Train intelligent identification method and device

Abstract

The present application relates to a kind of train intelligent identification method and device, the method comprises: S1, magnetic steel induction module senses train wheel magnetic field variation, identifies train wheel pair data;S2, RFID tag reader reads the RFID tag code of each train compartment of train;S3, RFID tag code and its reading order are matched with database, determine train type, marshalling and positive and negative information in combination with wheel pair data;S4, auxiliary calibration camera identifies scale calibration, establishes physical coordinate reference;S5, by the laser radar device of being set in the position capable of scanning train head scans train head contour, obtains train head contour point cloud data;Meanwhile, by the camera of being set in the position capable of scanning train head shoots train head image;S6, by control system, train head contour point cloud data and train head image are fused and handled, correct positioning deviation in combination with physical coordinate reference, calculate train parking position.Compared with prior art, the present application has the advantages of high efficiency, identification rate precision etc..

Description

Technical Field

[0001] This invention relates to the field of rail transit, and in particular to a train intelligent identification method and device based on RFID lidar and image recognition technology. Background Technology

[0002] With the deepening integration of four networks (high-speed rail, intercity rail, suburban rail, and urban rail transit), the scenario of intercity rail trains and subway trains running together is becoming increasingly common. However, many intercity rail trains are not equipped with vehicle-to-ground communication systems. When these trains enter urban rail transit platforms, the platform screen door system cannot obtain train information (train type, formation, forward / reverse formation, stopping position, etc.) through the traditional "vehicle-signal-platform screen door" linkage method. How to accurately align the platform screen doors with the train doors and achieve automatic opening has become a technical problem that urgently needs to be solved. Rail transit train identification technology is fundamental for achieving automatic platform door control and train operation monitoring. Existing technologies include lidar-based vehicle model identification schemes, which obtain model parameters by scanning vehicle outlines, but rely on data from a single sensor and suffer from insufficient accuracy when trains pass at high speeds; image recognition-based schemes identify car numbers by capturing images of trains, but their stability is poor due to lighting conditions and shooting angles; and electronic tag-based section positioning schemes locate trains by reading onboard tags, but cannot obtain train formation and forward / reverse formation information. Furthermore, existing solutions generally lack multi-source data cross-validation mechanisms, leading to a significant decrease in system reliability when a sensor malfunctions or displays abnormal data. Additionally, existing solutions are insufficiently adaptable to new train models, requiring tedious manual parameter adjustments and calibrations, which cannot meet the practical needs of frequent train model updates.

[0003] Therefore, how to realize a train intelligent identification scheme that integrates multi-source information, has data cross-validation capabilities, and automatic calibration functions has become a technical problem that needs to be solved. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the existing technology and provide a train intelligent identification method and device that is highly efficient, reliable, and accurate in recognition.

[0005] The objective of this invention can be achieved through the following technical solutions: According to a first aspect of the present invention, a train intelligent identification method is provided, the method comprising the following steps: Step S1: The magnetic steel induction module set on the train's entry path senses the change in the magnetic field of the train wheels, triggers the start of the control system, and identifies the train wheelset data. Step S2: After the magnetic steel sensing module is triggered, the RFID tag code of each carriage of the train is read by the RFID tag reader set on the train entry path behind the magnetic steel sensing module. Step S3: Match the RFID tag encoding and its reading order with the database, and combine the wheelset data to determine the train type, formation, and forward / reverse formation information; Step S4: Establish a physical coordinate reference by identifying the scale markings using an auxiliary calibration camera set up in the platform area; Step S5: After the train comes to a complete stop, the outline of the train's front is scanned by a lidar device positioned at the location capable of scanning the front of the train, and point cloud data of the front outline is obtained; at the same time, an image of the train's front is captured by a camera positioned at the location capable of scanning the front of the train. Step S6: The control system fuses the point cloud data of the train head contour and the train head image, corrects the positioning deviation by combining the physical coordinate reference, and calculates the train stopping position.

[0006] As a preferred technical solution, in step S1, a threshold range for the number of wheelsets is set. When the number of wheelsets identified by the magnetic steel sensing module exceeds the threshold range, it is determined to be abnormal and an alarm is triggered.

[0007] As a preferred technical solution, in step S3, when the grouping information of the wheelset data and the RFID tag data is inconsistent, an alarm signal is triggered.

[0008] As a preferred technical solution, in step S3, the train's forward and reverse configuration is determined based on the reading order of the RFID tag codes and the correspondence between the tag codes and carriages in the database.

[0009] As a preferred technical solution, in step S4, the auxiliary calibration camera establishes a physical coordinate reference by recognizing the scale markings, which is used to correct the positioning deviation of the lidar device. When a new vehicle model is detected to be entering the station for the first time, the calibration mode is automatically activated. The vehicle model deviation data and error correction parameters are calculated using the scale and the auxiliary calibration camera, and the vehicle model deviation data and the error correction parameters are stored in the database.

[0010] As a preferred technical solution, in step S6, the point cloud data of the vehicle front contour scanned by the lidar device is matched with the vehicle front image captured by the camera to cross-validate the vehicle model data.

[0011] As a preferred technical solution, in step S6, a weighted fusion algorithm is used to perform phased weighted fusion of the wheelset data, the RFID tag code, the vehicle front contour point cloud data and the vehicle front image, and output the recognition result with the highest comprehensive confidence. When the deviation of a certain sensor data from other sensor data exceeds a preset threshold, the abnormal data is automatically removed, and the remaining sensor data is used for fusion calculation.

[0012] According to a second aspect of the present invention, a train intelligent identification device is provided, comprising: The magnetic steel induction module is installed on the train's entry path to the station to sense changes in the magnetic field of the train wheels, trigger the start of the control system, and identify the train wheelset data. An RFID tag reader is installed on the train's entry path behind the magnetic steel sensing module to read the RFID tag codes of each carriage of the train. The lidar device is set up at a position that can scan the front of the train to scan the outline of the train front and obtain point cloud data of the front outline. A high-resolution camera is positioned to scan the front of the train and capture images of the train's front. A scale and an auxiliary calibration camera are installed at the train window position in the platform area. The scale is used to provide a physical coordinate reference, and the auxiliary calibration camera is used to identify the scale graduations. The control system is connected to the magnetic steel sensing module, the RFID tag reader, the lidar device, the high-resolution camera, and the auxiliary calibration camera, respectively. The control system includes a processor, a memory, and a communication interface. The processor is used to receive data from each module and execute data fusion algorithms. The memory is used to store the database and calculation results. The communication interface is used to communicate with each module.

[0013] As a preferred technical solution, the magnetic steel induction module is installed next to the rail at a distance of 10-50 meters from the turnout, and the installation height is 5-20cm above the rail surface.

[0014] As a preferred technical solution, the RFID tag reader is set in the center of the rail 5-20 meters away from the magnetic steel sensing module, with the antenna facing the direction of the train entering the station.

[0015] As a preferred technical solution, the lidar device and the high-resolution camera are positioned at a height of 3-8 meters directly above the train's stop position, and the scanning angle covers the outline of the train's front.

[0016] As a preferred technical solution, the scale is a vertically set scale with a scale interval of 1-5cm. The lens of the auxiliary calibration camera is aligned with the scale, and the scale and the auxiliary calibration camera are fixed to the train window position in the platform area by chemical anchors.

[0017] According to a third aspect of the present invention, an electronic device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the program to implement the method described thereon.

[0018] According to a fourth aspect of the present invention, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the method described thereon.

[0019] Compared with the prior art, the present invention has the following advantages: (1) This invention is a train intelligent identification device based on advanced technologies such as RFID radio frequency identification, lidar, high-definition camera and induction magnet. For stations that have not built a vehicle-signal-platform door linkage system, this is a train identification solution with high efficiency, high reliability and high accuracy. It can reduce the workload of station staff, reduce the difficulty of manual observation, reduce the risk of human identification errors, solve the problem that manual verification of train parking positions is not possible on site, realize the historical problem of automatic identification of various information of stopped trains at the station, and provide a good solution for similar scenarios in the platform door industry and other professional trackside equipment that need to identify vehicle information. It is of great significance to the intelligentization and automation of the rail transit field. (2) By adopting a phased triggering mechanism of magnetic steel induction, RFID tag reading, laser radar scanning and camera image acquisition, as well as a multi-source cross-verification mechanism of magnetic steel wheel pair data and RFID tag data, laser radar contour data and camera image data, this invention overcomes the defects of existing technologies that rely on a single identification technology and have insufficient identification accuracy, and achieves a train model identification accuracy of over 99% and a parking position positioning accuracy of within ±5 centimeters. (3) By setting a ruler and an auxiliary calibration camera at the position of the train window, the present invention establishes a physical coordinate reference and realizes the automatic calibration function when the new model enters the station for the first time. No manual and tedious debugging is required, and the calibration time is shortened to less than five minutes, which significantly improves the system's adaptability to the new model. (4) By adopting a weighted fusion algorithm and an abnormal data removal mechanism, when a sensor fails or data is abnormal, the system automatically removes the abnormal data and uses the data from the remaining sensors for fusion calculation, ensuring that the system can still work normally when a single sensor fails, and the system availability reaches more than 96%, which significantly improves the system reliability. (5) In the scenario where there is no vehicle-to-ground communication at the station, the present invention uses RFID radio frequency identification technology to read the electronic tags of each carriage of the train and track magnets to identify the number of wheelsets that the train passes. This enables the accurate identification of train type, formation, forward and reverse formation and other data even when the train is traveling at high speed, providing data support for the stable operation of the platform door system or other trackside equipment that requires vehicle data. (6) This invention utilizes the characteristic of lidar to scan and identify fan-shaped areas, which can scan the outline of the train head, and can be used to check data such as train type and train formation, and accurately locate and identify the parking position of the vehicle, providing data support for the stable operation of the platform door system or other trackside equipment that requires vehicle data. (7) This invention uses a high-resolution camera to take pictures of parked vehicles. Through image acquisition technology and intelligent algorithms, it can accurately identify data such as train type, parking position, and formation. It can also be used as a supporting device for lidar to realize the identification of vehicle information by two sets of equipment, double verification, and ensure the stable and reliable operation of the device. (8) This invention can be adapted to scenarios where trains are passing through stations at high speed or stopping at stations, and scenarios that require information data such as train type, formation, forward and reverse formation, and stopping position. Attached Figure Description

[0020] Figure 1 This is a flowchart illustrating the specific process of the method of the present invention; Figure 2 This is a schematic diagram of the structure of the device of the present invention; Figure 3 This is a schematic diagram of the planar arrangement of the lidar and camera of the present invention; Figure 4 This is a schematic diagram of the cross-sectional arrangement of the lidar and camera of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0022] This embodiment provides a complete intelligent train identification method. By using RFID radio frequency identification technology to read the tag identification code data of the vehicles entering the depot, the method automatically identifies the train type information of the station. In the track area, the method collects the wheel set data of the vehicles to identify the train formation information. In the stopping area, the method uses lidar detection and image recognition technology to accurately locate the train stopping position information. This invention can provide guidance for the platform door system to open the doors directly facing the train, and can also be used for other trackside equipment that requires accurate information such as train type, formation, and stopping position.

[0023] like Figure 1 As shown, the present invention provides a complete intelligent train identification method, which specifically includes the following steps: Step S1, Magnet induction trigger: When the train passes the turnout and determines the platform where it will stop, the magnet induction module 101 senses the change in the magnetic field of the train wheels, triggers an electrical signal, and starts the control system 106 and its functional modules. The magnet induction module 101 identifies the number of wheelsets that the train has passed and reports the wheelset data information to the control system 106.

[0024] Specifically, the magnetic steel induction module 101 can use a Hall sensor. When a train wheel passes by, the change in the magnetic field of the wheel is detected by the Hall sensor, generating a pulse signal. The control system 106 counts the number of pulse signals, with each two pulses corresponding to one wheelset. For example, for an 8-car train, there are 32 wheelsets, and the magnetic steel induction module 101 identifies 32 pulse signals.

[0025] Step S2, RFID tag reading: When the magnetic steel sensing module 101 is triggered, the RFID tag reader 102 starts working and reads the RFID tag code of each carriage of the passing train; the RFID tag reader 102 reads the tag codes of each carriage in sequence according to the direction of train travel, for example, reading the tag codes T001, T002, T003, T004, T005, T006, T007, T008 in sequence.

[0026] Step S3, Database Matching: The control system 106 matches the RFID tag codes and their reading order with the database, and combines the wheelset data identified by the magnetic steel sensing module 101 to determine the train type, formation, and forward / reverse formation information; for example, according to the reading order of tag codes T001 to T008, the system matches the train type information in the database to determine that the train is an 8-car formation and a forward formation.

[0027] Specifically, the database of the control system 106 pre-stores the label codes of each vehicle type and their corresponding carriages. For example, the label code sequence for an 8-car train is T001, T002, T003, T004, T005, T006, T007, T008. If the reading sequence matches the database, it is determined to be a forward train formation; if the reading sequence is T008 to T001, it is determined to be a reverse train formation.

[0028] The control system 106 also compares the wheelset data identified by the magnet with the grouping information of the RFID tag data to verify the consistency of the train grouping. For example, an 8-car group corresponds to 32 wheelsets. If the magnet sensing module 101 identifies 32 wheelsets and the RFID tag reader 102 reads 8 tag codes, the verification is successful. If the wheelset data and the tag data are inconsistent, an alarm signal is triggered.

[0029] Step S4, Coordinate reference establishment: A physical coordinate reference is established by identifying the scale markings using an auxiliary calibration camera set in the platform area; the scale is a vertically set scale with a scale interval of 1 to 5 centimeters. The lens of the auxiliary calibration camera is aligned with the scale, and a physical coordinate system is established by identifying the scale markings, which is used to correct the positioning deviation of the lidar device 103.

[0030] Step S5, LiDAR scanning: After the train enters the station and comes to a complete stop, the LiDAR device 103 starts the scanning function to scan the outline of the train's front and obtain the point cloud data of the front outline; the LiDAR device 103 can be a two-dimensional LiDAR with a scanning angle of 180 degrees, a scanning frequency of 10 Hz, a scanning distance of 0 to 30 meters, and a point cloud data resolution of 0.25 degrees.

[0031] Specifically, the lidar device 103 scans the outline of the train's front to acquire three-dimensional geometric data of the front. For example, for a train, the point cloud data of the front outline includes parameters such as the height, width, and tilt angle of the front. The control system 106 matches the point cloud data with the vehicle model outline data in the database to verify the vehicle model data.

[0032] Step S6, Camera Image Acquisition: Simultaneously, the high-resolution camera 104 captures an image of the train's front end; the high-resolution camera 104 is an industrial camera with a resolution of 2048×1536 pixels, a frame rate of 30 frames per second, and a lens focal length of 8 mm; the image of the train's front end captured by the camera includes information such as the appearance features of the train's front end, the train number, and the position of the windows.

[0033] Step S7, Data Fusion Calculation: The processor of the control system 106 fuses the point cloud data of the train's front profile scanned by the lidar device 103 and the image of the train's front captured by the high-resolution camera 104. It then combines this data with the physical coordinate reference established by the auxiliary calibration camera 105 to correct positioning deviations and calculate the train's stopping position. Specifically, the control system 106 employs a weighted fusion algorithm to perform feature matching on the point cloud data and image data, outputting the recognition result with the highest overall confidence level.

[0034] The specific steps of the weighted fusion algorithm are as follows: First, extract the vehicle front contour feature points from the point cloud data, such as the apex of the vehicle front, the bottom edge points of the vehicle front, and the side edge points of the vehicle front; second, extract the vehicle front appearance features from the image data, such as the color of the vehicle front, the position of the windows, and the vehicle number; then, match the point cloud feature points with the image features and calculate the matching degree; finally, assign different weights according to the matching degree and fuse them to calculate the parking position.

[0035] For example, the weight of point cloud data is 0.6, and the weight of image data is 0.4. If the parking position calculated from the point cloud data is 3 cm in front of the parking marker, and the parking position calculated from the image data is 5 cm in front of the parking marker, then the fused parking position is 3.8 cm in front of the parking marker. The control system 106 outputs parking position data for use by the platform screen door system or other trackside equipment.

[0036] Through the above steps, this embodiment realizes a complete identification process for trains from entering the station to coming to a complete stop. By employing a phased triggering mechanism involving magnetic induction, RFID tag reading, LiDAR scanning, and camera image acquisition, data from each stage is cross-verified, solving the problem of insufficient accuracy of single-technology identification. Magnetic wheelset data and RFID tag data are cross-verified to verify train model grouping; LiDAR contour data and camera image data are cross-verified to verify the stopping position; and scale calibration data corrects LiDAR positioning deviations.

[0037] As a further preferred embodiment, the present invention achieves automatic calibration and error correction when a new vehicle model enters the station for the first time, eliminating the need for tedious manual debugging, thereby improving the system's adaptability to new vehicle models and positioning accuracy. The specific implementation process is as follows: In step T1, the control system 106 detects the arrival of a new vehicle model at the station. When the tag code read by the RFID tag reader 102 has no matching record in the database, the control system 106 determines that the new vehicle model is entering the station for the first time and automatically starts the calibration mode. For example, if the RFID tag reader 102 reads tag codes N001 to N008, and there is no record of such tag codes in the database, the control system 106 starts the calibration mode.

[0038] Step T2: The auxiliary calibration camera 105 identifies the scale markings and establishes a physical coordinate reference. The auxiliary calibration camera 105 captures images of the scale and uses an image recognition algorithm to identify the scale markings and establish a physical coordinate system. For example, if the camera identifies scale markings from 0 cm to 200 cm, the origin of the physical coordinate system is set at the 0 cm position on the scale, the vertical axis represents the direction of the scale markings, and the unit is centimeters.

[0039] In step T3, the lidar device 103 scans the outline of the vehicle's front end to acquire point cloud data of the front end; the lidar device 103 also scans the front end outline to acquire three-dimensional geometric data of the front end. For example, the scan shows that the height of the front end is 3.8 meters, the width is 3.2 meters, and the tilt angle is 15 degrees. The control system 106 converts the point cloud data into coordinate values ​​in the physical coordinate system.

[0040] Step T4: Calculate the vehicle model deviation data and error correction parameters. The control system 106 calculates the deviation data and error correction parameters for the vehicle model by comparing the correspondence between the actual position of the vehicle's front end and the scale coordinates. For example, if the coordinates of the vehicle's front apex scanned by the lidar device 103 in the physical coordinate system are 180 cm, and the actual position of the vehicle's front apex is 182 cm on the scale, then the deviation data is 2 cm. The control system 106 calculates the error correction parameter to be ±2 cm.

[0041] Step T5: Store the vehicle model deviation data and error correction parameters to the database; the control system 106 stores the vehicle model's tag code, vehicle name, deviation data, and error correction parameters to the database. For example, it stores tag codes N001 to N008, vehicle name CR500AF, deviation data of 2 cm, and error correction parameter of ±2 cm. Subsequently, when this vehicle model enters the station, the lidar device 103 automatically corrects the positioning deviation based on the stored correction parameters.

[0042] Step T6: Calibration complete, exit calibration mode; control system 106 outputs calibration complete signal, exits calibration mode, and resumes normal recognition mode; calibration time is less than 5 minutes. For example, from the arrival of a new vehicle model to calibration completion, the total time is 4 minutes and 30 seconds. Subsequently, when the same vehicle model enters the station, control system 106 automatically corrects the lidar positioning deviation based on the calibration parameters in the database, improving the positioning accuracy to within ±3 cm.

[0043] Through the aforementioned automatic calibration function, this embodiment achieves automatic calibration and error correction when a new vehicle model enters the station for the first time, eliminating the need for tedious manual adjustments. Calibration time is reduced to less than 5 minutes, significantly improving the system's adaptability to new vehicle models. Subsequently, when the same vehicle model enters the station, the lidar automatically corrects the positioning deviation based on the calibration parameters, improving positioning accuracy to within ±3 centimeters, further enhancing the accuracy of parking position measurement.

[0044] Actual test data shows that the train model recognition accuracy of this embodiment reaches 99.7%, the train formation recognition accuracy reaches 99.3%, and the parking position positioning accuracy reaches ±3.2 cm. Compared with the prior art, the recognition accuracy is improved by 3.4%, and the positioning accuracy is improved by 62%, significantly improving the system's recognition accuracy and reliability.

[0045] The above is an introduction to the method embodiments. The following describes the solution of the present invention further through device embodiments.

[0046] like Figure 2-4 As shown, the train intelligent identification device in this embodiment includes a magnetic steel sensing module 101, an RFID tag reader 102, a lidar device 103, a high-resolution camera 104, an auxiliary calibration camera 105, and a control system 106. The control system 106 is connected to each functional module through a signal connection line, receives data from each module, executes a data fusion algorithm, and outputs train identification results and parking position data.

[0047] The magnetic steel induction module 101 is installed next to the rail after the station turnout, 10 to 50 meters away from the turnout, and the installation height is 5 to 20 centimeters above the rail surface; the RFID tag reader 102 is installed in the center of the rail 5 to 20 meters behind the magnetic steel induction module 101, with the antenna facing the direction of the train entering the station.

[0048] The lidar device 103 and the high-resolution camera 104 are positioned 3 to 8 meters above the train stop position, with a scanning angle covering the outline of the train's front. The auxiliary calibration camera 105 is positioned at the train window in the platform area and is fixed to the platform by chemical anchors.

[0049] The specific implementation process of the device in this embodiment is as follows: After the train enters the station, the driver will stop the train by aligning the driver's side window with the station's ground parking markings. The train's intelligent identification device should accurately identify the train's model, formation, forward / reverse formation, and precise stopping position relative to the parking markings, and transmit the relevant information to the platform screen door system. Based on the received vehicle information, the platform screen door system calculates the accurate coordinates of all doors in the station, enabling the sliding doors to open directly opposite the train doors.

[0050] The magnetic steel induction module is installed after the station turnout, next to the rails of the effective waiting platform. When the train passes the turnout and the platform where the train is about to stop is determined, the magnetic steel induction module senses the change in the magnetic field of the wheels, triggers an electrical signal, starts the control system, and starts detection of each functional module. The magnetic steel induction module can report the identified train wheelset data information to the control system of the device. The RFID tag reader is located in the center of the rail after the magnetic steel induction module. After the device is triggered by the magnetic steel signal, the tag reader starts working and reads the tag code of each carriage of the passing train. Through RFID radio frequency identification technology, it identifies the tags on each carriage, matches the read tag data and sequence with the database, and combines the wheelset data information identified by the magnetic steel to determine the accurate information such as the train type, formation, and forward / reverse formation of the train that is about to stop.

[0051] A lidar device and a high-resolution camera are installed at the train's stopping position. When the train comes to a complete stop at the station, the lidar device above the train starts scanning. Based on the train's front profile scanned by the radar, the train model data is verified. The camera takes an image. Through the radar cloud point map and the camera image correction data, the device automatically calculates the exact stopping position of the train.

[0052] A ruler and auxiliary calibration camera need to be installed at the train window position. The ruler can provide a physical coordinate reference for determining the train's stopping position. The ruler and auxiliary calibration camera are fixed to the platform area by chemical anchors. When the train enters the station for the first time, the ruler and calibration camera can automatically calculate the train type deviation data and error correction, and synchronize the results to the lidar device to provide the control system with a calibration value. After subsequent trains enter the station, the radar will automatically calculate the stopping deviation value of the train relative to the stopping mark based on the system calibration value, and provide accurate stopping data information to the platform screen door system or similar trackside professional equipment that requires train information.

[0053] The main advantages of this embodiment include: 1) An innovative intelligent train identification device based on advanced technologies such as RFID radio frequency identification, LiDAR, high-definition cameras, and inductive magnets is proposed. For stations without a vehicle-signal-platform door linkage system, this is a highly efficient, reliable, and accurate train identification solution. It can reduce the workload of station staff, reduce the difficulty of manual observation, reduce the risk of human error, and solve the problem of the inability to quickly verify the train's stopping position on-site. It solves the historical problem of automatically identifying various information of stopped trains at the station. It provides a good solution for similar scenarios in the platform door industry and other professional trackside equipment that need to identify vehicle information. It is of great significance to the intelligentization and automation of the rail transit field.

[0054] 3) In scenarios where there is no vehicle-to-ground communication at the station, RFID radio frequency identification technology is used to read the electronic tags of each carriage of the train, and track magnets identify the number of wheelsets that the train passes. This enables accurate identification of train type, formation, forward and reverse formation, and other data even when the train is traveling at high speed, providing data support for the stable operation of platform screen door systems or other trackside equipment that requires vehicle data.

[0055] 3) Utilizing the characteristic of LiDAR to scan and identify fan-shaped areas, it can scan the outline of the train head, which can be used to verify data such as train type and train formation, and accurately locate and identify the parking position of the vehicle, providing data support for the stable operation of platform screen door systems or other trackside equipment that require vehicle data.

[0056] 4) High-resolution cameras can be used to photograph parked vehicles. Through image acquisition technology and intelligent algorithms, data such as train type, parking position, and formation can be accurately identified. It can also be used as a supporting device for lidar to enable two sets of equipment to identify vehicle information, perform dual verification, and ensure the stable and reliable operation of the device.

[0057] 5) Due to the continuous operation optimization and adjustment of the domestic intercity railway and high-speed rail network, and the replacement of new train models, the train models and formations that may stop at each station are constantly being updated. With the assistance of RFID tag readers, LiDAR, calibration cameras and rulers, when a train enters the station for the first time, the device can autonomously complete the functions of train model scanning and recognition, train outline data collection, parking calibration benchmark positioning and formation recognition, etc., without the need for cumbersome manual debugging.

[0058] 6) Other trackside equipment application scenarios, whether it is a high-speed train passing through a station or a train stopping at a station, can be adapted to scenarios that require information data such as train type, formation, forward and reverse formation, and stopping position.

[0059] 7) Based on the performance characteristics of RFID radio frequency identification technology, LiDAR scanning technology, and high-definition camera image recognition technology, the above functions can be flexibly combined using one, two, or more of these technologies to still achieve the target function or part of the function of the device. This invention includes, but is not limited to, the above combination schemes, all of which are within the scope of this invention.

[0060] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the described module can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0061] This invention also provides an electronic device including a central processing unit (CPU), which can perform various appropriate actions and processes according to computer program instructions stored in a read-only memory (ROM) or loaded from a storage unit into a random access memory (RAM). The RAM may also store various programs and data required for device operation. The CPU, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0062] Multiple components in the device are connected to the I / O interface, including: input units such as keyboards and mice; output units such as various types of displays and speakers; storage units such as disks and optical discs; and communication units such as network interface cards (NICs), modems, and wireless transceivers. The communication unit allows the device to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0063] The processing unit executes the various methods and processes described above, such as methods S1 to S7. For example, in some embodiments, methods S1 to S7 may be implemented as computer software programs tangibly contained in a machine-readable medium, such as a storage unit. In some embodiments, part or all of the computer program may be loaded and / or installed on the device via ROM and / or a communication unit. When the computer program is loaded into RAM and executed by the CPU, one or more steps of methods S1 to S7 described above may be performed. Alternatively, in other embodiments, the CPU may be configured to execute methods S1 to S7 by any other suitable means (e.g., by means of firmware).

[0064] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0065] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0066] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0067] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A train intelligent identification method, characterized in that, The method includes the following steps: Step S1: The magnetic steel induction module set on the train's entry path senses the change in the magnetic field of the train wheels, triggers the start of the control system, and identifies the train wheelset data. Step S2: After the magnetic steel sensing module is triggered, the RFID tag code of each carriage of the train is read by the RFID tag reader set on the train entry path behind the magnetic steel sensing module. Step S3: Match the RFID tag encoding and its reading order with the database, and combine the wheelset data to determine the train type, formation, and forward / reverse formation information; Step S4: Establish a physical coordinate reference by identifying the scale markings using an auxiliary calibration camera set up in the platform area; Step S5: After the train comes to a complete stop, the outline of the train's front is scanned by a lidar device positioned at the location capable of scanning the front of the train, and point cloud data of the front outline is obtained; at the same time, an image of the train's front is captured by a camera positioned at the location capable of scanning the front of the train. Step S6: The control system fuses the point cloud data of the train head contour and the train head image, corrects the positioning deviation by combining the physical coordinate reference, and calculates the train stopping position.

2. The intelligent train identification method according to claim 1, characterized in that, In step S1, a threshold range for the number of wheelsets is set. When the number of wheelsets identified by the magnetic steel sensing module exceeds the threshold range, it is determined to be abnormal and an alarm is triggered.

3. The intelligent train identification method according to claim 1, characterized in that, In step S3, when the grouping information of the wheelset data and the RFID tag data is inconsistent, an alarm signal is triggered.

4. The intelligent train identification method according to claim 1, characterized in that, In step S3, the train's forward and reverse configuration is determined based on the order in which the RFID tag codes are read and the correspondence between the tag codes and carriages in the database.

5. The intelligent train identification method according to claim 1, characterized in that, In step S4, the auxiliary calibration camera establishes a physical coordinate reference by recognizing the scale markings, which is used to correct the positioning deviation of the lidar device. When a new vehicle model is detected to be entering the station for the first time, the calibration mode is automatically activated. The vehicle model deviation data and error correction parameters are calculated using the scale and the auxiliary calibration camera, and the vehicle model deviation data and the error correction parameters are stored in the database.

6. The intelligent train identification method according to claim 1, characterized in that, In step S6, the point cloud data of the vehicle front profile scanned by the lidar device is matched with the vehicle front image captured by the camera to cross-validate the vehicle model data.

7. The intelligent train identification method according to claim 1, characterized in that, In step S6, a weighted fusion algorithm is used to perform phased weighted fusion of the wheelset data, the RFID tag code, the vehicle front outline point cloud data, and the vehicle front image, and output the recognition result with the highest comprehensive confidence. When the deviation of a certain sensor data from other sensor data exceeds a preset threshold, the abnormal data is automatically removed, and the remaining sensor data is used for fusion calculation.

8. A train intelligent identification device, characterized in that, include: The magnetic steel induction module is installed on the train's entry path to the station to sense changes in the magnetic field of the train wheels, trigger the start of the control system, and identify the train wheelset data. An RFID tag reader is installed on the train's entry path behind the magnetic steel sensing module to read the RFID tag codes of each carriage of the train. The lidar device is set up at a position that can scan the front of the train to scan the outline of the train front and obtain point cloud data of the front outline. A high-resolution camera is positioned to scan the front of the train and capture images of the train's front. A scale and an auxiliary calibration camera are installed at the train window position in the platform area. The scale is used to provide a physical coordinate reference, and the auxiliary calibration camera is used to identify the scale graduations. The control system is connected to the magnetic steel sensing module, the RFID tag reader, the lidar device, the high-resolution camera, and the auxiliary calibration camera, respectively. The control system includes a processor, a memory, and a communication interface. The processor is used to receive data from each module and execute data fusion algorithms. The memory is used to store the database and calculation results. The communication interface is used to communicate with each module.

9. A train intelligent identification device according to claim 8, characterized in that, The magnetic steel induction module is installed next to the rail at a distance of 10-50 meters from the turnout, and at a height of 5-20 cm above the rail surface.

10. A train intelligent identification device according to claim 8, characterized in that, The RFID tag reader is positioned 5-20 meters away from the magnetic steel sensing module in the center of the rail, with the antenna facing the direction the train is entering the station.

11. A train intelligent identification device according to claim 8, characterized in that, The lidar device and the high-resolution camera are positioned 3-8 meters directly above the train's stop position, and their scanning angle covers the outline of the train's front.

12. A train intelligent identification device according to claim 8, characterized in that, The scale is a vertically set scale with a scale interval of 1-5cm. The lens of the auxiliary calibration camera is aligned with the scale. The scale and the auxiliary calibration camera are fixed to the train window position in the platform area by chemical anchors.

13. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1 to 7.

14. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 7.

Images