KR100876661B1 - Vision based monitoring system in railway station - Google Patents

Abstract

A vision based monitoring system of platform capable of monitoring critical regions of platform is provided to minimize damage of human life and property by rapidly determining a risk factor at platform. A vision based monitoring system of platform capable of monitoring critical regions of platform comprises an information collecting part(100), an information fusion part(200), and an information supplying part(300). The information collecting part generates monitoring information by sensing object and train inside a train area of platform(1) line(2) side. The information fusion part generates a warning message by using the monitoring information generated in the information collecting part. The information supplying part transmits the warning message information of the information fusion part to a corresponding receiver.

Description

Vision based monitoring system in railway station

1 is a conceptual diagram of image-based platform monitoring according to the present invention,

2 is a block diagram of an image-based platform monitoring system according to the present invention,

3 is a view showing a detection process of the image-based platform monitoring system according to the present invention,

4 is a state table of the train entering and leaving the platform,

5 is a state diagram showing the transition state of the train entering and exiting the platform,

6 is a state transition table of FIG. 5,

7 is a diagram illustrating an area of a camera.

*** Explanation of symbols for main parts of drawing ***

1: platform 2: track

10: camera area 11: train area

12: danger zone 100: information collector

110: camera 120: A / D converter

130: detection processor 132: train detection module

134: object detection module 136: object recognition module

138: object tracking module 200: information fusion unit

300: information providing unit 400: train operation total control device

402: train

The present invention relates to an image-based platform monitoring system. More specifically, the present invention relates to a system for monitoring a dangerous area of a platform so as to quickly and accurately recognize and cope with the emergency and emergency situation in the platform.

In general, railways are a popular and convenient means of transportation on a large scale.

However, because railroads are used by many passengers, many casualties can occur with only one inadequate safety management.

In particular, recent accidents caused by falls on the platform of subways, accidents in which clothes or body parts get stuck in the doors of trains during commute times, and the gap between the platform and railroad cars. Safety accidents on railroads are becoming more frequent, such as accidents where some are missing.

Of course, in order to monitor such safety accidents in the platform, the current CCTV is installed and operated in the subway station yard, but the CCTV must be monitored continuously by the operator in the command room or the station office, so the immediate recognition and emergency It was difficult to establish countermeasures, and was operated passively and limitedly, which made it difficult to recognize and respond quickly to emergency situations.

On the other hand, in recent years, the installation of screen doors (PSDs) has been expanded in order to prevent passengers from crashing in advance. However, it is economically expensive to install them on all platforms, especially in the case of ground platforms. It also has the drawback that it can block passengers and increase their frustration.

In addition, when the train does not stop in the right position may cause problems in the opening and closing of the screen door, it is difficult to prepare for the case where the human body or object is caught in the gap between the screen door and the train.

Accordingly, the present invention has been made to solve such a problem, the object of which is to detect the passenger risk factors in the platform using the image processing technology to respond quickly and accurately in the event of emergencies and emergency situations. To provide an image-based platform monitoring system that can minimize property damage.

It is another object of the present invention to provide an image-based platform monitoring system which is low in installation and maintenance costs and installable without being distinguished by the characteristics of the platform.

The present invention for achieving the above object;

An information collecting unit which detects trains and objects in the platform using image information obtained by using a camera and determines whether objects are falling and generates monitoring information; An information fusion unit for generating an alarm message using the monitoring information of the information collection unit; characterized in that consisting of.

Here, the information providing unit for transmitting the alarm message information of the information fusion unit, characterized in that it is further provided.

The information collecting unit uses a camera installed in a platform, an A / D converter for converting image information collected by the camera into a digital signal, and a digital image information through the A / D converter. Detects and detects the movement of the object is characterized by consisting of a detection processor for generating a monitoring information by determining the fall object by back tracking the image information.

The detection processor may further include a train detection module that determines whether a train exists in the image information of the camera, an object detection module that detects movement of an object when a train does not exist when determining the train detection module, and When the movement of the object is detected in the object detection module is characterized by consisting of an object tracking module for determining whether the object falls by trace back the frame of the previous image information.

The image information of the camera is divided into a train area into which a train enters and a danger area in which an object can fall within the train area, and the object detecting module determines whether a train exists in the train area. The object detecting module detects the movement of the object within the danger zone, and when the object detecting module detects the movement of the object within the danger zone, the object detection module detects the frame of the previous image information and traces the danger outside the danger zone. When the movement of the object inside the area is traced, it is characterized by the fall object.

The apparatus may further include an object recognition module configured to determine whether the detected object is a predetermined size or more when the movement of the object is detected in the object detecting module.

And, the information collecting unit is characterized in that one or more installed in the landing.

In addition, the receiving device is characterized in that any one of the train operation comprehensive control device (TTC), the train, the terminal of the station attendant.

Hereinafter, with reference to the accompanying drawings of the image-based platform monitoring system according to the present invention will be understood by the embodiments described in detail.

1 is a conceptual diagram of an image-based platform monitoring system according to the present invention, FIG. 2 is a configuration diagram of an image-based platform monitoring system according to the present invention, and FIG. 3 is a detection process of the image-based platform monitoring system according to the present invention. 4 is a state diagram of a train entering and exiting a platform, and FIG. 5 is a state diagram showing a transition state of trains entering and leaving a platform, and FIG. 6 is a state transition table of FIG. 7 is a diagram illustrating an area of a camera.

Referring to Figures 1 and 2, the image-based platform monitoring system according to the present invention to immediately recognize the platform passenger risk factors through the image processing technology and to quickly respond to the situation (1) track ( 2) using the information collecting unit 100 for detecting the train and the object in the train area to generate the monitoring information, and the monitoring information generated by detecting the train and the object collected by the information collecting unit 100 An information fusion unit 200 for analyzing the situation and generating an alarm message according to the situation analysis, and an information provider 300 for transmitting the alarm message information of the information fusion unit 200 to a corresponding receiver.

Hereinafter, each component will be described in more detail.

First, the information collecting unit 100 detects and recognizes a dangerous situation such as a fire in the platform 1 or a track crash of a passenger. To this end, the information collecting unit 100 detects a train, detects an object, and performs a series of tasks such as object recognition and object tracking.

The information collecting unit 100 is an A / D converter that converts the camera 110 installed to face the track 2 of the landing 1 and the image information collected by the camera 110 into a digital signal. A detection processor for detecting trains and objects by using the digitalized image information via the A / D converter 120 and performing a series of arithmetic operations related to object recognition and object tracking to generate monitoring information. 130).

In this case, the detection processor 130 detects the movement of an object when a train does not exist in the train detection module 132 and the train detection module 132 determines whether a train exists in the camera 110 region. When the object detection module 134, the object detection module 134 detects the movement of the object, the object recognition module 136 for detecting whether the object is a predetermined size or more, and the object in the object detection module 136 When the motion is detected, it is composed of an object tracking module 138 that identifies the falling object when the motion of the previous image information is traced back to the danger area from the outside of the danger area.

In addition, the information collecting unit 100 is installed in at least one platform (1) to obtain the image information about the train and the object for each specific location and using this to detect the track fall or danger, in particular the object The recognition module 136 may determine whether an object is an obstacle that may cause an abnormality in a person or a train progress when the object is larger than or equal to a predetermined size, and may perform control and incident response of an entry train as necessary.

Meanwhile, the information fusion unit 200 receives monitoring information from one or more of the information collecting units 100 for comprehensive situation analysis in the platform 1 and extracts more meaningful and high level information. Do this. That is, the information fusion unit 200 comprehensively analyzes the situation by using the monitoring information of the information collecting unit 100, and alerts each of the command room staff, the station attendant, and the entry train driver through the results. Will generate a message.

In addition, the information providing unit 300 is an alarm message generated in the information fusion unit 200 with a standard operation procedure (SOP; Standard Operation Procedure) corresponding to different recipients, that is, the command room staff, station attendant, driver of the entering train And it serves to immediately transmit a video on the dangerous situation.

In this case, the information providing unit 300 is information to the receiving device such as the TTC Total Traffic Control 400 of the command room and the terminal (not shown) of the train 402 and the station attendant 404. Send it to make a quick response.

On the other hand, it is natural that the information fusion unit 200 generates an alarm message by subdividing it according to the characteristics of a receiver (for example, a command room employee, a station attendant, and an entering train driver).

Hereinafter, the processing of the detection processor 130 will be described in detail.

First, referring to FIG. 3, the train detection module 132 of the detection processor 130 detects a train in a corresponding camera 110 region.

At this time, it is determined whether or not there is a train in the landing (1), that is, whether or not (OFF) mode (S110), if the train is present in the landing (1) in the determination process, that is, in the landing (1) The above step may be performed in any of a state in which a train is entering, an ON state in which a train is stopped at a platform, and an OUT state in which a train is departing from a platform. Branching to (S100) to determine the current state of the train.

In the step S110, when there is no train in the landing 1, that is, in the OFF mode, the object detecting module 134 detects an object (S120).

At this time, the object detection module 134 detects the movement when detecting the object (S130), if the movement of the object is not detected in the detection process of the movement branched to the step (S100) and the current of the train The state is determined.

On the other hand, if the movement of the object is detected in step S130, the object recognition module 136 determines whether the object is greater than or equal to a predetermined size (S140), and if it is less than or equal to the reference size, then to step S100. It branches to determine the current state of the train.

If the object is larger than a predetermined size in step S140, the object is recognized as an object (S150).

If it is recognized as an object through the step (S150), the object tracking module 138 traces the frame of the previous image information to determine whether the object movement in the danger zone is traced outside the danger zone and determines whether it is an emergency situation or not. In step S160, if the movement of the object from the danger zone to the danger zone is not tracked, the process branches to the step S100 to determine the current state of the train.

On the other hand, if the movement of the object in the dangerous area is tracked in the step S160 in the step (S160) is determined as the falling object to generate the monitoring information for controlling the train and coping with it to transmit to the information fusion unit 200. (S170)

At this time, the motion is detected by the difference between the image frame obtained from the camera 100 and the background separation, and the object is recognized when the size of the object is larger than a predetermined size. Therefore, if it is recognized as an object, it is necessary to determine whether it is an obstacle that may cause an abnormality in people or train progress so that it is possible to perform control of an ingress train and to deal with an accident as necessary.

On the basis of such a condition, the detection processor 130 first determines the state of the train in the train detection module 132 to prevent a mistake in performing the passenger fall detection of the train.

At this time, the state of the train obtained through the at least one or more cameras 110 installed in the platform 1, as shown in Figure 4, the off (state) state that there is no train in the platform 1, the platform ( 1) In state where the train enters, IN state where the train is stopped at the platform, ON state, OUT state where the train is leaving the platform Have

Therefore, detection of an object or a person among the four states of the train is performed only when the OFF state, that is, when the train is not in the detection area of the camera 110.

On the other hand, in order to determine the danger of the object falling on the platform, first the current state of the train must be accurately calculated. The current state of the train can be determined according to the transition state diagram as shown in FIG. 5, and each state transition thereof can be defined as shown in FIG. 6.

The state of a train can be regarded as the movement of a train if it detects movement within the danger zone and a large movement occurs. In this case, the sudden brightness change of the image is not considered, and when a large motion occurs continuously for 5 frames or more in consideration of noise, it is determined as a train movement. Through this image processing technique, the four states of the train can be detected completely.

That is, referring to FIGS. 5 and 6, the current state of the train is the OFF-IN state when a large movement occurs five times in a vehicle area, and a large movement in the vehicle area is not detected for five consecutive frames. If not, it is in the in-on state, if 5 consecutive movements occur in the vehicle area, if it is ON-OUT state, if 5 consecutive movements are not detected in the vehicle area, It can be known as OUT-OFF state.

On the other hand, since the train is not detected, the object detection module 134 detects the movement within the danger area of the landing 1 obtained through the camera 110.

At this time, when the motion is detected in the danger area of the landing 1, it may be divided into a case where a falling object exists, when a train movement exists, and a sudden brightness change of an image appears.

Therefore, the object discrimination technique of the object tracking module 138 first considers only the movement in the danger zone in the OFF state in which the train is not present in the landing 1. In addition, by back-tracking the movements in the previous frames, it may be determined as a fall object in the case of movement from outside the danger zone to the danger zone.

At this time, the object tracking stores the motion information of the previous frame, and if there is a motion in the OUT state, the back track, and if the back tracked object is out of the danger zone, it is determined as a fall object. It also preserves the object's position even if the fallen object stops and there is no movement.

That is, when the movement of the object is detected in the danger area 12 in the entire camera area 10 as shown in FIG. 7, it is possible to detect whether or not the object is falling through a back tracking technique.

At this time, the danger zone 12 exists inside the train zone 11 and the zone 13 of the object exists. Therefore, when the object enters the danger zone 12, the motion information of the previous frame is traced back to determine whether the object movement is from the outside of the danger zone 12 to the inside of the danger zone 12.

Hereinafter, an operation process of the image-based platform monitoring system according to the present invention will be described with reference to FIGS. 1 to 7.

The image information obtained through the camera 110 installed in the boarding point 1 is converted into digitized image information through the A / D converter 120 and input to the detection processor 130.

The detection processor 130 determines whether a train exists by determining whether a train exists in the train area 11 of the image information of the camera 110.

In this case, when the train does not exist, the object detection module 134 detects the object, and in this case, the motion is detected through the image processing by the frame difference and the background separation.

As such, when the movement of the object is detected by the object detection module 134, the object tracking module 138 traces the frame of the previous image information to track the movement of the object into the dangerous area 12 outside the dangerous area 12. Will be determined.

If the movement of the object is tracked out of the danger area 12 into the danger area 12 through the back trace, it is determined that the object is falling.

On the other hand, when the movement of the object is detected in the object detection module 134, the object recognition module 136 determines whether or not the detected object is a specific size or more, and if it is more than a specific size it is transmitted to the information fusion unit 200. Will be sent.

Such information is used as an analysis data necessary for controlling the entry train and performing an accident response in the information fusion unit 200.

The monitoring information generated by the processing of the image information of the information collection unit 100 is transmitted to the information display unit 200 to the standard operating procedures (SOP) and alarm messages corresponding to the command room staff, station attendants, drivers of the entry train and Image information about a dangerous situation is generated, and such information is transmitted from the information providing unit 300 to an alarm message such as an image and an alarm to a terminal (not shown) of the corresponding receiver.

Although the preferred embodiment of the present invention has been described above, the scope of the present invention is not limited thereto, and the scope of the present invention extends to the scope of the present invention to be substantially equivalent to the embodiment of the present invention. Various modifications can be made by those skilled in the art without departing from the scope of the present invention.

As can be seen from the above description, according to the present invention, when a hazard of a passenger is detected, such as when an object enters a train area into which a train in a platform enters and a danger area where an object can fall, the camera image frame is used. It has the advantage of minimizing the damage of life and property by quickly and accurately coping with the risk of falling of the object by tracking back the motion information.

In addition, when the present invention is installed and maintained in the platform, the cost is low, there is an advantage that can be installed without being separated in the characteristics of the platform.

Claims (8)

The camera installed in the platform, an A / D converter that converts the image information collected by the camera into a digital signal, and the train and the object are detected using the digitized image information through the A / D converter. An information collector configured to detect the falling object by back-tracking the image information and to detect the falling object and to generate monitoring information when the detected information; Consists of an information fusion unit for generating an alarm message using the monitoring information of the information collector, Image information of the camera is divided into a train area into which a train enters and a danger area where objects can fall inside the train area, The detection processor includes a train detection module for determining whether a train is present in the image information of the camera, an object detection module for detecting movement of an object when a train does not exist when the train detection module is detected, and the object detection. When the motion of the object is detected in the module, an object tracking module which tracks the frame of the previous image information and traces the movement of the object from the outside of the danger area to the inside of the danger area is identified as a falling object, and the motion of the object in the object detection module Image-based platform monitoring system, characterized in that consisting of an object recognition module for determining whether the detected object is a predetermined size or more when detected. The method of claim 1, Image-based platform monitoring system further comprises an information providing unit for transmitting the alarm message information of the information fusion unit to the receiving device. delete delete delete delete The method of claim 1, Image-based platform monitoring system, characterized in that one or more information collection unit is installed in the platform. The method of claim 2, The receiving device is a train-based integrated control system (TTC), trains, image-based platform monitoring system, characterized in that any one of the station staff.

Images