FR3091117A1 - VIDEO SURVEILLANCE MODULE AND SYSTEM, PARTICULARLY FOR SITE SECURITY - Google Patents

Abstract

VIDEO SURVEILLANCE MODULE AND SYSTEM, PARTICULARLY FOR SITE SECURITY The module (MVS) comprises at least one surveillance camera (CAM) and first, second and third image analysis units (UA1, UA2, UA3), the first image analysis unit being configured to detect an intrusion by modeling an area in the images and to emit a first primary alert signal (AL1), the second image analysis unit being configured to detect a movement in modeling a grid in the images and for emitting a second primary alert signal (AL2), and the third image analysis unit being configured to detect a line crossing by modeling an analysis line in the images and for issue a third primary alert signal (AL3). According to the invention, the module also includes a combinational logic unit (LOG) producing a secondary alert signal (ALS) from a logical combination of type "AND" between at least two of the primary alert signals. Figure for the abstract: Fig. 1

Description

Description

Title of the invention: VIDEO SURVEILLANCE MODULE AND SYSTEM, PARTICULARLY FOR SITE SECURITY

The present invention relates generally to the field of security of goods, infrastructure and people. More particularly, the invention relates to a surveillance module comprising at least one surveillance camera and means for image analysis and generation of alerts. The invention also relates to a video surveillance system comprising at least one module for monitoring the invention.

The need for protection against multiple malicious acts, such as theft and damage, economic and industrial espionage, as well as attacks against people, has led in recent years to the development of surveillance techniques. sites using video surveillance.

[0003] Thus, for example, video surveillance systems using thermal cameras in association with image analysis means allow continuous monitoring, day and night, of large areas, such as exterior infrastructures, car parks. , industrial sites and others.

[0004] Real time panoramic images are delivered by the cameras and are processed by digital image analysis means. Image analysis processing is performed automatically on the flow of video images and makes it possible to detect intrusion events, or attempted intrusion, in a monitored site.

When an intrusion attempt is detected, the monitoring system controls different types of action depending on its configuration, such as raising an alert signal to a remote monitoring center, sending a text message (for example, SMS type) to warn people concerned, the activation of a siren, a light projector, etc., aimed at deterring intruders.

In known video surveillance systems, image analysis typically allow three detection functions for the generation of alert signals, namely, an intrusion detection function, a motion detection function and a line crossing detection function. The intrusion detection function detects a change in a predefined area in the image. The motion detection function detects movement in a predefined area in the image. The line crossing detection function detects a crossing of a predefined line in the image. Each of these three detection functions produces its own warning signals. All the warning signals are sent back by the system to a monitoring center where an operator checks the validity of the alert triggers on the images and validates, for example, a request for intervention by the forces of order or intervention of a security guard.

The image analysis used in video surveillance systems uses image processing, such as segmentation by regions, contours and pixel thresholds, which do not guarantee total reliability of the detections. Large numbers of false alarms can be generated by various factors, such as changing weather conditions, light beams from motor vehicles, animals, insects, vegetation and the like, and result in additional operating costs. In addition, the accuracy in detecting real intrusions is around 80% in current systems and therefore presents a significant margin for improvement.

It is desirable to provide a solution to the aforementioned drawbacks of the prior art by providing a module and a video surveillance system offering high performance in terms of the number of false alerts and detection accuracy.

[0009] According to a first aspect, the invention relates to a video surveillance module comprising at least one surveillance camera and first, second and third image analysis units, the first image analysis unit being configured for detecting an intrusion by modeling an area in images supplied by the camera and for emitting a first primary alert signal, the second image analysis unit being configured to detect a movement by modeling a grid in images supplied by the camera and to emit a second primary alert signal, and the third image analysis unit being configured to detect a line crossing by modeling an analysis line in images supplied by the camera and to emit a third signal primary alert. In accordance with the invention, the video surveillance module also comprises a combinational logic unit producing a secondary alert signal from a logical combination of “AND” type between at least two primary alert signals among the first, second and third primary warning signals.

According to a particular embodiment, the video surveillance module comprises at least one volumetric type surveillance sensor and the combinational logic unit produces the secondary alert signal from a logical combination of “AND” type between at least two primary alert signals from the first, second and third primary alert signals, and an auxiliary primary alert signal produced from the volumetric type monitoring sensor.

According to a particular characteristic, the volumetric type surveillance sensor is of the infrared, microwave and / or ultrasonic type.

According to another particular characteristic, the combinational logic unit produces the secondary alert signal from a logical combination of “AND” type between the first and second primary alert signals and the primary alert signal auxiliary.

According to yet another particular characteristic, the surveillance camera is a thermal type camera.

According to yet another particular characteristic, the first, second and third image analysis units and the combinational logic unit are implemented in an electronic control unit in the form of at least one software module.

According to another aspect, the invention also relates to a video surveillance system comprising video surveillance means by camera, video image storage means, surveillance sensors, an electronic control unit, a transmitter alert, audible and / or light alarm means and uninterruptible power supply means. According to the invention, the video surveillance system comprises at least one monitoring module as briefly described above.

According to yet another particular characteristic, the alert transmitter is arranged to receive the secondary alert signal produced in the monitoring module and to transmit the latter to a remote monitoring center through a radio transmission network.

According to yet another particular characteristic, the video surveillance system also comprises an internal data communication network supported by a router which is connected to the Internet network through a mobile telephone network.

According to yet another particular characteristic, the means for storing video images comprise a "SSD" type hard disk.

Other advantages and characteristics of the present invention will appear more clearly on reading the detailed description below of several particular embodiments of the invention, with reference to the accompanying drawings, in which:

[Fig.l]

The Eig.l is a simplified block diagram of a particular embodiment of a video surveillance module according to the invention.

[Fig.2]

Eig. 2 shows an example of the implementation of an intrusion detection function in the video surveillance module of Eig.l.

[Fig.3]

Eig.3 shows an example of implementation of a motion detection function in the video surveillance module of Eig.l.

[Fig.4]

Fig. 4 shows an example of implementation of a line crossing detection function in the video surveillance module of Fig. L.

[Fig.5]

Fig.5 is a simplified block diagram of a particular embodiment of a video surveillance system according to the invention.

Referring to Figs.l and 2 to 4, there is now described by way of example a particular MVS embodiment of a video surveillance module according to the invention. Note that the MVS video surveillance module is designed to be integrated into a video surveillance system, possibly with several other similar surveillance modules.

In this exemplary embodiment, as visible in FIG. 1, the MVS video surveillance module essentially comprises a surveillance camera CAM, three image analysis units UA1, UA2 and UA3 and a logic logic unit LOG.

Note that the MVS module will typically be powered by uninterruptible power supply means (not shown). These power supply means will for example include a rechargeable electric battery housed in a housing of the CAM camera, as well as an uninterruptible power supply unit connected to the power supply sector. The CAM camera will be connected to the UA1, UA2 and UA3 image analysis units via a wired or wireless data communication network (not shown). Thus, for example, the data communication network could be a radio network of the WiFi type (registered trademark) based on a router connected to the Internet network through a mobile telephone network of the GPRS, EDGE, 3G, 4G type (registered trademarks). ) or others.

As shown in FIG. 1, the surveillance camera CAM, for example of the thermal type, delivers video images VD which are supplied as input to each of the image analysis units UA1 to UA3. In other embodiments, the video surveillance module may include several cameras, for example, three cameras providing images with different angles of view to the image analysis units UA1 to UA3.

The image analysis units UA1 to UA3 perform different detection functions on the VD video images. Thus, the image analysis units UA1, UA2 and UA3 respectively execute an intrusion detection function F1, a motion detection function F2 and a line crossing detection function F3.

Referring also to FIG. 2 which illustrates an example of implementation of the intrusion detection function F1 on an image IM1, the intrusion detection function F1 is executed in a detection area of ZI intrusion which was modeled in the image.

The ZI intrusion detection zone is modeled by a closed contour CO formed by several border lines B1 to B4. The intrusion detection function F1 monitors the pixel changes in the border lines B1 to B4 in order to detect any possible intrusion, or attempted intrusion, in the zone ZI. A pixel thresholding technique is used to detect any significant change in pixels in and near the boundary lines B1 to B4. The intrusion detection function F1 detects a border line, among B1 to B4, on which a change has occurred but does not detect the position of the change. Any detected change inducing an incoming direction DE, that is to say, from the outside to the inside of the zone ZI, generates a primary alert signal ALI.

With reference also to Lig.3 which illustrates an example of implementation of the motion detection function E2 on an IM2 image, the motion detection function E2 is executed in a motion detection zone ZM.

The ZM area is represented by a GR motion detection grid which has been modeled in the image. The GR motion detection grid defines a plurality of small boxes CA. The motion detection function E2 monitors the significant changes of pixels in each of the boxes CA by a pixel thresholding technique. A movement and its direction are deduced from the detection of a progression of a change of pixels from an initial box to an adjacent box. Any movement detected from one box to another generates an AL2 primary alert signal.

With reference also to Eig.4 which illustrates an example of implementation of the line crossing detection function E3 on an IM3 image, the line crossing detection function E3 is executed in an area of l image in which an analysis line LA and a line crossing direction DE have been modeled.

The line crossing detection function E3 monitors the pixel changes on the LA analysis line, and in the vicinity thereof, in order to detect any possible crossing, or attempted crossing. A pixel thresholding technique is used to detect a change of pixels representative of a crossing, or attempted crossing, of the analysis line LA, in the direction DE defined, but does not detect the position of the change of pixels. Any crossing, or attempted crossing, detected in the defined direction DE generates a primary alert signal AL3.

As shown in Eig.l, the primary warning signals ALI, AL2 and AL3, produced by the image analysis units UA1, UA2 and UA3 are supplied to inputs el, e2 and e3 of l logic unit LOG, respectively. An auxiliary primary alert signal AL4 is supplied to an input e4 of the combinatorial logic unit LOG. The primary alert signals ALI, AL2 and AL3 and the auxiliary primary alert signal AL4 are binary signals in which a logic state "1" indicates an active alert and a logic state "0" indicates the absence of an alert.

The LOG unit performs a combinational logic function of the so-called "AND" type on the primary alert signals ALI, AL2, AL3 and the auxiliary primary alert signal AL4 applied to the inputs el, e2, e3 and e4. The LOG unit delivers a secondary alert signal ALS whose logic state depends on the result of the combinational logic function "AND" performed.

The secondary warning signal ALS is defined by the following combinatorial logic equality:

ALS = (IN1 + AL1). (IN2 + AL2). (IN3 + AL3). (IN4 + AL4)

With the symbols ". "And" + "which correspond to the logical functions" AND "and" OR "in combinatorial logic; and

With INI, IN2, IN3 and IN4 which are binary invalidation inputs which authorize (in the state "0"), or not (in the state "1"), the taking into account of the signals ALI , AL2, AL3 and AL4 corresponding to produce the outgoing ALS alert signal.

The truth tables (tables 1 and 2) of the logical functions "AND" and "OR" for two inputs El and E2 are recalled below.

[Tables 1]

Summer

E2 0 00

01

[Tables2]

Summer

E2 0 00

01

According to the invention, the secondary alert signal ALS is the outgoing alert signal delivered by the MVS module and which gives rise to an effective triggering of an alert. In addition to taking into account the primary warning signals ALI, AL2 and / or AL3 from the image analysis means UA1, UA2 and UA3 of the MVS module, one or more auxiliary primary warning signals, originating from a or several auxiliary surveillance sensors and represented by AL4 in this exemplary embodiment, may also be taken into account to produce the secondary alert signal ALS.

The tests carried out by the inventive entity have shown that taking into account an auxiliary primary alert signal AL4 delivered by a volumetric type monitoring sensor for obtaining the secondary alert signal ALS allows a significant improvement in performance in terms of the number of false alerts and detection accuracy.

Table 3 below shows, for reference, performances obtained with the El intrusion detection, F2 motion detection and F3 line crossing detection functions when these are used separately for produce warning signals.

[Tables3]

NFA P Fl 85 0.65 F2 42 0.58 F3 47 0.62

In Table 3 above, NFA and P correspond respectively to the number of false alerts and to the detection accuracy. The values indicated for NFA and P correspond to actual measurements taken over a day at a test site.

Table 4 below shows performances obtained in different configurations of the MVS module according to the invention. The auxiliary primary AL4 signal considered is provided by a volumetric sensor.

[T ableaux4]

INI IN2 IN3 IN4 ALS NFA P 1 0011 ALS = AL1.AL2 10 0.98 2 1001 ALS = AL2.AL3 8 0.92 3 0101 ALS = AL1.AL3 7 0.85 4 0001 ALS = AL1.AL2.AL3 5 0.95 5 0010 ALS = AL1.AL2.AL4 7 1 6 1000 ALS = AL2.AL3.AL4 7 0.92 7 0100 ALS = AL1.AL3.AL4 7 0.94 8 0000 ALS = AL 1. AL2. AL3. AL 4 7 0.98

In table 4 above, the values indicated for NFA and P correspond to real measurements taken over a day on a test site and with measurement conditions comparable to those of table 3.

The comparison of Tables 3 and 4 above shows the significant improvement in performance brought by the configurations recommended by the present invention, in terms of reducing the number of false alerts and increasing the detection accuracy.

In the configurations of lines 1 to 4 of table 4, which operate without a volumetric sensor (IN4 = "1"), the configuration of line 1 gives the best detection precision, namely P = 0.98, with a number of false NFA alerts = 10.

The configuration of line 1 requests the intrusion detection function F1 and the motion detection function F2 to produce a secondary alert signal ALS = AL1.AL2, that is to say, a signal ALS secondary alert which is only active (ALS = "1") if the F1 and F2 functions deliver active alert signals AL1 = "1" and AL2 = "1".

In the configurations of lines 5 to 8 of table 4, which operate with a volumetric sensor (IN4 = "1"), the configuration of line 5 gives the best detection precision, namely P = 1, with a number of false NFA alerts = 7.

The configuration of line 5 requests the intrusion detection function F1 and the motion detection function F2, like the configuration of line 1, as well as a volumetric sensor to produce the secondary warning signal ALS = AL1.AL2.AL4.

Referring now more particularly to Fig.5, it is described below, by way of example, a particular embodiment S VS of a video surveillance system according to the invention comprising a video surveillance module MVSb analogous to the MVS module described above.

In the SVS video surveillance system, the MVSb module comprises a plurality of CAMb surveillance cameras which deliver several VDb video image streams. The MVSb module also includes image analysis units UAlb to UA3b and a combinatorial logic unit LOGb which are analogous to the units UA1 to UA3 and LOG of the MVS module.

The UAlb, UA2b and UA3b image analysis units respectively produce primary alert signals ALlb, AL2b and AL3b which are supplied to respective inputs of the combinatorial logic unit LOGb.

In addition to the primary warning signals ALlb, AL2b and AL3b, the combinational logic unit LOGb also receives an auxiliary primary warning signal AL4b as input which is supplied by a volumetric sensor CV, or a set of volumetric sensors. The CV volumetric sensor is for example an infrared, microwave or ultrasonic type sensor. The combinatorial logic unit LOGb outputs a secondary alert signal ALSb produced by a logic function AND from the primary alert signals.

It will be noted that the image analysis units and the combinational logic unit will typically be carried out in the form of one or more software modules hosted in a memory of an electronic control unit ECU which manages the operation of the entire SVS system.

The SVS video surveillance system also includes here other detection sensors, such as for example one or more switches SW and a microphone MIC, for the generation of other auxiliary primary alert signals which may also be taken into account. by the system. An HP alarm speaker and a DIR flashing siren are also provided.

As also visible in Fig. 5, the SVS video surveillance system typically comprises a STOR video image storage unit, a ROUT router, a NET data communication network, a TRANS alert transmitter and a unit uninterruptible power supply ALIM.

The STOR video storage unit allows the recording of video image streams and may include, for example, an SSD type hard disk, for "Solid-State Drive" in English.

The ROUT router is for example a router connected to the Internet network through the GPRS, EDGE, 3G, 4G (registered trademarks) or other mobile telephone network, allowing remote data communication, for example with a data center. monitoring, for system control and viewing of video stream images. The ROUT router also supports the NET communication network of the system, in a wired or radio form.

The TRANS transmitter is typically a radiofrequency transmitter of a specialized alarm network ensuring the raising of alert signals, such as the ALSb signal, to a monitoring center, as well as possibly a transmission of text messages. .

The uninterruptible power supply unit ALIM is connected to the power supply sector and comprises an electric battery BAT coupled to an alternative power supply device which provides the power supply to the system in the event of a power failure.

Of course, the invention is not limited to the particular embodiments which have been described here by way of example. Those skilled in the art, depending on the applications of the invention, may make different modifications and variants falling within the protective field of the invention.

Claims (1)

Claims [Claim 1] Video surveillance module (MVS) comprising at least one surveillance camera (CAM) and first, second and third image analysis units (UA1, UA2, U A3), said first image analysis unit (UA1 ) being configured to detect an intrusion by modeling an area (ZI) in images (IM1) supplied by said camera (CAM) and to emit a first primary alert signal (ALI), said second image analysis unit (UA2) being configured to detect movement by modeling a grid (GR) in images (IM2) provided by said camera (CAM) and to emit a second primary alert signal (AL2), and said third analysis unit of images (UA3) being configured to detect a line crossing by modeling an analysis line (LA) in images (IM3) supplied by said camera (CAM) and to emit a third primary alert signal (AL3) , wherein said video surveillance module (MSV) also comprises a combined logic unit roof (LOG) producing a secondary alert signal (ALS) from a logical combination of type "AND" between at least two said primary alert signals among said first, second and third primary alert signals (ALI , AL2, AL3). [Claim 2] Video surveillance module according to claim 1, comprising at least one volumetric type surveillance sensor, in which said combinational logic unit (LOG) produces said secondary alert signal (ALS) from a logic combination of type "AND" between at least two said primary alert signals among said first, second and third primary alert signals (ALI, AL2, AL3), and an auxiliary primary alert signal (AL4) produced from said at least one volumetric type monitoring. [Claim 3] A video surveillance module according to claim 2, in which said volumetric type surveillance sensor is of the infrared, microwave and / or ultrasonic type. [Claim 4] A video surveillance module according to claim 2 or 3, in which said combinational logic unit (LOG) produces said secondary alert signal (ALS) from a logical combination of type "AND" between said first and second alert signals. primary (ALI, AL2) and said auxiliary primary alert signal (AL4). [Claim 5] CCTV module according to any one of the claims 1 to 4, wherein said at least one surveillance camera (CAM) is a thermal type camera. [Claim 6] Video surveillance module according to any one of claims 1 to 5, in which said first, second and third image analysis units (ALI, AL2, AL3) and said combinational logic unit (LOG) are implemented in an electronic unit control in the form of at least one software module. [Claim 7] Video surveillance system (SVS) comprising video surveillance means by camera (CAMb), video image storage means (STOR), surveillance sensors (CV, SW, MIC), an electronic control unit (ECU) , an alert transmitter (TRANS), audible (HP) and / or light alarm (GIR) means and uninterruptible power supply (ALIM) means, in which said video surveillance system comprises at least one module for surveillance (MVSb) according to any one of claims 1 to 6. [Claim 8] A video surveillance system according to claim 7, in which said alert transmitter (TRANS) is arranged to receive said secondary alert signal (ALSb) produced in said monitoring module (MVSb) and to transmit it to a monitoring center. remote through a radio transmission network. [Claim 9] Video surveillance system according to claim 7 or 8, also comprising an internal data communication network (NET) supported by a router (ROUT) which is connected to the Internet network through a mobile telephone network. [Claim 10] A video surveillance system according to any one of claims 7 to 9, in which said video image storage means (STOR) comprise a "SSD" type hard disk. 1/4