RU2584816C1 - Method and system for reducing positioning error of ptz chamber - Google Patents

Abstract

FIELD: telecommunications.

SUBSTANCE: invention relates to video surveillance, in particular, to video surveillance using rotary (PTZ) cameras. Disclosed is a method of reducing positioning error of PTZ camera, characterised by that it comprises obtaining a target position of camera rotation, then determining at least one intermediate camera and its coordinates based on data on target position of camera rotation, after which camera is successively rotated to target position through said intermediate positions.

EFFECT: reduced positioning error of camera and increasing repeatability of positioning.

21 cl, 4 dwg

Description

FIELD OF TECHNOLOGY

This technical solution relates to the field of video surveillance, in particular to video surveillance using PTZ cameras.

BACKGROUND OF THE INVENTION

Currently, PTZ cameras are widely used to monitor large areas. They are a device that supports remote control of viewing direction and zoom. PTZ cameras are actively used in video conferencing, being a necessary attribute of a conference room or meeting room, when building security systems and other video surveillance systems. Due to their high prevalence at the moment, the range of cameras is quite extensive, cameras have different characteristics and cost.

When using a PTZ camera for video surveillance, an important characteristic of its functioning is the camera positioning error, it depends on the degree of wear of the mechanism, its initial accuracy, and camera calibration.

The prior art article is known [1], which describes the approaches that are used in solving the problem.

Sinha and Polleface proposed a positioning method for PTZ cameras [2], in which the camera is first calibrated at low zoom, and then the internal parameters of the camera are calculated with increasing zoom. Since calibration is performed discretely from one zoom value to another, piecewise linear interpolation is used to calculate the internal parameters. Using this method provides the need for a large number of calibration steps to mitigate noise, which significantly increases the operating time.

Also known is the method [3] of Sarkis, in which the internal parameters of the camera are dependent on the lens settings based on the moving least-squares method. However, this method is computationally demanding.

ESSENCE

This technical solution is aimed at eliminating the disadvantages inherent in solutions known from the prior art.

The technical result is to reduce the camera positioning error and increase the repeatability of positioning.

This technical result is achieved through the use of intermediate positioning points, which reduce the effect of inertia and "flight" of the required position. By reducing the positioning error, the positioning accuracy is improved.

Another technical result is the repeatability of positioning results. As a result, the positioning error becomes systematic and the same.

A method of reducing the positioning error of a PTZ camera, characterized in that the target position of the camera rotation is obtained, then at least one intermediate position of the camera and its coordinates are determined based on the target rotation position of the camera, and then the camera is rotated sequentially to the target position through the above intermediate positions.

In some embodiments of the technical solution, one intermediate position and its coordinates are determined by testing the camera.

In some embodiments, the implementation of the technical solution, at least one intermediate position and its coordinates are determined based on ideas about the method of operation and the characteristics of the rotary mechanism of the camera.

In some embodiments of the technical solution, the intermediate points can be determined through absolute coordinates or through relative coordinates.

In some embodiments of the technical solution, the number and coordinates of intermediate positions depend on the required speed and positioning accuracy.

In some embodiments of the technical solution, the relative coordinates of the intermediate position are determined so that the intermediate position is as close to the desired position.

In some embodiments of the technical solution, computer vision algorithms are used to determine positioning errors.

In some embodiments of the technical solution, at least two intermediate positions are determined, the first position being different by A in the panoramic angle and the angle of inclination from the desired position, and the second intermediate position in the panoramic angle is the same as the desired position, and in the angle of inclination different by BUT.

In some embodiments of the technical solution, at least two intermediate positions are determined, wherein the first position differs by A in the panoramic angle and the angle of inclination from the desired position, and the second intermediate position in the angle of inclination coincides with the desired position, and in the panoramic angle differs by BUT.

In some embodiments of the technical solution, the speed of movement from the starting point to the first intermediate, between the intermediate and between the last intermediate point and the target point may be different.

BRIEF DESCRIPTION OF THE DRAWINGS

The signs and advantages of this technical solution will become apparent from the following detailed description and the accompanying drawings, in which:

In FIG. 1 - shows a block diagram of the proposed method;

In FIG. 2 - shows a block diagram of an example implementation of the inventive system;

In FIG. 3 - an example of the implementation of a technical solution with one intermediate point is given.

In FIG. 4 is a graph showing the experimental dependence of the vertical and horizontal positioning errors in pixels on the distance of the intermediate point from the target point in degrees.

DETAILED DISCLOSURE OF TECHNICAL SOLUTION

Below will be described the concepts and definitions necessary for the detailed disclosure of the ongoing technical solution.

A panoramic angle (precession angle, Pan) is one of the Euler angles that describes the rotation of an object around the Z axis (for more details, see [9]). This angle corresponds to the rotation of the object in its own horizontal plane.

The angle of inclination (nutation angle, Tilt) is one of the Euler angles that describes the rotation of an object around the Y axis (for more details, see [9]). This angle corresponds to the rotation of the object in its own vertical plane.

Positioning error - the difference between the target position of the camera and the actual position of the camera after positioning in the target position. It is expressed by two angles: panoramic (corresponding to the difference between the panoramic angles of the actual and target position) and the angle of inclination (corresponding to the difference between the angle of inclination of the actual and target position).

Repeatability of a result is a characteristic that reflects the probability of a repetition of the result of an experiment subject to a certain set of initial conditions.

The repeatability of the positioning result is a characteristic that reflects the probability of achieving the same values of the positioning error while maintaining the target position of the camera and changing the initial position.

Computer vision is the theory and technology of creating machines that can detect, track, and classify objects.

Computer vision algorithms are algorithms for processing graphic and video data to identify useful information, such as those described in the information source Forsyth, D., and D. Pons. "Computer Vision. A Modern Approach. William" Moscow 928 (2004).

In standard positioning mode, if necessary, rotate the camera from the starting position, the rotation command is sent to the target camera:

,

,

where Pos ₁ (Pan ₁ ; Tilt ₁ ) is the initial position of the camera, which is characterized by the panoramic angle Pan ₁ and the angle of inclination Tilt ₁ , Pos _D1 (Pan _D + Pan _Er1 ; Tilt _D + Tilt _Er1 ) is the camera position when the camera moves from a point 1 to the target position, it differs from the required target position Pos _D (Pan _D ; Tilt _D ) by the value of the positioning error.

In this case, turning to the same target position from a different starting position generally leads to other positioning errors:

,

,

Pos ₂ (Pan ₂ ; Tilt ₂ ) - the second initial position of the camera orientation, different from the first values of the panoramic angle and tilt angle, Pos _D2 (Pan _D + Pan _Er2 ; Tilt _D + Tilt _Er2 ) - the orientation of the camera when it _comes to the target position from starting position 2, with other error values for the panoramic angle and angle of inclination. The resulting position also differs from the desired target direction of the camera view.

With this approach, the scatter of Pan _Er values; Tilt _Er can reach significant values, and for cameras with different rotation mechanisms, the error will be from 0.05 degrees to 0.5 degrees. Various rotation mechanisms are described in [4].

This technical solution proposes the introduction of at least one additional position through which the camera passes before reaching the target position, which allows to reduce the positioning error:

Similar to the previous one:

Pos ₁ (Pan ₁ ; Tilt ₁ ) - the first starting position.

Pos _Int1 (Pan _Int + Pan _Er1 ; Tilt _Int + Tilt _Erl ) - an intermediate position with its own positioning error when repositioning from the first starting position.

Pos _D1 (Pan _D + Pan _MEr1 ; Tilt _D + Tilt _MErl ) - the end (target) position when repositioning from an intermediate position.

Pos ₂ (Pan ₂ ; Tilt ₂ ) - second starting position, different from the first starting position.

Pos _Int2 (Pan _Int + Pan _Er2 ; Tilt _Int + Tilt _Er2 ) - an intermediate position with its own positioning error when repositioning from the second initial position.

Pos _D2 (Pan _D + Pan _MEr2 ; Tilt _D + Tilt _MEr2 ) - the end position when coming from an intermediate position.

Because to the end point, the camera is positioned from an intermediate point, the coordinates of which in the two described movement routes can differ by no more than a positioning error (from 0.05 to 0.5 degrees), the final error is Pan _MEr ; Tilt _MEr will be significantly smaller, and for cameras with a positioning accuracy of 0.05 degrees, it can already be about 0.01 degrees.

The necessary effect is achieved at the cost of some loss of time required for intermediate positioning and control of the installation of the rotary mechanism in the intermediate position. However, for many tasks, accuracy, which is achieved by reducing positioning error, is a priority. In addition, modern cameras have a very high positioning speed, which reduces the time spent on additional positioning to a minimum.

A method of increasing the accuracy of positioning a PTZ camera, characterized in that:

Get the target camera rotation position.

The target position is understood as the camera position defined by the camera control program or the operator.

At least one intermediate position of the camera and its coordinates are determined based on the target rotation position of the camera.

Intermediate positions and their coordinates are determined by testing the camera, or based on ideas about the device of the rotary mechanism of the camera. During testing, to obtain optimal parameters of intermediate points, global optimization algorithms can be used [6, 7, 8].

Intermediate positions can be defined through absolute coordinates or through relative coordinates.

The number of intermediate positions depends on the required speed and positioning accuracy. The higher the positioning speed is required, the less intermediate positioning points should be, and in some cases one intermediate positioning point may be sufficient. It is also obvious from general considerations that increasing the number of intermediate points beyond a certain limit will not increase the accuracy of positioning.

The speed and accuracy of positioning can be set automatically or by the operator, depending on the specific application of the camera.

In some embodiments, the definition of intermediate points build an automatic procedure for calculating the resulting positioning accuracy at a given speed and vice versa, the resulting positioning speed at the required accuracy.

In a particular case, testing to determine intermediate positions can occur as follows.

1. Randomly determine the target position of the camera.

2. The camera is sent in an arbitrary direction and returned to the target position according to a specific positioning algorithm, measuring accuracy and positioning speed.

3. Change the positioning algorithm and again measure the accuracy of positioning.

When changing the positioning algorithm, it implies both the choice of another algorithm and the setting of the current one.

In this case, various strategies for changing the positioning algorithm can be used, including the simplest ones. Here is an example of a simple strategy for determining a positioning algorithm: one intermediate position is selected that differs from the target by the panoramic angle and angle of inclination by A. Then, using the global optimization algorithm [6, 7, 8], for a one-dimensional function, choose a value A that corresponds to the minimum positioning error.

To determine the positioning error, computer vision algorithms can be used, for example, which implement the selection of control points on two frames and the determination of the displacement of control points between frames. The search for reference points can be performed as indicated in the information source [5]. In the particular case, two images can be used - the first obtained at the target point at the initial moment of time, and the second obtained at the target position after positioning from an intermediate position. Next, these images are compared and a positioning error is determined.

As an algorithm for computer vision, the following algorithm may be used, but not limited to. For example, the shift can be determined by a correlation method, i.e. when determining the shift between two images, take the center of the first of the images and calculate the two-dimensional correlation of this element of one image with the second image, while the coordinates of the maximum correlation value will correspond to the position of the same section as the center of the first image in the second image, its offset relative to the center of the image and there will be a shift in the image, based on which you can determine the positioning error.

Algorithms of linear (nonlinear) filtering, searching for lines, searching for angles, borders, etc., can also be used as computer vision algorithms.

Consistently rotate the camera to the target position through the aforementioned intermediate positions.

After receiving the target position and calculating a set of intermediate positions, the camera is first rotated to the first intermediate position, then, if there is one, to the second, and so on. The final step is to rotate the camera to the target position.

In one embodiment, this technical solution can be made in the form of a system for increasing the accuracy of positioning of a PTZ camera, including: a PTZ camera, one or more data processing devices, one or more data storage devices, one or more programs, where one or more programs are stored on one or more data storage devices and executed on one or more data processing devices, and one or more programs includes the following instructions: what is the target position of the camera rotation, then determine at least one intermediate position of the camera and its coordinates based on data on the target position of the camera’s rotation, after which the camera is successively rotated to the target position through the aforementioned intermediate positions.

In some embodiments, said method is physically implemented on a camera.

According to FIG. 2, an exemplary system for implementing a technical solution includes a data processing device 200. The data processing device 200 may be configured as a client, server, mobile device, or any other computing device that interacts with data in a network-based collaboration system. In the most basic configuration, the data processor 200 typically includes at least one processor 201 and a data storage device 202. Depending on the exact configuration and type of computing device, system memory 202 may be volatile (e.g., random access memory (RAM, RAM)), non-volatile (for example, read-only memory (ROM)), or some combination thereof. A data storage device 202 typically includes one or more application programs 203 and may include program data 204. The present technical solution as a method described in detail above is implemented in application programs 203.

The data processing device 200 may have additional features or functionality. For example, the data processing apparatus 200 may also include additional data storage devices (removable and non-removable), such as, for example, magnetic disks, optical disks, or tape. Such additional storages are illustrated in FIG. 1 through removable storage 207 and non-removable storage 208. Computer storage media may include volatile and non-volatile, removable and non-removable media implemented in any way or using any technology for storing information, such as machine-readable instructions, data structures, program modules or other data. Storage device 202, removable storage 207, and non-removable storage 208 are examples of computer storage media. Computer storage media includes, but is not limited to, random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact ROM a disc (CD-ROM), universal digital disks (DVDs) or other optical storage devices, magnetic tapes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that may be used on to store the desired information and which can be accessed by the data processing device 200. Any such computer storage medium may be part of the device 200. The data processing device 200 may also include an input device (a) 205, such as a keyboard, mouse, pen , a voice input device, a touch input device, and so on. Output device (a) 206, such as a display, speakers, printer, and the like, may also be included in the device.

The data processing apparatus 200 comprises communication connections that allow the device to communicate with other computing devices, for example over a network. Networks include local area networks and wide area networks along with other large, scalable networks, including, but not limited to, corporate networks and extranets. Communication connection is an example of a communication environment. Typically, a communication medium can be implemented using computer-readable instructions, data structures, program modules or other data in a modulated information signal, such as a carrier wave, or in another transport mechanism, and includes any information delivery medium. The term "modulated information signal" means a signal, one or more of its characteristics are changed or set in such a way as to encode information in this signal. By way of example, but without limitation, communication media include wired media such as a wired network or a direct wired connection, and wireless media such as acoustic, radio frequency, infrared, and other wireless media. The term “machine-readable medium”, as used herein, includes both storage media and communication media.

EXAMPLES OF IMPLEMENTATION

Let there be a camera model that allows positioning at different speeds. It is also known that due to the inertia of the mechanism, an attempt to position the camera at a point (Pan, Tilt) leads to positioning at a point (Pan + Perr, Tilt + Terr), where (Perr, Terr) is a positioning error, the magnitude of which depends on the speed the camera’s movement to the positioning point, and the direction depends on the motion vector that the camera possessed when it reached the target position. Thus, positioning at maximum speed results in a maximum error. The direction of the error is also unpredictable, because it depends on the position in which the camera was before the start of positioning. Positioning at a low speed would reduce the error, but would significantly increase the positioning time if the initial position of the camera and the target position are significantly different. During testing the capabilities of the camera’s rotary mechanism, it turned out that in order to achieve the minimum positioning error, it is enough to determine one intermediate point that differs from the target position by one degree in the panoramic angle and angle of inclination and move from the intermediate position to the target one with a minimum speed. Moreover, an increase in the distance between the target position and the intermediate one does not lead to a decrease in the error, but, of course, leads to an increase in the positioning time. At the same time, a further reduction in distance leads to an increase in positioning error. Thus, a distance of one degree at two angles is optimal in terms of reducing positioning errors.

In the course of the experiments, it was found that when using consistent income by the panoramic angle, then by the angle of inclination, with a decrease in the distance from intermediate points to the target point, the error changes nonlinearly.

In FIG. 4 along the X axis, the distance in degrees from the intermediate position to the target, along the Y axis, positioning error (along the panoramic angle and angle of inclination) (averaged over many experiments).

As can be seen from FIG. 4, as the distance from the intermediate position to the target decreases, the positioning error begins to increase (repeatability decreases), and this is due to the peculiarities of the mechanism of a specific camera model.

In this case, a point with a coordinate along X = 0 corresponds to the absence of the use of intermediate points.

As seen in FIG. 4, even the introduction of intermediate positions at a small distance from the target has already achieved a technical result, but the minimum error is achieved with an income angle of 10 degrees, and after it an increase in the income angle practically does not affect the error.

Thus, from the point of view of achieving a technical result, income by 10 and 20 degrees does not differ (the error is the same), but income by 10 degrees is preferable, because it uses less mechanical resource of the camera and it can be done faster (because the camera needs less distance to scroll the lens).

The following is an example implementation of a technical solution (Fig. 3), in which there is one intermediate point (Pan - 1 °, Tilt - 1 °). The definition of the intermediate position (Pan - 1 °, Tilt - 1 °) is given in the absolute coordinates of the camera, but can also be determined in relative coordinates. At the same time, if the initial position of the camera was (Fig, Tix), then the relative position of the intermediate point will be (Pan-1-Fig, Tilt-1-Tis) and will correspond to the offset by which the camera must be rotated to reach the intermediate position. Similarly, the coordinates of the target position relative to the intermediate position will be (1, 1).

Previously, get the target position (Pan, Tilt). Next, one intermediate position is determined (Pan - 1 °, Tilt - 1 °), after which the camera is rotated at maximum speed from the initial position to the intermediate. From the intermediate position, turn the camera to the target position with minimal speed.

The result of this sequence of actions will be a reduction in positioning error at the target point, since the movement into it was carried out at a minimum speed. At the same time, the positioning time will increase insignificantly, since with a minimum speed the camera moved only a short interval of the positioning trajectory, namely, the path of one degree along the panoramic angle and tilt angle. In addition, the direction of the reduced error will be the same, since the camera’s vector of movement at the moment of reaching the target position will be the same, which will allow translating the error into a systematic category, taking into account and thus completely leveling.

In an example implementation with two intermediate points, the target position is first obtained, after which two intermediate positions are determined. The first position (Pan - 1 °, Tilt - 1 °), the second position (Pan, Tilt - 1 °). Then they are positioned in the first intermediate position with maximum speed, after which they are positioned in the second intermediate position with minimum speed. As a result, they are positioned at the target position with minimal speed.

One skilled in the art will appreciate that the technical result is achieved using one intermediate point. As the number of intermediate points increases, the positioning error decreases.

In addition to the technical result described for the method with one intermediate point, this method results in a further reduction in positioning error due to the fact that in the last two steps the movement is carried out only in one of the angles and there is no error associated with inaccurate synchronization of the positioning mechanism drives movement along each of the corners.

The present detailed description is made up of various non-limiting and exhaustive embodiments. At the same time, for specialists having an average level of competence in the considered field of technology, it is obvious that various replacements, modifications or combinations of any of the embodiments disclosed herein (including partially) can be reproduced within the scope of this technical solution. Thus, it is understood and understood that the present description of the technical solution includes additional embodiments, the essence of which is not set forth here in an explicit form. Such embodiments may be obtained, for example, by combining, modifying, or transforming any actions, components, elements, properties, aspects, characteristics, limitations, etc., related to the embodiments presented herein and not being restrictive.

USED SOURCES

1. “Keeping a Pan-Tilt-Zoom Camera Calibrated)), authors: Ziyan Wu, Richard J. Radke, published: IEEE Trans. Pattern Anal. Mach. Intell - 2013.

2. S.N. Sinha and M. Pollefeys. Pan-tilt-zoom camera calibration and high-resolution mosaic generation. Computer Vision and Image Understanding, 103 (3): 170-183, Sept. 2006.

3. M. Sarkis, C. Senft, and K. Diepold. Calibrating an Automatic Zoom Camera With Moving Least Squares. IEEE Transactions on Automation Science and Engineering, 6 (3): 492-503, July 2009.

4. Internet resource: http://www.aktivsb.ru/article-info1052.html

5. Computer Vision A Modern Approach Computer Vision: A Modern Approach Authors: David A. Forsyth, Jean Pons. Translators: A. Nazarenko, I. Doroshenko. Languages: Russian Publisher: Williams ISBN 5-8459-0542-7, 0-13-085198-1; 2004 year

6. Strongin R.G. Numerical methods in multiextremal problems. "Optimization and investigation of operations", The main edition of the physical and mathematical literature of the publishing house "Science", M., 1978, 240 pp.

7. Papadimitriou X., Steiglitz K. Combinatorial optimization: Algorithms and complexity. M .: Mir, 1985.

8. Batishchev D.I. Genetic algorithms for solving extreme problems. Ed. Lvovich Ya.E .: Textbook. allowance. Voronezh, 1995, 64 p.

9. Internet resource: https://ru.wikipedia.org/wiki/Angle_Eilepa

Claims (21)

1. A method of reducing the positioning error of a PTZ camera, characterized in that:
- get the target position of rotation of the camera;
- determine at least one intermediate position of the camera and its coordinates based on the data on the target position of rotation of the camera;
- sequentially rotate the camera to the target position through the aforementioned intermediate positions. 2. The method according to claim 1, characterized in that at least one intermediate position and its coordinates are determined by testing the camera. 3. The method according to p. 1, characterized in that at least one intermediate position and its coordinates are determined on the basis of ideas about the method of operation and the characteristics of the rotary mechanism of the camera. 4. The method according to p. 1, characterized in that the intermediate points can be determined through absolute coordinates or through relative coordinates. 5. The method according to p. 1, in which the number and coordinates of intermediate positions depend on the required speed and accuracy of positioning. 6. The method according to claim 1, in which the relative coordinates of the intermediate position are determined so that the intermediate position is as close as possible to the desired position. 7. The method of claim 2, wherein computer vision algorithms are used to determine positioning errors. 8. The method according to claim 1, characterized in that at least two intermediate positions are determined, wherein the first position differs by A in the panoramic angle and the angle of inclination from the desired position, and the second intermediate position in the panoramic angle coincides with the desired position, and by tilt angle differs by A. 9. The method according to p. 1, characterized in that at least two intermediate positions are determined, the first position being different by A in the panoramic angle and the angle of inclination from the desired position, and the second intermediate position in the angle of inclination coincides with the desired position, and The panoramic angle differs by A. 10. The method according to p. 1, characterized in that the speed of movement from the starting point to the first intermediate, between the intermediate and between the last intermediate point and the target point can be different. 11. The PTZ camera positioning error reduction system includes:
- positioned remotely controlled PTZ camera;
- one or more data processing devices;
- one or more data storage devices;
- one or more programs,
i. where one or more programs are stored on one or more data storage devices and executed on one or more data processing devices, and one or more programs includes the following instructions:
- get the target position of rotation of the camera;
- determine at least one intermediate position of the camera and its coordinates based on the data on the target position of rotation of the camera;
- sequentially rotate the camera to the target position through the aforementioned intermediate positions. 12. The system for reducing the positioning error of a PTZ camera according to claim 11, in which the data processing device includes an instruction receiving unit and a control unit configured to display information in the data output unit. 13. The system for reducing the positioning error of a PTZ camera according to claim 11, in which at least one intermediate position and its coordinates are determined by testing the camera. 14. The system for reducing the positioning error of the PTZ camera according to claim 11, in which at least one intermediate position and its coordinates are determined based on ideas about the method of operation and the characteristics of the rotary mechanism of the camera. 15. The system for reducing the positioning error of the PTZ camera according to claim 11, in which the intermediate points can be determined through absolute coordinates or through relative coordinates. 16. The system for reducing the positioning error of a PTZ camera according to claim 11, in which the number and coordinates of intermediate positions depend on the required speed and positioning accuracy. 17. The system for reducing the positioning error of the PTZ camera according to claim 11, in which the relative coordinates of the intermediate position are determined so that the intermediate position is as close as possible to the desired position. 18. The system for reducing the positioning error of a PTZ camera according to claim 11, wherein computer vision algorithms are used to determine the positioning error. 19. The system for reducing the positioning error of the PTZ camera according to claim 11, wherein at least two intermediate positions are determined, the first position being different by A in the panoramic angle and the angle of inclination from the desired position, and the second intermediate position in the panoramic angle coincides with the required position, and the angle of inclination differs by A. 20. The system for reducing the positioning error of the PTZ camera according to claim 11, wherein at least two intermediate positions are determined, the first position being different by A in the panoramic angle and the angle of inclination from the desired position, and the second intermediate position in the angle of inclination coincides with the desired position, and in the panoramic angle differs by A. 21. The system for reducing the positioning error of the PTZ camera according to claim 11, in which the speed of movement from the starting point to the first intermediate, between the intermediate and between the last intermediate point and the target point can be different.

Images