A turnout identification method, system, and track cleaning vehicle

Abstract

The present invention provides a turnout identification method, system, and track cleaning vehicle. The turnout identification method obtains an infrared image of a track, performs preprocessing, edge detection, track profile screening, and Hough transformation on the infrared image in sequence, thereby obtaining a track line image, and then performs track line fitting and Kalman filtering on the track line image. Since the current track is fixed in position on the infrared image, the existence of the turnout can be determined based on the Kalman coefficient, and then the position of the turnout is determined by position offset information and the triangulation principle. The turnout identification system of the present invention implements the above turnout identification method, and the track cleaning vehicle of the present invention is equipped with the above turnout identification system. The present invention has the characteristics of being simple and feasible, having a small amount of calculation, and having good reliability.

Description

Technical Field

[0001] The present invention relates to the field of image recognition, and in particular to a turnout recognition method, system and track cleaning vehicle. Background Art

[0002] Modern trams are the backbone of transportation in small and medium-sized cities. Their lines are mainly laid on the road surface, mainly ground lines. Among them, turnouts are one of the core components of the track structure. The realization of tram interconnection benefits from the extensive use of cross turnouts.

[0003] At present, track cleaning vehicles are gradually being widely used in track cleaning. For the cleaning of turnouts, the camera on the track cleaning vehicle often collects images of the turnouts, and the driver determines whether to clean the turnouts. This manual judgment method is inefficient, and if the existence of the turnout cannot be determined in time, when passing the turnout, the cleaning tools on the track cleaning vehicle are easy to hit the turnout and be damaged, and the sweeping disc is easy to wear. Therefore, a solution that can realize automatic identification of turnouts is needed.

[0004] The Chinese invention patent with the announcement number CN112149622A discloses a railway turnout identification method and system. The scheme uses a camera to identify the color image of the railway turnout, extracts the edge information of the track line in the color image, obtains the edge binary image, and then uses Hough transform to determine the turnout main line, determines all turnout side lines through the inclination threshold range and pixel point threshold, and determines the type of turnout according to the coordinate numerical relationship between the turnout main line and the turnout side line. Although the scheme can identify the turnout, it uses a camera to collect color images, so it cannot identify the turnout in bad weather or at night when visibility is low. And because the scheme focuses on identifying the specific turnout type, the speed of turnout identification is slow. Summary of the invention

[0005] The first object of the present invention is to provide a turnout identification method capable of quickly and automatically identifying a turnout.

[0006] A second object of the present invention is to provide a turnout identification system.

[0007] A third object of the present invention is to provide a track cleaning vehicle that can quickly identify a turnout and promptly lift the cleaning tool.

[0008] In order to achieve the above-mentioned first purpose, the present invention provides a turnout identification method, wherein: an infrared image of a track is acquired; the infrared image is preprocessed to obtain a binary image; the binary image is processed by an edge detection algorithm to obtain an edge binary image; the edge binary image is subjected to a Hough transform to obtain a track line image; the track line image is subjected to track line fitting; a Kalman filter is performed on the track line image after the track line fitting; it is determined whether the Kalman coefficient is greater than a first preset threshold value, and if so, it is determined that a turnout exists; after determining that a turnout exists, the position of the turnout is determined based on the received position offset information and the triangulation principle.

[0009] As can be seen from the above scheme, the present invention processes the collected infrared image of the track to obtain the track line image, and determines the existence of the turnout according to the Kalman coefficient being greater than the first preset threshold when the turnout exists. After determining the existence of the turnout, the turnout is determined to obtain the position through the position offset information and the triangulation principle. Since the present invention uses an infrared thermal imager to collect track images, it can work in bad weather and at night, avoiding the problem that ordinary cameras cannot normally shoot images under the influence of bad environments such as rain and fog in bad weather such as rain and fog. When encountering rainy and foggy days, it is easier to accumulate garbage such as fallen leaves and stones on the track, and the track often needs to be cleaned; and it avoids the problem that ordinary cameras cannot shoot clearly at night, and has the characteristics of wide application range and good reliability. In addition, since it is judged whether the turnout exists by whether the Kalman coefficient is greater than the first preset threshold, and the turnout position is determined by the position offset information and the triangulation principle, it is not necessary to train for different types of turnouts, which improves versatility, reduces the difficulty of recognition, and improves the recognition speed. It has the characteristics of simple implementation method and fast calculation speed.

[0010] A further solution is to obtain the infrared image of the track by using a high-speed infrared thermal imager.

[0011] A further solution is to pre-process the infrared image and, when obtaining a binary image, perform the following steps: sequentially perform grayscale processing, median filtering, histogram equalization processing, and threshold segmentation processing on the infrared image to obtain a binary image.

[0012] It can be seen that the acquired image can be enhanced, noise interference can be reduced, and a clear binary image can be obtained.

[0013] A further solution is to process the binary image by an edge detection algorithm to obtain an edge binary image, and then perform the following steps: process the binary image by a Canny edge detection operator to obtain an edge binary image.

[0014] It can be seen that the amount of data can be reduced, information irrelevant to the edge can be eliminated, and the important structure of the image can be retained.

[0015] A further solution is to perform the following steps after obtaining the edge binary image and before performing Hough transform on the edge binary image: performing track line fitting on the track line image.

[0016] It can be seen that the amount of calculation can be reduced and the accuracy of turnout identification can be increased.

[0017] A further solution is to perform Hough transform on the edge binary image to obtain the track line image, and then perform the following steps: perform PPHT Hough transform on the edge binary image to obtain the track line image.

[0018] It can be seen that the track outline is highlighted in the image, and the lines are basically composed of single-pixel points, which is convenient for subsequent identification.

[0019] A further solution is to perform the following steps when fitting the track line image: fitting the track line through a parabola model and a least squares method.

[0020] It can be seen that the expression of the track line image can be made more intuitive and the complexity of the fitting algorithm can be reduced.

[0021] A further solution is to determine whether the Kalman coefficient is greater than the first preset threshold value, and at the same time determine whether the number of parabolas fitted by the fitting parabola model is greater than the second preset threshold value. If the above determination results are all yes, it is determined that the turnout exists. It can be seen that the accuracy of turnout identification can be further increased.

[0022] In order to achieve the above-mentioned second purpose, the present invention provides a turnout identification system, which includes: an infrared image acquisition module and a main control module, wherein the infrared image acquisition module is connected to the main control module; the infrared image acquisition module includes an infrared thermal imager and an infrared transmitter; the infrared thermal imager is used to acquire infrared images of tracks and transmit the infrared images to the main control module; the infrared transmitter is used to send infrared beams to the turnouts, receive reflected beams, output imaging offset information and transmit it to the main control module; the main control module is used to: pre-process the infrared image to obtain a binary image; process the binary image by an edge detection algorithm to obtain an edge binary image; perform track contour screening on the edge binary image; perform Hough transform on the edge binary image after track contour screening to obtain a track line image; perform track line fitting on the track line image; perform Kalman filtering on the track line image after track line fitting; judge whether the Kalman coefficient is greater than a first preset threshold value, and if so, determine that a turnout exists; after determining that a turnout exists, determine the position of the turnout according to the received position offset information and the triangulation principle.

[0023] It can be seen from the above scheme that the present invention collects infrared images through the infrared thermal imager in the infrared image acquisition module, and the infrared transmitter collects position offset information. The main control module obtains the above infrared images and position offset information, thereby being able to identify and determine the position of the turnout. It has the characteristics of all-weather operation, strong anti-interference, fast computing speed, and wide versatility.

[0024] In order to achieve the third purpose mentioned above, the present invention provides a track cleaning vehicle, which includes: an infrared image acquisition module, a main control module, and a cleaning module. The infrared image acquisition module is connected to the main control module, and the main control module is connected to the cleaning module; the infrared image acquisition module is arranged at the central position of the front cab of the track cleaning vehicle; the infrared image acquisition module includes an infrared thermal imager and an infrared transmitter; the infrared thermal imager is used to acquire infrared images of the track and transmit the infrared images to the main control module; the infrared transmitter is used to send infrared beams to the turnout, receive reflected beams, output imaging offset information and transmit it to the main control module; the main control The module is used for: preprocessing the infrared image to obtain a binary image; processing the binary image through an edge detection algorithm to obtain an edge binary image; performing track contour screening on the edge binary image; performing Hough transform on the edge binary image after track contour screening to obtain a track line image; performing track line fitting on the track line image; performing Kalman filtering on the track line image after track line fitting; judging whether the Kalman coefficient is greater than a first preset threshold value, and if so, determining the existence of a turnout; after determining the existence of the turnout, determining the position of the turnout according to the received position offset information and the triangulation principle.

[0025] It can be seen from the above scheme that the present invention has the characteristics of simple setting, automatic identification of switches and fast identification of switches. After the main control module identifies the switch, the cleaning module can be controlled to lift the cleaning tool in time, so that the cleaning module can work in a safe environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a system framework diagram of an embodiment of the turnout identification system of the present invention.

[0027] Figure 2 It is a flow chart of an embodiment of the turnout identification method of the present invention.

[0028] The present invention is further described below in conjunction with the accompanying drawings and embodiments. DETAILED DESCRIPTION

[0029] The turnout recognition system of the present invention collects infrared images through an infrared image acquisition module, and performs preprocessing, edge detection, track contour screening, and Hough transformation on the infrared images through a main control module to obtain a clearer track line image, and then performs track line fitting and Kalman filtering on the track line image, and determines the existence of the turnout through the Kalman coefficient, and then determines the position of the turnout through position offset information and triangulation principle. The track recognition system of the present invention is carried on a track cleaning vehicle, which can realize that the track cleaning vehicle automatically adjusts the cleaning tool according to the turnout to prevent the cleaning tool from being damaged by hitting the turnout.

[0030] The turnout identification system embodiment of the present invention:

[0031] Reference Figure 1 The turnout recognition system 1 includes an infrared image acquisition module 11 and a main control module 12 , and the infrared image acquisition module 11 is connected to the main control module 12 .

[0032] The infrared image acquisition module includes an infrared thermal imager 111 and an infrared transmitter 112. The image acquisition speed of the infrared thermal imager 111 needs to match the running speed of the track cleaning vehicle. The infrared thermal imager 111 is used to acquire infrared images of the track and transmit the infrared images to the main control module 12. The infrared transmitter 112 is used to send infrared beams to the switch and receive reflected beams, and output imaging offset information to the main control module 12.

[0033] The main control module 12 can realize the following functions:

[0034] Preprocess the infrared image to obtain a binary image; process the binary image through an edge detection algorithm to obtain an edge binary image; perform track contour screening on the edge binary image; perform Hough transform on the edge binary image after track contour screening to obtain a track line image; perform track line fitting on the track line image; perform Kalman filtering on the track line image after track line fitting; determine whether the Kalman coefficient is greater than a first preset threshold, and if so, determine that a turnout exists; after determining that a turnout exists, determine the position of the turnout according to the received position offset information and the triangulation principle. The functions implemented by the main control module 12 will be described in detail later.

[0035] Embodiments of the track cleaning vehicle of the present invention:

[0036] The track cleaning vehicle is equipped with the same track identification system as the above-mentioned track identification system embodiment, wherein the infrared image acquisition module is arranged in the central position of the cab of the front of the track cleaning vehicle, and the infrared thermal imager measures the projection of the ground position to determine that it is exactly located in the middle between the two grooved rails. The rest is the same as the above-mentioned track identification system embodiment, and will not be repeated here. In addition, the track cleaning vehicle also includes a cleaning module, which includes cleaning tools such as a sweeping disc, a high-pressure water gun, and a fan, and the cleaning module is used to clean the track. When the main control module determines the position of the turnout, it sends a control instruction to the cleaning module, and the cleaning module lifts the cleaning tool according to the control instruction to prevent the cleaning tool from colliding with the turnout and being damaged, so that the cleaning tool can perform cleaning work in a safe manner. The cleaning work can be to control the fan to form a strong negative pressure at the garbage bin to suck in the garbage of the turnout to achieve the purpose of cleaning. If the operation of lifting the cleaning tool fails, the cleaning module will also send an alarm message to the main control module to remind the operator to pay attention.

[0037] The turnout identification method embodiment of the present invention:

[0038] Reference Figure 2 When the track cleaning vehicle starts to run and the turnout recognition system starts to work, step S1 is first performed to obtain an infrared image of the track. Specifically, the infrared images of both sides of the track are obtained by a high-speed infrared thermal imager carried on the track cleaning vehicle, and the infrared images are transmitted to the main control module of the turnout recognition system in real time.

[0039] Then, step S2 is executed to pre-process the infrared image to obtain a binary image. Specifically, the main control module performs grayscale processing on the infrared images on both sides of the track to obtain a grayscale image. Since the acquired infrared image is not clear enough, in order to reduce the noise interference in the infrared image, the median filter is used for denoising. The median filter can smooth the impulse noise, protect the sharp edges of the image, and avoid edge blur during the processing of the infrared image. Then, the image after the median filter is processed by histogram equalization. The purpose of the histogram equalization is to further optimize the image quality and make the contrast of the image more prominent. That is, the grayscale histogram after the above processing reflects the probability of the occurrence of grayscale pixel values. Since the grayscale distribution of the infrared image is uneven, in order to change the grayscale level of the image, the histogram equalization process is required. Finally, the grayscale image after the histogram equalization process is subjected to threshold segmentation. Threshold segmentation is a method based on grayscale threshold. By classifying the grayscale of the pixel points, and then setting a threshold, the pixel points with a grayscale value less than the threshold are set to 0, and the remaining points are set to 1, thereby realizing the binarization of the image. Since the track lines in the binary image obtained by the automatic threshold method have some disconnected small gaps, which will affect the subsequent recognition processing results, this embodiment adopts a method of manually setting the threshold to obtain the binary image.

[0040] Execute step S3, process the binary image by edge detection algorithm, and obtain edge binary image. Edge detection algorithms include detection operators such as Sobel, Laplace, and Canny. The image edge processed by Sobel detection operator is strong or weak, and has good noise resistance. Laplace detection operator is more sensitive, so that some noise will enter the processed image. Canny detection operator has the advantages of optimal detection without additional response, minimum deviation between actual edge and detected edge position, and limiting the change and positioning of a single edge point. In this embodiment, Canny detection operator is used for edge detection. The specific algorithm process of Canny detection operator for edge point is to first remove noise from the image by filtering, then calculate the gradient in the x and y directions, then perform non-maximum suppression, and finally use high and low thresholds to extract edges. Through the edge detection algorithm, the image edge detection greatly reduces the amount of data, proposes edge-irrelevant information, and retains the important structural attributes of the image.

[0041] Execute step S4 to perform track contour screening on the edge binary image. After obtaining the edge binary image, in order to further reduce the amount of calculation and improve the accuracy of recognition, track contour screening can be performed before the subsequent Hough transform. Specifically, each contour of the edge binary image is extracted, and the contour parameters such as area, perimeter, and radian are calculated and compared with the predetermined track contour parameters. If the parameters of the predetermined basic track contour and turnout contour are not met, the contour is removed, thereby eliminating redundant images.

[0042] Execute step S5, perform Hough transform on the edge binary image after the track contour screening, and obtain the track line image. The purpose of performing Hough transform is to identify some lines with characteristic values, and at the same time further improve the continuity of the lines. In this embodiment, the PPHT Hough transform based on OPENCV is adopted. The PPHT Hough transform is the Hough transform of cumulative probability. It can not only identify and repair lines, but also indicate the coordinate positions of the pixel points of the endpoints of these lines, and can also eliminate some interference coordination, which greatly highlights the position of the track in the image. It is more efficient than the traditional Hough transform in processing some longer linear segmentations, and it is easier to detect straight lines. Since the turnout has obvious characteristics such as a large number of joints and complex curves compared with the track, and its track surface area, material, grayscale, etc. have similar characteristics to the track, so the track line image is obtained by PPHT Hough transform, not only the turnout will not be filtered as a redundant line, but also the turnout image in the track line image will be clearer, the track contour is more prominent in the whole background, and all lines are basically composed of single-pixel points, which is convenient for subsequent pixel-based turnout recognition.

[0043] Execute step S6 to fit the track line image. After obtaining the track line in the track line image detected above, in order to make the track line appear more intuitive, it is necessary to fit the track line. Since the infrared thermal imager collects track images in real time, the shape of the current track of the track cleaning vehicle on the infrared image is fixed, so the position of the current track can be identified. The first current rail and the second current rail respectively represent one of the two grooved rails opposite to each other in the track on which the track cleaning vehicle travels. According to the image position relationship of the lines in the track line image, the track lines represented by the first current rail and the second current rail can be distinguished. Since the track line on the structured track has a small curvature of the curve, in order to reduce the complexity of the fitting algorithm, the track line model is fitted using a parabolic model and the least squares method. The parabolic fitting model is y=Ax 2 +Bx+C, where A, B, and C are the three parameters of the parabola fitting model. The expression of the error square sum E is

[0044] Execute step S7 to perform Kalman filtering on the track line image after track line fitting. In the video formed by the infrared image, before the turnout appears, the position of the track line in two adjacent frames of the video will not change significantly, and this change occurs continuously, that is, the previous and next frames are interrelated. Therefore, the coordinates of the two track lines in the current frame can be predicted and tracked by the detection results of the previous frame, thereby improving the real-time and robustness of the detection. In this embodiment, the Kalman filter is mainly used to track the parameter values ​​A, B, C of the track line model and their changes ΔA, ΔB, ΔC.

[0045] State vector x = [A, B, C, ΔA, ΔB, ΔC] T ;

[0046] State transition matrix

[0047] The status is predicted by the following formula:

[0048]

[0049]

[0050] in, It is expressed as the estimated value of the state at the previous moment; F is the state transfer matrix; B is the control matrix, because there is no control quantity in the process of track line tracking and identification, so B = 0; is the covariance estimate at time t; Q t-1 is the noise covariance matrix of the prediction model itself.

[0051] The Kalman coefficient is calculated to measure the weight coefficient between the predicted quantity and the observed quantity as much as possible, and the Kalman coefficient is used to update the state quantity and covariance estimate. The update process is shown in the following formula:

[0052]

[0053]

[0054]

[0055] in, is the observation matrix at time t, which measures the transition between the state itself and the observed state. The observed state is the three parameters of the trajectory model; A, B, C, R t represents the observation noise covariance matrix at time t; K t is the Kalman coefficient.

[0056] Execute step S8 to determine whether the Kalman coefficient is greater than the first preset threshold. When a turnout appears, a new line other than the first current track and the second current track will appear in the track line image obtained by the PPHT Hough transform. At this time, after fitting and tracking, the track line will have an intuitive fitting result of transforming from two lines to multiple lines. The existence of a turnout can be determined by the morphological changes of the current track on the track line image, that is, whether the two current tracks are differentiated into multiple track lines. Specifically, when a turnout appears, the track line image obtained by the PPHT Hough transform shows that the track lines of the first current track and the second current track above are split into multiple lines including the turnout and the first current track and the second current track. At this time, the observed value suddenly changes, resulting in the Kalman coefficient used to represent the weight coefficient between the predicted value and the observed value being greater than the first preset value. At this time, it can be indicated that a turnout exists, and step S9 can be continued. If the Kalman coefficient does not meet the condition of being greater than the first preset value, it returns to step S6. In addition, when a turnout appears, during the process of track line fitting, due to the emergence of new lines that meet the characteristics of the track line, a new parabola fitting model will appear, and the number of fitted lines will increase. At this time, it can be simultaneously judged whether the number of parabolas fitted by the parabola model is greater than the second preset threshold. If so, it can also be used as one of the sufficient conditions for the existence of a turnout, that is, the Kalman coefficient is greater than the first preset value and the number of parabolas fitted by the parabola model is greater than the second preset threshold, which indicates that a turnout exists, and step S9 is continued to be executed, otherwise return to step S6.

[0057] Execute step S9 to determine the position of the turnout based on the received position offset information and the principle of triangulation. Specifically, when the existence of the turnout is determined, the infrared transmitter on the track cleaning vehicle emits an infrared beam toward the turnout at a preset angle and receives the infrared beam reflected back from the turnout. When the track cleaning vehicle continues to move forward to the turnout, the position offset information of the two points of the received infrared beam on the imaging surface can be used to calculate the displacement required to reach the turnout, thereby obtaining the position of the turnout. The track cleaning vehicle can then adjust the cleaning tools on the track cleaning vehicle according to the position of the turnout to prevent them from being damaged by hitting the turnout, thereby working in a safe environment.

[0058] In summary, the turnout identification method of the present invention obtains an infrared image of the track, performs preprocessing, edge detection, track contour screening, and Hough transformation on the infrared image in sequence, thereby obtaining a track line image, and then performs track line fitting and Kalman filtering on the track line image. Since the current track is fixed in position on the infrared image, the existence of the turnout can be determined based on the Kalman coefficient, and then the position of the turnout from the track cleaning vehicle is determined through position offset information and the principle of triangulation. The turnout identification system of the present invention implements the above turnout identification method, and the track cleaning vehicle of the present invention is equipped with the above turnout identification system. The present invention has the characteristics of simplicity, feasibility, small amount of calculation, and good reliability.

Claims

1. A turnout identification method, characterized in that: Acquire infrared images of the track; Preprocessing the infrared image to obtain a binary image; Processing the binary image by an edge detection algorithm to obtain an edge binary image; Performing PPHT Hough transform on the edge binary image after track contour screening to obtain a track line image; Performing orbit line fitting on the orbit line image by using a parabola model and a least square method; Performing Kalman filtering on the track line image after track line fitting; Determining whether the Kalman coefficient is greater than a first preset threshold, and determining whether the number of the parabolas fitted by the parabola model is greater than a second preset threshold, if the above determination results are both yes, determining that the turnout exists; After determining that the turnout exists, the position of the turnout is determined based on the received position offset information and the triangulation principle.

2. A turnout identification method as claimed in claim 1, characterized in that: When acquiring the infrared image of the track, the infrared image is acquired by a high-speed infrared thermal imager.

3. A turnout identification method as claimed in claim 1, characterized in that: When the infrared image is preprocessed to obtain the binary image, the following steps are performed: The infrared image is subjected to grayscale processing, median filtering, histogram equalization processing, and threshold segmentation processing in sequence to obtain the binary image.

4. A turnout identification method as claimed in claim 1, characterized in that: When the binary image is processed by an edge detection algorithm to obtain an edge binary image, the following steps are performed: The binary image is processed by using a Canny edge detection operator to obtain an edge binary image.

5. A turnout identification method as claimed in claim 1, characterized in that: After obtaining the edge binary image, before performing Hough transform on the edge binary image, the following steps are performed: Performing track line fitting on the track line image.

6. A turnout identification system, characterized in that: include: An infrared image acquisition module and a main control module, wherein the infrared image acquisition module is connected to the main control module; The infrared image acquisition module includes an infrared thermal imager and an infrared transmitter; the infrared thermal imager is used to acquire infrared images of the track and transmit the infrared images to the main control module; the infrared transmitter is used to send infrared beams to the turnout, receive reflected beams, and output imaging offset information to the main control module; The main control module is used for: Preprocessing the infrared image to obtain a binary image; Processing the binary image by an edge detection algorithm to obtain an edge binary image; Performing track contour screening on the edge binary image; Performing PPHT Hough transform on the edge binary image after the track contour screening to obtain a track line image; Performing orbit line fitting on the orbit line image by using a parabola model and a least square method; Performing Kalman filtering on the track line image after track line fitting; Determining whether the Kalman coefficient is greater than a first preset threshold, and determining whether the number of the parabolas fitted by the parabola model is greater than a second preset threshold, if the above determination results are both yes, determining that the turnout exists; After determining that the turnout exists, the position of the turnout is determined based on the received position offset information and the triangulation principle.

7. A track cleaning vehicle, characterized in that: include: Infrared image acquisition module, main control module, cleaning module, the infrared image acquisition module is connected to the main control module, The main control module is connected to the cleaning module; The infrared image acquisition module is arranged at the central position of the front cab of the track cleaning vehicle; The infrared image acquisition module includes an infrared thermal imager and an infrared transmitter; the infrared thermal imager is used to acquire infrared images of the track and transmit the infrared images to the main control module; the infrared transmitter is used to send infrared beams to the turnout, receive reflected beams, and output imaging offset information to the main control module; The main control module is used for: Preprocessing the infrared image to obtain a binary image; Processing the binary image by an edge detection algorithm to obtain an edge binary image; Performing track contour screening on the edge binary image; Performing PPHT Hough transform on the edge binary image after the track contour screening to obtain a track line image; Performing orbit line fitting on the orbit line image by using a parabola model and a least square method; Performing Kalman filtering on the track line image after track line fitting; Determining whether the Kalman coefficient is greater than a first preset threshold, and determining whether the number of the parabolas fitted by the parabola model is greater than a second preset threshold, if the above determination results are both yes, determining that the turnout exists; After determining that the turnout exists, determining the position of the turnout according to the received position offset information and the triangulation principle; Sending a control instruction to the cleaning module; The cleaning module is used to lift the cleaning tool according to the control instruction.

Images