JPH01301458A - Crossing obstructor detecting method - Google Patents

Abstract

PURPOSE:To easily and reliably detect an obstruction on a crossing by fixing a telecamera in a slightly oblique upper position that the crossing can be seen to the full width thereof on a display and processing the picture of the crossing taken after it is shut off. CONSTITUTION:A telecamera 2 is fixed over a crossing in a slightly oblique position to look down over it to the full width thereof. A picture memory 4 stores the pictures at specific intervals. After a specific period from receival of the signal from a crossing shut-off condition input 6, the framed picture of the crossing area taken after the shut-off is processed to make detection with a foreign substance detector 14 for any foreign substance on the crossing way. A lateral and longitudinal detector 15 then detects that the lateral and the longitudinal dimension of the foreign substance are more than the specified values or not. Using the result of the detection, a decision part 17 really identifies the existence of an obstruction on the crossing when deciding that the foreign substance has the dimensions more than the specific values and exists continuously during the specified period or more.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention relates to a method for detecting obstacles at a level crossing using image processing technology.

Conventional technology The method for detecting obstacles at railway level crossings, such as cars and people stopped within the level crossing after the level crossing has been blocked, is to send a signal across the track, install a light receiver oppositely, and detect a light from the light transmitter. There are two known types: a previous type that detects when an obstacle blocks the emitted light beam, and a loop type that detects based on changes in inductance caused by the obstacle.

Problems to be Solved by the Invention However, in the previous model, the transmitter and receiver must be installed at the height of the number above the ground + the height of the mouth, so in areas with deep snow, the transmitter and receiver are buried in the snow and cannot be detected. Sometimes it happened. On the other hand, loop type sensors cannot detect objects other than metals, and are susceptible to temperature changes, making stable detection impossible. Furthermore, the previous generation and loop type systems required in-track construction, which was accompanied by significant time constraints in relation to train operating hours and involved a large amount of cost.

Therefore, an object of the present invention is to provide a method that can easily and reliably detect obstacles at a level crossing without requiring any on-track construction.

Means for Solving the Problems In order to achieve the above object, the method according to the present invention fixes a television camera at a position looking down on the level crossing from slightly diagonally above so that the width of the level crossing can be recognized on the screen. In addition to projecting an image that always appears at fixed coordinates,
Images at regular intervals are stored in an image storage unit, and when the level crossing is closed, the presence or absence of a foreign object on the level crossing is detected by performing image processing on the inside of the level crossing based on the frame image after the level crossing is closed. If a foreign object is detected, it is verified whether the length of the foreign object in the horizontal direction is equal to or greater than a certain level, and furthermore, the lowest end of the foreign object is on the level crossing, and the uppermost end is continuous from the lowest end. Vertical verification is performed to determine whether or not the foreign object protrudes above the level crossing, and as a result of the verification, it is determined that the foreign object is larger than a certain size, and that the foreign object has remained in the same position continuously for a certain period of time. In this case, it is determined that there is an obstacle.

Function As mentioned above, the TV camera is fixed at a position looking down at the level crossing from slightly above so that the width of the level crossing can be recognized on the screen. If the foreign object is located on the screen, the lowermost end of the foreign object is located on the level crossing, and the uppermost edge of the foreign object is continuous from the lowermost edge and projects above the level crossing. In this way, the method of the present invention allows the foreign object to be moved into the level crossing when the lateral length of the foreign object exceeds a certain level and the state of protruding above the level crossing continues for a certain period of time as described above. It is determined that it is a real obstacle that is stationary.

Example The present invention will be described below based on embodiments shown in the drawings.

FIG. 2 shows a system used to implement the detection method of the present invention, which comprises an image input section (3) equipped with a television camera (2), an image storage section (4), and a computer. It consists of a central processing unit (lO). (
6) is a level crossing closing condition input unit connected to a track circuit or the like.

The television camera (2) is fixed at a position looking down at the level crossing slightly diagonally upward so that the width of the level crossing can be recognized on the screen, and as shown in FIG. Projects an image in real time that always appears at a fixed position (coordinates). Note that, hereinafter, the term "level crossing (and)" means the level crossing on the screen.

The angle at which the television camera (2) looks down at the level crossing (C) is such that the width of the level crossing (C) can be displayed on the screen as the length (h) in the vertical direction on the screen, and ), the angle is such that if an obstacle (A) is present above the length (h), the obstacle projects above the length (h). That is, when the height (H) of the obstacle (A) is greater than the length (h), the obstacle (A) always appears on the screen as a railroad crossing (
9) Projects upward.

In FIG. 3, (R) is a rail part, and (20) is a barrier that opens and closes the level crossing (C). In addition, (22) is a basic monitoring area set over the entire area of the level crossing (C), and (24) is a basic monitoring area that is set above and adjacent to the level crossing (C) and is similar to the basic monitoring area (22). This is an extended monitoring area having the same length (in the horizontal direction on the screen).

The image input section (3) sequentially converts the images projected by the television camera (2) into video signals, converts them into digital signals corresponding to the gradation of brightness, and outputs the digital signals to the image storage section (4). . In this way, the image storage section (4) stores images at fixed time intervals. Hereinafter, the image stored in this manner will be referred to as a "frame image."

The central processing unit (10) shown in FIG.
an image binarization unit (13) equipped with a foreign object detection unit (12);
14), a flat knitting testing section (15), and a determining section (17).

The method according to the invention will now be explained with reference to FIGS. 1 and 2. FIG.

First, as mentioned above, the TV camera (2) is fixed,
Images as shown in FIG. 3 are sequentially projected, and images at fixed time intervals are stored in the image storage section (4).

Next, a single reference brightness level is predetermined by the reference brightness level determination unit (12) of the image binarization unit (13) based on the -frame image stored in the image storage unit (4). This determination is based on, for example, the brightness level measured within the track near the level crossing (tan), which is not obscured by anything such as cars or people, when the level crossing is open and no trains are approaching. It is done. In this way,
The image binarization unit (13) uses the threshold value obtained by processing the reference brightness level to determine whether the level crossing is closed after a certain period of time after the signal from the level crossing closing condition input unit (6) is input. Basically, each pixel in the extended monitoring area (22°24) is determined based on the frame image after the railroad crossing is closed.
Image processing is performed to binarize the image into ``white'' or ``black'' (or ``1'' or ``0'').

Next, the foreign object detection unit (14) checks the continuous occupied area of "white" (or "1") in the basic monitoring area (22) after binarization, and detects the area on the level crossing (C). Detects the presence or absence of foreign objects.

When a foreign object on the level crossing (C) is detected in this way,
In order to remove small-sized noise (pebbles, dust, etc.), a horizontal knitting verification section (15) verifies whether the size of the foreign matter in the horizontal and vertical directions is larger than a certain level. That is, first, the horizontal direction verification is performed by measuring the length of the foreign object in the horizontal direction (left and right direction on the screen) and checking whether the measured value is equal to or larger than a predetermined minimum value. Furthermore, the vertical verification performed next detects the top and bottom edges of the foreign object on the screen, and confirms that the bottom edge is definitely on the level crossing (C) and that the top edge This test verifies whether or not it protrudes continuously from the lowest end above the level crossing (and), that is, into the extended monitoring area (24).

Therefore, in the determination section (17), based on the result of the test,
Determine whether the foreign object is larger than a certain size and select “Y”.
ES", it is further determined whether the foreign object has been continuously present at the same position for a certain period of time (for example, 5 to 6 seconds) or more.

In this way, the foreign object has a size of a certain size or more,
If it is determined that the obstacle exists for a certain period of time or more, it is determined that the obstacle (A) is truly present.

Although not shown, the image binarization section (13) in the central processing section (10) of the system shown in FIG. 2 may be replaced with an image multi-value conversion section. That is, by using the image multi-value conversion unit in this way, as shown by the broken line in FIG. , inside the railroad crossing (C) in the image multilevel conversion unit, that is,
A foreign object on the level crossing (C) is detected by visually differentiating the basic monitoring area (22) and performing edge tracking.

Thereafter, the size in the horizontal and vertical directions will be verified in the same way as the method shown in FIG.

Furthermore, in the method according to the present invention, the above-mentioned two
A foreign object on the level crossing (C) may be detected by using a colored image instead of a valued or multivalued image. That is, in this colored image, a color change due to the foreign object is detected.

Effects of the Invention As described above, in the method of the present invention, a television camera is fixed at a position looking down at a level crossing from slightly diagonally above so that the width of the level crossing can be recognized on the screen, and the level crossing is always at fixed coordinates. At the same time, the images are stored at fixed time intervals in the image storage unit, and when the level crossing is closed, image processing is performed on the inside of the level crossing based on the frame image after the level crossing is closed, thereby detecting foreign objects on the level crossing. detect the presence or absence of
When the foreign object is detected, it is verified whether the length of the foreign object in the lateral direction is equal to or greater than a certain value, and further, the lowermost end of the foreign object is located on the level crossing, and the uppermost end is located on the level crossing. Vertical verification is performed to determine whether or not the foreign object continuously protrudes above the level crossing, and as a result of the verification, it is determined that the foreign object is larger than a certain size, and that the foreign object remains in the same position continuously for a certain period of time or more. If there is an obstacle, it is determined that there is an obstacle.

In this way, the present invention can easily and reliably detect obstacles at a level crossing without requiring any construction work on the track.

[Brief explanation of the drawing]

1 is a flowchart of the method of the invention; FIG. 2 is a block diagram of the system used to carry out the method; and FIG. 3 is a diagram showing images captured by a television camera. (2): TV camera, (4): Image storage unit, (A):
Obstacle, (C): Level crossing. Patent applicant Nippon Signal Co., Ltd.

Claims (1)

[Scope of Claims] [1] A television camera (2) is fixed at a position looking down at the level crossing from slightly diagonally above so that the width of the level crossing can be recognized on the screen, and the level crossing (¥C¥) is always fixed. In addition to projecting an image that corresponds to the specified coordinates, images at regular intervals are stored in the image storage unit (4), and when the level crossing is closed, the level crossing road (￥C￥ ) detects the presence or absence of a foreign object on the level crossing (￥C￥) by image processing, and if the foreign object is detected, tests whether the lateral length of the foreign object is longer than a certain value, Furthermore, it is determined whether the lowest end of the foreign object is on the level crossing (¥C¥) and whether the top end of the foreign object is continuous from the lowest edge and protrudes above the level crossing (¥C¥). Vertical verification is performed, and as a result of the verification, the foreign object is larger than a certain size;
A method for detecting an obstacle at a level crossing using image processing technology, characterized in that, if the foreign object continues to exist at the same position for a certain period of time or more, it is determined that "an obstacle (A) is truly present".