WO2021105211A1

WO2021105211A1 - Device and method for autonomously monitoring a level crossing

Info

Publication number: WO2021105211A1
Application number: PCT/EP2020/083368
Authority: WO
Inventors: Philippe LAVIRON; Ying Zhao
Original assignee: Thales
Priority date: 2019-11-27
Filing date: 2020-11-25
Publication date: 2021-06-03
Also published as: CN114929549A; FR3103442B1; EP4065448A1; FR3103442A1

Abstract

The invention relates to a device for autonomously monitoring a level crossing, the level crossing forming an intersection between a rail first way and a second way. The monitoring device comprises: - a video-recording unit comprising at least three monitoring units, each of the monitoring units being equipped with a visible camera and an infrared camera, - a digital processing unit configured to receive and process in real-time sequences of video recording delivered by the video-recording unit, and - a warning unit configured to trigger warnings of a risk of a collision involving a rail vehicle moving over the railway.

Description

DESCRIPTION

Title of the invention: Device and method for autonomous monitoring of a level crossing

Prior art

The present invention relates to a device and a method for monitoring a level crossing formed of an intersection between a constraining transport lane and a road or pedestrian lane.

[0002] The present invention applies in particular to railway networks comprising level crossings designating the intersection between a railway line and a road or pedestrian route in which the risks of collisions involving a railway vehicle moving on the railway line are considerable. Such risks of collision generally exist for unprotected level crossings which do not use any device, such as gatekeepers, capable of preventing access to the railway track for a time interval corresponding to the passage of the vehicle. railway. Access control devices to a railway line are conventionally equipped with traffic lights, gate guards and bells. However, such devices have high installation and maintenance costs. Moreover, their deployment is generally limited to level crossings with significant rail and road traffic. The deployment of such devices reduces the risk of collisions without being able to cancel them, as for example in situations in which a road vehicle or a pedestrian are stuck between barrier guards.

Whether it is a protected level crossing, that is to say provided with gatekeepers, or unprotected, the state of such a level crossing is generally unknown to the vehicle. train preparing to cross it. The rail vehicle's leeway in terms of braking distance is reduced as it approaches the level crossing. It is known practice to equip the level crossing with a video surveillance system connected to a data network capable of reporting to a processing center the instantaneous state of the level crossing. For example, in patent application CN108749862 A, a video surveillance system is proposed to trigger the opening and closing of a level crossing equipped with gate guards. Video surveillance-based solutions proposed in the state of the art are generally only suitable for level crossings protected by gatekeepers and require the setting up of a remote processing center to process the videos supplied by the surveillance cameras. However, sending such video contents requires a high transmission rate and low latency which are not guaranteed in current data networks.

[0004] Furthermore, unguarded level crossings generally have no means of alerting road vehicles and / or pedestrians of the approaching rail vehicle. The approach of a rail vehicle is noticeable visually or audibly only at short distances generally involving an inevitable collision due to the high speed of movement of the rail vehicle and the insufficient time available to the road vehicle or the pedestrian. to react. For example, for a rail vehicle traveling at a perceived speed of 160 km / h 50 meters from a level crossing, the occupant of the level crossing has only 2 seconds to react. The perception of the approach of a railway vehicle in the absence of dedicated devices is highly dependent on weather conditions, the level of lighting, the level of noise pollution, etc. It is known practice to deploy signaling beacons near the railway track and on both sides of the level crossing in order to detect the presence of a railway vehicle and to trigger an alert of the approaching railway vehicle. Such an alert may appear near the level crossing by audible signals or traffic lights. Such solutions are expensive to set up and maintain. Moreover, their failure rate is generally high due to the short lifespan and the low resistance to weather conditions of the various electronic components used.

[0005] There is therefore a need for a device and a method for monitoring an improved level crossing capable of detecting the approach of a railway vehicle and of alerting road vehicles and / or pedestrians located near the Railroad Crossing.

[0006] A first document [US 2014/339374] discloses a system for monitoring a level crossing. A second document [US 2014/218482] discloses a recognition system capable of detecting a railway vehicle approaching a level crossing. General definition of the invention

The invention provides a device for autonomous monitoring of a level crossing, the level crossing forming a cross between a first railway line and a second track. Advantageously, the device comprises:

- a video recording unit comprising at least three monitoring units, each of the monitoring units being equipped with a visible camera and an infrared camera,

- a digital processing unit configured to receive and process in real time video recording sequences supplied by the video recording unit, and

- an alert unit configured to trigger alerts of risk of collision involving a railway vehicle moving on the railway track.

[0007] In one embodiment, the autonomous monitoring device may comprise an alert unit configured to trigger an alert intended for a rail vehicle, in response to the detection of the presence of an object in the passage to level, likely to cause a collision involving the railway vehicle.

[0008] Advantageously, the alert unit can also be configured to trigger an alert for a user of the second track, the alert notifying the passage of the rail vehicle.

[0009] As a variant, the monitoring units can comprise:

- At least one wide viewing angle monitoring unit capable of covering a risk area corresponding to the level crossing, and

- At least two low viewing angle monitoring units capable of detecting the approach of the railway vehicle towards the risk area.

[0010] In one embodiment, the autonomous monitoring device may further comprise an electric power supply unit capable of continuously supplying over time an electric current suitable for the operation of the units forming the monitoring device.

[0011] In one embodiment, the autonomous monitoring device may comprise an automatic archiving unit configured to save in a non-volatile memory video recording sequences supplied by the video recording unit, the trigger and the backup settings video recording sequences being controlled by the digital processing unit.

[0012] As a variant, the digital processing unit may include resources suitable for the execution of one or more algorithms for processing video sequences.

[0013] There is also provided a method for monitoring a level crossing formed by a crossing between a railway track and a road or pedestrian track, a railway vehicle being capable of moving on the railway track, configured to trigger alerts. Advantageously, the method comprises the steps consisting in:

- Film a level crossing using a video recording unit comprising at least three monitoring units, each of the monitoring units being equipped with a visible camera and an infrared camera,

- Process in real time video recording sequences supplied by the video recording unit, and

- Raise collision risk alerts involving the rail vehicle.

[0014] According to embodiments of the invention, the step of real-time processing of video recording sequences can comprise the sub-steps consisting of:

- Receive a video sequence comprising a video recording taken by a visible camera and another video recording taken by an infrared camera,

- Measure the level of brightness in the video recording taken by the visible camera,

- Compare the measured luminosity level with a luminosity threshold,

- Process the recording taken by the visible camera if the measured luminosity level is greater than or equal to the luminosity threshold,

- Process the recording taken by the infrared camera if the measured brightness level is below the brightness threshold.

Brief description of the figures

Other characteristics, details and advantages of the invention will emerge on reading the description given with reference to the accompanying drawings given by way of example and which represent, respectively: [0016] [Fig.1] Figure 1 shows the architecture of a monitoring device, according to one embodiment of the invention;

[0017] [Fig. 2] Figure 2 shows the detailed architecture of a video recording unit, according to one embodiment of the invention;

[0018] [Fig. 3] Figure 3 shows the architecture of a monitoring device, according to another embodiment of the invention;

[0019] [Fig. 4] Figure 4 shows an unprotected level crossing in which a monitoring device is deployed, according to one embodiment of the invention;

[0020] [Fig. 5] Figure 5 shows another unprotected level crossing in which a monitoring device is deployed, according to another embodiment of the invention;

[0021] [Fig. 6] Figure 6 shows a method of monitoring an unprotected level crossing according to one embodiment of the invention; and

[0022] [Fig. 7] FIG. 7 represents the sub-steps implemented to process a video sequence according to one embodiment of the invention.

detailed description

Figure 1 shows the general architecture of a monitoring device 100 of a level crossing, according to one embodiment of the invention. As shown in Fig. 1, the monitoring device 100 includes a video recording unit 101, a digital processing unit 102 and an alert unit 103. The video recording unit 101 is configured to perform video recording. previously identified areas of interest. Such areas of interest may include a risk zone 204 corresponding to a level crossing forming a crossing between a first railway line 202 and a second track 203. The level crossing is thus located at the intersection of the railway line 201 containing the railway line 202 and the second track 203. The railway line 202 delimits a traffic lane for a railway vehicle. On the second track 203, one or more users can circulate, such as a vehicle or a pedestrian. The second lane 203 can thus be a road or a pedestrian lane. The zones of interest can include two supervision zones 2051, 2052 covering the railway tracks 202. The supervision zones 2051, 2052 can extend on either side of the risk zone 204.

Figure 2 details the architecture of a video recording unit 101 may advantageously include three monitoring units 1011, 1012, 1013 configured to perform video recordings (that is to say film) the area to risk 204 and the supervision zones 2051, 2052. The supervision units 1011, 1012, 1013 of the video recording unit 101 can be configured to be active, continuously over time. Each monitoring unit implemented in the video recording unit 101 can incorporate two types of cameras: a visible camera 101a and an infrared camera 101b. The visible camera 101a and the infrared camera 101b of each monitoring unit can operate simultaneously so as to provide two video sequences representing each scene filmed. Alternatively, the surveillance unit can be configured to alternately operate the visible camera 101a and the infrared camera 101b so as to have a single video sequence for each scene filmed. For example, the monitoring unit can only use the infrared camera 101b when the brightness and / or the weather conditions are degraded. Such an operation has the advantage of being economical in terms of energy consumed and computing power.

The digital processing unit 102 can be configured to receive in real time the video sequences supplied by the monitoring units 1011, 1012,

1013 of the video recording unit 101. The frame rate in number of frames per second at which the video recordings are received by the digital processing unit 102 can be adjusted. In some embodiments, the rate of reception of the video recordings can be greater than ten images per second. The digital processing unit 102 can be placed in the immediate vicinity of the video recording unit 101. Such proximity makes it possible to reduce the response time necessary for the monitoring device 100 to trigger alerts. The digital processing unit 102 can comprise one or more electronic cards incorporating resources, for example hardware and / or software, suitable for the execution of one or more video processing methods. Such video processing methods can be performed independently or cooperatively using the sequences as inputs. video received by the digital processing unit 102. The video processing methods applied to the video sequences supplied by the surveillance unit of the risk zone 204 can comprise steps implemented to provide a state of the risk zone 204. corresponding. Video processing methods applied to the video sequences supplied by the surveillance units 1011, 1012, 1013 of the supervision zones 2051, 2052 can also comprise or in addition to the steps implemented to provide a state of the corresponding supervision zone.

The alert unit 103 can be configured to receive the state of the risk zone 204, as determined by the digital processing unit 102. The alert unit can be configured to trigger alerts intended for railway vehicles preparing to cross the supervised risk zone 204 if it is detected that collisions are likely to occur. The alert message sent by the alert unit 103 may include information identifying the characteristics of the object present in the risk zone 204. The alert unit 103 may further be configured to receive the states of the. supervision zones 2051, 2052, as provided by the digital processing unit 102. The alert unit can be configured to trigger alerts intended for users, for example road vehicles and / or pedestrians, located at proximity to the corresponding 204 risk zone. In this embodiment, the alert message sent may include information identifying the characteristics of the railway vehicle which is approaching the risk zone 204.

The alert unit 103 can be connected to one or more data networks making it possible to route the alert messages to their identified recipients. Alternatively, direct links can be established to route some or all of the alert messages to their recipients. The method of establishing such direct links can exploit the short distances between the monitoring device 100 and some of the identified recipients, which can be for example railway vehicles, road vehicles, pedestrians, etc.

FIG. 3 represents another embodiment of the invention in which the monitoring device 100 further comprises an automatic archiving unit 104 and a power supply unit 105. The archiving unit automatic 104 can be configured to save in a non-volatile memory video sequences supplied by the video recording unit 101. The recording of video sequences can be triggered each time a railway vehicle is about to cross the zone. at risk 204 is detected by the digital processing unit 102. Such a recording can be carried out by all the active surveillance cameras implemented in the video recording unit 101. The duration of each video recording can correspond to 1 'interval of time during which the railway vehicle is visible through one of the surveillance cameras implemented in the video recording unit 101. The automatic archiving unit 104 can further be configured to record footage video corresponding to the presence or passage of a user (for example road vehicle or pedestrian) in the risk zone 204.

[0029] The automatic archiving unit 104 can further be connected to a data network in a wired or wireless manner to make the saved video sequences accessible to external users.

The power supply unit 105 can be configured to supply the various units, forming the monitoring device 100, with the electric current necessary for their operation. The power supply unit 105 can further be configured to produce electric current independently by harnessing an external power source such as, for example, the sun, wind or the like.

According to embodiments of the invention, the alert unit 103 is further configured to monitor the operation of one or more units forming the monitoring device 100. The alert unit 103 can further be configured to generate an alert notification and transmit it to a remote maintenance center when a failure affects one or more units forming the monitoring device 100. The alert message sent by the alert unit 103 , in this case, can specify the failed unit and the type of failure.

According to other embodiments of the invention, the alert unit 103 can also be configured to transmit only a subset of the alert messages, the alert messages linked to the presence of an object, fixed or mobile, in the risk zone 204 not being transmitted systematically. In such modes of implementation, alert messages linked to the presence of moving objects (such as a road or pedestrian vehicle) are only sent if the passage of a railway vehicle is imminent. The imminent passage of a railway vehicle may correspond to the detection of the approach of a railway vehicle. In one embodiment, an alert message linked to the presence of a fixed object in the risk zone 204 is sent by the alert unit 103 only if the dimensions of such a fixed object are greater. at predefined thresholds. Such embodiments have the advantage of reducing the volume of data generated by the monitoring device 100.

[0033] Figure 4 shows a level crossing 200 equipped with a monitoring device 100 according to one embodiment of the invention. The level crossing corresponds to the intersection of a railway line 201 comprising a railway line 202 and a road line 203. In this example, the level crossing is not guarded, that is to say that no barrier guard is in place. The location of the three monitoring units 1011, 1012, 1013 forming the video recording unit 101 relative to the crossing can be optimized to completely cover the risk area 204 and the two supervision areas 2051, 2052.

In the example of Figure 4, the monitoring units 1011, 1012, 1013 each include two cameras, namely a visible camera 101a and an infrared camera 101b. The two cameras of each of the surveillance units 1011, 1012, 1013 can be placed in one place and configured with viewing angles to cover the same geographical area of interest.

The risk zone surveillance unit 204 can be positioned to cover a geographical area including the risk zone 204. Such a configuration is obtained according to the embodiment of FIG. 4 by placing the surveillance unit 1011 at the same time. '' outside the risk zone 204, at a height of between 1.5 meters and 2.5 meters, the shortest distance separating the projection on the ground of the surveillance unit and the railway line 202 or the median axis of the road being less than ten meters.

Each of the monitoring units 1012, 1013 associated with the supervision areas 2051, 2052 can be configured to cover a sufficiently large geographical area to detect the arrival of a railway vehicle from a distance of at least 300 meters measured from the risk zone 204. Such a Coverage can be obtained for railways without curvature as shown in Figure 4 by cameras with a low aperture angle (also called 'viewing angle').

FIG. 5 shows another embodiment of the invention in which the railway track 202 associated with the level crossing 200 has curvatures on either side of the risk zone 204. In such an embodiment of the invention, one of the supervision zones 2051 is associated with a railway track 202 having a curvature greater than 10 degrees at a distance from the risk zone 204 of less than 300 meters. The other supervision zone 2052 is associated with two railway tracks 202 which may be used by a railway vehicle to cross the risk zone 204 associated with the level crossing. The visible and infrared cameras implemented in the surveillance units 1012, 1013 of the supervision zones 2051, 2052 can be configured to have an opening angle large enough so that a railway vehicle is detected at least 300 meters from the risk zone 204.

The monitoring device 100 according to the embodiments of the invention has a simplicity of installation which does not require any modification of the rail or road system deployed. Such a surveillance system can be adapted to the different configurations of level crossings by adjusting the opening angles of the various cameras installed. The processing of the video content from the various surveillance units 1011, 1012, 1013 by the digital processing unit 102, located in their immediate vicinity, makes it possible to reduce the volume of the data streams generated by such a surveillance system to messages. alert. Further, in embodiments where the monitoring device 100 is provided with a power supply unit based on renewable energy sources, the system can operate autonomously and be energetically efficient.

[0038] FIG. 6 represents a method of monitoring 300 of a level crossing according to one embodiment of the invention.

In step 301, the risk zone 204 and the supervision zones 2051, 2052 associated with a level crossing are filmed continuously over time. In step 302, the filmed video sequences are separated into two categories, depending on whether they are associated with a risk zone 204 (first category) or with a supervision zone 2051, 2052 (second category). Video sequences are received in the form of data frames, one data frame corresponding to an entire image. Such video sequences can be received at a rate in number of frames per second high enough to correspond to a predefined response time.

In step 303, the video sequences associated with the risk zone 204 are analyzed to detect objects and / or pedestrians liable to cause a collision with a rail vehicle.

[0042] In step 304, the video sequences associated with the two supervision zones 2051, 2052 are analyzed to detect the approach of a railway vehicle.

In step 305, notifications are received indicating whether an object and / or a pedestrian is detected in the risk zone 204. An alert message is then triggered for each notification received. Such an alert message can include information identifying the nature and characteristics of the object or the pedestrian detected in the risk zone 204.

[0044] In step 306, notifications are received indicating whether a railway vehicle about to cross the level crossing is detected and a corresponding alert message is triggered. Such an alert message can specify the direction of origin of the railway vehicle.

[0045] In step 307, alert messages are sent to recipients.

FIG. 7 illustrates the sub-steps implemented to analyze video sequences associated with the two supervision zones 2051, 2052 according to embodiments of the invention. In such embodiments of the invention, each of the video sequences received comprises a video recording obtained by means of a visible camera 101a and a video recording obtained by means of an infrared camera 101b for a same geographical area and a same time of recording.

In step 401, the level of brightness of the video recording obtained by the visible camera 101a is measured. In step 402, the brightness level as measured in step 401 is compared to a given brightness threshold. If the measured luminosity level is greater than or equal to the luminosity threshold, the processing of the video sequence is carried out in a mode called “day mode” in which only the video recording filmed by a visible camera 101a is processed. If the measured luminosity level is lower than the luminosity threshold, the processing of the video sequence is carried out according to a mode called “night mode” in which only the video recording filmed by an infrared camera 101b is processed.

[0049] In step 403, a video recording obtained by means of a visible camera 101a is processed to detect the approach of a railway vehicle. Processing of such a video recording includes identifying the track 202 on which the railway vehicle is traveling. The processing includes a comparison of the successive images forming the video recording. Such a processing uses the dimensions and the range of travel speeds of the rail vehicles likely to travel on the railway track 202. The processing methods implemented in such a sub-step can be optimized to detect a rail vehicle at a distance. greater than 300 meters, measured from the risk zone 204.

In step 404, a video recording obtained by means of an infrared camera 101b for detecting the approach of a railway vehicle is processed. In this step, the coordinates of the railroad 202 as identified by step 403 are used. The thermal characteristics of the railway vehicles likely to move on the railway line 202 can also be used.

[0051] The embodiments of the invention allow efficient monitoring of a level crossing, whether or not guarded by gatekeepers. They provide a railway vehicle with the status of a level crossing it is about to cross, compatible with placement independent of the devices already deployed and with the use of a low transmission rate. They also make it possible to detect the approach of a railway vehicle and to alert road vehicles and / or pedestrians in the vicinity of the level crossing.

Those skilled in the art will understand that the system or subsystems according to the embodiments of the invention can be implemented in various ways. manners by hardware, software, or a combination of hardware and software, in particular in the form of program code which can be distributed as a program product, in various forms. In particular, the program code may be distributed using computer readable media, which may include computer readable storage media and communication media. The methods described in the present description can in particular be implemented in the form of computer program instructions executable by one or more processors in a computer computing device. These computer program instructions can also be stored in a computer readable medium.

Furthermore, the invention is not limited to the embodiments described above by way of nonlimiting example. It encompasses all the variant embodiments which may be envisaged by those skilled in the art.

Claims

1. Autonomous monitoring device (100) of a level crossing (200), said level crossing forming a crossing between a first railway track (202) and a second track (203) and comprising a risk zone (204) and two supervision zones (2051, 2052), characterized in that it comprises:

- a video recording unit (101) comprising at least three monitoring units (1011, 1012, 1013), each of said monitoring units (1011,

1012, 1013) being equipped with a visible camera (101 a) and an infrared camera (101 b), at least one monitoring unit (1011) being able to film said level crossing (200) and at least one surveillance unit being able to film one of said supervision areas (2051, 2052),

- a digital processing unit (102) configured to receive and process in real time video recording sequences supplied by the video recording unit (101), and

- an alert unit (103) configured to trigger collision risk alerts involving a railway vehicle moving on the railway track (202).

2. Autonomous monitoring device (100) according to claim 1 wherein the alert unit (103) is configured to trigger an alert to a rail vehicle, in response to the detection of the presence of an object. in the level crossing, liable to cause a collision involving the said railway vehicle.

3. Autonomous monitoring device (100) according to one of claims 1 and 2, wherein the alert unit (103) is further configured to trigger an alert to a user of the second channel (203). ), said alert notifying the passage of the railway vehicle.

4. Self-contained monitoring device (100) according to any one of claims 1 to 3 wherein said monitoring units (1011, 1012, 1013) comprise:

- At least one wide viewing angle monitoring unit (1011) capable of covering said risk zone (204) corresponding to said level crossing, and

- At least two monitoring units (1012, 1013) at low viewing angle capable of detecting the approach of said rail vehicle towards said risk zone (204).

5. Autonomous monitoring device (100) according to any one of the preceding claims, wherein said alert unit (103) is connected to a wired or wireless communication network capable of conveying alert messages to said rail vehicle. , and to a user of the second channel (203).

6. Autonomous monitoring device (100) according to any one of the preceding claims further comprising an electric power supply unit capable of continuously supplying over time an electric current suitable for the operation of the units forming said monitoring device (100). ).

7. Autonomous monitoring device (100) according to any one of the preceding claims further comprising an automatic archiving unit (104) configured to save in a non-volatile memory video recording sequences supplied by said recording unit. video (101), the triggering and the parameters for saving said sequences being controlled by said digital processing unit (102).

8. A stand-alone monitoring device (100) according to any one of the preceding claims, in which the digital processing unit (102) comprises resources suitable for the execution of one or more algorithms for processing video sequences.

9. A method of autonomous monitoring (300) of a level crossing, said level crossing forming a crossing between a first railway track (202) and a second track (203), characterized in that it comprises the steps of: :

- Making a video recording of a level crossing using a video recording unit (101) comprising at least three monitoring units (1011, 1012, 1013), each of said monitoring units (1011, 1012, 1013) being equipped with a visible camera (101 a) and an infrared camera (101 b), which provides video recording sequences,

- Process the said video recording sequences in real time, and

- Trigger collision risk alerts involving a railway vehicle moving on the railway track (202).

10. The method of monitoring (300) a level crossing according to claim 9 wherein the step of real-time processing of the video recording sequences comprises the sub-steps consisting of: - Receive a video sequence comprising a video recording taken by a visible camera (101a) and another video recording taken by an infrared camera (101b),

- Measure the level of brightness in the video recording taken by the visible camera (101a),

- Compare the measured luminosity level with a luminosity threshold,

- Process the recording taken by the visible camera (101a) if the measured luminosity level is greater than or equal to the luminosity threshold,

- Process the recording taken by the infrared camera (101b) if the measured luminosity level is lower than the luminosity threshold.