KR20190056650A - Method for controlling PTZ camera using values obtained by analyzing images obtained from a fixed camera - Google Patents

Abstract

Provided is a method for generating a control value to control a pan/tilt/zoom (PTZ) camera in a wide-angle image processing apparatus which can control the rotation speed. To this end, provided is a method for generating a distance look-up table and a focus look-up table and generating a PTZ rotation speed value and a focal distance by using the distance look-up table and the focus look-up table.

Description

A method for controlling a PTZ camera using an analysis value of an image captured by a fixed camera, the method comprising:

The present invention relates to a method for controlling a PTZ camera for remote face recognition of a moving person by using an analysis value of an image captured by a fixed camera.

Conventional face recognition technology mainly uses a method of performing recognition with a facial image acquired through close-up photography within 1m. Recently, as the performance of equipment has improved, a technique of recognizing faces by enlarging a person at a distance of 1 m or more using a high-performance camera has appeared. Unlike the conventional face recognition technology, the remote face recognition must be recognized even when the subject is continuously moving, not in the stop state. Accordingly, when a plurality of recognition objects continuously move, many cameras having a PTZ (Pan / Tilt / Zoom) function are required in order to track and recognize each recognition object.

The method for controlling the camera in the remote face recognition based on the PTZ camera is implemented in a multi camera environment using a fixed camera. For the remote face recognition of the moving person, first, the PTZ camera rotates to track the moving person, and then the screen enlargement function is utilized to acquire a good face shape. However, existing PTZ camera control methods do not consider the moving speed of a person, and control a PTZ camera at a fixed rotation speed to track a person, so it is difficult to obtain a good face image in various environments.

 According to the existing PTZ camera control methods, as the screen of the PTZ camera is enlarged, the blurring phenomenon of the image due to the pan and tilt rotation speed becomes worse, and thus the face recognition function can not be performed. In order to solve this problem, the pan and tilt must be rotated at an appropriate speed according to the moving speed of the person, and the focus adjustment to the object to be detected must be performed accurately and quickly.

The present invention provides a PTZ camera control method capable of controlling the rotation speed as a technique capable of solving the above-mentioned problems and satisfying the above-mentioned requirements.

In accordance with an aspect of the present invention, there is provided a wide-angle image processing apparatus for acquiring an image output by a fixed camera having a fixed photographing direction and a photographing magnification and controlling the operation of the PTZ camera using the obtained image, And a control value for controlling the control value. Wherein the wide-angle image processing apparatus detects, for each of a plurality of image acquiring time points, a first object included in the image transmitted from the fixed camera, (X, y) corresponding to the image acquiring time; Wherein the wide-angle image processing apparatus obtains, for each of the N consecutive image acquiring time points, an approximate width that is a value closest to a horizontal width of an area occupied by the first object from a previously provided distance look-up table, A first obtaining step of obtaining the object distance corresponding to the approximate width from the distance lookup table as an object distance representing a distance to the first object; Wherein the wide-angle image processing apparatus comprises: a calculating step of calculating a moving speed of the first object using N object distances and N image acquiring times acquired for the N image acquiring time points; Wherein the wide-angle image processing apparatus comprises: a calculating step of calculating a rotation speed value of the PTZ camera using a moving speed of the first object; And a first providing step in which the wide-angle image processing apparatus provides the PTZ rotation speed value to the PTZ camera as the control value.

In this case, the wide-angle image processing apparatus acquires N focal lengths for the PTZ camera from a previously provided focus lookup table, and the N focal lengths are values corresponding to the N object distances , A second acquiring step; And the second wide-angle image processing apparatus may further include a second providing step of providing the N focal lengths as the control value to the PTZ camera.

At this time, before the first acquisition step, the wide-angle image processing apparatus confirms whether the distance lookup table associated with the fixed camera exists or not. And the step of generating and storing the distance lookup table associated with the fixed camera when the wide-angle image processing apparatus determines that the distance look-up table does not exist. Wherein the step of generating the distance lookup table comprises the steps of: using a view angle of the fixed camera, the object distance being a distance from the fixed camera to a subject, and a pixel length value per meter in the object distance, Calculating a number of pixels per pixel; And storing the number of pixels per meter-width in the distance look-up table.

In this case, the object distance may be a distance value used when calculating the approximate width, which is the number of pixels per meter-width.

At this time, before the second acquisition step, the wide-angle image processing apparatus confirms whether the focus look-up table associated with the fixed camera exists or not. And generating and storing the focus lookup table if it is determined that the focus lookup table does not exist, wherein the step of generating the focus lookup table comprises the steps of: Calculating a zoom value and a focal length for extracting a face image of the object from the PTZ camera at a lattice spacing of the chessboard using a chessboard as an object having the chessboard; And storing the zoom value and the focal length of the distance value of the grating interval in the focus lookup table.

It is possible to provide a wide-angle image processing apparatus that obtains an image output from a fixed camera having a fixed photographing direction and a photographing magnification fixed according to an aspect of the present invention and controls the operation of the PTZ camera using the obtained image . The wide-angle image processing apparatus according to claim 1, wherein the wide-angle image processing apparatus detects a first object included in an image transmitted from the fixed camera for each of a plurality of image acquiring time points, (x, y) corresponding to the image acquisition time; Acquiring, for each of the N consecutive N image acquiring time points, an approximate width that is a value closest to a width of an area occupied by the first object, from a previously provided distance lookup table; and determining a distance from the fixed camera to the first object Acquiring the object distance corresponding to the approximate width as the object distance from the distance lookup table; Calculating a moving speed of the first object using the N object acquisitions for the N image acquiring times and the N image acquiring times; A calculating step of calculating a rotational speed value of the PTZ camera using the moving speed of the first object; And a first providing step of providing the PTZ rotational speed value as a control value to the PTZ camera.

Acquiring N focal lengths for the PTZ camera from a previously provided focus lookup table, wherein the N focal lengths are values corresponding to the N object distances, respectively; And a second providing step of providing the N focal lengths as the control value to the PTZ camera.

At this time, the fixed camera may be included.

According to the present invention, the blur shape of an image can be solved by controlling the rotation speed through a PTZ camera control method.

FIG. 1 is a flowchart illustrating a method of generating a control value for controlling a PTZ camera in a wide-angle image processing apparatus according to an exemplary embodiment of the present invention. Referring to FIG.
2 is a view for explaining the number of pixels per meter according to the position of an object according to an embodiment of the present invention.
3 is a view for explaining the number of pixels per meter-width according to an embodiment of the present invention.
4 is a diagram for explaining step S40 and step S50 according to an embodiment of the present invention.
5 is a view for explaining a wide-angle image processing apparatus according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. However, the present invention is not limited to the embodiments described herein, but may be implemented in various other forms. The terminology used herein is for the purpose of understanding the embodiments and is not intended to limit the scope of the present invention. Also, the singular forms as used below include plural forms unless the phrases expressly have the opposite meaning.

A face recognition system according to an embodiment of the present invention may include one or more PTZ cameras and one fixed camera, and a processing device capable of receiving information from them and commanding them. The fixed camera may have the photographing direction, the zoom magnification, etc. fixed to a specific value. An apparatus including the above processing apparatus may be referred to as a wide-angle image processing apparatus.

FIG. 1 is a flowchart illustrating a method of generating a control value for controlling a PTZ camera in a wide-angle image processing apparatus according to an exemplary embodiment of the present invention. Referring to FIG.

In step S10, the wide-angle image processing apparatus can check whether a distance lookup table exists for the first image acquired through the fixed camera and whether the distance lookup table can be accessed.

The distance lookup table may be generated by matching a specific fixed camera. The distance lookup table includes a horizontal length field having a value relating to a horizontal width of an image or an image photographed by a fixed camera, and a subject object distance field indicating a distance between the object existing in the image or the image and the fixed camera . Hereinafter, the distances to the fixed camera and the object may be referred to simply as object distances in order to distinguish them from other distances.

For example, the horizontal width occupied by the object having the predetermined width in the image or the image may be changed according to the object distance from the object to the fixed camera. As the object distance increases for the object having a constant width, the width will decrease.

As will also be described later, the distance lookup table may be provided with a field indicating the number of pixels per pixel (delta) (pixel / m).

In step S20, if the wide-angle image processing apparatus has the distance look-up table and is accessible, the wide-angle image processing apparatus determines whether a focus lookup table exists for the acquired first image And whether or not the focus lookup table can be accessed.

Wherein the focus lookup table comprises an object distance field indicating an object distance between an object existing in an image or an image photographed by a fixed camera and the fixed camera and a focus object distance field indicating a focal distance of the fixed camera determined for each object distance , And a zoom value field indicating an optimized zoom value for each object distance.

In step S30, if the wide-angle image processing apparatus can access the distance look-up table and the focus look-up table, the wide-angle image processing apparatus can detect the first object included in the acquired first image have. The first object may mean, for example, the whole body shape of a person.

In step S40, the wide-angle image processing apparatus may store the center point position value (x, y) representing the position of the first object and the value for the time point in the array at predetermined time points. At this time, the value for the time may be stored in milliseconds.

For example, the first position value (x1, y1), which is the center position value of the first object detected at the first time (t1), is stored in the first element of the first arrangement, The second position value (x2, y2), which is the center point position value of the first object detected at time t2, is stored in the second element of the first array, and the third position (x3, y3), which is a center position value of the first object detected at time t3, is stored in a third element of the first array, and a fourth time t4 (X4, y4), which is the center position of the first object, detected in the first array is stored in the fourth element of the first array, and at a fifth time (t5) several ms later than the fourth time The detected fifth position value (x5, y5), which is the center position of the first object, may be stored in the fifth element of the first array.

When the number of the center point position values stored in the first arrangement reaches a predetermined number (for example, five) in step S50, the wide-angle image processing apparatus outputs the corresponding point of time (ex: t1, t2, t3 , t4, t5) of the first area defined by the first object as a rectangular area occupied by the first image. The horizontal width may be calculated using the horizontal and vertical number of pixels occupied by the first area among the pixels constituting the first image.

At this time, assuming that the shoulder width of the person representing the first object standing in a posture in which the face is recognizable in the photographed first image is assumed to be 40 cm, the calculated width is multiplied by 2.5, .

Next, the wide-angle image processing apparatus can find an approximate width (delta c1), which is the width (delta c) closest to the horizontal width, from the width field of the generated distance lookup table.

 Then, the wide-angle image processing apparatus can find a distance value corresponding to the approximate width (delta c1) from the object object distance field of the distance lookup table generated in advance. The distance value is a value indicating an object distance from the fixed camera to the first object.

That is, the object distance between the first object and the fixed camera at the consecutive five viewpoints (ex: t1, t2, t3, t4, t5) using the distance lookup table and the captured first image Can be determined. Hereinafter, the five object distance values are referred to as a first object distance value d1, a second object distance value d2, a third object distance value d3, a fourth object distance value d4, (d5).

In step S60, the wide-angle image processing apparatus calculates the predetermined number (e.g., five) of the object distances (e.g., t1, t2, t3, t4, (m / s) of the first object can be calculated using the distance (distance). For example, the equation for obtaining the moving speed may be expressed by Equation 1 below.

[Equation 1]

Here, the speed is a moving speed of the first object,

The time _i is a time point at which the image is photographed and identified as an index i,

The distance _i is an object distance between the first object and the fixed camera, and is a distance from a point identified by the index i.

In step S70, the wide-angle image processing apparatus may calculate the PTZ rotation speed value (deg / s) using the moving speed of the object. For example, the equation for obtaining the PTZ rotational speed value may be as shown in the following equation (2).

[Equation 2]

In step S80, the wide-angle image processing apparatus may obtain the distance values corresponding to the object distances d1 to d5 by searching the generated focus distance look-up table.

Then, the wide-angle image processing apparatus may provide the PTZ rotation speed value and the focal distances to the PTZ camera.

2 is a view for explaining the number of pixels (number of pixels) per meter according to the position of an object according to an embodiment of the present invention.

3 is a view for explaining the number of pixels per meter according to an embodiment of the present invention. 3 is a diagram for explaining the number of pixels per meter at a distance of 3 m, for example.

Hereinafter, description will be made with reference to FIGS. 2 to 3 together.

The distance lookup table may be newly generated and stored if it is determined that the distance lookup table does not exist in step S10. The distance lookup table may be stored corresponding to the fixed camera.

The distance lookup table may be provided with a field indicating the number of pixels per pixel (delta) (pixel / m) for a specific object distance.

In one embodiment of the present invention, the specific object distance represents the length of an extended radius line from the fixed camera to the object, and the 'meter width' of the number of pixels per meter-width is the horizontal line Lt; RTI ID = 0.0 > 1 < / RTI > Therefore, the number of pixels per meter-meter may indicate how many pixels are contained in a space of 1 meter existing in the object distance. That is, the number of pixels per meter-meter may indicate how many pixels a one meter horizontal line contains at the location of an object that exists at the corresponding radial distance (i.e., the object distance).

The number of pixels per meter-meter (?) Of a specific object distance may be different according to the angle of view (AoV, Angle of View) of the fixed camera and the resolution of the image.

Wherein the number of pixels per meter width is determined by using parameters such as an angle of view (AoV) of the fixed camera, an object distance to the fixed camera and an object, and a pixel length of the fixed camera resolution . At this time, the number of pixels per pixel (pixel / m) can be calculated for each object distance at the position where the fixed camera 21 is installed.

The image 1 obtained from the fixed camera 21 may be divided into a plurality of matrix-shaped regions (or grids). For example, a grid of m * n may be generated over the image image 1.

In this case, when the object 10 having the fixed horizontal value is located at the object distance of 1 m, the horizontal length value of the area occupied by the pixels representing the object 10 in the image may be a1. If the object 10 is located at a distance of 2m objects, the width value of the area occupied by the pixels representing the object 10 in the image may be a2. When the object 10 is located at a distance of 3 m, the width value of the area occupied by the pixels representing the object 10 in the image may be a 3.

If the object is a person, it is assumed that the average shoulder width of a person is about 45 cm. If a person is located at a point where the object distance is 1 m, the width of the area occupied by the person in the image is, If the object distance is 2 m, the width of the area occupied by the person in the image may be close to, for example, a2. The values of a1 and a2 may be changed according to the direction in which the person stands and the camera photographing angle.

In FIG. 2, for convenience of description, the object is only located at a distance of 3 m. However, in the preferred embodiment, the length of each object can be calculated from 1 m to 15 m at intervals of 1 m.

The angle of view? Covered by each pixel and the actual spatial distance value a indicated by each pixel may vary depending on the object distance from the object 10 to the fixed camera 21. For example, when the object 10 is located at an object distance of 3 m from the fixed camera 21, the angle of view represented by each pixel is? 3, and the actual spatial distance value a indicated by each pixel may be a3.

By applying the angle of view (AoV), the object distance (m) of the fixed camera, and the spatial distance a (pixel length) of each pixel in the object distance to the following equation 3, The number of pixels per meter width? Can be calculated.

[Equation 3]

The pixel length represents the width of the image output by the fixed camera. For example, when the fixed camera outputs an image of 1920 * 1080, the pixel length is 1920px. A method of calculating how many meters in the real world the transverse length of an image outputted by the fixed camera corresponds to can be understood by tan (AoV / 2) * 2 * distance in [Expression 3]. [Equation 3] divides the output image by the horizontal direction meter value of the real world. Therefore, by using [Equation 3], it is possible to find the number of pixels in the image 1 meter in the real world.

Meanwhile, if it is determined in step S20 that a focus lookup table does not exist for the acquired image, the focus lookup table may be generated and stored. At this time, the optimum zoom value and focal length of the object can be stored in the focus lookup table.

At this time, the optimum zoom value and the focal length for extracting the face image of the object at intervals of N (50 cm) from the PTZ camera using a predetermined chessboard of N * N (for example, 50 cm * 50 cm) Can be calculated.

4 is a diagram for explaining step S40 and step S50 according to an embodiment of the present invention.

(X, y) of the center point (12) of the object (10) and the corresponding time at a predetermined point in time using the image image (1) acquired from the fixed camera You can store the value in an array.

(W) of the area 11 occupied by the object 10 is multiplied by 2.5 when the number of the center point position values stored in the array reaches a predetermined number (for example, five) An approximate width delta c1 that is a value matched to the length of the distance lookup table can be found from the width field of the distance lookup table.

Next, a first object distance, which is an object distance corresponding to the found approximate width? C1, can be obtained from the object object distance field of the distance lookup table.

5 is a view for explaining a wide-angle image processing apparatus according to an embodiment of the present invention.

The wide-angle image processing apparatus 20 can generate a rotation speed value and a focal distance for controlling the PTZ camera 30 and provide the rotation speed value and the focal distance to the PTZ camera 30.

Then, the wide-angle image processing apparatus 20 can acquire information output from the fixed camera 21.

The wide-angle image processing apparatus 20 includes a step of detecting an object (an object to be recognized) with respect to an image acquired using the fixed camera 21, a step of detecting an object center point position value (x, y) Storing a value for a corresponding time point in an array, and when a number of the center point position values stored in the array reaches a predetermined number, Calculating a moving speed of the object using the predetermined number of distance values calculated for each corresponding time point, calculating a moving speed of the object using the distance values calculated for each corresponding time point, Calculating a PTZ rotation speed value using the movement speed, searching the obtained distance values in a previously generated 'focus lookup table' Obtaining the focal distances, and providing the PTZ rotational speed value and the focal distance as control values to the PTZ camera 30. [

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the essential characteristics thereof. The contents of each claim in the claims may be combined with other claims without departing from the scope of the claims.

Claims (8)

A wide-angle image processing apparatus for acquiring an image output from a fixed camera having a fixed photographing direction and a photographing magnification and controlling an operation of the PTZ camera using the obtained image, As a method of generating,
The wide-angle image processing apparatus according to claim 1, wherein the wide-angle image processing apparatus detects a first object included in an image transmitted from the fixed camera for each of a plurality of image acquiring time points, (x, y) corresponding to the image acquisition time;
Wherein the wide-angle image processing apparatus obtains, for each of the N consecutive image acquiring time points, an approximate width that is a value closest to a horizontal width of an area occupied by the first object from a previously provided distance look-up table, A first obtaining step of obtaining the object distance corresponding to the approximate width from the distance lookup table as an object distance representing a distance to the first object;
Wherein the wide-angle image processing apparatus comprises: a calculating step of calculating a moving speed of the first object using N object distances and N image acquiring times acquired for the N image acquiring time points;
Wherein the wide-angle image processing apparatus comprises: a calculating step of calculating a rotation speed value of the PTZ camera using a moving speed of the first object; And
The wide-angle image processing apparatus comprising: a first providing step of providing the PTZ rotation speed value to the PTZ camera as the control value;
/ RTI >
PTZ camera control method. The method according to claim 1,
Wherein the wide-angle image processing apparatus obtains N focal lengths for the PTZ camera from a previously provided focus-lookup table, wherein the N focal lengths are values corresponding to the N object distances, 2 acquisition phase; And
Wherein the wide-angle image processing apparatus includes a second providing step of providing the N focal lengths as the control value to the PTZ camera
≪ / RTI >
PTZ camera control method. The method according to claim 1,
Before the first acquisition step,
Wherein the wide-angle image processing apparatus comprises: checking whether the distance look-up table associated with the fixed camera is present; And
Wherein the wide-angle image processing apparatus further comprises a step of generating and storing the distance lookup table associated with the fixed camera when it is determined that the distance lookup table does not exist
Further comprising:
Wherein the step of generating the distance look-
Calculating a number of pixels per meter-width in the object distance using the angle of view of the fixed camera, the object distance as a distance between the fixed camera and the subject, and the pixel length per meter in the object distance; And
Storing the number of pixels per meter-width in the distance look-up table
/ RTI >
PTZ camera control method. 4. The method of claim 3, wherein the object distance is a distance value used when calculating the approximate width that is the number of pixels per meter-width. 3. The method of claim 2,
Before the second acquisition step,
Wherein the wide-angle image processing apparatus comprises: checking whether the focus lookup table associated with the fixed camera is present; And
If it is determined that the focus lookup table does not exist, generating and storing the focus lookup table
Further comprising:
Wherein the generating the focus lookup table comprises:
Calculating a zoom value and a focal length for extracting a face image of the object from the PTZ camera at a lattice interval of the chess board using a chess board having an N * N shape and having a predetermined width; And
Storing the zoom value and the focal length for the distance value of the grid interval in the focus lookup table
/ RTI >
PTZ camera control method. A wide-angle image processing apparatus for acquiring an image output by a fixed camera having a fixed photographing direction and a photographing magnification, and controlling an operation of the PTZ camera using the obtained image,
Detecting a first object included in an image transmitted from the fixed camera for each of a plurality of image acquiring time points and calculating a center position value (x, y) of an area occupied by the detected first object in the image, Storing the image corresponding to the image acquiring time;
Acquiring, for each of the N consecutive N image acquiring time points, an approximate width that is a value closest to a width of an area occupied by the first object, from a previously provided distance lookup table; and determining a distance from the fixed camera to the first object Acquiring the object distance corresponding to the approximate width as the object distance from the distance lookup table;
Calculating a moving speed of the first object using the N object acquisitions for the N image acquiring times and the N image acquiring times;
A calculating step of calculating a rotational speed value of the PTZ camera using the moving speed of the first object; And
A first providing step of providing the PTZ rotational speed value to the PTZ camera as a control value;
Lt; / RTI >
Wide - angle image processing device. The method according to claim 6,
Acquiring N focal lengths for the PTZ camera from a previously provided focus lookup table, wherein the N focal lengths are values corresponding to the N object distances, respectively; And
A second providing step of providing the N focal lengths as the control value to the PTZ camera
, ≪ / RTI >
Wide - angle image processing device. 7. The wide-angle image processing apparatus according to claim 6, comprising the fixed camera.

Images