Contact net deicing device and method thereof

Abstract

The invention relates to the technical field of measurement, in particular to an overhead line system deicing device and a method thereof.The overhead line system deicing device comprises a telescopic supporting column table of a whole device carrier, the telescopic supporting column table is provided with an equipment supporting table, the equipment supporting table comprises a guide rail installed on the telescopic supporting table, and hollow rotating tables are installed on the two sides of the guide rail; a deicing mechanism is arranged on the hollow rotating table, the deicing mechanism comprises a forward knocking rod and a reverse knocking rod which are installed on the hollow rotating table on the two sides, a deicing shovel and a vibration stick are further installed on the hollow rotating table, and the forward knocking rod and the reverse knocking rod transmit power through a first servo motor to enable the forward knocking rod and the reverse knocking rod to rotate at a high speed. The working angle and the walking speed can be autonomously changed according to operation target characteristics of different contact network target objects, different contact line icing working conditions of the same path can be adapted, and the practicability is greatly improved. And in the aspect of operation, only two parts of walking and identification are adopted, so that the operation is relatively intelligent and humanized.

Description

Technical Field

[0001] The invention relates to the field of measurement technology, in particular to a contact network deicing device and a method thereof. Background Art

[0002] As an important basic equipment for the rail transit line system to transmit electricity to electric vehicles, the contact network distribution device has a great impact on the transmission efficiency of rail vehicles and the safe operation of electrified railways. As the railway line extends, it will pass through areas with various climates. In areas with high altitudes and high humidity, icing is more likely to occur. When the contact network is covered with ice, the pantograph of the electric locomotive will not be able to draw current smoothly, and an arc will be generated between the pantograph and the network, forming harmonic currents, affecting the quality of the power grid and endangering power equipment. After the contact wire is covered with ice, wrapping the contact wire will hinder the pantograph from drawing current, causing arcing, increasing the wear of the contact wire and the pantograph, and causing greater damage to the contact network. This will affect the running speed of the train. In serious cases, the line needs to be blocked, which will cause great interference to driving. Summary of the invention

[0003] The purpose of the present invention is to provide a compensation mainline deicing device operation positioning method, so as to overcome the shortcomings of inaccurate deicing operation positioning, loopholes in supplementary deicing operations, and low efficiency in identifying overhead line distribution line equipment.

[0004] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0005] A new type of contact network deicing device, comprising: an equipment support platform, a slide rail support, a hollow rotating platform, a deicing operation mechanism, and a telescopic support platform;

[0006] The deicing device mechanism is provided with a knocking rod, a deicing shovel, a guide groove, a vibrating rod, a hot and cold air blade, a water-repellent spray head, a carbon slide plate, a coaxial commutator and a first servo motor;

[0007] The equipment support platform is placed in the center of the de-icing device structure, and its extension platform is equipped with a semicircular binocular camera. The intelligent control device is installed inside the equipment support platform. The equipment support is installed on the telescopic support platform, and the two platforms are positioned and connected to ensure the stability of the overall mechanism.

[0008] The deicing workbench is provided with a hollow rotating platform, a connecting platform, a fixed base, and a first servo motor. The fixed base is arranged on the top surface of the equipment support platform, a stepper motor is arranged inside the fixed base, and the hollow rotating platform is arranged at the middle connecting position between the fixed base and the deicing device mechanism. The hollow rotating platform and the second private servo motor are rotated by gear meshing, and the telescopic support platform is connected by the stepper motor.

[0009] The intelligent control device is used to receive the collected information sent by the binocular camera, analyze and process the collected information, and then send control commands to the first servo motor, the first servo motor and the stepper motor, so as to control the rotation of the first servo motor to drive the forward and reverse knocking rod to knock the ice on the contact line.

[0010] The binocular cameras are calibrated respectively, and the images collected by the calibrated binocular cameras are used to realize the three-dimensional reconstruction of the contact line of the overhead contact network. The non-contact deicing operation posture learning and estimation method based on binocular vision is characterized in that a wide-angle binocular camera placed on the equipment support table is used, and the left and right cameras respectively obtain the relative position of the contact line relative to the deicing mechanism.

[0011] Furthermore, the binocular cameras are calibrated respectively, and the images collected by the calibrated binocular cameras are used to realize the three-dimensional reconstruction of the transmission contact line, including the following steps:

[0012] (1) Contact network identification part: collect images of the contact line, make training sets, validation sets and test sets, then establish a convolutional neural network model, iteratively train the collected image data in the convolutional neural network model to obtain the contact line feature recognition network model; use a binocular camera with pre-calibrated parameters to collect images respectively to obtain a first image and a second image;

[0013] (2) De-icing mechanism operation part: Through the operation of the de-icing mechanism device, the de-icing mechanism uses a binocular camera to establish a scene model according to the terrain. Based on the established scene model, the de-icing mechanism is controlled by the slide rail guided movement and the hollow rotating platform, so as to autonomously plan the walking route.

[0014] (3) Path planning: The de-icing device operates according to the autonomously planned walking route and uses a binocular camera to capture the contact network target object. After capturing the target, it determines whether the contact device of the de-icing device is in contact with ice. After determining that the knocking rod is in contact with ice, the lateral offset distance of the ice from the center axis of the knocking rod and the vertical distance of the contact line from the center axis are calculated. The intelligent control system is used to control the telescopic support platform and the hollow rotating platform based on the distances of the two, and the telescopic device is raised to the height where the knocking rod and the carbon slide plate are located on the contact line. The hollow rotating platform keeps the de-icing mechanism in contact with the contact line through the first servo motor, and the knocking rod rotates to operate. The de-icing device starts to operate, and the ice covering the contact line is always kept at the center axis of the forward and reverse knocking rods. When the contact line is exposed as the de-icing mechanism operates, the de-icing device continues to move along the direction of the contact line.

[0015] (4) Dynamic tracking of de-icing mechanism: obtain real-time video stream, transmit the video image to the front-end intelligent control system module and perform corresponding area segmentation according to pixels, match the contact line ID and coordinate information, and add a feature sequence module to the back end of the contact line feature recognition network model to dynamically track the contact line.

[0016] Preferably, in the above technical solution, the method of step (1) of contact network identification comprises:

[0017] (1) The characteristic image of the contact network is collected by the predicted azimuth-upward-looking pan-tilt servo method, and the contact network model prediction is first controlled. This is implemented in the hierarchical control structure of the predicted azimuth-upward-looking binocular camera system control, and then the first servo motor is controlled to provide the joint angle setting point for the hollow rotating table provided by the de-icing mechanism.

[0018] (2) Use annotation tools to annotate the key point features of the contact part between the contact network and the de-icing mechanism and the center part of the contact line in the image, add different types of labels to different types of feature targets, add the same type of labels to the same type of targets and generate a label file. The label file contains the coordinates of the central pixel point of the marked feature position, the width and height of the marking box, and the category to which the marked feature belongs;

[0019] (3) Establish a convolutional neural network model, load the configuration file and set the network parameters, input the feature recognition point label file into the established convolutional neural network model for iterative training, and obtain the optimal weight file after a certain number of training rounds. After the training is completed, the contact network feature recognition network model is obtained;

[0020] (4) The intelligent control system operates in direct mode to control the up and down movement of the telescopic support platform and the tilt joint angle of the hollow rotating platform. The image of the target is acquired, and the visual system generates the set value. The visual dynamic azimuth-upward PTZ controller recognizes and compares each set point signal, and finally transmits the correct output model to the control terminal;

[0021] Preferably, in the above technical solution, step (2) the operation of the deicing mechanism includes the following steps:

[0022] (1) The image information collected by the binocular camera and the spatial coordinate axis of the target object are generated. Then, the distance of the contact line and the distance between the center axis of the forward and reverse steering device and the contact line are calculated through the pixel points and coordinate position information of the central axis of the image, the camera parameters of the binocular camera, and the position distance information of the contact network in the image.

[0023] (2) The binocular camera takes real-time photos of the icing condition of the contact line and transmits them to the intelligent control system to calculate the ice surface parameters. That is, the diameter is calculated and generated through the icing surface parameters. The icing image information obtained in each frame is fitted into a dynamic icing network model, and a series of diameter parameters are arranged in sequence and transmitted to the intelligent control system.

[0024] (3) The image information generated by the real-time contact network in (2) is further fed back to the control system, and the error between the ice-covered diameter network model and the set diameter value is compared, and the obtained diameter is transmitted to the control system for pre-processing, and the positive and negative value signals of the diameter network model parameters are outputted by post-processing and transmitted to the control system, and the telescopic support platform and the hollow rotating platform based on the error value are coordinated to realize the continuous and accurate operation of the de-icing mechanism and the contact network;

[0025] Preferably, in the above technical solution, step (3) path planning includes the following steps:

[0026] (1) The de-icing mechanism generates a pre-contact position in the three-dimensional coordinate space through deep neural network processing of the intelligent control system. The spatial posture of the pre-processed position is encoded into a superimposed multi-channel target map. The posture comparison and evaluation is performed between the posture position parameters provided by the three-dimensional reconstruction network model and the actual position to predict the accuracy of the contact position. The intelligent control system autonomously makes the most optimal path plan based on the parameter error of the network model.

[0027] (2) The intelligent control system reads the image transmission data collected by the binocular camera, identifies the target contact network and derives its position and posture;

[0028] (3) Convert the two-dimensional coordinates of the contact network in the binocular camera system into three-dimensional coordinates relative to the optical center of the binocular camera, and finally convert them into three-dimensional coordinates relative to the equipment support platform;

[0029] (4) the intelligent control system outputs a signal to enable the first servo motor to control the hollow rotating platform and the stepper motor to control the telescopic support platform to coordinately achieve the optimal contact position of the de-icing mechanism;

[0030] Preferably, in the above technical solution, step (4) dynamic tracking of the deicing mechanism comprises the following steps:

[0031] The forward and reverse knocking rods of the de-icing mechanism operate on the contact network along the contact line, perform feature recognition and real-time tracking on the structure of each part of the contact network, filter the image information according to the pixel position, and thus divide it into a multi-region image information parameter set, and simultaneously track the target in each region;

[0032] The binocular camera records each identified contact network, marks it in its detection frame to generate the ID of that time, and records the ID parameter information accordingly to match it with the set coordinate information and store it;

[0033] The feature sequence module information is matched and calculated at the output terminal of the network model obtained for contact network feature recognition. The feature sequence is used to store image information features in the past. If the contact line features are temporarily blocked during the movement of the de-icing mechanism, the time series will record and feedback the contact lines that are not recognized at present, and transmit the three-dimensional spatial coordinate information of the relative size of the obstruction to the intelligent control system for identification, so as to adjust the position of the subsequent de-icing mechanism to cross the obstacle.

[0034] By means of the above technical solution, the present invention provides a contact network deicing device and method thereof, which have at least the following beneficial effects:

[0035] 1. The contact network deicing device and method can accurately locate the position of the deicing operation through self-tracking technology, reduce the deicing omission phenomenon caused by inaccurate positioning, and thus improve the operation efficiency.

[0036] 2. The contact network deicing device and method and the collaborative operation mechanism can enable multiple devices to cooperate effectively, optimize the operation process, shorten the time required for deicing, and enhance the overall operation efficiency.

[0037] 3. The overhead contact network deicing device and method thereof can realize effective identification of overhead contact network distribution line equipment, thereby improving the maintenance and monitoring efficiency of the equipment and timely discovering potential problems.

[0038] 4. The contact network deicing device and method can reduce the impact on the surrounding environment through efficient deicing operations, and alleviate equipment damage and power supply interruption caused by ice and snow accumulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The accompanying drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application:

[0040] Figure 1 It is a three-dimensional combined schematic diagram of a new type of overhead contact network deicing device;

[0041] Figure 2 This is a schematic diagram of the internal structure of the deicing mechanism of the new type of overhead line deicing device;

[0042] Figure 3 Process of a new device for deicing overhead wire Figure 1 ;

[0043] Figure 4 Process of a new device for deicing overhead wire Figure 2 ;

[0044] Figure 5 Schematic diagram of intelligent control system for new type of overhead line deicing device;

[0045] Figure 6Principle of a new device for deicing overhead contact network Figure 1 ;

[0046] Figure 7 Principle of a new device for deicing overhead contact network Figure 2 ;

[0047] Figure 8 Schematic diagram of coordinate system transformation for the new type of contact network de-icing device.

[0048] Reference numerals:

[0049] 100. De-icing mechanism; 101. Positive striking rod; 102. Reverse striking rod; 103. De-icing shovel; 104. Vibrating stick; 105. Hot air blade; 106. Cold air blade; 107. Carbon slide plate; 108. Guide groove; 109. Water repellent nozzle; 110. Coaxial commutator; 111. First servo motor; 2. Hollow rotating table; 3. Equipment support table; 301. Binocular camera; 302. Guide rail; 4. Telescopic support table. DETAILED DESCRIPTION

[0050] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] A contact network deicing device and method thereof provided by some embodiments of the present invention will be described below in conjunction with the accompanying drawings.

[0052] Embodiment 1:

[0053] Combination Figure 1-Figure 2 As shown, the present invention provides a contact network deicing device and method thereof, the walking device includes an equipment support platform 3, a slide rail support, a hollow rotating platform, a deicing operation mechanism, a telescopic support platform 4, a first servo motor 111, a first servo motor 111, a stepping motor, a positive knocking rod 101, a reverse knocking rod 102, a deicing shovel 103, a guide groove 108, a vibrating rod 104, a hot air knife disc 105, a cold air knife disc 106, a hydrophobic agent nozzle 109, a carbon slide plate 107 and a coaxial commutator 110, and the self-tracking technology can be used to accurately locate the position of the deicing operation, reduce the deicing omission phenomenon caused by inaccurate positioning, and thus improve the operation efficiency;

[0054] The forward and reverse striking rod transmits power through the first servo motor 111 to make it rotate at high speed, the de-icing shovel 103 is fixed on the de-icing device working table and behind the position of the striking rod, the hot and cold air blade 106 is fixed on the de-icing device working table and behind the position of the de-icing shovel 103, and the carbon slide plate 107 is fixed on the de-icing device working table and behind the position of the cold air blade 106.

[0055] Specifically, the guide groove 108 is fixed on the deicing device operating table and behind the position of the carbon slide plate 107. It is provided with a rolling bearing to enable the deicing device to move accurately along the contact line path. The hydrophobic agent nozzle 109 is fixed on the deicing device operating table and behind the position of the guide groove 108.

[0056] Furthermore, the first servo motor 111 is installed inside the hollow rotating table 2, the slide rail is installed on the base of the de-icing device workbench to realize its translation along the guide rail 302, and the stepper motor is installed inside the telescopic support platform 4 to provide up and down reciprocating motion for the de-icing device; the detection and identification device is composed of a binocular camera and an intelligent control device, and the binocular camera 301 divides the video information into image information through high-speed real-time video recording, thereby extracting image information.

[0057] According to the embodiments, the collaborative operation mechanism can enable multiple devices to cooperate effectively, optimize the operation process, shorten the time required for deicing, and enhance the overall operation efficiency.

[0058] Embodiment 2:

[0059] Combination Figure 2-Figure 8 As shown, a method for deicing a contact network comprises the following steps:

[0060] 1. Target acquisition and recognition part

[0061] First, based on the binocular camera 301 of the equipment support platform 3, the image information of the contact network is collected, and the deep learning convolutional neural network is used to identify the target feature information, and the feature points of the contact network are also extracted. The binocular camera is used to shoot in real time and screen the target objects and perform annotation set analysis on the core key features of the contact network prediction; when switching between the two binocular cameras, the de-icing mechanism 100 intelligent module detection compares the error feature points of the objects that receive tracking authority. In order to improve the success rate of detection, we propose an adaptive area detection method, and define the detection ROI as the area where the de-icing mechanism 100 should be detected. Through feature point extraction, the core key features of the contact network are annotated, which can help subsequent data analysis and processing, identify specific problem areas, and improve the targeted maintenance work.

[0062] 2. 3D network model reconstruction

[0063] According to the first and second image information collected and the features predicted by deep learning, fusion detection is performed. The intelligent control system is used to receive the collection information sent by the binocular camera 301. The two binocular cameras 301 on the equipment support platform 3 use the same parameters. The two binocular cameras 301 are calibrated based on the spatial coordinate system. The vertical distance from the contact network coordinate parameter information to the plane on the track and the center distance of the contact network from the center line of the track are used to establish the conversion relationship between the binocular vision sensor coordinates and the measurement coordinate system; the upper rail surface is used as the xoy plane of the measurement coordinate system, and the normal vector direction is used as the positive direction of the z axis; the vertical rail direction is used as the x axis, and the far point is selected on the center line of the track; the conversion relationship from the spatial coordinate system of the binocular camera 301 to the measurement coordinate system is expressed by the translation coordinate system Y and the rotation coordinate system R:

[0064] Among them, C is the common coordinate point of the two binocular cameras 301, and the conversion relationship between the binocular vision sensor coordinate system and the measurement coordinate system is established. Therefore, based on this conversion relationship, the coordinate system simulates the measurement control point on the track, such as Figure 8 As shown, the plane and inner side of the track are covered with reflective marking points, and the de-icing mechanism 100 is controlled by the intelligent control system to calibrate the control points to establish the track center coordinate system. The reflective markings are used to obtain the coordinate parameter set as the auxiliary coordinates of the line path, and then the three-dimensional space coordinate system of the track center coordinates and the binocular camera setting coordinates are respectively obtained, so as to establish a coordinate conversion relationship between the two.

[0065] Based on the fusion detection of the first and second image information and the features predicted by the deep learning algorithm, this method can improve the comprehensive judgment ability of the contact network status. When dealing with complex environments, fusing information from different sources will help enhance the robustness and reliability of detection.

[0066] By calibrating the spatial coordinate systems of the two binocular cameras, the geometric relationship between the contact network and the track can be accurately measured. This process ensures the accuracy of the detection data in practical applications and provides a standardized way to process spatial data.

[0067] 3. Mechanical structure and intelligent control function realization

[0068] like Figure 2The de-icing working principle of the internal mechanism of the de-icing mechanism 100 is mainly the positive knocking plate 101 and the reverse knocking plate 102, and the forward and reverse functions of the two are realized through the same speed coaxial steering gear; the position spatial coordinate information of the contact line is established through the binocular camera 301 and the intelligent control to establish a three-dimensional space network model, and the accuracy of the position estimation described above is measured, and the mean square error of the position estimation between each binocular camera 301 and the contact network feature point is estimated. Each mean square error is randomly calculated from the binocular camera shooting path, and the calculated error is analyzed and processed to transmit the signal to the first servo motor 111, and the servo motor drives the meshing gear to drive the forward and reverse knocking rod to rotate to achieve de-icing.

[0069] like Figure 3 and Figure 4 As shown, the binocular camera updates the real-time image information parameters in combination with the actual trajectory of the contact network path and the coordinate changes of the track reference centerline. The corresponding feature point information starts to change iteratively as the de-icing mechanism 100 operates. The real-time feature point parameter information is extracted to achieve registration to obtain a three-dimensional point cloud. Then, the spatial coordinate information of the feature points of the three-dimensional point cloud and the relationship with the reference coordinates are used to perform matching calculations to obtain a reconstructed three-dimensional network image. Finally, the acquired feature points are verified with the feature point parameter information of the acquired image, and error analysis processing is used to fit a conforming feature curve.

[0070] The spatial coordinate information of the contact line is obtained through the binocular camera 301, and a three-dimensional spatial network model is established with the intelligent control system. This process can capture the position changes of the contact network in real time, providing an accurate data basis for subsequent work.

[0071] 4. Path planning and tracking of de-icing mechanism 100

[0072] like Figure 6 The figure shows the principle diagram of the de-icing device. First, the coordinate parameters of the binocular cameras 1 and 2 are calibrated, and the image information collected by the binocular cameras 301 are classified and summarized through the intelligent control system. The near-end binocular camera 301 in the forward direction of the de-icing device detects the front contact line along the positive direction of the contact network path, and the far-end binocular camera 301 in the forward direction of the de-icing device detects the rear contact line along the reverse direction of the contact network path. The errors between the feature points of the reconstructed model and the feature points of the image are calculated and fitted respectively, and the corresponding three-dimensional target object curve segments 1 and 2 are output. The errors of the two fitted curves are aligned based on the benchmark feature point fitting curves. Finally, the complete three-dimensional feature curve is output, and data processing is performed through the intelligent control system to classify and summarize the collected image information for subsequent error calculation and three-dimensional model reconstruction.

[0073] Combination Figure 7 and Figure 8As shown, the de-icing device detection function is switching between the two binocular cameras 301. As the de-icing device moves, according to the height difference of the track centerline and the lateral offset numerical deviation based on the contact line sag, the corresponding telescopic support platform 4 provides a height difference output signal to always keep the de-icing device in contact with the contact line. The hollow rotating platform 2 provides a contact line lateral offset and a de-icing device platform angle deviation output signal, and the slide rail provides a lateral translation output signal to reduce the lateral offset error. During the movement, it can sense and adjust the relative position with the contact line in real time to ensure the effectiveness and safety of the operation. This mechanism not only improves the de-icing efficiency, but also ensures the maintenance quality of the track line, demonstrating the application potential of modern intelligent equipment in complex environments.

[0074] The error parameter information of the vertical-lateral translation-tilt signal parameters is analyzed and processed by the intelligent control system, and also changes continuously with the real-time position coordinate value. An object tracking algorithm based on the set coordinate calibration points is used to robustly track the deicing structure device in the dynamic image. The mechanism principle is based on the registration of the feature point coordinate information between the three-dimensional reconstructed network model and the candidate target, and the contact network feature point object tracking is matched by the error value setting and the authority average weighted method is used to make it more accurately close to the parameter value provided by the feature information of the detected object. This intelligent control system makes the contact network position information of the vertical-lateral translation-tilt position change always focused on the deicing structure device. The corresponding first and second servo motors and stepper motors control the mechanical structure frame by frame, so that the calibration coordinate axis parameter value set by the deicing structure device is always within the positioning error set by the three-dimensional reconstructed network model, thereby realizing the self-tracking and collaborative operation function of the deicing device.

[0075] It should be noted that the terms "comprises", "includes" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or apparatus that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or apparatus.

[0076] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A contact network deicing device, comprising a telescopic support platform (4) of the entire device carrier, characterized in that: The telescopic support platform (4) is provided with an equipment support platform (3); The equipment support platform (3) comprises a guide rail (302) installed on the telescopic support platform, a hollow rotating platform (2) is installed on both sides of the guide rail (302), and a deicing mechanism (100) is provided on the hollow rotating platform (2); The deicing mechanism (100) comprises a positive striking rod (101) and a reverse striking rod (102) mounted on two sides of a hollow rotating platform (2); a deicing shovel (103) and a vibrating rod (104) are also mounted on the hollow rotating platform (2); the positive striking rod (101) and the reverse striking rod (102) are driven by a first servo motor (111) to rotate at a high speed.

2. The contact network deicing device according to claim 1, characterized in that: A hot air knife disc (105) and a cold air knife disc (106) are installed in the middle of the hollow rotating table (2).

3. The contact network deicing device according to claim 2, characterized in that: The hollow rotating platform (2) is provided with a carbon slide plate (107), a guide groove (108) is provided on the hollow rotating platform (2), and a water repellent spray head (109) and a coaxial commutator (110) are also provided on the hollow rotating platform (2).

4. The contact network deicing device according to claim 3, characterized in that: The hot air blade (105) and the cold air blade (106) are fixed on the hollow rotating table (2) behind the position of the de-icing shovel (103), and the carbon slide plate (107) is fixed on the hollow rotating table (2) behind the position of the hot air blade (105) and the cold air blade (106).

5. The contact network deicing device according to claim 4, characterized in that: The guide groove (108) is fixed on the hollow rotating table (2) behind the position of the carbon slide plate (107), and is provided with a rolling bearing so that the deicing device can be precisely directed along the contact line path. The water repellent nozzle (109) is fixed on the hollow rotating table (2) behind the position of the guide groove (108).

6. The contact network deicing device according to claim 3, characterized in that: The first servo motor (111) is installed inside the hollow rotating platform (2), and the slide rail is installed on the telescopic support platform (4) to achieve translation along the direction of the guide rail (302).

7. The contact network deicing device according to claim 1, characterized in that: The slide rail is installed on the equipment support platform (3) to achieve translation along the guide rail (302), and the stepper motor is installed inside the telescopic support platform (4) to provide up and down reciprocating motion for the equipment support platform (3).

8. A method for deicing a contact network according to any one of claims 1 to 7, characterized in that: S1, placing the whole device on the track and setting the reference coordinates based on the center line of the track; S2, the binocular camera collects the contact network image information along the contact line path direction and compares the line sag and line lateral deviation with the set reference coordinate parameter standard value. When the height difference and lateral deviation of the collected target feature information exceed the standard value we set, the intelligent control system will analyze and process the error value and adjust the position of the running device relative to the contact network to make it within the error range; S3. If any working part of the de-icing device does not touch the contact line when the de-icing device is in operation, and it is detected that the contact network coordinate information deviates from the error set by the intelligent control, the operation target will be deviated from the set coordinate range accordingly, and the first servo motor and the stepper motor will start to work, so that the overall device calculates the inclination value and the height difference value based on the deviation error value starting angle and vertical direction respectively, and analyzes and processes the output signal to the first servo motor and the stepper motor through the intelligent control system; S4, the binocular camera (301) captures and processes the contact network parameter information in real time, extracts feature points from the image information and reconstructs the two-dimensional coordinate points through a convolutional neural network method, and calculates and draws a feature error curve on the reconstructed coordinate set; S5. Based on the feature point coordinate information and the reference coordinate information, the feature point coordinate parameters that meet the set error range are obtained by alignment, and a three-dimensional network model is reconstructed to enable the deicing device to continue to operate within the controllable error value range along the target object.

Images