Method for matching pantograph blade image and device therefor

Abstract

This invention discloses a method and apparatus for matching pantograph sliding plate images. The method includes the following steps: controlling a light-emitting device to emit light onto the pantograph sliding plate on the top of a train, and controlling a camera to capture a target roof image containing the pantograph sliding plate, wherein the train extends in a vertical direction in the target roof image; in the target roof image, searching for a first curve segment from bottom to top and a second curve segment from top to bottom, wherein the pixels in the first and second curves satisfy the following conditions: the gradient direction of the pixel in the vertical direction is a preset direction, and |grayscale value of the pixel above the pixel - grayscale value of the pixel below the pixel| > a preset threshold; the area between the first and second curve segments is the image corresponding to the pantograph sliding plate. This method can match the image corresponding to the pantograph sliding plate.

Description

Technical Field

[0001] This invention relates to the field of train inspection technology, and more particularly to a method and apparatus for matching pantograph slide images. Background Technology

[0002] Electric locomotives are typically equipped with pantographs, which draw current from the overhead contact line. During operation, the pantograph's contact plate is in direct contact with the overhead contact line, drawing current to power the locomotive. Understandably, this friction between the pantograph's contact plate and the overhead contact line causes wear on the contact plate. Therefore, the condition of the pantograph's contact plate directly affects the safe operation of the train. Excessive wear not only affects the normal power supply of the electric locomotive, but the resulting arcing further exacerbates the wear on both the contact plate and the overhead contact line. With the rapid development of high-speed railways, higher demands are placed on the reliable operation of pantographs, making intelligent detection of the pantograph's contact plate wear condition of great significance.

[0003] In the existing technology, one feasible solution is to take a picture of the pantograph plate and then determine whether the pantograph plate is damaged based on the obtained picture. Understandably, this requires matching the pantograph plate from the picture. Summary of the Invention

[0004] In view of this, the main objective of the present invention is to provide a method and apparatus for matching pantograph slide images.

[0005] To achieve the above objectives, the technical solution of the present invention is implemented as follows: a method for matching pantograph sliding plate images at a train detection point, wherein the train detection point includes: a track, a camera, a light-emitting device, and a ranging device, wherein the camera, the light-emitting device, and the ranging device are all located above the track, the camera is capable of capturing an image of the top of a train located on the track, and the light-emitting device is capable of emitting light towards the top of the train on the track; comprising the following steps:

[0006] The system controls the light-emitting device to be aimed at the pantograph plate on the top of the train and emit light, and controls the camera to capture a target roof image containing the pantograph plate. The target roof image is then converted to grayscale. In the target roof image, the train extends in the vertical direction.

[0007] In the target vehicle roof image, a first curve segment is found along the top-to-bottom direction. The first pixel on the first curve segment satisfies the following conditions: the gradient direction of the first pixel in the top-to-bottom direction is a preset direction, and |gray value of the pixel above the first pixel - gray value of the pixel below the first pixel| > a first threshold. A second curve segment is found along the bottom-to-top direction. The second pixel on the second curve segment satisfies the following conditions: the gradient direction of the second pixel in the top-to-bottom direction is a preset direction, and |gray value of the pixel above the second pixel - gray value of the pixel below the second pixel| > a second threshold. The area between the first and second curve segments is the image area corresponding to the pantograph sliding plate. The preset direction is either upward or downward, the first threshold > 0, and the second threshold > 0.

[0008] As an improvement to an embodiment of the present invention, the "converting the target roof image to grayscale" specifically includes: converting the target roof image to grayscale and then performing median filtering.

[0009] As an improvement to this embodiment of the invention, a ranging device is also provided above the track;

[0010] The specific steps of "controlling the light-emitting device to illuminate the pantograph plate on the top of the train and controlling the camera to capture an image of the target roof containing the pantograph plate" include: controlling the ranging device to measure the distance data between itself and the top of the train; fitting the top profile and the corresponding standard model of the train based on the distance data; obtaining the distance between the top profile and the pantograph based on the standard model; and waiting for a time T, where time T = distance / train speed. The standard model at least marks the outline and position information of each device on the top of the train. After that, controlling the light-emitting device to illuminate the pantograph plate on the top of the train and controlling the camera to capture an image of the target roof containing the pantograph plate.

[0011] As an improvement to an embodiment of the present invention, the step of "fitting the top profile and the standard model corresponding to the train based on the distance data" specifically includes: fitting the top profile based on the distance data, obtaining the train's front slope and average roof base plane based on the top profile, obtaining the train's model based on the front slope and average roof base plane, and obtaining the standard model corresponding to the model.

[0012] As an improvement to this embodiment of the invention, the "fitting the top contour based on the distance data" specifically includes: processing the distance data based on the DSCN neural network and fitting the top contour.

[0013] This invention also provides a device for matching pantograph contactor images at train detection points. The train detection point includes: a track, a camera, a light-emitting device, and a ranging device. The camera, light-emitting device, and ranging device are all located above the track. The camera can capture an image of the top of a train on the track, and the light-emitting device can emit light towards the top of the train on the track. The device includes the following modules:

[0014] The data acquisition module controls the light-emitting device to aim at the pantograph sliding plate on the top of the train and emit light, and controls the camera to capture a target roof image containing the pantograph sliding plate, and performs grayscale processing on the target roof image; in the target roof image, the train's extension direction is vertical; the processing module is used to find a first curve segment in the target roof image along the vertical direction, where the first pixel on the first curve segment satisfies the following conditions: the gradient direction of the first pixel in the vertical direction is a preset direction, and |pixels above the first pixel The gray value of a point minus the gray value of the pixel below the first pixel is greater than the first threshold. A second curve segment is found along the top-to-bottom direction. The second pixel on the second curve segment satisfies the following conditions: the gradient direction of the second pixel in the top-to-bottom direction is a preset direction, and |the gray value of the pixel above the second pixel minus the gray value of the pixel below the second pixel| is greater than the second threshold. The area between the first and second curve segments is the image area corresponding to the pantograph sliding plate. The preset direction is either upward or downward, the first threshold is greater than 0, and the second threshold is greater than 0.

[0015] As an improvement to this embodiment of the invention, the data acquisition module is further configured to: convert the target roof image to grayscale and then perform median filtering.

[0016] As an improvement of this embodiment of the invention, a ranging device is also provided above the track; the data acquisition module is further used to: control the ranging device to measure the distance data between itself and the top of the train, fit the top profile and the standard model corresponding to the train based on the distance data, obtain the distance between the top profile and the pantograph based on the standard model, and wait for a time T, where time T = distance / train speed, and the standard model is marked with the outline and position information of each device on the top of the train; then, control the light-emitting device to aim at the pantograph sliding plate on the top of the train and emit light, and control the camera to capture an image of the target roof containing the pantograph sliding plate.

[0017] As an improvement to an embodiment of the present invention, the data acquisition module is further configured to: fit a top profile based on the distance data, obtain the train's front slope and average roof base plane based on the top profile, obtain the train's model based on the front slope and average roof base plane, and obtain the standard model corresponding to the model.

[0018] As an improvement to this embodiment of the invention, the data acquisition module is further configured to: process the distance data based on the DSCN neural network and fit the top contour.

[0019] The method and apparatus for matching pantograph sliding plate images provided in this invention have the following advantages: This invention discloses a method and apparatus for matching pantograph sliding plate images. The method includes the following steps: controlling a light-emitting device to emit light onto the pantograph sliding plate on the top of a train, and controlling a camera to capture a target roof image containing the pantograph sliding plate, wherein the train's extension direction in the target roof image is vertical; in the target roof image, searching for a first curve segment from bottom to top and another from top to bottom, wherein the pixels in the first and second curves satisfy the following: the gradient direction of the pixel in the vertical direction is a preset direction, and |grayscale value of the pixel above the pixel - grayscale value of the pixel below the pixel| > a preset threshold; the area between the first and second curve segments is the image corresponding to the pantograph sliding plate. This method can match the image corresponding to the pantograph sliding plate. Attached Figure Description

[0020] Figure 1 This is a structural diagram of a train inspection point provided in an embodiment of the present invention;

[0021] Figure 2A , Figure 2B and Figure 3 A schematic diagram illustrating the method for matching pantograph slide plate images provided in an embodiment of the present invention;

[0022] Figure 4 A flowchart illustrating the method for matching pantograph sliding plate images provided in an embodiment of the present invention;

[0023] Figure 5 This is a schematic diagram illustrating the method for matching pantograph slide images provided in an embodiment of the present invention. Detailed Implementation

[0024] The present invention will now be described in detail with reference to the embodiments shown in the accompanying drawings. However, these embodiments do not limit the present invention, and any structural, methodological, or functional modifications made by those skilled in the art based on these embodiments are included within the scope of protection of the present invention.

[0025] The following description and accompanying drawings fully illustrate specific embodiments described herein to enable those skilled in the art to practice them. Some embodiments may include or substitute parts and features of other embodiments. The scope of the embodiments herein encompasses the entire scope of the claims and all available equivalents thereof. Throughout this document, the terms “first,” “second,” etc., are used only to distinguish one element from another without requiring or implying any actual relationship or order between the elements. Indeed, a first element can also be referred to as a second element, and vice versa. Furthermore, the terms “comprising,” “including,” or any other variations thereof are intended to cover non-exclusive inclusion, such that a structure, apparatus, or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a structure, apparatus, or device. Without further limitation, an element defined by the phrase “comprising one…” does not exclude the presence of other identical elements in the structure, apparatus, or device that includes said element. The various embodiments described herein are presented in a progressive manner, with each embodiment focusing on its differences from other embodiments; similar or identical parts between embodiments can be referred to interchangeably.

[0026] The terms "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer" used in this document to indicate orientations or positional relationships are based on the orientations or positional relationships shown in the accompanying drawings. They are used solely for the convenience of describing the document and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In the description herein, unless otherwise specified and limited, the terms "installed," "connected," and "linked" should be interpreted broadly. For example, they can refer to mechanical or electrical connections, or internal connections between two elements; they can be direct connections or indirect connections through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms according to the specific circumstances.

[0027] Embodiment 1 of the present invention provides a method for matching pantograph contactor images at train detection points. The train detection point includes: a track 1, a camera 2, a light-emitting device 3, and a ranging device. The camera 2, light-emitting device 3, and ranging device are all located above the track 1. The camera 2 can capture an image of the top of a train located on the track 1, and the light-emitting device 3 can emit light towards the top of the train on the track 1. Here, several train detection points can be set along the track 1, such as... Figure 1As shown, a gantry frame 4 is installed beside track 1. A light-emitting device 3 (e.g., a light bulb) and a camera 2 are installed on top of the gantry frame 4. A distance measuring device is installed on top of the gantry frame 4 between the front and rear sets of light-emitting devices 3 and cameras 2. In practice, the train will travel on track 1 and can also stop on track 1. Optionally, the camera 2 and the light-emitting device 3 are rotatable, thereby adjusting the shooting angle of the camera 2 and the light-emitting device 3 towards the pantograph slider.

[0028] like Figure 4 As shown, it includes the following steps:

[0029] Step 401: Control the light-emitting device 3 to align with the pantograph plate on the top of the train and emit light, and control the camera 2 to capture an image of the target roof containing the pantograph plate, and perform grayscale processing on the target roof image; in the target roof image, the extension direction of the train is vertical; here, as... Figure 2A and Figure 2B As shown, in the image on the left, the target vehicle roof image may be tilted (i.e., the train's extension direction is not vertical). Therefore, the target vehicle roof image can be rotated to make the train's extension direction vertical. Furthermore, as... Figure 2B As shown, the extension direction of the wire in the contact wire above the pantograph can be considered to be approximately parallel to the direction from the front of the vehicle to the rear. Therefore, the target vehicle roof image can be further rotated based on the extension direction of the wire.

[0030] In practice, some pantographs have one pantograph slide plate, while others have two. When there are two slide plates, the captured images can be segmented, and each pantograph image can be identified separately.

[0031] Step 402: In the target vehicle roof image, find a first curve segment from bottom to top. The first pixel on the first curve segment satisfies the following conditions: the gradient direction of the first pixel in the vertical direction is a preset direction, and |gray value of the pixel above the first pixel - gray value of the pixel below the first pixel| > a first threshold. Find a second curve segment from top to bottom. The second pixel on the second curve segment satisfies the following conditions: the gradient direction of the second pixel in the vertical direction is a preset direction, and |gray value of the pixel above the second pixel - gray value of the pixel below the second pixel| > a second threshold. The area between the first and second curve segments is the image area corresponding to the pantograph sliding plate. The preset direction is either upward or downward, the first threshold > 0, and the second threshold > 0.

[0032] like Figure 2A and Figure 2BAs shown, because the upper surface of the pantograph slide plate is constantly in a state of friction with the power line during pantograph raising, the upper surface of the pantograph slide plate is relatively smooth. This results in a significant difference in color between the upper surface of the pantograph and the surrounding environment. Furthermore, the pantograph slide plate is approximately horizontal (i.e., left-right). To reduce interference from other vertical edges, only horizontal edges can be matched. The difference in grayscale thresholds between the upper and lower edges of the pantograph is highly sensitive; when the difference in grayscale values ​​between the two sides of an edge exceeds a certain threshold, that point is defaulted to an edge point. Based on the overall imaging of the upper and lower edges of the pantograph slide plate, an appropriate threshold is selected to ensure that the pantograph edges can be successfully detected.

[0033] In practice, we can first locate a target pixel and then search left or right along this target pixel. If a break point occurs, we can use the slope of the previous speed limit point as a constraint to fit the current break point until the search is completed, and then we can obtain the first curve segment and the second curve segment.

[0034] In this embodiment, the "converting the target roof image to grayscale" specifically includes: converting the target roof image to grayscale and then performing median filtering. Here, averaging, max-min averaging, weighted averaging, binary image methods, and image inversion methods can be used to convert the target roof image to grayscale. Median filtering is a non-linear smoothing technique that sets the grayscale value of each pixel to the median of the grayscale values ​​of all pixels within a neighborhood window of that pixel, thereby enabling the filtering of image noise.

[0035] Here, as Figure 3 As shown, the gray values ​​of the pixels are segmented, and the gray value of each pixel can be described as... Where A and B are the grayscale values ​​of the pixels on both sides of the edge, and S A and S B Let A and B be the areas occupied by gray values ​​A and B within the pixel, respectively, and h be the pixel side length. The segmentation boundary is... By combining and iterating the formulas, the normalized normal vector can be obtained. The normal vector can be used as the gradient.

[0036] In this embodiment, a ranging device is also installed above the track; the step of "controlling the light-emitting device 3 to aim at the pantograph plate on the top of the train and emit light, and controlling the camera 2 to capture an image of the target roof containing the pantograph plate" specifically includes: controlling the ranging device to measure the distance data between itself and the top of the train, fitting the top outline and the standard model corresponding to the train based on the distance data, obtaining the distance between the top outline and the pantograph based on the standard model, and waiting for a time T, where time T = distance / train speed, and the standard model at least marks the outline and position information of each device on the top of the train; then, controlling the light-emitting device 3 to aim at the pantograph plate on the top of the train and emit light, and controlling the camera 2 to capture an image of the target roof containing the pantograph plate.

[0037] Here, during trial operation before the equipment is officially put into use, multiple sets of distance data between the train's top and the ranging device can be obtained. The DSCN algorithm is used to fit the train's top profile, and positioning parameters such as the front curvature (curve slope here), the basic top height (mode), and the pantograph height range are calculated. Trains with significantly different positioning parameters are defined as standard models and input into the database. Subsequently, during use, the distance between the top profile and the ranging device should be obtained, and the ranging device should directly output the distance.

[0038] Here, as Figure 1 As shown, in practice, a gantry 4 can be set on track 1, and a distance measuring device can be installed on the top of the gantry 4. When a train passes, the distance measuring device can be controlled to scan the top of the train, and at this time, distance measuring data (i.e., the real-time distance value between the distance measuring device and the roof of the train) will be returned.

[0039] Here, the ranging device can be a pulse radar ranging sensor. A pulse radar ranging sensor is a type of radar that can emit pulse signals and perform ranging functions. Its basic process is as follows: The transmitter in the radar generates sufficient electromagnetic energy, which is transmitted to the antenna via a transceiver switch. The antenna then radiates this electromagnetic energy into the atmosphere, concentrating it in a narrow direction to form a beam that propagates forward. When the electromagnetic wave encounters a target within the beam, it is reflected in various directions, with a portion of the electromagnetic energy reflected back to the radar and captured by the radar antenna. The energy captured by the antenna is sent to the receiver via the transceiver switch, forming the radar echo signal. Because electromagnetic waves attenuate with propagation distance, the radar echo signal is very weak, almost completely drowned out by noise. The receiver amplifies the weak echo signal, processes it through a signal processor, extracts the information contained in the echo, and sends it to a display to show the target's distance, direction, speed, etc., thus completing the ranging function.

[0040] Optionally, the pulse radar ranging sensor employs pulse microwave technology, in which the antenna system emits a pulse beam with a frequency of 6.3 GHz and a duration of 0.8 ns, followed by a 278 ns pause. During the pulse emission pause, the antenna system acts as a receiver, receiving the reflected wave and simultaneously processing the echo image data to plot the top profile.

[0041] In this embodiment, the step of "fitting the top profile and the corresponding standard model of the train based on the distance data" specifically includes: fitting the top profile based on the distance data; obtaining the train's front slope and average roof base plane based on the top profile; obtaining the train's model based on the front slope and average roof base plane; and obtaining the standard model corresponding to the model. Here, the train's front is usually conical (when both ends of the train are conical, both ends can be considered as the front), and the upper surface of the front is inclined upwards in the direction facing the train body. Therefore, the front slope can be obtained, and the front slope of different train models is usually different; in addition, the average roof base plane of different train models is also different; therefore, the train model can be determined based on these two attributes.

[0042] Here, distance data is sequentially input into the DSCN neural network, using the difference between the output value and the actual data as the objective function. Finally, the gradient descent method is used to solve for the DSCN neural network input with near-zero error. The output image is shown below. Figure 5 As shown, dark colors represent discrete real data, while light colors represent fitted continuous curves.

[0043] In this embodiment, "fitting the top contour based on the distance data" specifically includes: processing the distance data using a DSCN neural network and fitting the top contour. Here, after performing simple interval filtering on the received data, the remaining values ​​are sequentially used as input values ​​and input into the DSCN neural network, with the difference between the fitted data and the actual data used as the objective function. The number of hidden layer neurons in the DSCN neural network varies; it starts with a small network structure and gradually increases the number of nodes in the hidden layers, calculating external weight parameters until the model meets the pre-given conditions and completes the fitting of the existing data. As the amount of input data increases, the fitting curve will approach the roof contour.

[0044] DSCN (Deep Stochastic Configuration Networks) neural networks possess a basic neural network structure. They utilize hidden nodes generated by inequality constraints to automatically adjust the range of hidden parameter selection and use the least squares method to calculate the optimal output weights, thus exhibiting extremely fast learning speed and universal approximation capabilities. Using the obtained real-time train roof curve and vehicle type as input, different positioning parameters can be selected based on different vehicle types. The real-time train roof curve is input into the DSCN neural network with the corresponding parameters, allowing for the fitting of the real-time roof profile. This process eliminates outliers and interference terms. Finally, the pantograph's profile data is compared with the real-time roof profile to identify the corresponding positioning point.

[0045] Embodiment 2 of the present invention provides a device for matching pantograph sliding plate images at a train detection point. The train detection point includes: a track 1, a camera 2, a light-emitting device 3, and a ranging device. The camera 2, the light-emitting device 3, and the ranging device are all located above the track 1. The camera 2 can capture an image of the top of a train located on the track 1, and the light-emitting device 3 can emit light towards the top of the train on the track 1. The device includes the following modules:

[0046] The data acquisition module is used to control the light-emitting device 3 to aim at the pantograph plate on the top of the train and emit light, and to control the camera 2 to capture a target roof image containing the pantograph plate, and to perform grayscale processing on the target roof image; in the target roof image, the extension direction of the train is the up and down direction;

[0047] The processing module is used to find a first curve segment in the target vehicle roof image from bottom to top. The first pixel on the first curve segment satisfies the following conditions: the gradient direction of the first pixel in the vertical direction is a preset direction, and |gray value of the pixel above the first pixel - gray value of the pixel below the first pixel| > a first threshold. The module also searches for a second curve segment from top to bottom. The second pixel on the second curve segment satisfies the following conditions: the gradient direction of the second pixel in the vertical direction is a preset direction, and |gray value of the pixel above the second pixel - gray value of the pixel below the second pixel| > a second threshold. The area between the first and second curve segments is the image area corresponding to the pantograph sliding plate. The preset direction is either upward or downward, and the first threshold > 0 and the second threshold > 0.

[0048] In this embodiment, the data acquisition module is further configured to: convert the target roof image to grayscale and then perform median filtering.

[0049] In this embodiment, a ranging device is also provided above the track;

[0050] The data acquisition module is further configured to: control the ranging device to measure the distance data between itself and the top of the train; fit the top profile and the standard model corresponding to the train based on the distance data; obtain the distance between the top profile and the pantograph based on the standard model; and wait for a time T, where time T = distance / train speed. The standard model at least marks the profile and position information of each device on the top of the train. After that, control the light-emitting device 3 to aim at the pantograph sliding plate on the top of the train and emit light, and control the camera 2 to capture an image of the target roof containing the pantograph sliding plate.

[0051] In this embodiment, the data acquisition module is further configured to: fit a top profile based on the distance data, obtain the train's front slope and average roof base plane based on the top profile, obtain the train's model based on the front slope and average roof base plane, and obtain the standard model corresponding to the model.

[0052] In this embodiment, the data acquisition module is further configured to: process the distance data based on the DSCN neural network and fit the top contour.

[0053] It should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This way of describing the specification is only for clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

[0054] The detailed descriptions listed above are merely specific descriptions of feasible embodiments of the present invention, and are not intended to limit the scope of protection of the present invention. All equivalent embodiments or modifications made without departing from the spirit of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for matching pantograph sliding plate images at a train detection point, the train detection point comprising: a track (1), a camera (2), a light-emitting device (3), and a ranging device, wherein the camera (2), the light-emitting device (3), and the ranging device are all located above the track (1), the camera (2) is capable of capturing an image of the top of a train located on the track (1), and the light-emitting device (3) is capable of emitting light towards the top of the train on the track (1); a ranging device is also provided above the track; characterized in that, Includes the following steps: The distance measuring device is controlled to measure the distance data between itself and the top of the train. The distance data is processed based on the DSCN neural network and the top profile is fitted. Based on the top profile, the front slope of the train and the average base plane of the roof are obtained. Based on the front slope and the average base plane of the roof, the model of the train is obtained, and the standard model corresponding to the model is obtained. Based on the standard vehicle model, the distance between the top outline and the pantograph is obtained, and the waiting time T is calculated, where time T = distance / train speed. The standard vehicle model at least marks the outline and position information of each device on the top of the train. Then, the light-emitting device (3) is controlled to be aligned with the pantograph sliding plate on the top of the train and emit light, and the camera (2) is controlled to capture the target roof image containing the pantograph sliding plate. The target roof image is then converted to grayscale. In the target vehicle roof image, the train extends in a vertical direction; In the target vehicle roof image, a first curve segment is found along the top-to-bottom direction. The first pixel on the first curve segment satisfies the following conditions: the gradient direction of the first pixel in the top-to-bottom direction is a preset direction, and |gray value of the pixel above the first pixel - gray value of the pixel below the first pixel| > a first threshold. A second curve segment is found along the bottom-to-top direction. The second pixel on the second curve segment satisfies the following conditions: the gradient direction of the second pixel in the top-to-bottom direction is a preset direction, and |gray value of the pixel above the second pixel - gray value of the pixel below the second pixel| > a second threshold. The area between the first and second curve segments is the image area corresponding to the pantograph sliding plate. The preset direction is either upward or downward, the first threshold > 0, and the second threshold > 0.

2. The method for matching pantograph trolley images according to claim 1, characterized in that, The "converting the target vehicle roof image to grayscale" process specifically includes: The target vehicle roof image is converted to grayscale and then subjected to median filtering.

3. An apparatus for matching pantograph sliding plate images at a train detection point, the train detection point comprising: a track (1), a camera (2), a light-emitting device (3), and a ranging device, wherein the camera (2), the light-emitting device (3), and the ranging device are all located above the track (1), the camera (2) is capable of capturing an image of the top of a train located on the track (1), and the light-emitting device (3) is capable of emitting light towards the top of the train on the track (1); a ranging device is also provided above the track; characterized in that, Includes the following modules: The data acquisition module is used to control the ranging device to measure the distance data between itself and the top of the train, process the distance data based on the DSCN neural network, fit the top profile, obtain the train's front slope and the average base plane of the roof based on the top profile, obtain the train's model based on the front slope and the average base plane of the roof, and obtain the standard model corresponding to the model. Based on the standard vehicle model, the distance between the top outline and the pantograph is obtained, and the waiting time T is calculated, where time T = distance / train speed. The standard vehicle model at least marks the outline and position information of each device on the top of the train. Then, the light-emitting device (3) is controlled to be aligned with the pantograph sliding plate on the top of the train and emit light, and the camera (2) is controlled to capture a target roof image containing the pantograph sliding plate. The target roof image is then converted to grayscale. In the target roof image, the extension direction of the train is the up and down direction. The processing module is used to find a first curve segment in the target vehicle roof image from bottom to top. The first pixel on the first curve segment satisfies the following conditions: the gradient direction of the first pixel in the vertical direction is a preset direction, and |gray value of the pixel above the first pixel - gray value of the pixel below the first pixel| > a first threshold. The module also searches for a second curve segment from top to bottom. The second pixel on the second curve segment satisfies the following conditions: the gradient direction of the second pixel in the vertical direction is a preset direction, and |gray value of the pixel above the second pixel - gray value of the pixel below the second pixel| > a second threshold. The area between the first and second curve segments is the image area corresponding to the pantograph sliding plate. The preset direction is either upward or downward, and the first threshold > 0 and the second threshold > 0.

4. The apparatus for matching pantograph sliding images according to claim 3, characterized in that, The data acquisition module is also used to: convert the target roof image to grayscale and then perform median filtering.

Images