KR20090027410A - Stereo vision based monitoring system in railway station and method thereof - Google Patents

Abstract

A system for monitoring a platform based on a stereo image and a method thereof are provided to increase reliability of the system and to manage dangerous situation rapidly. A system for monitoring a platform based on a stereo image comprises the followings: an information collecting unit(100) comprising a stereo camera(110); the stereo camera acquiring image information about a platform monitoring area; and a information processing unit(200) including an object detection module(220) and an object recognition module(230). The object detection module determines falling of the object in a dangerous region based on the image information.

Description

Stereo image-based platform monitoring system and its method {STEREO VISION BASED MONITORING SYSTEM IN RAILWAY STATION AND METHOD THEREOF}

The present invention relates to a stereo image-based platform monitoring system and method, and more particularly, to a stereo image-based platform monitoring system and method for determining a dangerous situation in the platform based on the image information obtained by the stereo camera. It is about.

Passenger safety is the most important task in the railway sector, and despite the centralized management of railway safety, dozens of passengers are being killed every year due to the crash of the platform, which is emerging as the most urgent problem to be solved in the railway sector.

Recently, with the development of information technology (IT), an application system using a closed circuit TV or various types of vision sensors in a railway environment has been attempted. The subway closed circuit TV is currently installed and operated in the platform or passenger area of each station for the purpose of checking train operation and monitoring passenger movement and getting on and off.It is also used for maintenance of railway infrastructure including vehicles and track conditions. Various types of camera sensors are applied.

Whereas conventional closed-circuit TVs simply passively transmit video information to the driver or local command office, they can detect the location and movement of people or other obstacles and intelligently determine the dangerous situation and inform the relevant workers to respond quickly. It was necessary to try a form.

As such, the conventional monitoring system is simply installed by analogue closed circuit TV in the place to be monitored, so that the worker must monitor the situation that is displayed on the screen (sequentially showing several divided screens according to the number of CCTVs). And there is a problem that it is difficult to identify the correct situation and prompt response.

On the other hand, in recent years, the installation of screen doors has been expanded in order to prevent passengers from crashing in advance, but it is economically expensive to install on all platforms, especially in the case of ground platforms. There is a problem that increases the passenger's frustration by blocking the.

SUMMARY OF THE INVENTION An object of the present invention is to provide a platform monitoring system and method capable of quickly identifying and promptly responding to a dangerous situation in a platform and improving the reliability of the system.

The above object is, according to the present invention, the information collection unit having at least one stereo camera for obtaining image information for the platform monitoring area including the train area to enter the train; And an object detection unit determining whether or not the object falls within a danger area that is previously set as an area in which the object may fall and interfere with normal train operation, based on the image information obtained by the information collecting unit. And an object recognition module configured to determine whether an object fallen in the dangerous area is a dangerous object that impedes normal train operation based on a preset criterion based on the three-dimensional coordinate information of the object by the stereo camera. It is achieved by a stereo image-based platform monitoring system comprising an information processing unit.

In the present invention, the object recognition module, based on the three-dimensional coordinate information of the object by the stereo camera calculates the size of the object fall in the dangerous area, if the calculated size is more than the set size the dangerous object Judging by

The information processing unit may further include an object tracking module for assigning a unique identification code to the fallen object and tracking the movement of the object to which the identification code is assigned when the fallen object is determined to be the dangerous object. have.

When the object tracking module detects a movement of an object to which the identification code is assigned, the object tracking module continues tracking the object and determines that the object is out of the dangerous area during tracking of the object. You can delete it.

If the object tracking module does not detect the movement of the object to which the identification code is assigned, if it is determined that the object to which the identification code is assigned does not exist in the dangerous area by comparing the current image with a pre-stored background image, The assigned identification code can be deleted.

The object tracking module may track the movement of the object to which the identification code is assigned based on the image difference between the current frame and the previous frame of the image information.

The information processing unit may further include a train detection module that determines a state of a train in the train area.

The train detection module may determine a state of a train in the train area based on the image information.

The object detection module may be operated only when there is no train in the train area.

The object detection module may determine whether an object falls within the danger area based on an image difference between a current frame and a previous frame of the image information.

The object of the present invention is to obtain image information on a platform monitoring area including a train area into which a train enters using at least one stereo camera; Determining whether an object falls within a danger area defined as an area in which the object falls among the train areas to interfere with normal train operation based on the acquired image information; And determining whether an object fallen in the danger area is a dangerous object that impedes normal train operation based on a preset criterion, based on the three-dimensional coordinate information of the object by the stereo camera. It is achieved by stereo image based platform monitoring method.

In the present invention, the step of determining whether the dangerous object, the size of the object falling in the dangerous area based on the three-dimensional coordinate information of the object by the stereo camera, the calculated size is If the size is larger than the set may be determined as the dangerous object.

In the stereo image-based platform monitoring method according to the present invention, if it is determined that the fallen object is the dangerous object, assigning a unique identification code to the fallen object, and tracking the movement of the object to which the identification code is assigned. It may further include.

The tracking of the movement of the object may include: if the movement of the object to which the identification code is assigned is detected, if the object is out of the dangerous area during the tracking of the object and continues to track the object, the grant It may include the step of deleting the identified identification code.

In the tracking of the movement of the object, if the movement of the object to which the identification code is assigned is not detected, the object to which the identification code is assigned does not exist in the danger area by comparing a current image with a pre-stored background image. If it is determined that it may include the step of deleting the assigned identification code.

The tracking of the movement of the object may track the movement of the object to which the identification code is assigned based on the image difference between the current frame and the previous frame of the image information.

The stereo image-based platform monitoring method according to the present invention may further include determining a state of a train in the train area before determining whether the object has fallen.

In the determining of the state of the train, the state of the train in the train area may be determined based on the image information.

The determining of the fall of the object may be performed only when there is no train in the train area.

The determining whether the object falls may determine whether the object falls within the danger area based on an image difference between the current frame and the previous frame of the image information.

According to the present invention, by intelligently determining a dangerous situation in a platform based on three-dimensional image information obtained by a stereo camera, it is possible to quickly identify and quickly deal with a dangerous situation and to improve the reliability of a system.

Hereinafter, an embodiment of the present invention will be described with reference to FIGS. 1 to 5.

1 is a block diagram of a stereo image-based platform monitoring system according to an embodiment of the present invention, Figure 2 is a view for explaining the platform monitoring area 10 taken by the stereo camera 110 of FIG. 3 is a view for explaining a method of acquiring three-dimensional coordinate information using the stereo camera 110, FIG. 4 is a view showing the state transition of the train in the train area 12, Figure 5 is a view of FIG. It is a figure for demonstrating the method of detecting the movement of a train in the boarding point monitoring area | region 10. FIG.

Referring to FIG. 1, the stereo image-based platform monitoring system according to the present embodiment may be collected by the information collection unit 100 and the information collection unit 100 for obtaining image information of the platform monitoring area 10. The information processing unit 200 determines whether or not the object falls and whether the fallen object is a dangerous object in the dangerous area 14 based on the image information.

Referring to FIG. 2, the platform monitoring area 10 includes a train area 12 into which a train enters, and an area in which train objects 12 fall, which may interfere with normal train operation. The area 14 is set in advance. In this case, the 'hazardous area 14' specifically refers to an area around a train track.

1 and 2, the information collecting unit 100 is a stereo camera 110 installed in a platform to obtain image information on the platform monitoring area 10 including a train area 12 into which a train enters. It is provided. The stereo camera 110 is preferably installed at a position capable of photographing the train entering the platform and the entire platform situation (for example, in front of the platform with respect to the direction in which the train proceeds), and sufficiently monitoring the overall platform situation. For this purpose, a plurality of stereo cameras 110 may be installed at predetermined intervals along the track of the train.

Unlike the general mono camera, since the stereo camera 110 can obtain three-dimensional coordinates for each pixel of the obtained image, the stereo camera 110 can define the monitoring area as an arbitrary three-dimensional space, and thus Three-dimensional coordinate information can be obtained, which will be described later.

In the present embodiment, the information collecting unit 100 collects only image information about the platform monitoring area 10 and transmits the image information to the information processing unit 200. However, the present invention is not limited thereto, and the information collecting unit 100 may include an infrared sensor, a laser sensor, an ultrasonic sensor, a microwave sensor, and the like installed in the platform in order to obtain detection information for the platform monitoring area 10. The sensor module 120 may be further provided. The detection information on the platform monitoring area 10 collected by the sensor module 120 is transmitted to the information processing unit 200 as determination data on whether a dangerous situation occurs together with the image information collected by the stereo camera 110. do.

Referring to FIG. 1, the information processing unit 200 is an area in which an object may fall based on the image information acquired by the stereo camera 110 of the information collecting unit 100 and may interfere with normal train operation in advance. The object detection module 220 which determines whether the object falls in the set danger area 14 and the object which has fallen in the danger area 14 based on the three-dimensional coordinate information of the object by the stereo camera 110 are dangerous. It is provided with an object recognition module 230 for determining whether or not the object. Here, the "dangerous object" refers to an object that interferes with normal train operation among objects that have fallen in the danger area 14 and includes a person who has fallen in the danger area 14.

The object detection module 220 detects a motion in the danger area 14 based on an image difference between a current frame and a previous frame of the image information acquired by the stereo camera 110 of the information collecting unit 100. It is determined whether or not the object has fallen in (14). In the present exemplary embodiment, the stereo camera 110 photographs the platform monitoring area 10 at about 30 frames per second, and the object detection module 220 analyzes image information transmitted from the stereo camera 110 in units of frames. That is, the object detection module 220 detects the movement of the object into the danger area 14 outside the danger area 14 by analyzing the image difference between the current frame and the previous frame of the image information. As such, when the movement of the object into the danger area 14 is detected outside the danger area 14, the object detection module 220 determines that the object has fallen in the danger area 14.

On the other hand, objects that have fallen in the danger zone 14 are dangerous objects that suspend normal train operations, such as a person who has fallen off the platform in a platform for suicide or inadvertently or a large object having a large size, causing a dangerous situation, Small objects such as beverage cans can be classified as small objects that do not interfere with train operation even if they fall near train tracks. That is, even if it is determined that the object has fallen in the danger area 14 by the object detection module 220, it cannot be assumed that a dangerous situation has occurred uniformly.

Accordingly, in the present invention, when it is determined that an object has fallen in the danger area 14, an object that has fallen into the danger area 14 through the object recognition module 230 interferes with normal train operation (ie, Determine whether it is a small object that does not interfere with train operation. In detail, the object recognition module 230 determines whether an object that has fallen in the dangerous area 14 is a dangerous object based on three-dimensional coordinate information of the object by the stereo camera 110.

A method of obtaining 3D coordinate information by the stereo camera 110 will now be described with reference to FIG. 3. Due to the characteristics of the stereo camera 110, the positions of the images formed on the left and right cameras with respect to one object point (x, y, z) are respectively P _l (x ' _l , y' _l ) and P _r (x ' _r , y ' _r ). Given the baseline b at which the camera is located, and knowing the focal length f of the lens, the depth information Z of the object is represented by b * f / (x ' _l -x' _r ) by the triangular ratio. It is possible. That is, the depth information Z of the object is proportional to the distance b between the lenses of the left camera Cl and the right camera Cr and the focal length f of the lens, and the binocular difference x ' _l -x' _r ) inversely proportional to the value.

As such, since the stereo camera 110 can know depth information of an object to be photographed unlike a general mono camera, the stereo camera 110 can define a monitoring area as an arbitrary three-dimensional space, and thus three-dimensional coordinate information about an object in the monitoring area. Will be obtained.

By using the characteristics of the stereo camera 110, the object recognition module 230 calculates the size of the object falling in the danger area 14 based on the three-dimensional coordinate information of the object by the stereo camera 110. If the size of the object that falls in the danger zone 14 is greater than or equal to the set size, it is determined as a dangerous object that impedes train operation. At this time, the setting size is set in consideration of the possibility of affecting the normal operation of the train, it is preferable to set the size higher than the train track.

On the other hand, since the object recognition module 230 can recognize the shape of the object falling in the dangerous area 14 on the basis of the three-dimensional coordinate information of the object by the stereo camera 110, it falls in the dangerous area 14 It may be configured to determine whether or not it is a dangerous object based on the shape of one object. For example, a method of determining whether the object is a dangerous object may be applied by storing information on the shape of a person in advance and comparing it with the object that has fallen in the dangerous area 14.

Referring to FIG. 1, the information processing unit 200 further includes an object tracking module 240 for tracking a movement of an object determined as a dangerous object.

When the object falling down to the dangerous area 14 by the object recognition module 230 is determined to be a dangerous object, the object tracking module 240 assigns a unique identification code to the dangerous object and Track the movement. At this time, 'Identification code' means that a dangerous situation occurred in the platform. The object tracking module 240 uses an image processing technique that analyzes an image difference between a current frame and a previous frame of image information acquired by the stereo camera 110 to track the movement of an object to which an identification code is assigned.

The object tracking module 240 deletes the assigned identification code if the object to which the identification code is assigned leaves the danger area 14 as a result of the motion tracking. At this time, 'deletion of identification code' means that the dangerous situation in the platform has been released.

On the other hand, the object tracking module 240 compares the current image and the pre-stored background image when the movement tracking result, the movement of the object to which the identification code is assigned is not detected. This is because the object detection module 220 and the object recognition module 230 are dangerous in the danger area 14 due to the sudden change of brightness around the stereo camera 110 even though there is no danger object that falls in the danger area 14. This is to find out if the object is wrongly judged to have fallen. That is, when it is determined that an object assigned with an identification code does not exist in the danger area 14 by comparing the current image with a pre-stored background image, the object tracking module 240 deletes the assigned identification code.

In this way, the present invention, by having the object tracking module 240 for tracking the dangerous object, it is possible to determine whether or not to release the generated dangerous situation, as well as the object detection module 220 and the object recognition module 230 Correction errors can be corrected. Accordingly, the present invention can increase the reliability of the entire system, it is possible to prevent the situation that interferes with the smooth operation of the train by wrong judgment.

1, 4, and 5, the information processing unit 200 further includes a train detection module 210 that determines a state of a train in the train area 12.

In general, when motion is detected in the danger zone 14, when an object falls in the danger zone 14, when a train is present, and there is a sudden change in brightness around the stereo camera 110. It can be divided into cases.

Accordingly, the object detection module 220 that determines whether the object falls in the dangerous area 14 moves in the dangerous area 14 in the absence of a train in the train area 12 of the platform in order to increase the reliability of the determination. It is desirable to be set to judge only. That is, the object detection module 220 is preferably operated only when there is no train in the train area 12.

In order to determine the state of the train in the train area 12, the train detection module 210 of the present embodiment, the image of analyzing the image difference between the current frame and the previous frame of the image information obtained by the stereo camera 110 Using the processing technique, the principle is explained as follows.

TABLE 1

The condition of the train Explanation OFF No train in the platform IN Train is entering the platform ON Train is stopped in platform OUT Train is leaving from the platform

As described in the above [Table 1], the state of the train in the platform is an OFF state in which there is no train in the platform, an IN state in which the train is entering the platform, and the train in the platform. Is divided into four states: an ON state in which the state is stopped and an OUT state in which the train is departing from the platform. As described above, the object detection module 220 is determined whether the object falls in the danger zone 14 is performed only when the state of the train is determined to be off state by the train detection module 210.

The train detection module 210 analyzes the image difference between the current frame and the previous frame of the image information acquired by the stereo camera 110 to determine the transition between the states of the four trains as described above.

Referring to FIG. 4, in the present exemplary embodiment, the train detection module 210 determines that the train movement module 12 is an off-in transition when a large motion is detected continuously in the train area 12 for five frames. If a large motion is not detected continuously in the train area 12 after 5 transitions, it is determined as an IN-ON transition, and a large movement in the train area 12 is 5 frames after the in-on transition. If it is detected continuously, it is judged as an ON-OUT transition.If a large movement is not detected continuously in the train area 12 after five frames after the on-out transition, it is an OUT-OFF transition. Judging by. At this time, whether the large motion is continuously detected for five frames is a reference for changing the train state, which is determined by a sudden change of brightness around the stereo camera 110 or a determination error of the train detection module 210 caused by various noises. This is to minimize.

On the other hand, the train detection module 210, in order to distinguish between the movement of the train and the movement of other objects in the train area 12, as shown in Figure 5, the movement area 16 is detected in the train area 12 This train area 12 is included, the width w of the movement area 16 is 40% or more of the width W of the train area 12, and the height h of the movement area 16 is the train area ( Only 60% of the height (H) of 12) is regarded as the movement of the train.

As described above, the train detection module 210 of the present embodiment analyzes the image information obtained by the stereo camera 110 to determine the state of the train in the train area 12 of the platform. However, the present invention is not limited thereto, and the train detection module 210 may be configured to determine the state of the train in the train area 12 of the landing using various sensors such as an infrared sensor or a laser sensor installed in the landing. Of course.

Referring to FIG. 1, the information processing unit 200 further includes a result output module 250 for outputting a determination result of the object recognition module 230 and the object tracking module 240. In addition, the stereo image-based platform monitoring system according to the present embodiment further includes an information providing unit 300 for providing a dangerous message and the image of the dangerous situation to the train operator.

The result output module 250 is a result of the determination of the object recognition module 230 and the object tracking module 240, that is, whether a dangerous situation occurs due to the fall of the dangerous object in the dangerous area 14, the type of dangerous situation, and the risk. A signal corresponding to whether or not the situation is released is output. Based on the signal output from the result output module 250, a danger message informing of a dangerous situation is generated.

The information providing unit 300 may be implemented as various display devices such as a liquid crystal display (LCD), a personal digital assistant (PDA), and the like. According to the determination result of the 200, a danger message is provided to the train operator along with an image of a dangerous situation. Here, the train operator concerned includes a station employee of the corresponding history, a train driver, and an operator of the general control room.

Hereinafter, an embodiment of a stereo image-based platform monitoring method performed by the above-described stereo image-based platform monitoring system will be described with reference to FIGS. 6 to 8.

 FIG. 6 is a flowchart for describing a stereo image-based platform monitoring method by the system shown in FIG. 1, FIG. 7 is a flowchart for explaining in more detail the step S140 shown in FIG. 6, and FIG. 8 is shown in FIG. 6. It is a flowchart for explaining the step S150 shown in more detail.

Referring to FIG. 6, the stereo image-based platform monitoring method according to the present embodiment includes obtaining image information (S110), determining a state of a train in the train area 12 (S120), and a danger area 14. Determining whether the object falls (S130), determining whether the dangerous object (S140), tracking the dangerous object (S150) and outputting the determination result (S160), The steps are described sequentially.

First, the information collecting unit 100 obtains image information on the platform monitoring area 10 including the train area 12 into which a train enters using the stereo camera 110 (S110).

next. The information processor 200 determines a state of a train in the train area 12 based on the image difference between the current frame and the previous frame of the image information transmitted from the information collector 100 (S120).

As a result of the determination of the train state, if it is determined that there is no train in the train area 12, the information processing unit 200 determines whether an object falls in the danger area 14 based on the image difference between the current frame and the previous frame of the image information. Determine (S130).

If it is determined that the object has fallen in the dangerous area 14, the information processing unit 200 determines whether the object that has fallen in the dangerous area 14 is a dangerous object based on the three-dimensional information of the object by the stereo camera 110. It is determined (S140).

The step S140 of determining whether a dangerous object is described in detail with reference to FIG. 7 will now be described. First, the information processing unit 200 calculates the size of the object dropped in the danger area 14 based on the three-dimensional information of the object by the stereo camera 110 (S141). Next, the information processing unit 200 compares the calculated size with the set size, and if the calculated size is larger than the set size, determines that the fallen object is a dangerous object (S142).

If the object fallen in the dangerous area 14 is determined to be a dangerous object, the information processing unit 200 assigns a unique identification code to the fallen object and tracks its movement (S150). At this time, the information processing unit 200 tracks the movement of the object to which the identification code is assigned by analyzing the image difference between the current frame and the previous frame of the image information.

Referring to FIG. 8, the step S150 of tracking a dangerous object will be described in detail. First, when the object fallen in the dangerous area 14 is determined to be a dangerous object, the information processing unit 200 assigns a unique identification code to the fallen object (S151), and the image difference between the current frame and the previous frame of the image information. By detecting the detection of the movement of the dangerous object given the identification code (S152).

If the movement of the dangerous object to which the identification code is assigned is detected, in order to confirm whether the dangerous situation is released, the information processing unit 200 continues tracking the dangerous object (S153), and the dangerous object is the dangerous area 14. It is determined whether or not out of step (S154). If it is determined that the dangerous object is out of the dangerous area 14, the information processing unit 200 deletes the assigned identification code (S155).

On the other hand, if the movement of the dangerous object to which the identification code is assigned is not detected, in order to correct a determination error caused by the sudden change in brightness around the stereo camera 110, the information processing unit 200 stores the current image and the pre-stored background. By comparing the images (S156), it is determined whether there is a dangerous object in the dangerous area 14 (S157). If it is determined that there is no dangerous object in the dangerous area 14, the information processing unit 200 deletes the assigned identification code (S158).

Finally, the information processing unit 200 outputs a signal corresponding to the determination result and the tracking result of the dangerous object (S160). A message informing of a dangerous situation is generated based on the signal output as described above, and the generated dangerous message is provided to a train operator through an information providing unit 300 implemented by various display devices together with an image of the dangerous situation.

It is apparent to those skilled in the art that the present invention is not limited to the above-described embodiments, and that various modifications and changes can be made without departing from the spirit and scope of the present invention. Therefore, such modifications or variations will have to be belong to the claims of the present invention.

1 is a block diagram of a stereo image-based platform monitoring system according to an embodiment of the present invention.

FIG. 2 is a diagram for describing a platform monitoring area photographed by the stereo camera of FIG. 1.

3 is a diagram for describing a method of obtaining 3D coordinate information using a stereo camera.

4 is a view showing a state change of the train in the train area.

FIG. 5 is a diagram for describing a method of detecting a movement of a train in a platform monitoring area of FIG. 2.

6 is a flowchart illustrating a stereo image-based platform monitoring method by the system of FIG. 1.

FIG. 7 is a flowchart for describing the step S140 shown in FIG. 6 in more detail.

FIG. 8 is a flowchart for describing the operation S150 shown in FIG. 6 in more detail.

<Explanation of symbols for the main parts of the drawings>

100: information collecting unit 110: a stereo camera

120: sensor module 200: information processing unit

210: train detection module 220: object detection module

230: object recognition module 240: object tracking module

250: output module 300: information providing unit

Claims (20)

An information collector including at least one stereo camera for acquiring image information of a platform monitoring area including a train area into which a train enters; And Based on the image information obtained by the information collecting unit, the object detection module for determining whether the object falls in the danger zone that is preset to the area in which the object falls in the train area to interfere with normal train operation And an object recognition module for determining whether an object fallen in the dangerous area is a dangerous object that impedes normal train operation based on a preset criterion based on the three-dimensional coordinate information of the object by the stereo camera. Stereo image-based platform monitoring system comprising a processing unit. The method of claim 1, The object recognition module, The stereo image-based platform, characterized in that for calculating the size of the object falling in the dangerous area based on the three-dimensional coordinate information of the object by the stereo camera, and if the calculated size is more than the set size. Monitoring system. The method of claim 1, The information processing unit, If the falling object is determined to be the dangerous object, a stereo image further comprising: an object tracking module for assigning a unique identification code to the fallen object, and tracking the movement of the object to which the identification code is assigned. Based platform monitoring system. The method of claim 3, The object tracking module, When the movement of the object to which the identification code is assigned is detected, if the object is out of the dangerous area during the tracking of the object by continuing the tracking of the object, the assigned identification code is deleted. Stereo image based platform monitoring system. The method of claim 3, The object tracking module, If the movement of the object to which the identification code is assigned is not detected, comparing the current image with a pre-stored background image and determining that the object to which the identification code is assigned does not exist in the dangerous area, Stereo image-based platform monitoring system characterized in that the deletion. The method of claim 3, The object tracking module, And a platform for tracking the movement of the object to which the identification code is assigned based on the image difference between the current frame and the previous frame of the image information. The method of claim 1, The information processing unit, Stereo image-based platform monitoring system further comprises a train detection module for determining the state of the train in the train area. The method of claim 7, wherein The train detection module, Stereo image-based platform monitoring system for determining the state of the train in the train area based on the image information. The method according to claim 1 or 7, The object detection module, Stereo image-based platform monitoring system, characterized in that the operation only when there is no train in the train area. The method of claim 1, The object detection module, And determining whether an object falls within the dangerous area based on an image difference between a current frame and a previous frame of the image information. Obtaining image information on a platform monitoring area including a train area into which a train enters using at least one stereo camera; Determining whether an object falls within a danger area that is previously set as an area where objects fall in the train area to interfere with normal train operation based on the acquired image information; And And determining whether an object fallen in the dangerous area is a dangerous object that impedes normal train operation based on a preset criterion based on the three-dimensional coordinate information of the object by the stereo camera. Image-based platform monitoring method. The method of claim 11, Determining whether or not the dangerous object, The stereo image-based platform, characterized in that for calculating the size of the object falling in the dangerous area based on the three-dimensional coordinate information of the object by the stereo camera, and if the calculated size is larger than the set size. Monitoring method. The method of claim 11, And if the fallen object is determined to be the dangerous object, assigning a unique identification code to the fallen object and tracking the movement of the object to which the identification code is assigned. Monitoring method. The method of claim 13, Tracking the movement of the object, If the movement of the object to which the identification code is assigned is detected, continuing tracking the object and determining that the object is out of the dangerous area during tracking of the object, and deleting the assigned identification code. Stereo image-based platform monitoring method characterized in that. The method of claim 13, Tracking the movement of the object, If the movement of the object to which the identification code is assigned is not detected, comparing the current image with a pre-stored background image and determining that the object to which the identification code is assigned does not exist in the dangerous area, Stereo image-based platform monitoring method comprising the step of deleting. The method of claim 13, Tracking the movement of the object, 3. The method of claim 3, further comprising: tracking a movement of an object to which the identification code is assigned based on an image difference between a current frame and a previous frame of the image information. The method of claim 11, Before the step of determining whether the object falls, Stereo image-based platform monitoring method further comprising the step of determining the state of the train in the train area. The method of claim 17, Determining the state of the train, Stereo image-based platform monitoring method for determining the state of the train in the train area based on the image information. The method according to claim 11 or 17, Determining whether or not the object falls, Stereo image-based platform monitoring method characterized in that it is performed only if there is no train in the train area. The method of claim 11, Determining whether or not the object falls, And determining whether an object falls within the danger area based on an image difference between a current frame and a previous frame of the image information.

Images