RU2686020C1 - Method of object detection and recognition - Google Patents

Abstract

FIELD: security monitoring.SUBSTANCE: invention relates to methods of remote security monitoring of routes on the ground and can be used in case of joint use of television camera (TV camera) and passive optical-electronic means of detection (DM), controlling its operation modes. Method consists in deployment of TV camera and DM on section of area with bend of track; unfolding the DM on the outer side of the bend of the track so that the axis of its ZO lies on the straight section of the track, and at least 2/3 of the length of the ZO covers the section of the track; installation of a TV camera opposite to a straight section of the path at a distance which provides the size of the image of the object relative to the size of the entire frame of not less than 50 %; crossing with optical axis of TV-camera the section of trail closed by DM ZO; detecting a DM of an object moving through the bend of the track and transmitting a control command to the camera to switch it on to transmit a video image for a predetermined time interval to an information collection and processing system (ICPS); operator's recognition of an object on a video image on a ICPS monitor.EFFECT: technical result consists in providing conditions for successful recognition by an operator of an object on its video image using only one detection means for controlling a television camera.1 cl, 8 dwg

Description

The invention relates to methods for remote security monitoring of traffic routes on the ground and can be used in the case of joint use of a television camera (TV camera) and a passive optical electronic detection device (CO) that controls its operation modes when continuous transmission of a video image with a TV camera, feeding on an autonomous source is technically impractical.

One of the tasks of remote security monitoring of traffic routes on the ground is the recognition of objects detected by CO, which consists in determining the class of such objects [1]. As a rule, this problem is solved with the help of TV cameras. In practice, reconnaissance and signaling complexes (RSK) are widely used for security monitoring of the routes of violators on the ground with the transfer of information about the state of the radio equipment and video from TV cameras via radio [2, 3]. The characteristics of the TV cameras used for the recognition of objects in the RSK are known (Fig. 1). Taking into account the high power consumption of TV cameras in the video transmission mode and the limited capacity of the used self-contained power sources, their continuous operation time is no more than 12 hours, and when the infrared illuminator is turned on no more than 6 hours (Fig. 1). At the same time, the tactics of using the tools that make up the RSK means that these tools work without changing the power source for up to several months, so most of the time the TV camera is in standby mode, its inclusion in the video transmission mode occurs on the control command when it detects an object [2, 3].

TV cameras are installed on the most likely routes of violators - roads (paths) [4]. The capacity of the RAC radio network is limited to several dozens of means, so the decision maker on the deployment of the RGC is forced to rationally use the CO, and as a rule, it can use no more than one CO to control the TV camera [1, 2].

DGCs are deployed on an unprepared engineering area, on which artificial lighting is not used at night, and barriers are not erected to deter the actions of objects [2, 3]. Therefore, the following conditions lead to a decrease in the probability of successful recognition of objects on the ground:

- low illumination in the surveillance area of TV cameras, especially in the dark and in conditions of limited visibility (rain, snow, fog);

- free, without delay, the movement of the object on the ground, and therefore its short duration of stay in the zone of visibility of TV cameras.

According to the widely used method of assessing the effectiveness of security television, for an operator to successfully recognize an object detected on the ground based on its video image, it is necessary to simultaneously fulfill the following video image requirements:

- the size of the image of the object must be at least 50% relative to the size of the entire frame [5, 6];

- the video image should be transmitted during the entire time the object is in the observation zone (t, s).

The required size of the image of the object on the monitor screen is provided by the distance between the TV camera to the part of the path along which the object moves. For TV cameras from the RAC (matrix format 1/2 "and a focal length of 20 mm), this distance is no more than 10-15 m (Fig. 2). The value of the distance between the TV camera to the part of the path along which the object is moving is especially important in the dark when the visibility conditions are limited by the range of the built-in IR illumination (not more than 20 m).

Video transmission during the whole time the object is in the surveillance zone can be provided either by continuous operation of the TV camera (which is not acceptable for TV cameras as part of the RSK) or by its timely activation from the control command from the CO.

Taking into account the characteristics of TV cameras used as part of the RSK, and finding them in standby mode before sending a command to switch on from CO, to create conditions for successful recognition by an operator of an object on the ground by its video image, it is necessary to ensure:

- the distance from the TV camera to the part of the path along which the object moves no more than 10-15 m (Fig. 2);

- timely turn on the TV camera (before the object enters the surveillance zone).

There are several ways to detect and recognize an object that differ from each other in the degree of compliance with the required conditions and the cost of achieving them [1].

The known method consists in orientation of the TV camera perpendicular to the controlled direct part of the trail at a distance of 80 m and the operation of the TV camera most of the time in standby mode; the installation of a single CO outside the surveillance zone of a TV camera near its boundary, so that the detection zone (AO) would intersect the middle of the section of the pathway limited to the observation zone; detection of CR of an object moving along the path when it crosses the AOR and sending control commands to the TV camera for its inclusion in the video transmission mode; transmitting a video image of a TV camera during a predetermined time interval to the system for collecting and processing information; recognition by the operator of the object on the video image on the monitor of the system for collecting and processing information (Fig. 3).

Another known method is to orient the TV camera perpendicular to the controlled direct path section at a distance of 15 m and to operate the TV camera most of the time in standby mode; the installation of two COs outside the surveillance area of the TV camera, so that an object moving along the path in any direction could be detected by one of the COs before entering the surveillance area at a distance that the intruder would pass during the time the TV camera was turned on; the detection by one of the CR of an object moving along the path when it crosses the AOR and the control command is transmitted to the TV camera for its inclusion in the video transmission mode; transmitting a video image of a TV camera during a predetermined time interval to the system for collecting and processing information; recognition by the operator of the object on the video image on the monitor of the system for collecting and processing information (Fig. 4).

In the first mentioned method, the necessary conditions for successful recognition by the operator of the detected object by its video image are not met (Fig. 3, 5):

- the distance from the TV camera to the path section along which the object (S) moves is several times greater than the required one (no more than 10–15 m) and is up to 80 m (Fig. 2);

- The TV camera is turned on only after passing the object to the middle of its surveillance zone.

The second specified method provides the necessary conditions for successful recognition by the operator of the detected object by its video image:

- the distance from the TV camera to the part of the path along which the object moves (S) does not exceed the required (no more than 10-15 m);

- The TV camera is turned on before the object appears in its surveillance area. However, for control of TV camera operation modes

two WITH (Fig. 4, 5).

Also, in both methods, prior to the deployment of JI, an engineering preparation of a site is required, including work on cleaning the sector in which its AOR will be located, from vegetation: cutting down trees and shrubs, mowing up high grass. These works are labor-intensive, for example, cutting a hardwood tree with a diameter of 40 cm flush with the ground requires up to 0.41 people * hour, over 40 cm - up to 0.511 people * hours. [7]

The aim of the invention is to provide conditions for successful recognition by the operator of the object by its video image by installing a TV camera opposite the path section along which the object moves, at a distance of no more than 10-15 m, timely (before the object enters the monitoring area) turning on the TV camera using only one detection tool to control the TV camera and reduce labor costs for engineering preparation of a site.

To achieve this goal, a method for detecting and recognizing an object has been developed, which consists in deploying CO from the outside of the bend of the path so that its axis of the AOR lies on the straight part of the path and at least 2/3 of the length of the AOR covers the part of the path; installing a TV camera opposite the straight path section at a distance that provides the image size of an object relative to the entire frame size of at least 50%; the intersection of the optical axis of the TV camera in the middle of the section of the path closed by the ZO SB; TV-camera operation during the standby mode; detection of CR of an object moving along the path when it crosses the AOR and sending control commands to the TV camera for its inclusion in the video transmission mode; transmitting a video image of a TV camera during a predetermined time interval to the system for collecting and processing information; recognition by the operator of the object on the video image on the monitor of the system for collecting and processing information (Fig. 5, 6).

As a rule, a path that passes through a terrain is not absolutely straight, one of the most frequently encountered elements of the road network is a bend of the road or path (arc-shaped turn, curvature consisting of two adjacent straight lines connected by a circular curve) [8].

CO is deployed from the outside of the bend of the path in such a way that the axis of its AOR lies on the straight part of the path and at least 2/3 of the length of the AOR covers the part of the path. The lack of vegetation on the trail reduces the labor costs of engineering preparation of the site for the deployment of JI.

A TV camera is installed opposite the straight path section at a distance of 10-15 m, providing the image size of the object relative to the frame is at least 50%.

The optical axis of the TV camera monitoring zone is located in the middle of the trail section, closed by the AOR. Due to the proposed deployment scheme of the CO and the TV camera, in either of the two directions of movement along the path, the object is first detected by the CO, and then it enters the surveillance zone of the TV camera. After crossing the AOR, the object passes the distance AB or ED, depending on the direction of movement, which is not less (Fig. 6):

where AB (ED) is the distance that the object travels from the moment of crossing the ZO SO to its entry into the TV camera monitoring zone, m;

AU (EC) - distance from the intersection point of the AOR by the object to the intersection point of the optical axis of the TV camera with the path, m;

BD - width of the TV camera monitoring area at a distance of 10-15 m, m.

Taking into account the fact that the width of the TV camera monitoring area (BD) at a distance of 10-15 m is 5 m (Fig. 2, 6), the distance from the intersection point of the DZ object to the intersection point of the optical axis of the TV camera with the AC path (EU ) - not less than 17 m (with the AOR length - 50 m), the numerical value of the distance that the object travels from the moment of crossing the AOR SO to its entry into the TV camera monitoring zone will be:

The time from the moment of detecting a CO object to entering the surveillance area of a TV camera is:

where T is the time from the moment of the detection of the object CO to its entry into the surveillance zone of the TV camera, s;

V _MAX - the maximum possible speed of movement of the intruder for this part of the trail, m / s.

The dependence of the speed of movement of a person on the terrain conditions is practically determined, known and confirmed on the basis of experimental data (Fig. 7) [9]. The maximum speed of the intruder can be 1.39 m / s. Then the numerical value of the time from the moment of detecting the CO object to entering the surveillance zone of the TV camera will be:

In RAC, the inclusion of a TV camera in the video transmission mode by a CO command occurs with a time delay of up to 5 seconds from the moment the object is detected by the CO (t _C ), taking into account formula 4:

where t _C - the possible time delay of the transition of the TV camera to the video transmission mode, p.

That is, according to formula 5, the TV camera is turned on in a timely manner - before the object enters the surveillance zone.

The pre-set time interval of the TV camera operation in the video transmission mode is taken with a safety factor of 1.2 and is calculated by the expression:

where t _P - the time interval of the TV camera in the mode of video transmission, s;

AE - the length of the section of the trail, closed ZO, m;

V _MIN - the minimum speed of movement of the intruder for this part of the path, m / s.

The developed method includes two stages: preparatory and main.

The preparatory stage includes:

1. The installation of the detection means 1 and the television camera 5 on the bend of the path according to the developed scheme (Fig. 6).

2. Deployment of the information collection and processing system 3 (FIG. 8).

3. Setting the time interval of operation of the television camera 5 in the video image transmission mode (formula 6).

The main stage begins with the movement of the object through the selected bend of the path and the intersection of the AOR. It includes:

1. Detection by means of detecting an object 1 crossing its AOR, by transmitting by means of detecting a control command to switch on the television camera 5 to the video transmission mode.

2. Transition of the television camera 5 to the video transmission mode on the system 3 of information collection and processing.

3. Reception of the video image by the system 3 of collecting and processing information and outputting it to the monitor 4.

4. Video analysis and recognition of an object moving along a path by an operator 6.

The invention is illustrated graphic materials, which are presented on:

- FIG. 1 - the main characteristics of the television cameras that make up the reconnaissance-signaling complex;

- FIG. 2 - a table of the dependence of the image size of an object from a distance to a television camera;

- FIG. 3 is a diagram of the deployment of a single detection device and a television camera in a known manner;

- FIG. 4 is a diagram of the deployment of two detection means and a television camera in a known manner;

- FIG. 5 is a table comparing the characteristics of known and proposed methods;

- FIG. 6 is a diagram of the deployment of a detection device and a television camera in the proposed method;

- FIG. 7 - a table of the dependence of the speed of movement of a person on the terrain conditions;

- FIG. 8 is a block diagram of the relationship of the devices used, when implementing a method for detecting and recognizing an object.

The technical result consists in providing conditions for successful recognition by the operator of the object by its video image by installing a TV camera opposite the path section along which the object moves, at a distance of no more than 10-15 m, timely (before the object enters the surveillance zone) turning on the TV camera using only one detection tool to control the TV camera and reduce labor costs for engineering preparation of a site.

Information sources

1. Marshalov, T.A. Technical means of border protection: a textbook / TA. Marshalov, A.V. Gustov, I.M. Potapov. - Kaliningrad: KPI FSB RF, 2009. - 568 p.

2. Radio alarm television PC-TV. The operation manual of the PMEC. 424242. 9000 RE - M .: Polyus-ST, 2016. - 28 p.

3. Electronic resource - www.radiobarier.ru

4. Shumov, V.V. The use of mathematical methods and models to justify decisions on the protection of the state border: Scientific and practical guide. - Part 2. - M .: Enlightenment, 1996. - 196 p.

5. Gedzberg, Yu.M. Security TV. - M .: Hotline - Telecom, 2005. - 312 p.

6. Gaunt, A. Practical guide CCTV / A. Gaunt. - M .: Spetskniga, 2006. - 80 p.

7. Typical standards for felling, standard time for felling forest care in lowland conditions. - M .: 1999. - 83 p.

8. Psarev, A. A. Military Topography: Textbook. - M .: Military Publishing, 1986. - 384 p.

9. Balenko, NE Survival School. - M .: 1994. - 140 p.

Claims (1)

A method for detecting and recognizing an object, which consists in orienting a television camera to a controlled part of the path, operating a television camera in the main time in standby mode; detection by means of detecting an object moving along a path when it crosses the detection zone and transmitting a control command to the television camera to switch it on to the video transmission mode; transmission of video by a television camera during a predetermined time interval to the system for collecting and processing information; recognition by the operator of the object on the video image on the monitor of the information collection and processing system, characterized in that the detection means is deployed from the outside of the bend of the trail so that the axis of its detection zone lies on the straight portion of the trail and at least 2/3 of the detection zone length covers the trail portion ; a television camera is installed opposite the straight path section at a distance of 10–15 m, providing the image size of an object moving along the path relative to the size of the entire frame at least 50%; the optical axis of the television camera passes in the middle of the section of the path covered by the detection area of the detection means.

Images