JP2018002007A - Obstacle detection system for railway - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an obstacle detection system for a railway capable of improving reliability with a simple configuration.SOLUTION: An obstacle detection system 1 for a railway is mounted on a railway vehicle T to detect an obstacle on rails 11. The system 1 includes a polarization imaging camera 2 for photographing an image of rails 11, and an information processing device 3 for detecting an obstacle on the rails 11 on the basis of information of a photographed image. The information processing device 3 includes a brightness information extraction part 30 for extracting brightness information in each pixel of the image, and a polarization information calculation part 31 for calculating the polarization direction and the polarization degree of each pixel on the basis of the extracted brightness information. Further, the information processing device 3 includes a rail detection part 32 for making pixels whose polarization direction and polarization degree are respectively within the range of a prescribed range and equal to or more than a prescribed value be pixels corresponding to the rails 11, and an obstacle detection part 36 for detecting an obstacle on the rails 11 on the basis of the pixels caused to correspond to the rails 11.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an obstacle detection system for railways.

  In land transportation vehicles such as railway vehicles, there are obstacles on the running road surface, which may lead to a serious accident. For this reason, if it can be detected at an early stage whether there is an obstacle on the road surface, such an accident can be prevented in advance.

  Therefore, conventionally, an apparatus for detecting whether there is an obstacle on the road surface has been proposed. For example, a map is created by photographing with a polarization imaging device, a point where the polarization angle (polarization direction) is uniform and continuous is specified as a road surface on this map, and an obstacle located at the point specified as the road surface is detected. A road surface obstacle detection device has been proposed as an application example of a polarization imaging device (see, for example, Patent Document 1).

Japanese Patent No. 4974543

  By the way, since the road surface obstacle detection apparatus described in Patent Document 1 specifies the road surface only by the polarization angle (polarization direction), the road surface obstacle detection device lacks reliability. For this reason, if the vehicle traveling on the road surface is a relatively low speed vehicle that can avoid obstacles by steering left and right, the road surface failure described in Patent Document 1 is unreliable. Even an object detection device can be applied in practice.

  However, if the vehicle travels on the rail at a high speed and cannot avoid the obstacle by steering, the road surface obstacle detection device described in Patent Document 1 lacks reliability. If so, it may lead to a serious accident, so it cannot be applied. On the other hand, if the road surface obstacle detection apparatus described in Patent Document 1 is subjected to complicated calculations for improving reliability, such calculations may not catch up with the speed of the railway vehicle. There is also a possibility that it cannot be applied to railway vehicles.

  Then, an object of this invention is to provide the obstruction detection system for railroads which can improve reliability with a simple structure.

In order to solve the above-described problem, a railroad obstacle detection system according to a first aspect of the present invention is a railroad obstacle detection system that is mounted on a railroad vehicle and detects an obstacle on a rail that the railcar is about to travel. A system,
A polarization imaging camera for taking an image of the rail in front of the railway vehicle;
An information processing device that detects an obstacle on the rail based on information of an image photographed by the polarization imaging camera;
The information processing apparatus is
A luminance information extraction unit for extracting luminance information in each pixel of the image;
A polarization information calculation unit that calculates a polarization direction and a polarization degree of each pixel based on the luminance information extracted by the luminance information extraction unit;
A rail detection unit in which a polarization direction and a polarization degree calculated by the polarization information calculation unit are pixels within a predetermined value range and a predetermined value or more, respectively, and pixels corresponding to the rail;
And an obstacle detection unit that detects an obstacle on the rail based on the pixel that corresponds to the rail in the rail detection unit.

The railroad obstacle detection system according to the second invention includes an obstacle detection unit in the railroad obstacle detection system according to the first invention.
A continuity determination unit that determines that there is an obstacle on the rail if the pixel that is supposed to correspond to the rail in the rail detection unit is discontinuous;
A past information providing unit that provides pixels corresponding to the rail acquired in advance;
If the pixel corresponding to the rail in the rail detection unit and the pixel corresponding to the rail provided from the past information providing unit do not match each other beyond an allowable range, it is determined that there is an obstacle on the rail. And a past information matching unit.

  Furthermore, the obstacle detection system for railroad according to the third invention is the case where the obstacle detection unit in the railroad obstacle detection system according to the first or second invention detects an obstacle on the rail, It controls the speed at which the railway vehicle travels.

  According to the above-mentioned obstacle detection system for railroads, pixels corresponding to the rail are detected based on the polarization direction and the degree of polarization, so that the pixels corresponding to the rail are detected with high accuracy without performing complicated calculations. The reliability can be improved with a simple configuration.

1 is a schematic perspective view showing a railroad obstacle detection system according to an embodiment of the present invention. It is a schematic exploded perspective view of the polarization imaging camera with which the obstacle detection system for railroads is provided. It is a block diagram of the information processor with which the obstacle detection system for railroads is provided. It is the figure which showed typically an example of the image image | photographed with the same polarization imaging camera. FIG. 5 is a partially enlarged view of FIG. 4. It is a figure which shows the polarization direction in each pixel of FIG. It is a figure which shows the polarization degree in each pixel of FIG. It is the figure which performed the binarization process in FIG. It is a figure which shows the past information for collating with FIG. It is a flowchart which shows operation | movement of the obstacle detection system for railways.

Hereinafter, a railroad obstacle detection system according to an embodiment of the present invention will be described with reference to the drawings.
As shown in FIG. 1, the obstacle detection system for railway is mounted on a railway vehicle T, and the obstacle on the rail 11 (that is, the rail 11 in front of the railway vehicle T) on which the railway vehicle T is to travel. It detects objects.

  The railway obstacle detection system 1 is arranged in the cab C of the railway vehicle T, and the polarization imaging camera 2 that photographs the front through the window glass W of the cab C, and the polarization imaging camera 2 An information processing device 3 (for example, a laptop computer) that processes information of the obtained image and an alarm device 4 that issues an alarm when an obstacle is detected by the information processing device 3 are provided.

[Polarized imaging camera 2]
As shown in FIG. 2, the polarization imaging camera 2 includes a camera body 21 capable of continuously shooting a large number of still images (that is, moving images), and a solid-state image sensor such as a CCD image sensor provided in the camera body 21. 22 and a polarization imaging filter 23 covering the CCD image sensor.

  The camera body 21 is arranged so as to photograph a building limit area (including the rail 11 as a matter of course) in front of the railway vehicle T. The solid-state imaging device 22 only needs to be able to capture an image from which luminance information can be extracted by the information processing device 3, and may be a device capable of capturing a monochrome image (for example, a monochrome CCD imaging device). The polarization imaging filter 23 includes a polarizer array 5 in which polarizers 6 having different transmission axis directions 7 are arranged in two rows and two columns in order to obtain polarization information (polarization direction and polarization degree) of the light L passing therethrough. , And arranged in multiple rows and multiple columns (3 rows and 3 columns as an example in FIG. 2). For example, as shown in FIG. 2, the polarizer arrays 5 may be configured such that the transmission axis directions 7 of the polarizers 6 adjacent to each other in the vertical or horizontal direction are different from each other by 45 °. You may make it differ by an angle. The polarizer array 5 is configured in this manner because the information processing apparatus 3 uses the data fitting method or the like based on the intensity of the light L that has passed through the polarizer 6 of the polarizer array 5 to perform the light L. This is to obtain the polarization direction in each of the polarizer arrays 5. The method is not limited to the data fitting method, and any method may be used as long as it is a method for obtaining the polarization direction such as Fourier analysis.

[Information processing apparatus 3]
As shown schematically in FIGS. 1 and 3, the information processing apparatus 3 includes a luminance information extraction unit 30 that extracts luminance information in each pixel of an image from the polarization imaging camera 2, and the luminance information extraction unit. And a polarization information calculation unit 31 that calculates the polarization direction and the polarization degree of each pixel based on the luminance information extracted at 30. In addition, the information processing apparatus 3 further includes a rail detection unit 32 in which a pixel whose polarization direction and degree of polarization calculated by the polarization information calculation unit 31 are within a predetermined range and a predetermined value or more is a pixel corresponding to the rail 11. And an obstacle detection unit 36 that detects an obstacle on the rail 11 based on the pixel that is supposed to correspond to the rail 11 by the rail detection unit 32.

  The luminance information extraction unit 30 acquires information on an image 8 as shown in FIG. 4 in front of the railway vehicle T from the polarization imaging camera 2 and extracts luminance information on each pixel of the acquired image 8. . In the following, when the above image is described, it is difficult to understand based on the image 8 shown in FIG. 4 because the number of pixels is extremely large. Therefore, a part 80 of FIG. 4 is extracted and enlarged in FIG. It is based on the image 80 shown.

  The polarization information calculation unit 31 calculates the polarization direction of each pixel as shown in FIG. 6 and the degree of polarization of each pixel as shown in FIG. 7 based on the luminance information at each pixel of the image 80 as shown in FIG. Is to be calculated. As is clear from a comparison between FIG. 5 and the corresponding FIG. 6, the polarization direction of the pixel 81 corresponding to the rail 11 and some other pixels 82 has a polarization direction within a predetermined range (for example, 130 in the horizontal axis in FIG. 6). A feature within the range of (° to 140 °) is observed. On the other hand, as is clear from comparison between FIG. 5 and FIG. 7 corresponding thereto, the pixel 81 corresponding to the rail 11 has a characteristic that the degree of polarization is not less than a predetermined value (for example, 0.8 in FIG. 7). It is done.

  As shown in FIG. 3, the rail detection unit 32 includes a polarization direction comparison unit 33, a polarization degree comparison unit 34, and a binarization processing unit 35. The polarization direction comparison unit 33 temporarily detects the pixels 81 and 82 whose polarization directions are within the predetermined range as pixels that can correspond to the rail 11 by comparing the polarization direction of each pixel with the predetermined range. The polarization degree comparison unit 34 temporarily detects a pixel 81 having a polarization degree equal to or greater than the predetermined value as a pixel that can correspond to the rail 11 by comparing the polarization degree of each pixel with the predetermined value. The binarization processing unit 35 detects, as the pixel 81 corresponding to the rail 11, the pixel 81 that is temporarily detected as being able to correspond to the rail 11 in both the polarization direction comparison unit 33 and the polarization degree comparison unit 34. . Further, as shown in FIG. 8, the binarization processing unit 35 sets the pixel 81 detected as corresponding to the rail 11 to 1 (or 0) and sets the other pixels to 0 (or 1). That is, binarization processing is performed.

As shown in FIG. 3, the obstacle detection unit 36 includes a continuity determination unit 37, a recording unit 38 (which is an example of a past information providing unit), and a past information matching unit 39.
If there is a plurality of pixel groups 81u and 81d (see FIG. 8) corresponding to one rail 11 (see FIG. 5) in the binarized image, the continuity determination unit 37, that is, the pixel group. If 81u and 81d are discontinuous, it is determined that there is an obstacle 10 (see FIG. 5) on the rail 11, and the other is determined that there is no obstacle 10 on the rail 11. In the example shown in FIG. 8, since the pixel groups 81u and 81d corresponding to the rail 11 are divided into an upper side 81u and a lower side 81d, that is, since there are a plurality of pixels, it is determined that there is an obstacle 10 on the rail 11. The It should be noted that whether or not there are a plurality of pixel groups 81u and 81d may have an allowable range. For example, when the distance separating the plurality of pixel groups 81 u and 81 d corresponding to the rail 11 is several pixels or less, the plurality of pixel groups 81 u and 81 d corresponding to the rail 11 may be regarded as one pixel group 81.

  The recording unit 38 stores information (hereinafter referred to as past information) obtained by performing binarization processing on a past image when there is no obstacle 10 on the rail 11, and the past information is stored in the past. This information is provided to the information verification unit 39. The past information is stored in the recording unit 38 before the railway vehicle T travels.

  The past information collation unit 39 collates the information as shown in FIG. 8 from the binarization processing unit 35 with the past information as shown in FIG. 9 from the recording unit 38, and this collation shows in FIG. If the pixels 81 and 84 corresponding to the rail 11 do not match between the information and the past information as shown in FIG. 9, it is determined that there is an obstacle 10 on the rail 11, and the rest is on the rail 11. It is determined that there are no obstacles 10. In the example shown in FIGS. 8 and 9, the pixel group 83 corresponding to the obstacle 10 on the rail 11 shown in FIG. 8 is different from that in FIG. 9, so that they do not match and the obstacle 10 is on the rail 11. Determined. It should be noted that an acceptable range may be given as to whether they do not match. For example, when the number of pixels that do not match is equal to or less than a few percent of the total number of pixels to be matched, it may be considered that they do not match (match).

[Alarm 4]
If the continuity determination unit 37 or the past information collation unit 39 determines that the obstacle 10 is on the rail 11, the alarm device 4 issues an alarm by voice or video. In FIG. 1 and FIG. 3, the alarm device 4 is illustrated as a configuration different from the information processing device 3, but the information processing device 3 may include the alarm device 4. As an example in which the alarm device 4 is built in the information processing device 3, when the information processing device 3 is a laptop computer and the alarm is a sound, the alarm device 4 corresponds to a speaker built in the laptop computer. If the alarm is an image, the alarm device 4 corresponds to a display provided in the laptop computer.

Hereinafter, the operation of the railway obstacle detection system 1 will be described.
As shown in FIG. 1, when the railway vehicle T is about to travel or is traveling, a moving image in front of the railway vehicle T is captured by the polarization imaging camera 2 (S1 in FIG. 10). At this time, the aperture of the lens (not shown) of the polarization imaging camera 2 is adjusted according to the headlamp of the railway vehicle T and the sunshine conditions. The images 8 and 80 as shown in FIGS. 4 and 5 taken by the polarization imaging camera 2 are sent to the luminance information extraction unit 30 of the information processing device 3 as shown in FIG. In step 30, luminance information of each pixel is extracted (S2 in FIG. 10). Based on the extracted luminance information, the polarization information calculation unit 31 calculates the polarization information (polarization direction and polarization degree) of each pixel (S3 in FIG. 10). Based on the calculated polarization direction and polarization degree, the rail detection unit 32 detects the pixel 81 corresponding to the rail 11 and performs binarization processing (S4 in FIG. 10). The continuity determination unit 37 determines whether or not the pixel 81 corresponding to the rail 11 is discontinuous (S5 in FIG. 10). If discontinuous, the alarm device 4 issues an alarm (S7 in FIG. 10). If the pixel 81 corresponding to the rail 11 is not discontinuous (that is, if it is continuous), the information from the binarization processing unit 35 and the past information from the recording unit 38 do not match in the past information matching unit 39. (S6 in FIG. 10), if it does not match (that is, if it matches), it enters a loop for processing information of the next image sent from the polarization imaging camera 2, and if it does not match, the alarm device 4 An alarm is issued (S7 in FIG. 10). After issuing the warning, the railway vehicle T is decelerated (S8 in FIG. 10). Depending on the result of this deceleration, the railway vehicle T may be stopped.

  As described above, according to the railway obstacle detection system 1, the pixel 81 corresponding to the rail 11 is detected based on the polarization direction and the polarization degree, so that the pixel 81 corresponding to the rail 11 is high without performing complicated calculations. As a result, the reliability can be improved with a simple configuration.

In addition, since the past information providing unit (recording unit 38) and the past information collating unit 39 collate with the past information when detecting the obstacle, the reliability can be further improved with a simple configuration.
In the above embodiment, the obstacle detection unit 36 has been described as having the continuity determination unit 37, the recording unit 38 (an example of the past information providing unit), and the past information matching unit 39. What is necessary is just to detect whether the obstacle 10 exists on the rail 11 by information. For example, the obstacle detection unit 36 may be configured by only the continuity determination unit 37 or may be configured by only the past information providing unit (recording unit 38) and the past information matching unit 39.

  In the above embodiment, the recording unit 38 has been described as an example of the past information providing unit. However, any unit that provides past information to the past information matching unit 39 may be used. For example, the past information providing unit may be a receiving unit that receives past information from the outside of the information processing apparatus 3 by wire or wireless and provides it to the past information matching unit 39.

  Furthermore, in the above embodiment, the recording unit 38 has been described as storing past information in advance before the railway vehicle T travels, but binarization is performed when the obstacle 10 on the rail 11 is not detected. The information from the processing unit 35 may be overwritten at any time. With this configuration, since the recording unit 38 provides the latest past information of the rail 11 that changes according to the season and aging, the reliability can be further improved with a simple configuration.

  In addition, in the above-described embodiment, it has been described that the alarm device 4 gives an alarm when the obstacle detection unit 36 detects the obstacle 10 on the rail 11, but the railway vehicle T is used together with this alarm or instead of the alarm. May be configured to automatically decelerate or stop (that is, to control the traveling speed). In this case, the obstacle detection unit 36 is connected to a speed control device (not shown) of the railway vehicle T, and the brake of the speed control device is activated by detecting the obstacle 10 by the obstacle detection unit 36. Is done. With this configuration, when the obstacle 10 on the rail 11 is detected, the railway vehicle T is automatically decelerated or stopped, so that the reliability can be further improved with a simple configuration.

T Railway vehicle 1 Obstacle detection system for railway 2 Polarization imaging camera 3 Information processing device 4 Alarm device 5 Polarizer array 6 Polarizer 7 Transmission axis direction 21 Camera body 22 Solid-state imaging device 23 Polarization imaging filter 30 Luminance information extraction unit 31 Polarization Information calculation unit 32 Rail detection unit 33 Polarization direction comparison unit 34 Polarization degree comparison unit 35 Binarization processing unit 36 Obstacle detection unit 37 Continuity determination unit 38 Recording unit 39 Past information collation unit 81 Pixel corresponding to rail

Claims (3)

A railroad obstacle detection system that is mounted on a railcar and detects an obstacle on the rail that the railcar is about to travel,
A polarization imaging camera for taking an image of the rail in front of the railway vehicle;
An information processing device that detects an obstacle on the rail based on information of an image photographed by the polarization imaging camera;
The information processing apparatus is
A luminance information extraction unit for extracting luminance information in each pixel of the image;
A polarization information calculation unit that calculates a polarization direction and a polarization degree of each pixel based on the luminance information extracted by the luminance information extraction unit;
A rail detection unit in which a polarization direction and a polarization degree calculated by the polarization information calculation unit are pixels within a predetermined value range and a predetermined value or more, respectively, and pixels corresponding to the rail;
An obstacle detection system for railways, comprising: an obstacle detection unit that detects an obstacle on the rail based on a pixel that corresponds to a rail in the rail detection unit. Obstacle detection unit
A continuity determination unit that determines that there is an obstacle on the rail if the pixel that is supposed to correspond to the rail in the rail detection unit is discontinuous;
A past information providing unit that provides pixels corresponding to the rail acquired in advance;
If the pixel corresponding to the rail in the rail detection unit and the pixel corresponding to the rail provided from the past information providing unit do not match each other beyond an allowable range, it is determined that there is an obstacle on the rail. An obstacle detection system for railways, comprising:   The railway obstacle detection system according to claim 1 or 2, wherein the obstacle detection unit controls a speed at which the railway vehicle travels when an obstacle is detected on the rail.

Images