JPH05219421A - Television camera controller - Google Patents

Abstract

PURPOSE:To relive the working load of a camera man by devising a controller such that a camera position and a zoom distance are automatically controlled while tracking with the motion of an object. CONSTITUTION:The controller is provided with a key signal processing section 1 extracting a range occupied by an object and noticed from a video signal picked up by a television camera 11, a position size detection section 33 recognizing the object based on the extracted key signal and calculating the position and size on a screen, a control variable arithmetic operation section 5 calculating moving direction data and moving distance data of the object based on the layout and the field angle of the object preset to a layout field angle setting section 7 and the calculated position and size of the object, and a universal head controller 9 calculating the movement quantity of the universal head of the television camera and the zoom quantity of the camera lens based on the moving direction data and moving distance data of the object and executing zoom control of the camera lens 15 and the moving control of a universal head 13.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a device for recognizing a subject in a video signal and a device for automatically controlling the movement of a television camera using the recognition result.

SUMMARY OF THE INVENTION The present invention provides a control device suitable for capturing a scene with a relatively small motion such as a conference or a lecture with a television camera. For example, it is possible to focus on the face portion of a performer, detect the movement in the video signal, and automatically control pan, tilt, and zoom of the television camera.

[0003]

2. Description of the Related Art Conventionally, as a control device for a television camera, a television camera remote control system for controlling a camera platform so that a fixed angle of view is known.

In this television camera remote control system, the camera platform can be controlled so that the camera faces the designated direction.

[0005]

As described above, the conventional television camera control device cannot specify the angle of view in advance, but cannot automatically control the television camera by following the movement of the subject. It relied on the manual operation of the cameraman, and the development of this type of automatic control device was eagerly awaited.

The present invention has been made in view of the above circumstances, and an object thereof is to automatically control the position and zoom distance of a camera by following the movement of a subject of interest, thereby reducing the work load on the cameraman. Another object of the present invention is to provide a control device for a television camera capable of performing the above.

[0007]

In order to achieve the above object, the present invention provides a means for extracting a range occupied by a subject of interest as a key signal from a video signal picked up by a television camera, and the extracted key. A means for recognizing the subject based on the signal and calculating the position and size on the screen,
Means for calculating moving direction data and moving distance data of the subject from preset composition and angle of view of the subject and the calculated position and size of the subject, and calculated moving direction data and moving distance data of the subject From the television camera to calculate the pan head movement amount of the television camera and the zoom amount of the camera lens to execute the pan head control and the zoom control of the camera lens.

[0008]

According to the above structure, for example, the face of the performer is focused on, the angle of view and composition thereof are stored in advance, and the range occupied by the face on the screen is extracted as the key signal from the image signal of the subject. .. The area occupied by the face on the screen is calculated from the extracted key signal, and the moving direction and distance thereof are calculated by comparing with the stored angle of view and composition. Then, the pan head movement amount of the television camera and the zoom amount of the camera lens are calculated from the calculated movement direction and distance, and the pan head and the camera are controlled based on the calculated data. Thus, the movement amount of the pan head and the zoom amount of the camera lens can be automatically controlled by following the movement of the face without human intervention.

[0009]

1 is a block diagram showing the configuration of an embodiment of a control device for a television camera according to the present invention.

The apparatus of this embodiment comprises a key signal processing section 1, an image processing section 3, a controlled variable calculating section 5, a composition / angle of view setting section 7.
And a pan head controller 9, and detects the movement of the image captured by the television camera 11 and follows the movement.
The zoom amount of 5 is controlled.

The key signal processing unit 1 inputs a video signal picked up by the television camera 11 through the camera controller 17 and, in the same way as the key processing of a chroma key or a level key, determines the range occupied by the skin color portion in the subject. It is taken out as a key signal and supplied to the image processing unit 3. For example, when key processing is performed on an image of a performer who has placed both hands on a desk, as shown in FIG. 2A, images of only the skin-colored face and hands are extracted.

The image processing section 3 comprises a mask processing section 31 and a position / size detecting section 33.

The mask processing section 31 specifies the range occupied by the subject image signal by mask processing and supplies the image signal to the position / size detecting section 33.

The position / size detecting section 33 detects the position and size of the object on the image from the supplied mask-processed object image signal and supplies it to the control amount calculating section 5.

The control amount calculation unit 5 can be constituted by a personal computer, and compares the supplied position data and size data with the composition data and angle of view data preset in the composition / angle of view setting unit 7. The deviation signal is generated and supplied to the platform controller 9.

The platform controller 9 inputs the supplied deviation signal, generates a platform control signal and a zoom control signal, and supplies them to the zoom mechanism of the platform 13 and the lens 15.

Next, the operation of this embodiment will be described.

In this embodiment, an example will be described in which the television camera 11 is made to follow the movement of the face of a performer who has placed both hands on a desk. Prior to automatic tracking of the television camera 11, target composition data and view angle data are set in advance in the composition / view angle setting unit 7. In this example,
The initial center position data (x ₀ ,
y ₀ ) as the composition data, and the horizontal / vertical initial distance data L _x0 and L _y0 occupied by the face image are input as the angle-of-view data.

In this state, when image pickup is started, a video signal is sent from the camera controller 17 to the key signal processing unit 1.
Entered in. In the key signal processing unit 1, the range occupied by the face and both hands is extracted as a key signal from the video signal for each frame. The extracted image signals of the face and both hands are supplied to the mask processing section 31 and subjected to the following mask processing.

That is, in the case where, for example, a part other than the face of the performer, such as both hands, is displayed in the same screen in the extracted image signal as shown in FIG. The areas other than the face are masked to limit the range to be recognized. As a result, unnecessary images such as the parts of both hands can be excluded from the recognition target, and the recognition efficiency is improved.

With respect to the image thus masked, the position / size detecting unit 33 sets a rectangle circumscribing the face image and sets the center position data ( x ₁ , y ₁ ) and the horizontal and vertical distance data (size data) L _x1 and L _y1 occupied by the face image are calculated. As can be seen from FIG. 2B, the mask range is set wider than the actual distance data L _x1 and L _y1 because the subject may move.

Calculated center position data (x ₁ , y ₁ )
And the horizontal and vertical distance data L _x1 and L _y1 are supplied to the control amount calculator 5. As described above, at the time of initial setting, the initial center position data (x ₀ ,
y ₀ ) is set in the composition / view angle setting unit 7 as composition data and initial distance data L _x0 and L _{y0 in the} horizontal and vertical directions are set as view angle data.

In the control amount calculator 5, the composition data (x ₀ , y ₀ ) and the angle-of-view data L _x0 , which are stored in advance, are stored.
L _y0 is compared with the center position data (x ₁ , y ₁ ) and distance data L _x1 , L _y1 to calculate the deviation. This deviation means a deviation from the position of the face in the initial setting screen.

The calculated deviation signal is used as the platform controller 9
The pan / tilt amount of the platform 13 is calculated from the deviation of the center position data by the platform controller 9, and the platform control based on the calculated amount is executed. Further, the zoom amount of the lens 15 is calculated from the deviation of the view angle data, and the zoom control based on the zoom amount is executed.
The above process is executed every three frames. This is because the time required for three frames is required for image processing.
Even for each frame, it is a sufficient speed considering the operating speeds of the platform 13 and the lens 15.

As described above, according to the present embodiment, the television camera can be automatically tracked so that the desired composition and angle of view can be set in advance even if the subject moves in any direction during image pickup.

As described above, in the present embodiment, the composition change is recognized based on the shift of the center position data of the subject image to be noticed. This is because the shift of the center of gravity position data of the subject image and other points is recognized. You may make it monitor and request. Also, although the face of a person is shown as an example of the subject of interest, it is needless to say that the subject is not limited to this. Further, the mask area in the mask processing can be set also from the predicted position and size of the subject estimated from the target composition, and in the case of a simple image, the mask processing may be omitted.

Further, as shown in FIG. 3, in the case of a so-called two-shot image in which two performers are lined up, unnecessary parts such as hands are masked within a range including two faces, The same recognition / tracking processing as in the above-described embodiment may be performed on the mask range.

[0028]

As described above, according to the present invention, the position of the camera and the zoom distance can be automatically controlled by following the movement of the subject of interest, and the work load on the cameraman can be reduced.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an embodiment of a control device for a television camera according to the present invention.

FIG. 2 is an explanatory diagram showing an operation of the embodiment.

FIG. 3 is an explanatory diagram showing mask processing in the case of a two-shot image.

[Explanation of symbols]

 1 Key signal processing unit 3 Image processing unit 5 Control amount calculation unit 7 Composition / angle of view setting unit 9 Camera platform controller 11 Television camera 13 Camera platform 15 Lens 31 Mask processing unit 33 Position / size detection unit

Claims (1)

[Claims] 1. A means for extracting, as a key signal, a range occupied by a subject of interest from a video signal picked up by a television camera, and a position and a size on a screen for recognizing the subject based on the extracted key signal. A unit for calculating the moving direction data and moving distance data of the subject from the preset composition and angle of view of the subject and the calculated position and size of the subject, A television comprising: means for calculating a pan head movement of a television camera and a zoom amount of a camera lens from movement direction data and movement distance data to execute pan head control and camera lens zoom control. Control device for John camera.