JP2012121483A - Apparatus for detecting situation of platform - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain an apparatus for detecting situations of a platform, which divides a region to be monitored at the platform into a plurality of detection areas and performs determination according to each of the detection areas.SOLUTION: The apparatus for detecting situations of the platform includes: an image input part for photographing an entrance/exit gate of a train, which is an opening of a fixed fence of the platform, and the periphery thereof and inputting platform image information; and a detection control part for setting the plurality of detection areas in the platform image information from the image input part, performing image processing according to each of the detection areas and determining situations of each of the detection areas. The plurality of detection areas includes a first area being a passenger waiting area, a second area being passenger getting on/off area and a third area being a space area between a railway car and the edge of the platform. The detection control part determines, in the first area, whether there is a passenger making a dash for the train and whether waiting passengers form a line, detects, in the second area, the approach of an object to be detected to the railway car, and detects, in the third area, the size of the space area between the railway car and the edge of the platform.

Description

  The present invention relates to a home status detection device that provides information to station facility users based on an image obtained by imaging a platform of a station to be monitored from above.

  The following techniques are disclosed as conventional techniques for performing alarm detection when a person or an object is detected in a dangerous area on a platform of a station. For example, from the image, the speed of the train vehicle and the presence / absence of a person in the home are determined and alarm detection is performed (see, for example, Patent Document 1). From the image, the vehicle is stopped, the door of the movable fence is opened and closed, A technique for detecting a person or an object and detecting an alarm (for example, see Patent Document 2), and a technique for detecting an obstacle or a passenger between a movable fence of a home and a vehicle by an image (for example, see Patent Document 3). and so on.

  Moreover, the following techniques are disclosed as conventional techniques for determining the degree of congestion on the home. For example, a technique for calculating the degree of congestion based on the length of the waiting queue of the home and the density of the waiting from the image (see, for example, Patent Document 4), and the number of people on the home or the area occupied by the person is determined from the image. In addition, there is a technique for performing congestion prediction by outputting congestion information and recording and analyzing data (for example, see Patent Document 5).

JP 10-338125 A JP 2004-066694 A Patent No. 3980305 Japanese Patent Laid-Open No. 3-138274 JP 2007-201556 A

According to the conventional technology, it is not possible to detect the rush-in boarding, the waiting line alignment, the gap between the home and the vehicle, and there is little information for determining the environment of the home.
The present invention has been made to solve the above-described problems, and determines various situations of the home such as the position, number of passengers, and movement status of passengers based on images from a camera installed above the platform. An object is to obtain a home status detection device capable of performing the above.

  The home condition detection device according to the present invention photographs the train entrance / exit and the vicinity of the train entrance / exit which are openings of a fixed fence provided on the platform along the end of the platform from above the platform of the station. Then, a plurality of detection areas are set for the image input unit for inputting home image information and the home image information from the image input unit, and image processing corresponding to each of the detection areas is performed. A detection control unit for determining the situation, the detection area is a first area that becomes a waiting waiting area closer to the inside of the platform from the fixed fence, a train getting on and off area from the train entrance to the platform end A second area, a third area that is a gap area between the train vehicle that has entered the station and the home end, and the detection control unit includes the home image information in the first area. In addition to performing the rush-on detection that determines whether there is a rush-in boarding from the moving direction information and the moving speed information of the detected object, the waiting-line alignment detection that determines the alignment state of the waiting customer from the position information of the detected object, From the position information of the detected object included in the home image information of the second area, proximity detection is performed to determine an approach state of the detected object to the train vehicle passing through the home, and the third area The size of the gap between the train vehicle and the platform is detected from the home image information.

  According to the home situation detection device of the present invention, a plurality of detection areas are set, image processing is performed according to each detection area, and the situation of each passenger is moved or aligned in order to determine the situation for each detection area. Can be determined in detail.

It is an arrangement plan of a home situation detection device in a station platform of the present invention. It is a station platform bird's-eye view which shows the several area detected by the home condition detection apparatus of this invention. It is a block diagram of the home condition detection apparatus of this invention. It is a detection flowchart of the last boarding detection of the home condition detection apparatus of this invention. It is a detection flow figure of waiting unalignment detection of the home condition detection device of the present invention. It is a platform top view for demonstrating the waiting line alignment state in a station platform of this invention. It is a detection flowchart for detecting the space | interval of the platform and a train vehicle of the home condition detection apparatus of this invention. It is a detection flow figure of fixed fence going over detection of the home situation detection device of the present invention. It is a station platform bird's-eye view which shows the several area detected by a home condition detection apparatus at the time of providing the movable platform fence of this invention.

Embodiment 1.
FIG. 1 shows a layout of the home status detection device according to the present invention in a station platform. FIG. 1 is a view of the platform of a station facility observed from the side. In FIG. 1, the direction of train entry (the direction in which the rail extends) matches the depth direction from the front of the drawing. As shown in FIG. 1, a camera unit (corresponding to an image input unit) 3 constituting a home status detection device is arranged on the back side of a roof 2 on a platform 1 of a station. An image is taken looking down at the entrance / exit and the area around it. The image and sound (home image information) photographed by the camera unit 3 are input by the detection control unit 4 constituting the home state detection device, and various determination processes are performed according to a plurality of detection areas. When the detection control unit 4 grasps the situation of passengers at the home 1 and requires an alarm, the determination information is output to the alarm device 5. When an image display to the passenger is necessary, for example, the home 1 The determination information is output to a display device 7 attached to a fixed fence (not a movable platform fence but a fence where the train entrance / exit is always open) 6 installed on the top.
In the arrangement example of the camera unit 3 in FIG. 1, the camera unit 3 is arranged above the home 1 and close to the train vehicle 8 that enters the home 1. You may arrange the camera part 3 in any upper position.

  In the present invention, as shown in FIG. 2, an area (detection area) photographed by the camera unit 3 is divided into a plurality of parts, and analysis determination processing corresponding to each detection area (A1 to A5) is performed. FIG. 2 is a bird's-eye view of the station platform 1 as viewed from above, showing the photographing range of the camera unit 3. As shown in FIG. 2, the fixed fence 6 stands along the direction in which the platform 1 extends, but the opening area where the fixed fence 6 is not arranged is a train to the train vehicle 8 of the passenger (person) 10. It becomes the entrance / exit 6a. The vehicle door 9 of the train vehicle 8 that has entered the station premises stops at the position of the train entrance 6 a of the platform 1. In the first embodiment, it is considered that the train entrance 6a is always open.

The detection area is, for example, a first detection area (A1) that is a waiting area in the range of the opening width of the fixed fence 6a (train entrance / exit 6a) opening width in the area closer to the inside of the platform 1 from the fixed fence 6. A second detection area (A2), which is a train getting-on / off area from the train entrance / exit 6a to the home end outside the platform 1, and a third area that is a gap area between the train vehicle 8 entering the station and the end of the platform 1 The detection area (A3) is included.
Furthermore, in addition to the above-described detection areas, as the fourth detection area (A4), the area of the door 9 portion of the train vehicle 8, and as the fifth detection area (A5), on both sides of the train entrance 6a on the platform 1. A region including the fixed fence 6 that is positioned (both sides of the first to fourth detection areas) can be set.

  FIG. 3 shows a configuration diagram of the home status detection apparatus of the present invention. As shown in FIG. 3, the home image information photographed by the camera unit 3 is output to the detection control unit 4 side. In the detection control unit 4, home image information is captured by the image input processing unit 11. An output signal from the signal input processing unit 12 is input to the image input processing unit 11. This signal input processing unit 12 receives a vehicle approach signal or a vehicle door opening / closing preparation start signal (several seconds before the door is closed). In response to the input of an external input signal such as a vehicle door open / close state signal, a signal notifying the timing of the detection process is output to the signal input processing unit 11 side.

  The home image information obtained by the image input processing unit 11 is associated with setting data (detection area reference image: an image showing a plurality of detection area allocation standards) 14 by the detection processing unit 13 for each detection area. Perform detection processing to determine the status of the home. Information and images determined by the detection processing unit 13 are recorded via the image storage unit 15 when, for example, an abnormal state is detected by distribution via a LAN or storage in a medium (CF card or the like). This makes it possible to grasp and analyze the situation.

In response to the output signal from the detection processing unit 13, the signal output processing unit 16 outputs the detected signal to an external display device / alarm device. Examples of output signals from the signal output processing unit 16 include, for example, a rush-in boarding detection signal, a waiting-for-unalignment detection signal, a home end intrusion detection signal, a home-to-vehicle excess (gap is greater than a predetermined value) detection signal, a vehicle It is a door scissors detection signal.
In response to the output signal from the detection processing unit 13, the audio output processing unit 17 outputs a detection signal to an external alarm device, and performs processing to output a sound (enlightenment sound or the like) corresponding to the detection signal. .

Next, a detection processing flow performed by the detection processing unit 13 in FIG. 3 will be described with reference to FIGS. First, the items to be detected in the first detection area (A1) are the presence / absence of the rush-in boarding and the determination of the waiting state alignment state.
At the time of rush-on boarding detection, as shown in the flowchart shown in FIG. 4, the speed of the person 10 heading from the inside of the home 1 to the train entrance / exit 6a outside the home 1 is detected and determined. First, the reference image and the current photographed image are compared (step 11 (S11)), and if there is a change in the image, the detected object in the change portion matches the registered person image pattern or is constant. Is determined (S12) and whether the temperature of the detected object is within a specified range (S13), and the person 10 is identified. The speed and direction of movement of the person 10 are calculated, heading toward the fixed fence 6 (S14), and when it is determined that the speed is equal to or higher than a certain speed (S15), as a rush-on ride (S16), a signal / audio output ( Sound and image display such as “Ride on board is dangerous”) and image accumulation. In step 17 (S17) following step 16 (S16), when the vehicle is operating or the door closing preparation signal is not input as an external input signal, or when No is determined in steps 12 to 16 Returns to step 12 and the process is repeated.

Next, in the waiting line alignment detection detected in the first detection area (A1), as shown in the flowchart of FIG. 5, the person 10 on the home 1 is detected to form a line. Determine whether or not. First, the reference image and the current photographed image are compared (S21). If there is a change in the image, whether the object of the change part matches the registered person image pattern or has a certain size. Further, it is determined (S22), and it is determined whether the temperature of the detected object is within a specified range (S23), and the person 10 is specified. When the number of persons 10 is confirmed to be equal to or greater than a certain number (threshold value) (S24), it is determined whether or not the stopped persons are lined up on the registered alignment pattern (S25). Then, it is detected that the waiting customers are not aligned (S26), signal / sound output (sound / image display such as “please wait in two rows”), and image accumulation are performed.
Here, as shown in the overhead view of the station platform in FIG. 6, when the trains are arranged in, for example, two rows before the train entrance 6a, the width for one row is considered as a constant interval, and the constant interval width is determined. When a certain number of people are lined up in the area, it is determined that they are lined up.

  In the second detection area (A2), the approach state of the person 10 to the train vehicle 8 passing through the home 1 from the position information of the person (detected object satisfying the size and temperature conditions) 10 included in the home image information. The approach detection can be performed, and when the train vehicle 8 is too close, appropriate guidance (“Train arrives”, “Train will depart”, etc.) can be performed. It becomes possible. Needless to say, the guidance can be performed by screen display in addition to the voice.

  Next, in the third detection area (A3), the size of the gap generated between the train vehicle 8 that has entered the station premises and the platform 1 by the home-vehicle gap detection flow as shown in FIG. Is detected. In this case, first, after the initial process (S31), the train vehicle 8 that has entered the track is detected, and the stop of the train vehicle 8 is determined by an image or an input signal (S32). When the train vehicle 8 stops, the size of the gap between the home end and the vehicle is determined (S33). If the value is equal to or greater than a certain value, it is determined that the home-vehicle gap is detected (S34), and the signal / audio output ("Home and Audio space and image display such as “There is a gap in the vehicle. Please be careful.”), Image accumulation.

  When the gap between the platform 1 of the station and the train vehicle 8 that enters the station is large, a device called a gap filler that serves as a horizontal step may be additionally provided on the platform to fill the gap. . When the gap filler fails or is not operating due to a power failure, the gap detection between the home and the vehicle is detected to detect the gap filler malfunction and make an appropriate announcement to the passengers. It is possible to take measures such as

  Further, in the fourth detection area (A4), from the position information of the detected object (including a person or luggage, a car such as a stroller) included in the home image information, the train door 9 closed before the train vehicle departs. The door pinching detection can be performed to determine whether there is a door pinching state where the sensing object is pinched. When the door pinching occurs, appropriate measures or announcements (“Door closes”, “Emergency stopped.” Etc.).

  Next, in the fifth detection area (A5), first, after the initial process (S41), the person 10 on the home 1 is detected (S42, S43) by the overcoming detection flow as shown in FIG. It is determined whether or not the vehicle has passed over the fixed fence 6 and has entered the train vehicle 8 (rail) side (S44). If it is determined that the vehicle has been crossed (S45), the vehicle is detected and the image storage unit 15 outputs a signal. Information is transmitted to the processing unit 16 and the audio output processing unit 17.

Embodiment 2.
In the above-described first embodiment, the case where the platform 1 is provided with the fixed fence 6 and the train entrance / exit 6a has no fence and is always open is shown. However, recently, as shown in FIG. 9, the movable platform fence 20 is provided at the train entrance / exit 6a, and the facility has been improved so that the movable platform fence 20 is opened only when the train gets on and off. Yes. When such a movable home fence 20 is provided and the movable home fence 20 is closed, a detection object that appears to be a person 10 is detected from the first detection area (A1) to the second detection area (A2). When moving, it is possible to detect that the passenger is going over the fence and going outside the platform 1.

  Thus, even if it is the home 1 in which the movable home fence 20 was installed, image processing according to each detection area divided | segmented based on the passenger's positional information by using the home condition detection apparatus of this invention. By performing signal input, image accumulation, and signal / sound output, warning, vehicle information, and riding manner information can be provided to the external sound device / display device. By using image processing, various sensors can be omitted, and it is possible to improve service to customers and improve safety.

DESCRIPTION OF SYMBOLS 1 Home 2 Roof 3 Camera part 4 Detection control part 5 Alarm apparatus 6 Fixed fence 6a Train entrance / exit 7 Display apparatus 8 Train vehicle 9 Train door 10 Passenger (person)
DESCRIPTION OF SYMBOLS 11 Image input process part 12 Signal input process part 13 Detection process part 14 Setting data 15 Image storage part 16 Signal output process part 17 Audio | voice output process part 20 A movable home fence.

Claims (4)

  Image input to photograph the train entrance / exit and the area around the train entrance / exit, which is an opening of a fixed fence provided on the platform along the end of the platform, from above the platform of the station, and input home image information A detection control unit that sets a plurality of detection areas for the home image information from the image input unit, performs image processing according to each of the detection areas, and determines the status of the detection areas; The detection area is a first area that becomes a waiting waiting area near the inside of the platform from the fixed fence, a second area that becomes a train getting on and off area from the train entrance to the platform edge, and enters the station. A third area that is a gap area between the train vehicle and the home end, and the detection control unit includes movement direction information and movement of the detected object included in the home image information of the first area. In addition to performing rush-on detection that determines whether or not there is a rush-in ride from the speed information, wait-alignment detection is performed to determine the alignment state of the waiters from the position information of the detected object, and the home image information in the second area is From the position information of the detected object included, proximity detection is performed to determine an approach state of the detected object to the train vehicle passing through the platform, and from the home image information of the third area, the train vehicle and A home condition detection device for detecting the size of a gap with the home.   The detection area includes a fourth area that is a region of the door portion of the train vehicle, and the detection control unit is configured to calculate the position of the detected object included in the home image information of the fourth area from the position information of the detected object. The home situation detection apparatus according to claim 1, wherein door pinching detection is performed to determine whether or not there is a door pinching in which the detection object is pinched by a train door.   A fifth area which is an area including the fixed fence located on both sides of the train entrance and exit on the platform, and the detection control unit includes the detection included in the home image information of the fifth area. 2. The home situation detection according to claim 1, wherein the vehicle is detected from the moving direction information and the position information of the object to detect whether or not the detected object moves over the fixed fence and moves outside the home. apparatus.   2. The home status detection according to claim 1, wherein a movable platform fence is installed at the train entrance of the platform, which is open when the train gets on and off, and is closed except when the train gets on and off. apparatus.

Images