as the image moves when compared with the position of the initially set object image will be described in detail below.

1. [Initial setting state] In this initial setting state, the subject image is distinguished into the subject region and the other regions based on the color information of the subject image photographed in the approximate center of the imaging screen, and the area information and the center of gravity of the subject region are displayed. The position information is calculated and each initial value is set, and the color information is held.

First, the subject is set and photographed in the center on the monitor screen by a viewfinder (not shown) or the subject image is designated by using a cursor or the like on the monitor screen. At this time, it is performed on the premise that the focus setting of the subject is completed by a well-known auto focus setting (so-called auto focus) mechanism not described in detail here, and the video camera according to the present invention performs the auto focus. It goes without saying that it has a configuration for performing. Further, it is performed on the premise that the well-known zoom control (which sets the zoom position (focal magnification) of the zoom lens) mechanism, which is not described in detail here, completes the zoom control for photographing the subject in a desired size. It is a thing. Further, for convenience of description, it is assumed that the video camera according to the present invention always captures an object on the monitor screen in a size that can be recognized.

Now, in this initial setting state, the video signal is processed according to the following procedures (1) to (3). That is, (1) extraction of color information and binarization based on color information (2) shaping of binarized information (3) collection of subject information (1) extraction of color information and binarization based on color information The digital video signal output from the pre-processing unit 4 is supplied to the binarization processing unit 5 and the color information of the point designated by the user is extracted in the image area based on the digital video signal. Be seen. This color information is the phase (hue) of the color signal included in the digital video signal,
This is a collection of data relating to each information of the amplitude (saturation) of the color signal, the amplitude (brightness) of the luminance signal, and the like, and is data necessary for performing binarization and fill processing described later.

When the color information is extracted, all the image areas based on the digital video signal based on the color information are extracted, and the area where the color information is present and the area where the other color information is present. And is divided into two (binarization). That is, as shown in FIG. 2, any one point 5A (or a plurality of points) in the area 5B where the subject designated by the user exists.
Is specified and the color detected at that point (hue, saturation,
(Brightness etc.) is used as the color information of the subject. Then, a binarized signal obtained by dividing all image areas into an area 5C having the color information and an area 5D having the other color information (for example, the area 5C specified as a subject is set to "1", and The area 5D is set to "0"). After this binarization, this color information is used as information for identifying the subject in the “imaging state” described later.

Although the user specifies the subject image by using the cursor or the like, the invention is not limited to this.
For example, the point where the diagonal lines of the area 5B described above intersect may be set as a point for extracting color information, and the image at this point may be automatically set as a subject image. (2) Shaping of Binarized Information As shown in FIG. 2, the image of the binarized signal has a plurality of isolated points caused by signal noise or the like in the vicinity of the region 5C.
A large number of areas 5E such as thin lines are generated. Since this area 5E must be subjected to complicated signal processing in later-described fill processing and data processing, this area 5E is removed,
The area 5C needs to be shaped (filtered), and this shaping is performed by the logical filter 6. The logical filter 6 searches the above-mentioned "1" area, and removes only the detected "1" isolated points. The isolated point is detected because the area around the "1" area is all "0" areas. In addition, since the thin line is connected to the area of "1" in one direction in the area around the area of "1" (upper, lower, left, and right), the direction in which the connected area of "1" is continuously detected. Then, the detected "1" area is sequentially removed to remove the thin line "1". Thus, as shown in FIG. 3, the logical filter 6 removes a large number of small areas existing in the vicinity of the area 5C having the color information of the subject, and shapes a relatively large area of "1",
Shaping 2 having large and small subject candidate regions 6A shown in FIG.
Output a digitized signal. The plurality of subject candidate regions 6A obtained at this time are large regions having subject color information (information of "1"). Therefore, as described above, this shaping is performed to identify the area corresponding to the object area 5C existing in the area 5B designated by the user by using the cursor or the like among the plurality of object candidate areas 6A obtained here. The binarized signal is supplied to the fill processing unit 7, and the fill processing described below is performed based on the color information of any one point 5A in the area 5B previously designated by the user. (3) Collection of subject information Here, in order to identify a region corresponding to the subject region 5C from the plurality of subject candidate regions 6A described above, based on the shaped binarized signal supplied from the logical filter unit 6, the user first The fill processing is performed only on the subject candidate area 6A corresponding to the subject area 5C according to the color information of any one point 5A within the designated area 5C, and the binarized signal output from the fill processing unit 7 in this way is data calculated. The area information and the center-of-gravity position information of the subject candidate area, which has been subjected to the fill processing and corresponds to the subject area 5C, are calculated by the unit 8. After that, by supplying the color information and the center-of-gravity position information of this area to the drive control unit 9 as the subject information, it is possible to perform the above-described automatic tracking control of the subject in the “imaging state” described later.

Now, the operation of the fill processing section 7 will be described in detail with reference to FIGS. 4 and 5. Here, the filling process means an area corresponding to the object candidate area 5C (a left area 6 in FIG. 4) of the two object candidate areas 6A shown in FIG.
This is a process of replacing all the areas including the starting point 7A described later in A) with the above-mentioned aggregate area containing only the color information (not to mention, there are many color information different from the above-mentioned color information in this area). Existing). This starting point 7A
May be located anywhere in the area 6A, or may be any one point (or a plurality of points) 5A in the area 5B previously designated by the user.

That is, first, the fill processing is performed from the starting point (pixel) 7A to the left to the left boundary of the area 6A, and then to the right to the right boundary of the area 6A. Next, the filling process is sequentially performed on the pixel rows above and below the column on which the filling process has been performed up to the left and right boundary lines.
When all the pixels in A have been scanned, this filling process ends. In this way, as shown in FIG. 5, an area for which the fill processing has been performed, that is, a subject area 7C is obtained.

The binarized signal that has been subjected to the fill processing in the above-mentioned fill processing section 7 is supplied to the above-mentioned data calculation section 8 where the total number of pixels subjected to the fill processing in the above-mentioned subject area 7C is integrated ( By performing the counting, the area information of the subject area 7C can be obtained. Further, the coordinates (X, Y) in the image area including the entire subject area 7C are virtually set, and the above-mentioned X value and Y at which the number of pixels becomes maximum.
By obtaining the respective values, the coordinate values at the center of gravity 7D of the subject area 7C can be obtained as the center of gravity position information. The area information and the barycentric position information thus obtained are supplied to the drive control unit 9 as an area information signal and a barycentric position information signal, respectively. The center-of-gravity position information signal is sequentially compared with the center-of-gravity position information signal of the subject information obtained in the “imaging state” below as the initial center-of-gravity position information (reference center-of-gravity position information) signal of the subject area 7C in the drive control unit 9. Memorized in. Further, the center-of-gravity position information signal is held as data for predicting the starting point for performing the fill process only on the true subject area in the “imaging state” described later.

2. "Shooting state" Now, "Initial setting state"
The state in which the normal video camera is in the shooting state after the initial setting is performed is referred to as “shooting state”. The operation performed by the video camera at this time will be described. In the "shooting state", the imaged image changes every moment, so the subject information set in the previous "initial setting state" is used as a reference, and the subject is constantly photographed in response to changes in the color and movement of this subject. Follow the shooting as you do. Here, after performing the signal processing on the video signal of the subject according to the above-described procedures (1) to (3) in the “initial setting state”, (4) resetting of subject information, which will be described later, (5) Each procedure of tracking control is sequentially repeated for each frame. (4) Resetting of subject information In this "shooting state", if the subject's color or shape changes momentarily, the color information, area information, and barycentric position information of the subject area (these three pieces of information will be referred to as subject information below) Change. Therefore, the color information and the barycentric position information are sequentially obtained from the subject information, and the color information and the barycentric position information of the current one frame period (current frame screen) and the color information of the immediately preceding one frame period (previous frame screen) are acquired. And the center-of-gravity position information are compared, and if this comparison result is within a range having a correlation with that of the previous frame screen, it is determined that the current frame screen is substantially the same as the previous frame screen, and recognition of the subject image is continued. At the same time, the subject information of the current frame screen is reset as the latest subject information. The video signal picked up in the “shooting state” is subjected to the same processing as “(1) Color information setting and binarization based on color information” described above, and then a subject area 7E (FIG. 6) described later. Shown in FIG.
For convenience, the subject area 7C in the “initial setting state”,
The subject area 7E in the "imaging state" is displayed on the same monitor screen). At this time, based on the color information stored and held in the drive control unit 9 (color information of the previous frame screen including the “initial setting state”), the hue and saturation closest to the range having a correlation with this color information, An area having color information composed of lightness (color information of the current frame screen) is determined. Then, the binarized signal in which the area determined at this time is obtained is supplied to the above-described logical filter 6 and the same processing as the above-mentioned “(2) Shaping of binarized information” is performed to thereby create an isolated point. And the thin lines are filtered and then supplied to the above-described fill processing unit 7,
The same processing as the above-mentioned “(3) Collecting subject information” is performed.

The operation of the fill processing section 7 when the subject information is reset will be described with reference to FIG. A subject area 7E is obtained through the binarization processing unit 5 and the logical filter 6 described above in the figure, and the subject area 7E is obtained.
In the case of performing the filling process of, the center of gravity of the subject area 7E is predicted in order to set the starting point of the filling process (prediction point 7).
F). The predicted point 7F is obtained by a function that satisfies or approximates the locus 7H (dashed line) of the center of gravity position information by sequentially accumulating the center of gravity position information of the subject area 7E obtained for each frame. Then, the filling process described above is performed with the predicted point 7F as a starting point. In addition, since the color information (color information including the hue, the hue, and the brightness forming the prediction point 7F) of the subject area 7E is obtained from the subject area 7E obtained at this time, the previous color information (the previous frame screen Color information of this subject area 7
The subject area 7 that is updated to the color information of E (color information of the current frame screen) and changes in the future via the data calculation unit 8 described above.
In order to predict the movement of E, the drive control unit 9 stores and holds the color information of the current frame screen.

Further, when there is no subject candidate area at the above-mentioned predicted point 7F (for example, when the subject moves rapidly or is large), the center of gravity 7 determined in the initial setting state is set.
It is also possible to obtain coordinate data of a point in the closest subject area 7E from the coordinate data of D (a point in the subject area 7E located at the shortest distance from the center of gravity 7D), and perform the above-described filling process starting from the point 7G. . Further, when the center of gravity 7D of the subject area 7C in the "initial setting state" is included in the current subject area 7E, for example, the subject in the "initial setting state" is compared with the method of obtaining the starting point for performing the filling process from the prediction point 7F. When the movement of the subject hardly changes, such as when the center of gravity 7D of the region 7C is included in the subject region 7E, the center of gravity 7D may be the starting point of the fill process.

The subject region 7E is filled with the thus obtained prediction point (start point) 7F or point 7G,
Similarly, since the area information and the center-of-gravity position information are obtained by the data calculator 8, the area information signal and the center-of-gravity position information signal are sent to the drive controller 9 in order to update these information together with the color information as the latest object information. Supply each. Then, the subject information of the previous frame screen stored in the drive control unit 9 is replaced with this subject information. In the drive control unit 9, the value of the center-of-gravity position information signal of the subject area 7E relative to the value of the center-of-gravity position information signal of the subject area 7C in the "initial setting state" (the value of the reference center-of-gravity position information signal, which is the reference value) ( Current barycentric position value)
Find the difference value between and. A change value of the current center-of-gravity position is obtained from this difference value. For example, when the X-coordinate value of the current center-of-gravity position decreases, the subject moves from the position in the “initial setting state” to the left on the imaging screen. If the X-coordinate value of the current center-of-gravity position increases, it is determined that the subject has moved from the position in the “initial setting state” to the right on the imaging screen, and the Y-coordinate of the current center-of-gravity position If the value decreases, the subject moves vertically from the position in the “initial setting state”, and if the Y coordinate value of the current center of gravity increases, the subject is in the “initial setting state”. It is determined that the position has moved upward from the position on the imaging screen. Then, a change value corresponding to a change in each value of the X and Y coordinates of these barycentric positions is calculated. (5) Tracking control From the change value of the value of the center-of-gravity position information signal calculated in this way, it is possible to find how much the value of the current center-of-gravity position has changed from the reference value. Therefore, the incidence of the video camera corresponding to this change value can be obtained. The amount of movement of the optical axis in the vertical and horizontal directions is obtained, and a control signal based on this is supplied to the drive unit 10. The drive unit 10 outputs an instruction signal for rotationally driving the platform with the rotation angle and the rotation speed based on the control signal.
Supply to. The drive motor 11 drives the pan head to rotate in response to the instruction signal. That is, the pan / tilt of the video camera installed on the platform is performed along with the rotational driving motion. For example, when the position of the subject on the current frame screen moves to the left of the position of the subject on the previous frame screen, the video camera is rotated so that the incident optical axis of the video camera is directed leftward about the rotation axis of the platform. In this way, the incident optical axis of the video camera is directed according to the moving direction of the subject, and the subject is always photographed.

Now, in the above-described fill processing, when an image having no correlation with the previous frame screen is obtained on the current frame screen (for example, a person in blue clothes was recognized as the subject up to the previous frame screen. , When this subject suddenly holds a red doll on the current frame screen, etc.),
Since the color information of the subject and the color information of another object inside the subject are different, a region corresponding to a hole is generated inside the subject region obtained by the above-described binarization processing, and even if the subject does not change, Since the area of this hole obtained by the filling process changes when the shape changes or the doll moves, the calculated center of gravity position changes. Therefore, the position of the center of gravity is erroneously predicted, or a change in the position of the center of gravity is erroneously detected, which may cause a malfunction in tracking shooting. However, it is assumed that the different area described above occurs inside the subject area.

Therefore, when there is a region (different region) having different color information inside the subject region, it is necessary to perform the fill process with priority given to the subject without being affected by the different region. As this method, the above-described filling process is performed stepwise as described below. First, after performing the above-described fill processing on the subject area excluding another area (referred to as fill processing 1),
Fill processing is performed on all areas of the screen other than the subject area (referred to as fill processing 2). At this time, the starting point of the fill processing 2 is, for example, a point in an area other than the subject area located at the shortest distance from the starting point of the fill processing 1. Further, at this point, the filling process is not performed on another region inside the subject region. Then, the area other than the area subjected to the filling processing by the above-described filling processing 2 is subjected to the filling processing (hereinafter referred to as filling processing 3). In this fill process 3, the subject region is specified by performing a fill process on the entire region including another region inside the subject region. Therefore, by performing the above-described fill processing 1 to fill processing 3, even when the above-described another area exists inside the subject area, the fill processing that prioritizes the subject area can be performed while absorbing the influence of the other area. If the predicted point 7E of the center of gravity of the subject is calculated from the area and the center of gravity of the subject region calculated by the data calculation unit 8 at the same time as the fill process 3 described above, the position of the center of gravity with respect to the subject region is obtained. The defect can be improved.

According to the filling processing of the subject area described above, the filling processing is performed every predetermined frame period to obtain the above-mentioned center-of-gravity position information, and the information is stored in the drive control unit 9, so that the center-of-gravity position of the subject area is stored. Since information is stored, it is possible to predict the position of the center of gravity of the subject,
Further, when the subject area is obtained, the center-of-gravity position information of the subject area is updated each time, so that the predicted value can be corrected, so that the position of the subject area can be accurately captured. Therefore, the subject was tracked and photographed according to the amount of change in the current frame screen compared to the value of the previous frame screen, but for example, in the "initial setting state" By setting the position at the center of the image capturing screen, it is of course possible to perform the tracking shooting of the subject such that the center of gravity of the subject is always located at the center of the image capturing screen.

According to the filling processing of the subject area described above, the filling processing can be performed every predetermined frame period to obtain the area information of the subject together with the center of gravity position information described above. By controlling the zoom lens of the optical system so that the area information becomes constant, it is of course possible to photograph the object while keeping the size of the object substantially constant.

By performing the fill process for each frame, the subject information can be obtained by following the ever-changing state of the subject. That is, since the average value of the color information of the subject obtained up to the previous frame period is used as the color information for specifying the subject area, when the color of the subject changes gradually, for example, the sun, a light bulb, a fluorescent lamp. Since the subject can be specified even when the same subject is imaged under various light sources such as, for example, the subject can be accurately tracked and photographed, and at the same time, for example, the area information of the subject can be fixed so that the area information of the subject becomes constant. It goes without saying that the size of the subject may be kept substantially constant and the subject may be photographed by controlling the zoom lens.

In this embodiment, the reference center-of-gravity position information signal of the subject is set in the initial setting state, the change of the subject's movement is captured by comparison with the current center-of-gravity position information signal of the subject, and the change in the subject's motion is detected according to the change. Although the pan / tilt control of the video camera was performed, the pan / tilt control is performed by sequentially comparing the change in the movement of the subject with the gravity center position information signal of the subject on the current frame screen and the gravity center position information signal of the subject on the previous frame screen. Needless to say,

Although the auto-zoom control of the video camera has been described in this embodiment, for example, the above-mentioned CCD
It goes without saying that automatic tracking control of the still camera or the like can be performed by adding the configuration of the present invention to the still camera or the like using the image pickup device as an image pickup medium.

[0040]

As described above, the video camera of the present invention binarizes a video signal obtained by capturing an image of a subject into a subject region and a region other than that, and a binarization unit for the subject region. Fill processing means for filling the value image signal, data calculation means for respectively outputting the area information signal and the center of gravity position information signal of the subject region in the fill processed image signal output from the fill processing means, the area information signal, The center-of-gravity position information signal and the color information signal related to the subject area are updated and held, and the held color information signal and center-of-gravity position information signal are output to the binarization processing unit and the fill processing unit, respectively. Drive control means for outputting a control signal based on a difference between the set center-of-gravity position information signal and the updated center-of-gravity position information signal; Since it has a driving means for performing pan-tilt control so that the image of the subject is always displayed on the image pickup screen based on the control signal, the subject is specified based on the color information signal of the imaged subject, and By obtaining the center-of-gravity position information signal, for example, based on the value of the center-of-gravity position information signal in the initial setting state, depending on how much the center-of-gravity position information signal has changed at present, the object and the motion of about the same size and about the same as the object Therefore, even if the child of the subject and the child dressed in a similar color are in the same imaging screen, the subject can be identified from the above color information. Since the object can be identified, it is possible to distinguish the subject from a child other than the subject, and there is an effect that the subject can be automatically tracked with high accuracy.

Further, even when an object having the same size and color as the object is moved so as to overlap with the object, the object can be specified with high accuracy by obtaining the position information of the center of gravity of the object. There is an effect that the subject can be automatically tracked.

Furthermore, for example, by covering a child as a subject with a hat of R; red, G; green, B; blue, etc., the color of this hat is present in any area on the image pickup screen. It is not necessary to specify the color of an object that is clearly distinguished from the color of an object other than the object in advance, and therefore the above-described drawbacks can be eliminated.

Further, since the average value of the color information obtained up to the previous image pickup screen is used as the color information for specifying the subject area, when the color of the subject changes gradually, for example, the subject changes from the shade to the shade. Since the subject can be specified even when entering or exiting, there is an effect that the subject can be automatically tracked with high accuracy.

[Brief description of drawings]

FIG. 1 is a block diagram for explaining an embodiment of a video camera according to the present invention.

FIG. 2 is a diagram for explaining the operation of a binarization processing unit.

FIG. 3 is a diagram for explaining the operation of the logical filter.

FIG. 4 is a diagram for explaining the operation of the fill processing means.

FIG. 5 is a diagram for explaining the operation of the fill processing means.

FIG. 6 is a diagram for explaining the current operation of the fill processing means after the initial state.

FIG. 7 is a block diagram for explaining a conventional video camera.

[Explanation of symbols]

 5 Binarization processing means 5C, 6A, 7C, 7E Subject area 6 Filter 7 Fill processing means 8 Data calculation means 9 Drive control means 10 Drive means

Claims (1)

[Claims] 1. A binarization processing unit for binarizing a video signal of a subject into a subject region and a region other than the subject region, and a fill processing unit for performing a fill process on a binary image signal relating to the subject region. Data calculation means for respectively outputting an area information signal and a barycentric position information signal of the subject area in a fill processed image signal outputted from the fill processing means, the area information signal and the barycentric position information signal, and color information related to the subject area The signals are updated and held respectively, and the held color information signal and barycentric position information signal are output to the binarization processing means and the fill processing means, respectively, and the barycentric position information signal initialized is used as a reference and updated. A drive control unit that outputs a control signal based on a difference from the held center-of-gravity position information signal, and an image capturing screen that constantly captures the image of the subject based on the control signal A video camera, comprising: a drive unit that performs pan-tilt control as shown above.