JP7368180B2 - Station platform safety support system and obstructive object determination method - Google Patents

Description

The present invention relates to a station platform safety support system and the like.

On station platforms, station staff and conductors visually check trains to prevent accidents. Specifically, on the platform, check whether people, wheelchairs, strollers, etc. are at a certain distance from the edge of the platform when a train is approaching, or from a stopped train when a train is departing. Prevent car accidents by checking. If a certain distance is not maintained, an announcement will be made to alert people. If a person, wheelchair, stroller, etc. is caught in the gap between the train and the edge of the platform, help them. These tasks are performed manually by station staff and conductors.

Platform fences are used as a system to reduce collision accidents and to supplement the safety support work of station staff and conductors. However, the platform fence is not installed right at the side edge of the platform, and there is a gap on the platform between the platform fence and the platform side edge. Therefore, simply installing platform fences does not completely prevent collisions.

Therefore, various safety support systems for station platforms using images captured by cameras have been proposed with the aim of preventing collision accidents without increasing the human burden on station staff and conductors (for example, Patent Document 1 ,2).

Patent No. 5748468 Japanese Patent Application Publication No. 8-282494

However, conventional methods including Patent Documents 1 and 2 are based on the premise that obstructive objects such as people, wheelchairs, and strollers can be detected from captured images. Therefore, when a foreign object appears in a photographed image, there is a risk that it may be mistakenly detected as an obstructive object such as a person, wheelchair, or stroller. A good example is when a newspaper or a collection of leaves are photographed.

The problem to be solved by the present invention is to propose a technology for reducing false detection of obstructive objects as a safety support system for a station platform using images captured by a camera.

The first invention for solving the above problem is:
A first bird's-eye camera (for example, the first bird's-eye camera 20a in FIG. 4) and a second bird's-eye camera are installed at different positions to photograph the same observed portion of the linear shadow that occurs in the gap between the train and the platform. an overhead camera (for example, the second overhead camera 20b in FIG. 4);
a processing device (for example, the processing device 200 in FIG. 1) that determines the presence or absence of an obstructive object based on a first image taken by the first bird's-eye camera and a second image taken by the second bird's-eye camera;
A station platform safety support system (for example, safety support system 2 in FIG. 1) equipped with
The processing device includes:
A first detection means (for example, a first inconsistency detection unit in FIG. 8 241) and
A second detection means (for example, a second inconsistency detection unit in FIG. 8 243) and
a determination unit (for example, the presence/absence determining unit 245 in FIG. 8) that determines the presence or absence of the obstructing object based on the difference between the first image ineligible location and the second image ineligible location;
Equipped with
It is a safety support system.

According to the first invention, bird's-eye cameras are installed at different positions in order to photograph the same observed portion of the linear shadow that appears in the gap between the train and the platform. Then, it is possible to detect image inconsistencies where the image of the observed portion does not correspond to a linear shadow from the images taken by each bird's-eye view camera, and to determine the presence or absence of an obstructive object based on the differences. According to this, it is possible to reduce false detection of obstructing objects using images taken by a camera, and it is possible to prevent accidents caused by trespassing vehicles at stations.

The second invention is
Different positions in order to photograph the same observation part of the linear part drawn or provided along the track on the side edge of the platform (for example, the Braille block 31 in FIG. 3 or the limit line 35 in FIG. 12) a first bird's-eye camera and a second bird's-eye camera installed in;
a processing device that determines the presence or absence of an obstructive object based on a first image taken by the first bird's-eye camera and a second image taken by the second bird's-eye camera;
A station platform safety support system equipped with
The processing device includes:
a first detection means for detecting a first image ineligible portion in the first photographed image where the image of the observed portion does not correspond to the linear portion;
a second detection means for detecting a second image ineligible portion in the second photographed image where the image of the observed portion does not correspond to the linear portion;
determining means for determining the presence or absence of the obstructing object based on a difference between the first image ineligible location and the second image ineligible location;
Equipped with
It is a safety support system.

According to the second invention, bird's-eye cameras are installed at different positions in order to photograph the same observed portion of the linear portion drawn or provided along the track on the side edge of the platform. Then, it is possible to detect locations where the image of the observed portion does not correspond to a linear portion from the images taken by each bird's-eye view camera, and determine the presence or absence of an obstructive object based on the difference. According to this, it is possible to reduce false detection of obstructing objects using images taken by a camera, and it is possible to prevent accidents caused by trespassing vehicles at stations.

A third invention is, in the second invention,
The linear portion is a line of a predetermined color or a Braille block,
It is a safety support system.

According to the third invention, a line of a predetermined color or a Braille block drawn or provided on a platform along a railway track can be used as a linear part.

A fourth invention is any one of the first to third inventions,
The determination means determines whether a positional deviation between a position of the first image-inconsistent part in the observed part and a position of the second image-inconsistent part in the observed part is a predetermined positional deviation. determining that the obstructing object exists if a condition is met;
It is a safety support system.

According to the fourth invention, it is possible to determine the presence or absence of an obstructive object based on the positional shift of the observed portion corresponding to each of the image inconsistencies seen from each bird's-eye camera.

The fifth invention is
The same observed part of the linear shadow that occurs in the gap between the train and the platform is photographed by a first bird's-eye camera and a second bird's-eye camera installed at different positions, and the first bird's-eye camera of the first bird's-eye camera is An obstructive object determination method by determining the presence or absence of an obstructive object based on a photographed image and a second photographed image of the second overhead camera, the method comprising:
a first detection step of detecting, in the first photographed image, a first image ineligible portion where the image of the observed portion does not correspond to the linear shadow;
a second detection step of detecting, in the second captured image, a second image inequivalent location where the image of the observed portion does not correspond to the linear shadow;
a determination step of determining the presence or absence of the obstructing object based on a difference between the first image ineligible location and the second image ineligible location;
This is an obstructing object determination method including the following.

According to the fifth invention, it is possible to realize an obstructing object determination method that provides the same effects as the first invention.

The sixth invention is
Photographing the same observed portion of the linear portion drawn or provided along the track on the side edge of the platform with a first bird's-eye camera and a second bird's-eye camera installed at different positions, and An obstructive object determination method by determining the presence or absence of an obstructive object based on a first photographed image of one bird's-eye view camera and a second photographed image of the second bird's-eye camera, the method comprising:
a first detection step of detecting, in the first captured image, a first image ineligible portion where the image of the observed portion does not correspond to the linear portion;
a second detection step of detecting a second image ineligible portion where the image of the observed portion does not correspond to the linear portion in the second photographed image;
a determination step of determining the presence or absence of the obstructing object based on a difference between the first image ineligible location and the second image ineligible location;
This is an obstructing object determination method including the following.

According to the sixth invention, it is possible to realize an obstructing object determination method that provides the same effects as the second invention.

The figure which shows the example of the whole structure of the safety support system of a station platform. A schematic diagram of the lead car of a train stopped at a station equipped with a safety support system, viewed from the front in the direction of travel. A schematic view from above of the station where the train stopped. FIG. 3 is a diagram for explaining the relationship between the installation position of each bird's-eye camera and the observation area. FIG. 3 is a diagram illustrating an obstructing object determination process. FIG. 7 is another diagram illustrating the obstructing object determination process. FIG. 7 is another diagram illustrating the obstructing object determination process. FIG. 2 is a block diagram showing an example of a functional configuration of a processing device. The figure which shows the data structure example of support object detection data. 5 is a flowchart showing the flow of processing by the processing device. 5 is a flowchart showing the flow of obstructing object determination processing. The schematic diagram seen from above of the platform in a modified example.

Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings. Note that the present invention is not limited to the embodiments described below, and the forms to which the present invention can be applied are not limited to the following embodiments. In addition, in the drawings, the same parts are denoted by the same reference numerals.

[overall structure]
FIG. 1 is a diagram showing an example of the overall configuration of a station platform safety support system 2 in this embodiment. FIG. 1 also shows the outline of a train 1 stopped at a station. Moreover, FIG. 2 is a schematic diagram of the lead car of the train 1 stopped at the station, viewed from the front side in the direction of travel, and FIG. 3 is a schematic diagram of the station, viewed from above.

Here, as shown in FIG. 3, the longitudinal direction of the platform 3 is defined as the x direction, and the transverse direction of the platform 3 is defined as the y direction. On the platform 3, a Braille block 31, which is an example of a linear part, is laid on the edge of the track side (the side edge of the platform 3) in the direction along the track, which is the x direction. In order to ensure the visibility of the Braille block 31, a color having a high luminance ratio with the surrounding road surface is used.

The safety support system 2 of this embodiment is provided for each target platform 3 (more specifically, a track) at the installation station, and as shown in FIG. 1, includes a plurality of overhead cameras 20, a processing device 200, Equipped with.

The plurality of bird's-eye cameras 20 are installed above the platform 3 at a bird's-eye view position overlooking the stopped train 1. Although the manner in which each bird's-eye view camera 20 is installed is not particularly limited, for example, as shown for one bird's-eye view camera 20 in FIG. 2, it is installed on the ceiling portion of the platform 3 with the photographing direction facing downward. Of course, other installation methods such as mounting on the pillars of the platform 3 may also be used.

Here, when the train 1 running on the rail R enters the platform 3, a characteristic shadow appears in the gap between the side of the car body of the train 1 and the edge of the platform 3. This shadow is called a linear shadow. Each bird's-eye view camera 20 is installed so that the entire area of the linear shadow 33 that occurs in the gap and the area of the Braille block 31 on the platform 3 are photographed by at least two bird's-eye cameras 20 partially overlappingly. Ru. FIG. 4 is a diagram for explaining the relationship between the installation position of each bird's-eye view camera 20 and the observation part, and a schematic diagram of the side of the train 1 seen from the platform 3 side is shown in the upper row (1), and the train 1 and A schematic diagram of the platform 3 seen from above is shown in the lower row (2).

In this embodiment, as shown in FIG. 4(2), an observation area 40 including the entire area of the linear shadow 33 and the entire area of the Braille block 31 is divided into a plurality of observation unit areas 41. Each is an observation part of two bird's-eye cameras 20 installed at different positions. Each bird's-eye view camera 20 is installed so that its own observation area is included in its photographing range.

For example, in the example of FIG. 4, each bird's-eye view camera 20 is installed at equal intervals along the x direction so that the observation unit area 41 immediately below and the observation unit areas 41 on both sides thereof are within the imaging range. . The two observation unit areas 41 at both ends of the three within the photographing range are the observation areas of the bird's-eye camera 20. Thereby, the entire area of the linear shadow 33 and the Braille block 31 can be photographed overlappingly by the two bird's-eye cameras 20 for each observation unit area 41. In addition, at this time, each bird's-eye camera 20 can photograph the two observation unit areas 41, which are the observation parts, from diagonally above. Therefore, for each observation unit area 41, an image of the observation unit area 41 taken from the diagonally upper left side and an image taken from the diagonally upper right side as viewed in FIG. 4 are obtained.

In terms of the observation unit area 41c, it is the observation area of the bird's-eye camera 20a and the bird's-eye camera 20b, and is photographed from the diagonally upper left by one of the bird's-eye cameras (first bird's-eye camera) 20a, and the other bird's-eye camera (first bird's-eye camera) 20a is photographed from diagonally above the left. The photograph is taken from diagonally above and to the right by the bird's-eye view camera 20b (No. 2).

In addition, in this embodiment, in the positional deviation determination process described later, the position in the captured image captured by each bird's-eye camera 20 (X value of the image coordinate system (X, Y); FIG. 5(1), FIG. 7(1) ), etc.) to the position on the platform 3 reflected there (x position along the edge of the track side of the platform 3) A table 253 (see FIG. 8) is prepared in advance.

Then, the processing device 200 performs an obstructing object determination process to determine the presence or absence of an obstructing object based on the captured images of each bird's-eye camera 20 .

[About obstructing object determination processing]
A possible case where some object exists in the gap between the train 1 and the platform 3 is a case where a person, wheelchair, stroller, etc. is caught in the gap. In that case, station staff and others will provide necessary rescue and assistance as appropriate to prevent the train from departing, thereby preventing a collision.

Furthermore, even if no person, wheelchair, stroller, etc. is caught in the gap, if a certain distance is not maintained between them and the train 1, an announcement is made to call attention to prevent collisions with vehicles. In this embodiment, when an object is present on the Braille block 31, an announcement is made to call attention to it.

However, just because an object exists in the gap between the train 1 and the platform 3 or on the Braille block 31, it does not necessarily mean that it is a person, wheelchair, stroller, etc.; for example, if it is a newspaper or an aggregate of leaves, etc. There is also. Therefore, it is important to determine whether the object present in the gap is a person or other object (object that requires some kind of measure to prevent accidents involving vehicles; an obstructive object) or another object such as newspaper. Become.

Therefore, in the obstructing object determination process, each observation unit area 41 is sequentially set as a target area, and the first photographed image of the first bird's-eye view camera 20a and the second photographed image of the second bird's-eye view camera 20b are set as target areas. The process of determining the presence or absence of an obstructive object in the target area based on the photographed image is repeated for each frame (every predetermined unit time). For example, if we focus on the observation unit area 41c in FIG. Determine whether it exists.

Here, FIG. 4 shows a case where a person 51 exists in the gap between the train 1 and the platform 3 within the observation unit area 41c. FIG. 5 is a diagram illustrating the obstructing object determination process in that case, with the first image taken by the first bird's-eye camera 20a shown in the upper row (1) and the second image taken by the second bird's-eye camera 20b. Examples are shown in the lower row (3), and a schematic diagram of the observation unit area 41c on the platform 3 is shown in the middle row (2).

Further, FIG. 6 is a diagram showing a case where a newspaper 53 exists in the same observation unit area 41c. FIG. 7 is a diagram illustrating the obstructing object determination process in that case, and shows a first example of an image taken by the first overhead camera 20a (1), a schematic diagram (2) of the observation unit area 41c, A second example (3) of an image taken by the second bird's-eye view camera 20b is shown.

The obstructive object determination process will be described below with a focus on the linear shadow 33, but the presence or absence of an obstructive object on the Braille block 31 can also be determined in a similar manner.

If any object exists in the gap between the train 1 and the platform 3, the object will overlap the linear shadow 33 in the photographed image, causing the linear shadow 33 to be partially missing or partially thick. The contour shape is distorted. For example, in the first captured image shown in FIG. 5(1) and the second captured image shown in FIG. 5(3), there is a gap between the train 1 and the platform 3 in the images 61 and 63 of the observation unit area 41c. Since the person 51 existing in the area overlaps the linear shadow 33, that area appears as non-corresponding areas 611 and 631 that do not correspond to the linear shadow 33. Similarly, in the first photographed image shown in FIG. 7(1) and the second photographed image shown in FIG. 7(3), the newspaper 53 existing in the gap between the train 1 and the platform 3 overlaps with the linear shadow 33. , those portions appear as inconsistencies 651 and 671 that do not correspond to the linear shadow 33.

However, objects such as people, wheelchairs, and strollers that can become obstacles all have a certain height. In this embodiment, the first bird's-eye view camera 20a and the second bird's-eye view camera 20b are used to observe the observation unit area 41 (for example, the observation unit area 41c), which is the same observation part, from diagonally above on both the left and right sides of the figure. Take a photo. Therefore, in the first photographed image (FIG. 5(1)) obtained by photographing the observation unit area 41c from diagonally upward to the left, the portion of the linear shadow 33 on the right side is blocked by the presence of the person 51, and the irrelevant portion 611 becomes. On the other hand, in the second photographed image (FIG. 5 (3)) obtained by photographing the observation unit area 41c from the diagonally upper right direction, the linear shadow 33 on the left side is obstructed by the presence of the person 51, resulting in an unsuitable area. It becomes 631.

Therefore, when the ineligible location 611 in the image 61 (first image ineligible location) and the ineligible location 631 in the image 63 (second image ineligible location) are converted to their positions on the platform 3, A positional shift occurs between the two. Specifically, as shown in FIG. 5(2), a part 711 in the observation unit area 41c corresponding to the first image-inconsistent part 611 and an observation unit area corresponding to the second image-inconsistent part 631 The position is shifted from the part 713 in 41c.

On the other hand, objects such as newspapers and aggregates of leaves that are not obstructing objects have almost no height compared to the obstructing objects. Therefore, an ineligible location (first image ineligible location) 651 in the image 65 of the observation unit area 41c in the first captured image shown in FIG. 7(1) and a second captured image shown in FIG. 7(3) Even if the observation unit area 41c in the image and the non-corresponding portion (second image non-corresponding portion) 671 in the image 67 are converted to the respective positions on the platform 3, there is no or small positional deviation. Specifically, as shown in FIG. 7(2), a part 731 within the observation unit area 41c corresponding to the first image-inconsistent area 651 and an observation unit area corresponding to the second image-inconsistent area 671 It roughly matches the site 733 in 41c.

Therefore, by determining the amount of positional deviation between the two (for example, A1 in FIG. 5(2)), it is possible to determine whether the first image mismatch point and the second image mismatch point are caused by an obstructing object. can. In addition, in FIG. 5, the distance between the left ends of each part 711, 713 is shown as the positional deviation amount A1, but it is also possible to calculate the distance between the centers of each part 711, 713 and use it as the positional deviation amount, or the distance between the right ends The distance may be used as the amount of positional deviation. For example, it is determined that "the determined positional displacement amount exceeds a predetermined threshold" as a predetermined positional displacement condition, and if the positional displacement condition is satisfied, it is determined that an obstructive object exists. The threshold value can be set based on the height of the three-dimensional object desired to be detected. According to this, it is possible to reduce false detection of obstructive objects such as people, wheelchairs, strollers, etc. that exist in the gap between the train 1 and the platform 3 or on the Braille block 31.

[About notification control processing]
If it is determined in the above-described obstructive object determination process that an obstructive object exists, a notification control process is performed to notify passengers on the platform 3, station staff, conductors, etc. of this fact. Although the mode of notification is not particularly limited, it can be realized, for example, by outputting a predetermined announcement sound from a speaker installed at an appropriate location.

In the present embodiment, whether an obstructive object on the linear shadow 33 is detected (is it determined that an obstructive object exists on the linear shadow 33) or whether an obstructive object on the Braille block 31 is detected (on the Braille block 31)? An announcement is made in accordance with whether it is determined that there is an obstructive object. More specifically, for example, when the train 1 is not at the station, an alert announcement is made to passengers on the platform 3 upon detection of an obstructive object on the Braille block 31 (hereinafter referred to as "first alert announcement"). ). This can be achieved by performing control to output an announcement sound prepared in advance from a speaker installed on the platform 3 or the like.

On the other hand, if the train 1 is on the line and an obstructive object on the linear shadow 33 is detected before the door of the train 1 opens, an alert announcement for passengers (hereinafter referred to as "second alert announcement") is made. I do. This second attention-seeking announcement may be an announcement with different content from the above-described first attention-seeking announcement. This is because when train 1 is on the line, you should be more careful about a collision accident.

In addition, if an obstructive object on the linear shadow 33 is detected before departure while the train 1 is on the line, an alert announcement for passengers (hereinafter referred to as the "third alert announcement") is made. In addition, announcements for station staff and conductors (hereinafter referred to as ``report announcements'') will be made as appropriate. Specifically, the third warning announcement is made in response to the detection of an obstructive object on the linear shadow 33 after passengers have finished getting on and off the vehicle. Then, if the obstructing object remains detected even after a predetermined period of time has elapsed, a notification announcement is made. In that case, the position of the observation unit area 41 in which it is determined that the obstructing object exists may also be notified. This is because it will help station staff and others when they go for rescue. In this embodiment, the timing at which the presence or absence of an obstructive object on the linear shadow 33 is detected is immediately after the door of the train 1 is closed. Whether or not the door of train 1 has been closed can be determined, for example, by the processing device 200 receiving a signal from the outside indicating that the door has been closed in train 1, or by receiving a signal from a separate device that detects the completion of door closing. This can be determined by the processing device 200 receiving a signal indicating that the door has been closed.

Note that the announcement itself may be made by a station staff member or the like. In that case, notification may be made to station staff or the like by outputting a warning sound or displaying a message.

[Functional configuration]
FIG. 8 is a block diagram showing an example of the functional configuration of the processing device 200. As shown in FIG. 8, the processing device 200 includes an operation section 210, a display section 220, a communication section 230, a control section 240, and a storage section 250, and can be configured as a type of computer.

The operation section 210 includes push switches, dials, and the like, and the display section 220 includes an LED, a small liquid crystal display, and the like. The operating section 210 and the display section 220 are mainly used by workers during maintenance.

The communication unit 230 is a wired or wireless communication device realized by, for example, a wireless communication module, a router, a modem, a TA, a wired communication cable jack, a control circuit, etc. communicate between. Photographed images input frame by frame from each bird's-eye view camera 20 via this communication section 230 are stored in the storage section 250 as photographed image data 255.

The control unit 240 is realized by electronic components such as a processor such as a CPU (Central Processing Unit) or a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), or an FPGA (Field-Programmable Gate Array), and is connected to each part of the device. Data input/output control is performed between. Then, it performs various arithmetic processing based on predetermined programs, data, data input via the communication unit 230, etc., and controls the operation of the processing device 200. The control section 240 includes a first ineligible point detection section 241 , a second ineligible point detection section 243 , an existence/absence determination section 245 , and a notification control section 247 . These functional units may be processing blocks realized as software by executing a program, or may be circuit blocks realized by hardware circuits such as ASIC or FPGA. This embodiment will be described as a processing block realized as software when the control unit 240 executes the station platform safety support program 251.

The first ineligible point detection unit 241 detects a point where the image of the target area does not correspond to a linear shadow in the first image based on the first photographed image of the first bird's-eye camera 20a whose observation area is the target area. Detected as an inappropriate location. In the present embodiment, a first mismatch detection process is performed to detect a first linear shadow mismatch location and a first braille block mismatch location, which will be described later, as a first image mismatch location.

The second non-corresponding portion detection unit 243 detects a portion where the image of the target region does not correspond to a linear shadow in a second image based on the second photographed image of the second overhead camera 20b whose observation portion is the target region. Detected as an inappropriate location. In this embodiment, a second non-corresponding portion detection process is performed to detect a second linear shadow non-corresponding portion and a second braille block non-corresponding portion, which will be described later, as second image non-corresponding portions.

The presence/absence determining unit 245 determines the presence or absence of an obstructing object in the target area based on the difference between the ineligible location in the first image and the ineligible location in the second image. In this embodiment, positional deviation determination processing is performed for each frame to convert the first image ineligible portion and the second image ineligible portion into positions (x positions) on the platform 3, and the amount of positional deviation between them is satisfies the positional deviation condition, it is determined that an obstructive object exists in the target area.

The notification control unit 247 is a functional unit that performs notification control processing, and when the presence/absence determination unit 245 determines that there is an obstructive object on the linear shadow 33 or the Braille block 31, the notification control unit 247 is a functional unit that performs notification control processing. Control is performed to output announcements from an external sound output device (speakers installed on the platform 3, etc.).

The storage unit 250 is realized by a storage medium such as an IC memory, a hard disk, or an optical disk. In this storage unit 250, programs for operating the processing device 200 and realizing various functions of the processing device 200, data used during the execution of the program, etc. are stored in advance, or are stored each time processing is performed. Memorized temporarily. For example, the storage unit 250 stores a station platform safety support program 251, a platform position replacement calculation table 253, photographed image data 255, and obstructive object detection data 257. In addition, audio data such as alert announcements and reporting announcements are also stored.

The station platform safety support program 251 causes the control unit 240 to function as a first ineligible point detection unit 241, a second ineligible point detection unit 243, an existence determination unit 245, and a notification control unit 247, and performs station platform safety support processing. This is a program for executing.

The home position substitution calculation table 253 is prepared for each bird's-eye camera 20 in association with the bird's-eye camera number for identifying each bird's-eye camera 20. One platform position replacement calculation table 253 stores the correspondence between the X value of the photographed image taken by the corresponding overhead camera 20 and the x position on the platform 3 captured therein.

The obstructive object detection data 257 stores data regarding the presence or absence of an obstructive object, which is determined for each frame of each observation unit area 41 . This obstructing object detection data 257 is generated for each frame by the presence/absence determination unit 245 and is accumulated.

Specifically, as shown in FIG. 9, one obstructive object detection data 257 includes a linear shadow presence/absence flag (ON: Exists/OFF: Does not exist) and Braille block presence/absence flag (ON: Exists/OFF: Does not exist) are set. If there is an observation unit area 41 in which it is determined by the presence/absence determining unit 245 that an obstructive object exists on the linear shadow 33, the presence/absence flag on the linear shadow corresponding to the observation unit area number is set to "ON". Similarly, when there is an observation unit area 41 in which it is determined that an obstructing object exists on the Braille block 31, the Braille block presence/absence flag corresponding to the observation unit area number is set to "ON".

[Processing flow]
FIG. 10 is a flowchart showing the flow of station platform safety support processing performed by the processing device 200. The processing described here can be realized in the processing device 200 by the control unit 240 reading out the station platform safety support program 251 from the storage unit 250 and executing it.

As shown in FIG. 10, in the station platform safety support process, first, an obstructing object determination process is started (step S1). FIG. 11 is a flowchart showing the flow of the obstructing object determination process.

As shown in FIG. 11, in the obstructing object determination process, each observation unit area 41 is sequentially set as a target area, and the process of loop A is executed (steps S101 to S115).

In loop A, the first inconsistency detection unit 241 detects a first image inconsistency location from an image of the target area in the first photographed image (hereinafter referred to as "first target area image"). First ineligible point detection process; step S103).

As a specific procedure, first, the linear shadow 33 and the Braille block 31 in the first target area image are determined. The linear shadow 33 is determined by, for example, extracting a dark color part (pixels with a brightness value that satisfies a predetermined dark color brightness condition) from the first target area image, and extracting a linear contour with a constant width from the extracted dark color part. A method of determining whether a dark-colored portion forming a shape exists can be adopted. Alternatively, a pattern matching method may be adopted in which a linear dark-colored partial image having a predetermined width is detected from the first target area image. The same applies to the determination of the Braille block 31. For example, a bright color portion (pixel with a brightness value that satisfies a predetermined bright color brightness condition) is extracted from the first target area image, and a certain width is selected from the extracted bright color portion. It is possible to adopt a method of determining whether there is a light colored portion having a linear contour shape. Alternatively, a pattern matching method may be adopted in which a linear bright color partial image having a predetermined width is detected from the first target area image.

Next, as the first image-inconsistent locations, an image-inconsistent location related to the linear shadow 33 (hereinafter referred to as "first linear shadow ineligible location") and an image-inconsistent location related to the Braille block 31 (hereinafter referred to as " 1) is detected. Specifically, as described above, a location where the outline of the determined linear shadow 33 is collapsed is detected as a location that is ineligible for the first linear shadow, and a location where the outline of the determined Braille block 31 is collapsed, as described above, is detected. Detected as a location that does not correspond to the first Braille block.

Subsequently, the second non-corresponding portion detection unit 243 detects a second image non-corresponding portion from the image of the target region of the second captured image (hereinafter referred to as “second target area image”). 2 non-corresponding location detection process; step S105).

The procedure is similar to the first inconsistency detection process, and an image inconsistency location related to the linear shadow 33 (hereinafter referred to as "second linear shadow inconsistency location") is detected from the image of the second target area image. , an image-inconsistent location related to the Braille block 31 (hereinafter referred to as a "second Braille block-inconsistent location") are detected as a second image-inconsistent location.

After the second inconsistency location detection process is completed, it is then determined whether a pair of the corresponding first image inconsistency location and second image incompatibility location has been detected (step S107). That is, when a first linear shadow incommensurate part is detected by the first incommensurate part detection process and a corresponding second linear shadow incommensurate part is detected by the second incommensurate part detection process, or If a first braille block ineligible point is detected by the first ineligible point detection process and a corresponding second braille block ineligible point is detected in the second ineligible point detection process, an affirmative determination is made in step S107 (S107 : YES), the process moves to step S109 and a positional deviation determination process is performed.

In the positional deviation determination process, first, the presence/absence determination unit 245 calculates the amount of positional deviation for each pair of the first image ineligible location and the second image ineligible location detected in step S103 and step S105. (Step S109). For example, the home position permutation table 253 for the first bird's-eye view camera 20a is used to convert the X position of the uncorresponding part in the first image to the x position on the platform 3, and the home position for the second bird's-eye view camera 20b is converted to the x position on the platform 3. Using the conversion table 253, the X position of the second image mismatch is converted to the x position on the platform 3, and the difference therebetween is determined as the amount of positional deviation.

Then, if there is a pair of the first image ineligible location and the second image ineligible location whose calculated amount of misalignment satisfies the misalignment condition (step S111: YES), it is determined that an obstructive object exists there. (Step S113). That is, if the set is a pair of the first linear shadow-inappropriate location and the second linear shadow-inappropriate location, it is determined that an obstructive object exists in the gap between the train 1 and the platform 3 in the target area. Furthermore, if the set is a pair of the first Braille block non-corresponding location and the second Braille block non-corresponding location, it is determined that an obstructing object exists on the Braille block 31 in the target area.

After performing the process of loop A for all target areas, the presence/absence determining unit 245 generates the obstructing object detection data 257 in which the linear shadow presence/absence flag and the braille block presence/absence flag are set for each observation unit area 41. is generated (step S117).

Thereafter, a termination determination is made, and until the present process is terminated (step S119; NO), the process returns to step S101 and the above-described process is repeated.

Return to FIG. 10. After starting the obstructing object determination process, the notification control unit 247 branches the process depending on whether or not the train 1 is present at the station (target platform 3). If it is detected that the train 1 is on the corresponding track of the platform 3 (from when the train 1 enters the station until it leaves) (step S3: YES), the process moves to step S9. On the other hand, when the train 1 is not on the line (step S3: NO), the process moves to step S5.

Then, in step S5, the latest obstructing object detection data 257 at that time is referred to, and if there is an observation unit area 41 whose braille block presence/absence flag is "ON" (step S5: YES), a first alert announcement is made. Control is performed to output from speakers such as the platform 3 (step S7). Note that the output control of the first attention-calling announcement may be performed when the number of obstructing objects on the Braille block 31 is equal to or greater than a predetermined number set in advance.

Moreover, in step S9, it is determined whether the door of the train 1 on the platform 3 is opened, and if it is not opened (step S9: NO), the process moves to step S11. Determination as to whether the door is open is made using a signal received from the outside. For example, the processing device 200 may receive a signal indicating that a door has been opened on train 1 from the outside, or the processing device 200 may receive a signal indicating that a door has been opened from a separate device that detects the completion of the door opening. 200 may be received. Then, referring to the latest obstructive object detection data 257 at that time, if there is an observation unit area 41 whose linear shadow presence/absence flag is "ON" (step S11: YES), a second alert announcement is sent to the platform 3. Control is performed to output from speakers such as (step S13).

On the other hand, after the door of the train 1 is opened (step S9: YES), the system remains in a standby state until the door of the train 1 on the platform 3 is closed (step S14: NO). The determination as to whether the door is closed can be made using a signal received from the outside, similar to the determination as to whether the door is open. For example, the processing device 200 may receive a signal indicating that a door has been closed on train 1 from the outside, or the processing device 200 may receive a signal indicating that a door has been closed from a separate device that detects the completion of door closing. 200 may be received.

When the door of the train 1 is closed (step 14: YES), the latest obstructing object detection data 257 at that time is referred to, and if there is an observation unit area 41 whose linear shadow presence/absence flag is "ON", If (step S15: YES), control is performed to output the third alert announcement from the speaker of the platform 3 or the like (step S17). In this case, after a predetermined period of time has elapsed, the latest obstructing object detection data 257 at that time is referred to again to check the result of determining the presence or absence of an obstructing object, and the observation unit area where the linear shadow presence/absence flag is "ON" is checked. 41 (step S19: YES), control is performed to output the notification announcement from the speaker of the platform 3 or the like (step S21).

Thereafter, the process waits until there are no more observation unit areas 41 whose linear shadow presence/absence flags are "ON". Then, if it is gone (step S23: YES), the advance of train 1 from the station is monitored. Then, until the train 1 finishes advancing (step S25: NO), the presence/absence flag on the braille block is monitored, and if there is an observation unit area 41 that is "ON" (step S27: YES), a second alert announcement is issued. Control is performed to output from speakers such as the platform 3 (step S29).

If the train 1 has finished advancing (step S25: YES), an end determination is made. Then, until this process is finished (step S31: NO), the process returns to step S3 and the above-described process is repeated.

As explained above, according to the present embodiment, it is possible to reduce false detection of obstructive objects such as people, wheelchairs, strollers, etc. that exist in the gap between the train 1 and the platform 3 or on the Braille block 31, and to detect the presence or absence of obstructive objects. can be appropriately determined. If it is determined that there is an obstructive object, it is possible to alert passengers or notify station staff or conductors, thereby preventing accidents where people, wheelchairs, strollers, etc. come into contact with train 1. It can be prevented.

Note that the form to which the present invention can be applied is not limited to the above-described embodiments, and components may be added, omitted, or changed as appropriate.

For example, in the above embodiment, the case where the presence or absence of an obstructive object on the Braille block is determined is illustrated, but instead of the Braille block, a white line, which is another example of a linear part on the platform, is used to detect the presence or absence of an obstructive object on the white line. It can also be configured to determine the presence or absence of an object. In that case, it is sufficient to determine the white line in the photographed image (in the image of the observation unit area) and detect the inappropriate portion.

Alternatively, a detection line (hereinafter referred to as "limit line") may be separately attached to the platform. FIG. 12 is a schematic diagram of the platform 3B seen from above in this modification. As shown in FIG. 12, a Braille block 31 is installed along the track on the edge of the platform 3B on the track side, and a limit line 35 is drawn along the track further outside the Braille block 31. The position of this limit line 35 is drawn at the limit position where the presence of people, wheelchairs, strollers, etc. is allowed when the train enters or exits the platform 3B.

In this modification, a limit line 35, which is another example of a linear portion, is determined instead of the linear shadow described in the above embodiment, and a portion where the contour of the limit line 35 collapses is determined as an inappropriate portion. By detecting it, the presence or absence of an obstacle on the limit line 35 is determined. According to this, it is possible to detect an obstructive object that exists near the outer track side edge of the Braille block 31.

In addition, in the notification control process of this modification, when a train is not on the line, an alert announcement (first alert announcement) is made in response to detection of an obstructive object on the Braille block 31. On the other hand, when a train is on the line, an alert announcement (second alert announcement or (third alert announcement)) or notification announcement is made in response to the detection of an obstructive object on the limit line 35. Then, the warning announcement when the train is not on the line and the warning announcement when the train is on the line are made different, and the alert is called for more strongly than when the train is on the line.

Furthermore, the above embodiments and modifications can be similarly applied to station platforms where platform fences are installed. The platform fence is placed at the edge of the platform on the track side and separates the platform from the track side. By installing the safety support system 2 exemplified in the above-described embodiments and modifications, it is possible to separate the platform fence from the train. It is possible to detect people, wheelchairs, strollers, etc. that have gotten between the vehicle and the vehicle. Therefore, similar to the above embodiment, it is possible to prevent a collision accident.

1 train, 2 safety support system, 20 bird's-eye camera, 20a first bird's-eye camera, 20b second bird's-eye camera, 200 processing device, 240 control section, 241 first ineligible point detection section, 243 second ineligible point detection section, 245 presence/absence determination section, 247 notification control section, 250 storage section, 251 station platform safety support program, 253 platform position replacement calculation table, 255 photographed image data, 257 obstructive object detection data, 3 platform, 31 Braille block, 41 observation unit Area, 35 limit line

Claims (6)

a first bird's-eye camera and a second bird's-eye camera installed at different positions to photograph the same portion of a linear shadow created in a gap between the train and the platform due to the presence of the train ;
a processing device that determines the presence or absence of an obstructive object based on a first image taken by the first bird's-eye camera and a second image taken by the second bird's-eye camera;
A station platform safety support system equipped with
The processing device includes:
A first detection of detecting a first image inconsistency location where the same portion of the linear shadow is partially occluded and becomes a portion that does not correspond to the linear shadow in the first photographed image. means and
A second detection of detecting a second image inconsistency location where the same portion of the linear shadow is partially occluded and becomes a portion that does not correspond to the linear shadow in the second photographed image. means and
Based on the difference in position along the longitudinal direction of the platform between the first image-inconsistent location in the first captured image and the second image-inconsistent location in the second captured image determining means for determining the presence or absence of the obstructive object;
Equipped with
Safety support system. A first bird's-eye view camera and a second bird's-eye view camera installed at different positions to photograph the same part of the linear part drawn or provided along the track on the side edge of the platform;
a processing device that determines the presence or absence of an obstructive object based on a first image taken by the first bird's-eye camera and a second image taken by the second bird's-eye camera;
A safety support system for a station platform equipped with
The processing device includes:
A first detection of detecting a first image inconsistency location in which the same portion of the linear portion is partially occluded and becomes a portion that does not correspond to the linear portion in the first photographed image. means and
A second detection of detecting a second image inconsistency location in which the same portion of the linear portion is partially occluded and becomes a portion that does not correspond to the linear portion in the second photographed image. means and
Based on the difference in position along the longitudinal direction of the platform between the first image-inconsistent location in the first captured image and the second image-inconsistent location in the second captured image determining means for determining the presence or absence of the obstructive object;
Equipped with
Safety support system. The linear portion is a line of a predetermined color or a Braille block,
The safety support system according to claim 2. The determining means determines that the obstructing object does not exist if the difference does not satisfy a predetermined positional deviation condition, and determines that the obstructing object exists if the difference does not satisfy a predetermined positional deviation condition.
The safety support system according to any one of claims 1 to 3. A first bird's-eye camera and a second bird's-eye camera installed at different positions photograph the same part of the linear shadow that occurs in the gap between the train and the platform due to the presence of the train , and An obstructive object determination method by determining the presence or absence of an obstructive object based on a first photographed image of a camera and a second photographed image of the second overhead camera, the method comprising:
A first detection of detecting a first image inconsistency location where the same portion of the linear shadow is partially occluded and becomes a portion that does not correspond to the linear shadow in the first photographed image. step and
A second detection of detecting a second image inconsistency location where the same portion of the linear shadow is partially occluded and becomes a portion that does not correspond to the linear shadow in the second photographed image. step and
Based on the difference in position along the longitudinal direction of the platform between the first image-inconsistent location in the first captured image and the second image-inconsistent location in the second captured image a determination step of determining the presence or absence of the obstructive object;
Obstacle object determination method including. A first bird's-eye camera and a second bird's-eye camera installed at different positions photograph the same part of the linear part drawn or provided along the track on the side edge of the platform, and An obstructive object determination method by determining the presence or absence of an obstructive object based on a first photographed image of an overhead camera and a second photographed image of the second overhead camera, the method comprising:
A first detection of detecting a first image inconsistency location in which the same portion of the linear portion is partially occluded and becomes a portion that does not correspond to the linear portion in the first photographed image. step and
A second detection of detecting a second image inconsistency location where the same portion of the linear portion is partially occluded and becomes a portion that does not correspond to the linear portion in the second photographed image. step and
Based on the difference in position along the longitudinal direction of the platform between the first image-inconsistent location in the first captured image and the second image-inconsistent location in the second captured image a determination step of determining the presence or absence of the obstructive object;
Obstacle object determination method including.

Images