JP5597057B2 - Passenger drag detection system at home - Google Patents

Description

The present invention relates to a passenger dragging detection system at a home, and in particular, a plurality of MEMS sensors ( area type distance sensors ) that function as distance sensors for the detection areas set in each are arranged in a row in an upper region along the home edge. In addition, the present invention relates to a passenger dragging detection system at a home that detects a passenger dragging accident using the detection operation characteristics of the plurality of MEMS sensors.

As conventional techniques for detecting a fall from a home, there are techniques described in Patent Documents 1 and 2.

The home fall detection device disclosed in Patent Document 1 is arranged at predetermined intervals along the track as shown in FIG. 1, and the track from the home toward the track as shown in FIG. A captured image of each fall detection area is acquired by a plurality of image capturing means provided at an upper position in such an arrangement posture as to look into the track side obliquely. Based on the image data related to the captured image relating to each fall detection area acquired by the plurality of image pickup means, first set the fall detection sample data, and then acquire the current image data, and the current image data and the fall A difference image with the detected sample data is obtained, and the fall of the object is detected based on the number of pixels in the difference area of the difference image. The image imaging means is a so-called imaging camera.

In the home fallen person detection method disclosed in Patent Document 2, a plurality of stereo cameras are arranged at predetermined intervals along the edge of the home, and each detection area of the plurality of stereo cameras is imaged. In the image processing using the image signal obtained in step 1, the falling object from the home is detected, the size of the falling object is obtained, and whether the falling object is a person is determined from the size. .
At this time, in the image processing, the presence of a falling object is determined on the condition that the height is lower than a predetermined height set below the top surface of the home, and the distance from the stereo camera to the falling object is obtained, and the distance is calculated. Based on this, the height of the fallen object is obtained and the number of screens occupied by the fallen object is obtained.

JP 2003-276605 A (FIGS. 1 and 2) JP 2003-246268 A

In the home fall detection device based on the above Patent Documents 1 and 2, an imaging camera (including a stereo camera) is used as detection means to detect an accident such as a fall that occurs on the home based on a captured image. Yes. Since the fall accident is determined based on the number of pixels on the captured image using information related to the image captured by the camera, the person or object that falls on the track is detected, and the fact that it has fallen is detected. However, it was not possible to quickly and accurately detect an accident in which a passenger was dragged with the traveling operation of a train, for example, by being caught in a train door.
In addition, image data obtained by imaging with a camera has a problem that it is not easy to detect a passenger who is affected by disturbance light and may be dragged.
Automatic detection of passenger drag accidents at the time of departure from a train is technically difficult with conventional fall detection devices, and has been performed artificially based exclusively on visual confirmation work.

In view of the above-described problems, an object of the present invention is to detect an accident in which a passenger is dragged by a train along a platform edge on a station platform and can automatically and accurately detect this. It is to provide a passenger drag detection system at home.

In order to achieve the above object, a passenger dragging detection system at home according to the present invention is configured as follows.

The passenger drag detection system at the first home (corresponding to claim 1)
Train is Installation along the station home edge of the incoming lines, for each meter measurement time, a plurality of area type distance sensor for outputting a detection image of the detection area is set,
And that SL憶means to store data relating to passenger image based on the detection image area type distance sensor outputs,
Change data related to the movement of the passenger image with the passage of time of different area type distance sensors based on the data related to the passenger image is obtained, and the change data indicates the movement change along the station platform edge compared to the determination data and determining the processing means and the passenger dragging state occurs when,
It is characterized by having.

In the above-mentioned passenger drag detection system at the platform, an area type distance sensor such as a plurality of MEMS sensors each having a detection area set at a predetermined location on the edge of the platform where a train enters or stops at the station platform is arranged. Based on the height information of the detection area obtained by the mold distance sensor, the movement of the passenger in the detection area is detected, and a movement pattern peculiar to passenger dragging is used as judgment data (movement change pattern along the station platform edge) It is possible to prepare, compare, and detect the passenger drag state accurately and quickly.

The passenger dragging detection system at the second home (corresponding to claim 2) is preferably configured such that, in the above-described configuration, trains are arranged in parallel with the detection images of two adjacent area-type distance sensors in the plurality of area-type distance sensors. The change data of the passenger image is acquired based on the data related to the passenger image moving in the running direction. Since the movement of the train in the parallel running direction occurs in the passenger dragging state, it is possible to detect a movement pattern in a more suitable passenger dragging state by using detection by two adjacent area type distance sensors. .

In the passenger drag detection system at the third home (corresponding to claim 3), the storage unit preferably relates to the train image based on the detection image output from each of the plurality of area type distance sensors. storing data, processing means obtains the changed data according to the movement of passengers image based on data relating to passenger image, the passenger image and trains image based on the data relating to the change data and train image passenger image The positional relationship data is acquired, and it is determined that a passenger dragging state has occurred when the positional relationship data and the determination data are compared to indicate a movement change along the station platform edge . By detecting based on the positional relationship between the movement of the passenger image and the movement of the train image, it is possible to detect the drag state of the passenger with higher detection accuracy.

Fourth home Passenger drag detection system (corresponding to claim 4), in the above configuration, preferably, the boundary of the detection area of each of the two area type distance sensor adjacent in the plurality of area type distance sensors Are in contact with each other or have an overlapping portion. In setting the detection areas of two adjacent area type distance sensors , the determination algorithm can be easily created by touching or overlapping the boundary portion.

The passenger dragging detection system at the fifth home (corresponding to claim 5) is preferably configured as described above, wherein the determination data is the passenger movement in the parallel running direction of the train in the danger area image in the detection image. It is the data concerning an image, It is characterized by the above-mentioned.

According to the present invention, the area including the platform edge of the station platform is two-dimensionally scanned with a laser beam using area type distance sensors such as a plurality of MEMS sensors, and the height information at the platform edge is obtained. By identifying the passenger image from the change in information and associating it with the movement information of the train, the occurrence of the passenger drag state can be automatically and accurately detected from the movement pattern of the passenger image. For this reason, when an accident occurs in which a passenger is dragged by a train along a platform edge on a station platform, this can be automatically and quickly detected with high accuracy.

It is a top view which shows the concrete arrangement configuration in the home of the sensor system used with the passenger dragging detection system in the home which concerns on this invention. It is a front view which shows the specific arrangement configuration in the home of a sensor system. It is a figure which shows an example of the detection image obtained by the measurement of one MEMS sensor. It is a block diagram which shows the structure of the detection and determination system of the passenger dragging detection system in the home which concerns on this invention. Three examples of detection images obtained by measurement operation of each MEMS sensor (A) When in a safety area, (B) When in a danger area, (C) A case of suddenly moving from the danger area to the safety area It is. It is a top view which shows the condition where the passenger has stayed in the vicinity of the said train when the train which entered and stopped at the platform starts again and starts moving. It is a figure which shows the change state of the detection image of two MEMS sensors (SE1, SE2) with progress of measurement time t1, t2, t3 in a time axis (t). It is a flowchart which shows the flow of the detection operation and determination operation | movement of a passenger drag state of the passenger drag detection system in a home. It is a figure which shows an example of the preservation | save table at a certain measurement time preserve | saved at a status memory (41). It is a figure which shows an example of the preservation | save table at a certain measurement time preserve | saved at a status memory (42).

  DESCRIPTION OF EMBODIMENTS Preferred embodiments (examples) of the present invention will be described below with reference to the accompanying drawings.

A passenger drag detection system at a home according to the present invention is a sensor system for detecting a passenger that may cause a passenger drag accident, and whether or not a drag state has occurred with respect to the detected passenger accurately and quickly. And a determination system having a determination algorithm related to accident detection that is automatically detected.

With reference to FIG. 1 and FIG. 2, the structure regarding the specific arrangement | positioning in the home of a sensor system is demonstrated. FIG. 1 shows a plan view and FIG. 2 shows a front view.

In FIG. 1 and FIG. 2, reference numeral 11 is a station platform viewed from above, and 12 is a train entering the platform 11 in the direction of the arrow as viewed from above. 11A is a home edge formed at a position along the track along which the train 12 enters. The home edge 11A is a boundary between the home 11 and the track. A plurality of (for example, n) sensors SE1, SE2,..., SEn are arranged at predetermined equal intervals in a line along the home edge 11A, preferably at a position inside the home edge 11A. Installed based on relationships. In FIG. 1 and FIG. 2, seven sensors are shown for convenience. The n sensors SE1, SE2,..., SEn are MEMS sensors. Hereinafter, it is described as “MEMS sensors SE1, SE2,..., SEn”. Here, “MEMS” means the well-known “Micro Electro Mechanical System”. In addition, the “MEMS sensor” includes a configuration of a detection unit that two-dimensionally scans laser light and a configuration of a distance measurement device based on an optical pulse time-of-flight measurement method, and in a specific detection area (two-dimensional region) that is allocated. This is an area sensor that can acquire three-dimensional information ("height information" in this embodiment) based on a distance measuring function. In other words, the “MEMS sensor” is an area type distance sensor. In this embodiment, an example using a MEMS sensor as a typical area type distance sensor will be described. However, the area type distance sensor used in the present invention is not limited to the MEMS sensor, and various types of area type distance sensors are used. be able to.

A plurality of MEMS sensors are mounted on a mounting frame 13 provided upward along the home edge 11A.
Are installed at regular intervals. Each MEMS sensor S installed on the mounting frame 13
E1, SE2,..., SEn are arranged with their detection parts facing downward (or obliquely downward),
A rectangular detection area 14 is set below. The detection area 14 is set at a desired height position from the floor surface of the home 11. Each of the plurality of MEMS sensors SE1, SE2,..., SEn is arranged corresponding to the entrance door 12A of each vehicle of the train 12 which enters the platform 11 and stops. Therefore, the detection area 1 corresponding to each MEMS sensor SE1, SE2,.
4 is normally set so as to include the area of the entrance / exit of each entrance / exit door 12 </ b> A of each vehicle of the train 12. The entrance / exit area of the entrance / exit door 12A is an area in which a drag accident is likely to occur due to the clothes, bags, or the like being caught by the passenger's door when the passenger gets in / out. Therefore, the detection area 14 is set so as to include the entrance / exit area. Multiple MEMS
The sensors SE1, SE2,..., SEn function as area sensors for obtaining height information (two-dimensional distance information) of the corresponding detection areas 14.

In FIG. 2, two MEMs adjacent to each other by a plurality of MEMS sensors SE1, SE2,.
The relationship between each detection area 14 of S sensor is shown. In this illustrated example, the boundary portion between two adjacent detection areas 14 is set to contact. The boundary part of the detection area 14 can also be set so as to have a common overlapping area. In FIG. 2, 14A shows the scan profile (scan contour) of the outermost laser beam. With respect to the positional relationship between two adjacent detection areas 14, the common area of the boundary portion can be set to almost zero (contact), or the detection areas 14 can be set apart from each other. 2 adjacent
When setting so that the detection areas of two MEMS sensors are separated from each other, a prediction process is required by a determination algorithm based on height information described later.

Each of the n MEMS sensors SE1, SE2,..., SEn installed as described above
At a detection time (measurement time) set at a predetermined time interval, a detection operation related to height information is individually performed as a distance measuring sensor device for the corresponding detection area 14. The height information about the corresponding detection area 14 obtained by the detection operation at each detection time by each MEMS sensor is sent to the processing computer, and based on the height information in the detection area 14 at each detection time by the processing computer. The detected object information for each detection area 14 is acquired. Based on the acquired height information, it is possible to detect the presence of a passenger in the detection area 14, the presence of a passenger car of the train 12, and the like using a threshold value. The detected object information is presence information such as passengers. The acquired height information and detected object information are stored in the memory in a table format as detection information for each of the n MEMS sensors SE1 to SEn for each detection time.

Further, according to the detection information obtained by each of the n MEMS sensors SE1, SE2,..., SEn (detection image), on the processing computer side, regarding the corresponding detection area 14,
Based on the above-mentioned passenger presence information and the pre-position information regarding the installation position, the home edge 1
The position of 1A, the position of the dangerous area near the home edge 11A, and the positions of other safety areas are known.

FIG. 3 shows an example of a detected image obtained by measurement with one MEMS sensor. 21
It is the detection image which showed the corresponding detection area 14 as a two-dimensional rectangular area. This detected image 2
1, 22 is a line image showing the position of the home edge 11 </ b> A, 23 is a danger area image, 24 is a safety area image, and 25 is a passenger image existing in the detection area 14.
In FIG. 3, the upper side of the line image 22 is actually an area on the track track side. In practice, the danger area image 23 is usually an area between the home edge 11A and the approach danger line.

Next, with reference to FIG. 4, the structure of the detection / judgment system of the passenger dragging detection system in the home which concerns on this invention is demonstrated.

In the passenger dragging detection system at home, n MEMS sensors SE1, SE2, ...
, SEn time series image data is acquired, and the presence or absence of a passenger's drag state is automatically detected based on changes in the image content of these image data as time elapses when the train 12 arrives and departs It has a functional part of a detection / determination system.

The detection / judgment system accurately and promptly determines whether or not a dragged state has occurred for a passenger based on the memory and the memory that extracts and stores the passenger and train presence information detected by the sensor system. It is comprised from the arithmetic processing part (CPU) which performs the determination algorithm which concerns on the accident determination automatically detected and determined.

In the configuration shown in FIG. 4, as an example, seven MEMS sensors SE1 to SE7 are shown for convenience. Actually, the output lines of the n MEMS sensors SE1 to SEn are connected to the switch unit 3.
1 is connected. A signal related to the height information of each detection area 14 output from each of the n MEMS sensors SE1 to SEn is input to the switch unit 31 having an input selection type switch function. Output signals of the plurality of MEMS sensors input to the switch unit 31 are selectively captured by the switch unit 31 to the arithmetic processing device 32 side at an appropriate timing. The arithmetic processing unit 32 has three types of memories 41, 42, and 43 and five functional units realized by executing a predetermined program by the CPU, that is, a passenger tracking unit 51, a train home approach detection unit 52, and a train. A stop detection unit 53, a train departure detection unit 54, a passenger drag detection start unit 55, and a passenger drag determination unit 56 are provided. Further, an output device 61 is provided on the output side of the arithmetic processing device 32.
Is provided.

Each of the memories 41 to 43 is a status memory (state storage memory). The status memory 41 is a memory for storing a state relating to the presence and movement of a passenger (person). The status memory 42 is a memory that stores a state relating to the presence and movement of a train. The status memory 43 stores the combined state of the storage contents of the status memories 41 and 42, and further has a positional relationship between the train and a specific passenger (temporal state change or state transition). Is a memory for storing. The stored data relating to the movement having this relationship is used as basic data for subsequent passenger drag determination.

  The function contents (processing contents) of the five function units (51 to 56) are as follows.

The passenger tracking unit 51 takes in the detection image 21 output from each of the n MEMS sensors SE1 to SEn at each detection time, and based on the passenger image 25 appearing in the plurality of detection images 21, the passenger image 25 is identical. And tracking the movement of the passenger based on the change in the position of the passenger image 25. Status information relating to the passenger tracked by the passenger tracking unit 51 (specific information of each passenger, position information, position change information, etc.) is stored in the status memory 41.

The train home entry detection unit 52 captures the detection images 21 output from each of the n MEMS sensors SE1 to SEn at each detection time, and a plurality of the detection images 21 located on the train entry side.
2 has a function of detecting the home approach of the train 12 based on the train image appearing on the track side of the line image 22. Normally, the home approach of the train 12 is detected based on the movement of the height information of the front part of the front vehicle of the train 12. The status information related to the train platform approach detected by the train platform entry detector 52 is stored in the status memory 42.

The train stop detection unit 53 captures the detection image 21 output from each of the n MEMS sensors SE1 to SEn at each detection time, and for example, a line image in the plurality of detection images 21 existing at positions corresponding to the head of the train. It has a function of detecting the stop of the train 12 based on the stop operation state of the train image appearing on the 22 track tracks side. Normally, the stop of the train 12 is detected based on the stop operation state based on the height information of the front portion of the front vehicle of the train 12. Train stop detection part 5
The state information relating to the stop of the train detected in 3 is stored in the status memory 42.

The train departure detection unit 54 captures a detection image 21 output from each of the n MEMS sensors SE1 to SEn for each detection time, and for example, a line image 22 in a plurality of the detection images 21 existing at positions corresponding to the head of the train. The train 12 has a function of detecting the departure of the train 12 based on the departure operation state of the train image appearing on the track track side. Usually, the departure of the train 12 is detected based on the operation start state based on the height information of the front portion of the front vehicle of the train 12. Train departure detection unit 54
The status information relating to the departure of the train detected in (1) is stored in the status memory 42.

The passenger dragging detection start unit 55 includes state information (state data) relating to the tracking target passenger stored in the status memory 41, and state information of the train 12 entering the platform, stopping, and starting, stored in the status memory 42. (State data) is taken in, and processing for associating the state / motion of the passenger on the platform 11 with the state / motion of the train on the track is executed.
The association process by the passenger dragging detection start unit 55 is executed on condition that the departure signal SI1 related to the departure of the train 12 is input. The process of the association by the passenger dragging detection start unit 55 is a precondition for starting detection of a state in which the passenger is dragged. Passenger and train 1 created by the association process by the passenger dragging detection start unit 55
The two related state information is stored in the status memory 43.

The passenger dragging determination unit 56 compares the passenger and train relation state information stored in the status memory 43 with the determination data (danger pattern data) 71, so that the train 1
It is determined whether or not a passenger dragging state according to No. 2 has occurred, and when it matches the determination data 71, detection / determination that a passenger dragging state has occurred is performed. When the passenger dragging state is detected, the passenger dragging determination unit 56 outputs a detection signal to the output device 61. The output device 61 is, for example, an alarm signal generator or a stop signal generator. The alarm signal generator outputs an alarm signal and generates an alarm sound and an alarm light in a related department or a related peripheral area. The stop signal generator outputs a stop signal such as urgently stopping the train 12 that has started departure at the platform 11 to the train operation control system.

In the description of the present embodiment, because of the configuration, detection / determination of passenger dragging is as described above.
This is performed by using the passenger and train related state information (information of each position and movement (position change) and relationship) stored in the storage means. On the other hand, in the danger area image 23 of the detected image 21 It can also be configured to detect / determine passenger drag by using only the information on passenger position and movement (position change).

Next, the operation of the detection / determination system of the passenger dragging detection system at the home according to the present invention will be described with reference to FIGS.

As described in FIG. 3, according to the measurement operation of each of the n MEMS sensors SE <b> 1 to SEn, the output information is based on the assumption that the passenger 12 exists in each corresponding detection area 14. Thus, the detection image 21 as shown in FIG. 3 can be obtained. At this time, if a train is present in the detection area 14, a train image is also displayed.

An example of a detection image obtained based on the measurement operation of each of the n MEMS sensors SE1 to SEn is shown again in FIG.

The detected image 21 shown in FIG. 5A has exactly the same contents as the detected image described in FIG. In this case, since the passenger image 25 exists in the safety area image 24, it can be determined that the passenger is in a safe state in the positional relationship with the train 12 entering the home 11.

In the detection image 21 shown in FIG. 5B, the passenger image 25 remains in the danger area image 23 for a certain period of time and the danger image is associated with the passenger image 25 even though the train 12 starts moving. The stay state in 23 continues. Such a detected image 21 is displayed on the home 11.
This means that there is a high possibility that a dangerous situation has actually occurred with respect to the passenger on the top.
In this case, in particular, when a state in which the passenger runs in the same direction as the train (indicated by the arrow 72) is generated in the detection image 21, a dangerous state (passenger dragging state) has occurred. To be judged.

On the other hand, as shown in the detection image 21 in FIG. 5C, when the passenger image 25 existing in the danger area image 23 moves toward the safety area image 24 (the movement is indicated by an arrow 73). Is
For example, it is determined that a passenger getting off from the train 12 has normally moved from a dangerous area near the entrance of the train 12 to the safety area, and is determined to be safe.

The three detection images 21 shown in FIGS. 5A to 5C show detection images obtained by measurement based on a single MEMS sensor. (A) is one detected image at a certain measurement time. (C) makes a determination based on a state (arrow 73) in which the passenger image 25 has moved from the danger area image 23 to the safety area image 24, so that the determination is made by comparing the detected images at at least two measurement times. It is. In the case of (B), as in (C), the passenger image 25
Therefore, it is necessary to compare the detected images at at least two measurement times. However, in the movement (arrow 73) in the case of (C), since the passenger is moving in a direction away from the home edge 11A, it can be determined from a detection image obtained by measurement of a single MEMS sensor. On the other hand, in the movement (arrow 71) in the case of (B), since the passenger moves in the direction parallel to the train, it cannot always be determined by the detection image obtained by the measurement of the single MEMS sensor, and the movement destination It is necessary to make a judgment based on a combination with a detected image obtained by measurement of another adjacent MEMS sensor and a comparison.

On the premise of the above measurement operation based on a sensor system composed of n MEMS sensors SE1 to SEn, the operation of the detection / determination system of the passenger dragging detection system at home according to the present invention is executed.

FIG. 6 shows a situation in which the passenger 74 stays in the vicinity of the train 12 when the train 12 that has entered or stopped the platform 11 starts again and starts moving. In FIG. 6, three measurement times (detection times) t1, t2, and t3 are shown as an example regarding the detection of the train 12 and the passenger 74. In this situation, it is assumed that the passenger 74 is dragged by the train 12. According to the passenger drag detection system at the home according to the present invention, the occurrence state of the passenger drag is detected and determined as follows.

In the case of detection / determination of passenger dragging in the example shown in FIG. 6, in the detection operation of the n MEMS sensors SE1 to SEn at the measurement times t1 to t3, the movement position of the passenger 74 (locations 74A, 74B, 74C). ), Especially two adjacent MEMS sensors SE1, SE2
The data related to the detected image by the measurement is used.

FIG. 7 shows how the detected images of the two MEMS sensors SE1, SE2 change with the passage of the measurement times t1, t2, t3 on the time axis (t).

At the measurement time t1, since the passenger 74 is at the location 74A, the passenger 74 exists in the detection area 14 of the MEMS sensor SE1, but does not exist in the detection area 14 of the MEMS sensor SE2. At the measurement time t1, the detection image 21 of the MEMS sensor SE1 is an image 81 including a passenger image 25.
The detected image 21 of the MEMS sensor SE2 becomes an image 91 that does not include a passenger image. In the image 81, the passenger image 25 exists in the danger area image 23.

At the measurement time t2, since the passenger 74 is in the next place 74B, the MEMS sensor SE still remains.
1 exists in the detection area 14, but does not yet exist in the detection area 14 of the MEMS sensor SE1. At the measurement time t2, the detection image 21 of the MEMS sensor SE1 is a passenger image 25.
And the detected image 21 of the MEMS sensor SE2 does not change and becomes an image 92 that does not include a passenger image. In the image 82, the passenger image 25 exists in the danger area image 23, but has moved in the direction of the arrow 72 along the line image 22 as compared with the previous image 81.

At the measurement time t3, since the passenger 74 is further in the next place 74C, the MEMS sensor SE
The existence state in one detection area 14 disappears, and on the other hand, it exists in the detection area 14 of the MEMS sensor SE2. At the measurement time t3, the detection image 21 of the MEMS sensor SE1 is as an image 83 that does not include the passenger image 25, and the detection image 21 of the MEMS sensor SE2 is as an image 93 that includes the passenger image 25. In the image 93, the passenger image 25 exists in the danger area image 23.

As described above, the two MEMS sensors SE1, S in the lapse of time t1, t2, t3.
Since a change state occurs in the detection image 21 obtained by E2, the change state of the detection image 21 of the two MEMS sensors SE1 and SE2 is monitored, and the presence or absence of the passenger dragging state is detected by detecting the characteristic change. Can be determined. That is, in this case, based on the detection data of the MEMS sensor SE1 having the detection image 21 including the passenger image 25 in the danger area image 23, the change state of the detection image 21 in the MEMS sensor SE2 adjacent to the MEMS sensor SE1 is shown. See if there is a passenger dragging state. More specifically,
Hazardous area image 2 of detected images (images 81 and 82) at times t1 and t2 of the MEMS sensor SE1
3 and there is a passenger image 25 moving in the parallel running direction (arrow 72) and the detected image at time t3 (
The passenger image 25 disappears in the danger area image 23 of the image 83), the passenger image 25 does not exist in the detection images (images 91 and 92) of the MEMS sensor SE2 at the times t1 and t2, and the detection image (image 93) of the time t3. ) Is determined as “occurrence of passenger dragging state” on the basis of the characteristic change state that the passenger image 25 is generated in the danger area image 23 of FIG.

  The contents of the reference for the characteristic change state are the determination data 71 described with reference to FIG.

With reference to the flowchart shown in FIG. 8, the flow of the detection operation and determination operation | movement of the passenger drag state of the passenger drag detection system in a home is demonstrated.

By the steps S11 and S12, n MEMS sensors SE1, SE2,.
The signal detected in is captured. The detection signal is taken in repeatedly at every measurement time set at a predetermined time interval in step S11.

In step S13, a passenger is detected as a detection object based on the height information for each detection area 14 obtained by each of the n MEMS sensors SE1 to SEn, and the state information (specific information of each passenger) related to the passenger is detected. , Position information, position change information, etc.) are stored in the status memory 41 in a table format. FIG. 9 shows an example of a storage table 101 stored in the status memory 41 at a certain measurement time. In this storage table 101, n MEMS sensors SE1 to S1.
For each En, for example, “position”, “detected object”, and “height information” in the detection area 14 are recorded. Here, the “position” is any one of the dangerous area, the safety area, and the outside of the home edge (track track side) in the detection area 14. "Detected object"
Is the presence or absence of a passenger image. “Height information” is height information of the “detected object”, and is used, for example, for determination of an adult or a child. The storage table 101 is created for each measurement time.

Based on the steps S11, S12, S13, the n MEMS sensors SE1, SE
2,..., SEn realizes the above passenger tracking unit 51 for tracking a passenger in a dangerous area or the like set in the vicinity of the home edge 11 </ b> A of the home 11.

Steps S14 to S18 are executed in a parallel processing relationship with respect to step S13. These steps S14 to S18 take in height information for each detection area 14 obtained by each of the n MEMS sensors SE1 to SEn, and n MEMS sensors SE1 to S18.
The train 12 based on the change in the height information with the passage of time in each detection area 14 of En.
Presence / absence (presence) and movement of the train 12 are detected. The detection of the presence of the train 12 is performed by detecting a change in height existing in the region on the track track side outside the line image 22 in the detection image 21 related to the detection area 14 described above. n MEMS sensors SE
The height information for each detection area 14 obtained in each of 1 to SEn is stored in a temporary storage memory for the detected images 21 obtained at a predetermined number of continuous measurement times (step S1).
4) Based on the information related to the height image stored in the storage memory, detection of the home approach operation of the train 12 (step S15), detection of the stop operation (step S16), detection of the departure operation (step S17) Is executed. Information on the detected home approach operation, stop operation, and departure operation relating to the train 12 is stored in a table format in the status memory 42 in association with each of the n MEMS sensors SE1 to SEn as a positional relationship ( Step S18
). FIG. 10 shows an example of the storage table 102 stored at a certain measurement time stored in the status memory 42.

In the above, the train platform approach detection unit 52 described above is realized in step S15, the train stop detection unit 53 described above is realized in step S16, and the train departure detection unit 54 described above is realized in step S17. Thus, in the status memory 42, the status information related to the train platform approach, the status information related to the stop of the train, and the status information related to the departure of the train are associated with the passage of time, and the n MEMS sensors SE1 to SEn. It is stored in a table format in association with each position of each detection area.

Next, it is determined whether a passenger dragged state has occurred on the condition that a departure signal for the train is generated.

In step S19, it is determined whether or not a departure signal for the train 12 has been generated. Step S1
In the case of YES at 9, the passenger dragging detection process is started in the next step S <b> 20.
In this step S20, a process for synthesizing the first and second storage tables 101 and 102 stored in the two status memories 41 and 42 is performed to form a third storage table. The third storage table is stored in the status memory 43. In the third storage table stored in the status memory 43, n MEMS sensors SE1 to S1.
Information related to the positional change relationship between the passenger image 25 and the train image is included in at least one detection image of En. For this reason, in the next step S21, the positional change relationship information between the passenger image and the train image is extracted based on the record information of the third storage table stored in the status memory 43.

In the next determination step S22, the extracted positional change relation information is compared with the above-described determination data 71 to determine whether or not they match. If NO in the determination step S22, it is determined that no passenger dragging has occurred, and the detection process is terminated. Y in determination step S22
If it is ES, an alarm signal and a stop signal are output to an external device (step S23). Thereafter, the detection process ends.

In the above, the passenger drag detection start unit 55 is realized by step S20, and the passenger drag determination unit 56 described above is realized by steps S21 and S22.

According to the above, in the home fall detection system, passengers are detected using the detected image information of the required number of MEMS sensors SE1, SE2,..., SEn arranged at a predetermined upper position corresponding to the home edge 11A of the home 11. Image and train image, and information related to the relationship between passenger image and train image is acquired, and this is compared with judgment data (the relationship shown in FIG. 7) for accurate and rapid passenger dragging. Automatic detection is possible, and the occurrence of an accident can be prevented by performing an emergency safety operation.

As described above, the position and movement of the passenger in the danger area image 23 of the detection image 21 (
It is also possible to detect / determine passenger drag using only information on (position change).

The configurations, shapes, sizes, and arrangement relationships described in the above embodiments are merely schematically shown to the extent that the present invention can be understood and implemented. Therefore, the present invention is not limited to the described embodiments, and can be variously modified without departing from the scope of the technical idea shown in the claims.

A passenger dragging detection system at a home according to the present invention is configured such that a plurality of MEMS sensors are installed at a required arrangement position with respect to a detection area along a home edge of a platform where a train at a station enters, and ranging of the MEMS sensors is performed. The height information of the corresponding detection area is acquired using the function, and information related to passenger movement is obtained based on the height information to detect the passenger drag state. Since the passenger dragging detection system at the home uses the MEMS sensor of the two-dimensional area sensor to detect the object based on the height information in the detection area, the passenger trains along the edge of the platform on the platform. This is used to automatically and quickly detect an accident that is dragged by a high accuracy.

DESCRIPTION OF SYMBOLS 11 Home 11A Home edge SE1-SEn MEMS sensor 12 Train 13 Mounting frame 14 Detection area 14A Scan profile 21 Detection image 22 Line image 23 Hazardous area image 24 Safety area image 25 Passenger image 31 Switch part 32 Calculation processing part 41,42,43 Status memory 51 Passenger tracking section 52 Train platform approach detection section 53 Train stop detection section 54 Train departure detection section 55 Passenger drag detection start section 56 Passenger drag detection section 61 Output device 71 Data for determination 74 Passenger 81, 82, 83 Image (MEMS) Detection image of sensor SE1)
91, 92, 93 images (detected images of MEMS sensor SE2)
101 Storage table 102 Storage table

Claims (5)

Train is Installation along the station home edge of the incoming lines, for each meter measurement time, a plurality of area type distance sensor for outputting a detection image of the detection area is set,
Said area type distance sensor is the detection to that Symbol憶means stores data related to passenger image based on the image to be output,
Acquires the change data according to the movement of the passenger image with the passage of time of said different area type distance sensor based on data relating to prior Symbol passenger image, the change data by comparing the verification data is the station platform edge Processing means for determining that a passenger dragging state has occurred when indicating a movement change along ;
A passenger dragging detection system at home characterized by comprising: Obtaining the change data of the passenger image based on the data relating to the passenger image moving in the parallel running direction of the train, with detection images of two adjacent area type distance sensors in the plurality of area type distance sensors. The passenger dragging detection system at home according to claim 1. The storage means may store data relating to train image based on the detection image, each of said plurality of area type distance sensor outputs,
Said processing means obtains the changed data according to the movement of the passenger image based on data relating to prior Symbol passenger image, and the passenger image on the basis of the data relating to the change data and the previous SL train image of the passenger image Acquiring positional relationship data with the train image, and comparing the positional relationship data with the determination data to determine that a passenger dragging state has occurred when a movement change along the station platform edge is indicated. 3. The passenger dragging detection system at a home according to claim 1 or 2.   The boundary part of the detection area of each of the two adjacent area-type distance sensors in the plurality of area-type distance sensors is in contact with each other or has an overlapping portion. The passenger drag detection system at the home described in the section.   5. The determination data according to claim 1, wherein the determination data is data related to a passenger movement image in a parallel running direction of the train in a danger area image in the detection image. 6. Passenger drag detection system at home.

Images