A train approach warning method

Abstract

This invention relates to a train approach warning method, comprising the following steps: acquiring and preprocessing an image of a circular sign; performing a uniquely tuned circular Hough transform to obtain the center and radius of the circular area of ​​the sign, and calibrating the color gamut of the circular area; calculating the actual color gamut of the circular area of ​​the sign, and determining the direction of train approach.

Description

Technical Field

[0001] This invention belongs to the field of rail transit and relates to a train approach warning method. Background Technology

[0002] Track crossing monitoring is a typical labor-intensive industry. Currently, there are two main methods for train approach warning: one is to rely on manual inspection, handwritten records of various information, and 24 / 7 on-duty personnel; the other is to receive independent train approach warnings from track sensors, which requires the installation of a large number of expensive track sensors, resulting in a huge engineering workload and difficulty in implementation.

[0003] However, due to the rapid development of urban rail transit, the number of trains has increased dramatically, and operating conditions have become increasingly complex. This has led to shorter train intervals and longer operating times, placing immense pressure on level crossing management. Typically, a single rail line has dozens of level crossings. Manual monitoring is outdated, automation is low, and traditional track sensors are expensive and require extensive installation. Furthermore, there is a lack of data accumulation and interoperability, which falls far short of the vision of sustainable equipment and automated functions that intelligent operation and maintenance aims to achieve. Summary of the Invention

[0004] Based on this, in order to address the current predicament, this invention proposes a train approach warning method to replace manual monitoring and traditional track sensors, and to enable relevant personnel to monitor the approach status information of rail transit trains in real time.

[0005] To achieve the above objectives, the solution of this invention is as follows: acquire and preprocess the image of the circular sign; perform a uniquely tuned circular Hough transform to obtain the center and radius of the circular area of ​​the sign, and calibrate the color gamut of the circular area; calculate the actual color gamut of the circular area of ​​the sign, and determine the direction in which the train is approaching.

[0006] Further, step S2 involves performing a uniquely tuned circular Hough transform to obtain the accurate center and radius of the circular area of ​​the sign, and calibrating the color gamut of the circular area. This specifically includes the following steps:

[0007] S21) The blurred region of the sign in the image is manually determined. Setting the coordinates of the blurred center and the blurred radius can reduce the computational complexity of the uniquely tuned circular Hough transform. Let the blurred parameters be (x′0, y′0, r0′), where x′0, y′0, and r0′ represent the abscissa, ordinate, and blurred radius of the blurred center of the sign, respectively. Preferably, the circular region represented by the set blurred parameters should encompass the desired circular region.

[0008] S22) Using the position parameters (x′0, y′0, r0′) from S21, perform a uniquely tuned circular Hough transform on the preprocessed edge map to obtain the accurate center position parameters (x0, y0, r0) of the circular area of ​​the sign. Process the edge map I obtained from S1 according to the fuzzy position parameters (x′0, y′0, r0′). The pixel I(x, y) in the edge map takes the value 0 or 1. When its value is 1, it is an edge point; when its value is 0, it is a non-edge point. For a pixel with pixel coordinates (x, y), if the following conditions are met:

[0009]

[0010] Then set the value of I(x,y) to 0. It is easy to see that the range of x-coordinates for the region defined by the fuzzy circle is [x′0-r0′, x′0+r0′], and the range of the corresponding point indices is denoted as [P]. x Q x Similarly, the range of y-coordinates corresponding to the index range is denoted as [P]. y Q y ].

[0011] In the interval [P] x Q x On the [above], when x = x1, there exists y = y1 if the following conditions are met:

[0012]

[0013] Then an tunable circular Hough transform will be performed, that is:

[0014] y1sinθ=ρ

[0015] By iterating through θ∈[0,π], we obtain the range of parameters for ρ when x=x1. In [P x Q x By iterating through x, the tuning parameter range of ρ is obtained. After frequency statistics, the unique parameter pair with the highest frequency (θ) is selected. y ,ρ y If the ordinate of the center of the circle is y0, then the required ordinate of the circle is ρ. y / sinθ y Similarly, in [P] y Q y By iterating through y, we can find the x-coordinate of the circle's center, x0.

[0016] In [P] x Q x Iterate through x, in [P y Q y Iterate through y; if I(x,y) is 1, then calculate the candidate radius. The tuned circular Hough transform, i.e.:

[0017] rsinθ=ρ

[0018] After frequency statistics, select the unique parameter pair (θ0, ρ0) with the largest frequency, and the radius of the circle is r0 = ρ0 / sinθ0.

[0019] S23) Based on the center and radius of the circular region obtained in S22, define the color gamut of the circular region. Traverse the preprocessed image M(x,y) (with the same dimensions as the edge map described in S22) to form a set of pixel coordinates T, satisfying... Calculate the calibrated color gamut and calculate the average pixel value of the three channels (R, G, and B channels).

[0020]

[0021] Among them, M R (x,y) represents the pixel value of M(x,y) in the R channel, M G (x,y) represents the pixel value of M(x,y) in the G channel, M B (x,y) represents the pixel value of M(x,y) in the B channel, and |T| represents the number of elements in the set.

[0022] Further, step S3, calculating the actual color gamut of the circular area of ​​the sign and determining the direction in which the train is approaching, specifically includes the following steps:

[0023] S31) Based on the center pixel coordinates and pixel radius of the circular area of ​​the signboard in S2, and referring to step S23, calculate the current color gamut of the circular area, and calculate the average pixel value of the three channels (R, G, and B channels). The value of δ. Set the threshold δ. R ,δ G ,δ B ,like If both conditions are met, it is assumed that no train has passed at the current time, and step S31 is repeated; otherwise, proceed to step S32.

[0024] S32) Compare the calibrated color gamut with the current real color gamut, count the pixels that have changed, and calculate the center coordinates of the changed areas. Count the pixel coordinates of the occluded circular areas to form a set of pixel coordinates T0, satisfying... And the pixel values ​​of each channel of the points in the set satisfy... Any one of the three conditions. Let time t be the moment of T0, then the x-coordinate of the center of T0 is... t for

[0025]

[0026] Where |T0| represents the number of elements in the set. Let time t be a certain moment. a ,satisfy Let time t be a certain moment. b ,satisfy Let time t be a certain moment. c ,satisfy

[0027] S33) Analyze the center coordinates calculated in S32, and determine the direction of oncoming traffic based on the captured image. Statistically analyze the sequence of center horizontal coordinates at different times. If t... a <t b <t c If t is satisfied, then the train is coming from left to right in the image; a >t b >t c If so, the train is coming from right to left in the image. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of the scenario involved in this invention.

[0029] Figure 2 This is a block diagram illustrating the working principle of the present invention. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] To facilitate understanding by those skilled in the art, an embodiment is presented for the acquisition and preprocessing of images from the equipment cabinet: the camera is fixedly installed facing the circular sign, and the preprocessing includes, but is not limited to, grayscale conversion, equalization, and edge reduction. The original image, grayscale image, and edge image are stored in variable matrices M, G, and I, respectively, for use in subsequent steps.

[0032] To facilitate understanding by those skilled in the art, an embodiment is presented below to illustrate how to obtain the center and radius of a circular region on a sign using a uniquely tuned circular Hough transform, and how to define the color gamut of that circular region:

[0033] Specifically, the fuzzy parameters are set to (1013, 540, 460). The range of x-coordinates of the area defined by the fuzzy circle corresponds to the index range of points in the range of [553, 1473], and the range of y-coordinates corresponds to the index range of [80, 1000]. The x-coordinate of the circle center is found to be 1010, the y-coordinate to be 545, and the radius to be 450.

[0034] To facilitate understanding by those skilled in the art, an embodiment is presented below to illustrate how to calculate the actual color gamut of the circular area of ​​a signboard and determine the direction of train approach:

[0035] Specifically, the actual color gamut of the circular area of ​​the sign is obtained. The threshold values ​​are 240, 40, and 30, respectively, with a threshold δ. R ,δ G ,δ B The percentages are 5%, 8%, and 8%, respectively. In this embodiment, t is calculated. a =10.2s,t b =11.6s,t c =12.8s. Determined that the train is approaching from left to right in the captured footage.

Claims

1. A train approach warning method, characterized in that, Includes the following steps: S1. Acquire and preprocess the image of the circular sign; S2. Perform a uniquely tuned circular Hough transform to obtain the center and radius of the circular area of ​​the sign, and calibrate the color gamut of the circular area. The steps include: S21. Manually specify the blurred position area of ​​the sign in the image; S22. Using the blurred position area in S21, perform a uniquely tuned circular Hough transform on the preprocessed edge image to obtain the accurate center and radius of the circular area of ​​the sign; S23. Based on the center and radius of the circular area obtained in S22, calibrate the color gamut of the circular area. S3. Calculate the actual color gamut of the circular area of ​​the signboard and determine the direction of the approaching train. The steps include: S31. Calculate the actual color gamut of the circular area based on the center pixel coordinates and pixel radius of the circular area of ​​the signboard in S2; S32. Compare the calibrated color gamut with the current actual color gamut, count the pixels that have changed, and calculate the center coordinates of the changed area; S33. Determine the center coordinates calculated in S32 to determine the direction of the approaching train.

Images