Urban rail transit automatic measurement system and method

Abstract

An urban rail transit automatic measurement system disclosed by the present invention comprises a trolley, two sides of the lower part of the trolley are respectively and fixedly provided with a group of uniformly arranged rail test assemblies, each rail test assembly comprises a third sensor and a fourth sensor, and a rail is tested through the third sensor and the fourth sensor. The rail testing assembly further comprises a fixing base which is in an inverted U shape. The invention relates to the technical field of automatic rail measuring equipment, in particular to an automatic urban rail transit measuring system and method. The automatic measuring system can monitor a plurality of key parts of the track in real time through a third sensor (a laser displacement sensor) and a fourth sensor (an infrared distance sensor) in the track testing assembly and a second sensor (a pull rod type displacement sensor) in the auxiliary testing assembly.

Description

Technical Field

[0001] This invention relates to the field of automatic track measurement equipment, and more specifically, to an automatic measurement system and method for urban rail transit. Background Technology

[0002] With the continuous expansion of urban scale and the acceleration of urbanization, urban rail transit, as an efficient, convenient, and environmentally friendly mode of public transportation, has developed rapidly. The safe operation of rail transit is directly related to the travel safety of the general public and the normal operation of the city; therefore, the monitoring and maintenance of rail and related facilities are of paramount importance.

[0003] Traditional urban rail transit track measurement methods have many drawbacks. In track inspection, the past often relied on manual inspections combined with simple measuring tools, such as using vernier calipers to measure contact wire wear. This approach is not only inefficient but also highly susceptible to human error, making data accuracy difficult to guarantee. Taking the Beijing-Hong Kong Metro as an example, traditional manual contact wire wear inspections only selected 5-7 points every 250 meters, failing to comprehensively reflect the actual wear condition of the contact wire. In track deformation monitoring, traditional methods often involve measuring a cross-section at regular intervals within the tunnel, ignoring data changes between cross-sections and posing safety risks. Furthermore, traditional measurement methods require the use of multiple independent devices to measure various aspects of the track, making real-time comprehensive data analysis impossible and failing to meet the high efficiency, comprehensiveness, and accuracy requirements of modern urban rail transit for track safety monitoring. In addition, with the increase in the operating mileage of rail transit, the cost of manual inspections continues to rise, making it difficult to meet the maintenance needs of large-scale lines. Summary of the Invention

[0004] The technical problem this invention aims to solve is to provide an automatic measurement system and method for urban rail transit. The automatic measurement system utilizes a third sensor (laser displacement sensor) and a fourth sensor (infrared distance sensor) in the track testing assembly, as well as a second sensor (rod-type displacement sensor) in the auxiliary testing assembly, to monitor multiple key components of the track in real time. The laser displacement sensor accurately measures the distance change between the wheel and the connecting frame, the infrared distance sensor acquires real-time distance data from the connecting frame to the side of the wheel, and the rod-type displacement sensor monitors the distance change between the ends of each pair of top blocks. These sensors work together to achieve comprehensive acquisition of multi-dimensional track data. Compared to traditional single-point acquisition methods, the acquired data is more continuous and comprehensive, avoiding measurement blind spots.

[0005] An automatic measurement system for urban rail transit includes a trolley. A set of evenly arranged track testing components are fixedly mounted on both sides of the lower part of the trolley. Each track testing component includes a third sensor and a fourth sensor, which test the track. The track testing components also include a fixed base, which is inverted U-shaped and fixed to the lower side of the trolley. Symmetrical connecting rods are rotatably connected to one end of each pair of connecting rods on both sides of the fixed base. The other end of each pair of connecting rods is rotatably connected to both ends of one side of a rotating frame. The center of the rotating frame is rotatably connected to the central axis of a measuring wheel, and one end of the central axis of the measuring wheel is fixed. The output shaft of the motor is connected, and the motor is fixed inside the rotating frame. The upper middle part of the rotating frame rotates to connect to the lower end of the gas spring housing. The gas spring housing matches the gas spring telescopic rod. The upper end of the gas spring telescopic rod is fixedly connected to a swing block. The swing block is rotatably connected to a fixed block. The fixed block is fixedly connected to the fixed seat. The third sensor is fixed on the side of the gas spring housing corresponding to the fixed block and is used to monitor the distance between the gas spring housing and the swing block. The fourth sensor is fixedly connected to the side of the connecting frame corresponding to the measuring wheel and is used to monitor the distance from the connecting frame to the side of the measuring wheel.

[0006] It also includes an auxiliary testing component, which includes a fixed frame fixed to one end of the trolley. Guide blocks are fixedly connected to both ends of the lower side of the trolley. The guide blocks are screwed to screw rods. One end of the screw rod is fixedly connected to both ends of a knob. The other end of the screw rod is rotatably connected to a movable block. The movable block is slidably connected to the fixed frame. The outer end of one side of the movable block is rotatably connected to the center of two support rods. The two support rods are arranged in an X-shape. The outer ends of the support rods are fixedly connected to a top block. The surface of the top block is arc-shaped. A second connecting post is fixedly connected to one side of the movable block corresponding to the guide block. The other end of the support rod is fixedly connected to a first connecting post. The first and second connecting posts are connected by an elastic connector.

[0007] Furthermore, the elastic connector is a spring, and the two ends of the spring are respectively fixedly connected to the first connecting post and the second connecting post;

[0008] The track includes a rail head tread, a rail head side, and a rail bottom upper surface. The rail bottom upper surface can abut against the measuring wheel, the rail head tread abuts against the center roller of the measuring wheel, and the rail head side corresponds to the side of the measuring wheel.

[0009] Furthermore, the track also includes a lower jaw at the rail head and an upper jaw at the rail bottom. The lower jaw at the rail head can abut against the top block on the lower side, and the upper jaw at the rail bottom can abut against the top block on the upper side. A second sensor is fixedly installed at a corresponding position on each pair of support rods.

[0010] Furthermore, it also includes a controller installed inside the trolley. A connecting plate is fixed to the other end of the trolley, and a second camera is fixedly connected to the lower side of the connecting plate. The second camera is used to eliminate obstacles. When an obstacle is encountered, the controller processes the video data using an obstacle recognition algorithm based on contour feature matching to determine whether it is an obstacle and controls the motion motor to stop moving forward in advance. A first camera is fixedly connected to the rear side of the connecting plate. The first camera is used to capture the track and uses a rail head deformation detection algorithm based on shape deviation analysis to determine whether the outer surface and shape of the rail head have changed. A comprehensive anomaly assessment algorithm based on multi-source data fusion comprehensively judges whether a signal requiring maintenance has been generated by combining the captured content and the sensor-sensed content.

[0011] Specifically, the obstacle recognition algorithm based on contour feature matching extracts feature parameters such as the contour area, perimeter, and vertex angle of abnormal objects in video frames and performs weighted matching calculations with a preset obstacle feature library to determine obstacles; the rail head deformation detection algorithm based on shape deviation analysis models the rail head edge contour, calculates the root mean square deviation between the actual shape and the standard model, and analyzes the degree of rail head deformation; the comprehensive anomaly assessment algorithm based on multi-source data fusion calculates a comprehensive anomaly index by fusing image analysis results with multi-sensor monitoring data to achieve accurate generation of maintenance signals.

[0012] Furthermore, the controller transmits data with the third, second, and fourth sensors, and the second and first cameras transmit data with the controller. An alarm is also provided, which is controlled by the controller and fixed on the vehicle. The first and second cameras transmit monitoring images to the controller in real time. The third, second, and fourth sensors perform real-time monitoring and transmit the monitoring data to the controller. The controller transmits the monitoring data to the cloud. The controller determines whether the monitoring data exceeds a set threshold. If it does, the controller activates the alarm.

[0013] Furthermore, the second sensor is a displacement sensor, specifically a rod-type displacement sensor. Its rod end is connected to one of the top blocks, and its housing end is connected to the other top block. It detects the distance change between the ends of each pair of top blocks in real time by sensing the extension and retraction of the rod, and transmits the data to the controller in real time. When the distance change exceeds a preset range, such as 1mm, the controller sends a warning signal. The third sensor is a displacement sensor, specifically a laser displacement sensor. It emits a laser beam to the surface of the swing block and receives the reflected signal to detect the distance change between the wheel and the connecting frame in real time, and transmits the data to the controller in real time. When the distance change exceeds a preset range, such as 1mm, the central processing unit sends a warning signal. The fourth sensor is a distance sensor, specifically an infrared distance sensor. Its transmitting end is fixed to the connecting frame, and its receiving end faces the side of the measuring wheel. It calculates the distance using the time difference of infrared beam reflection. When lateral movement or swinging occurs, the sensor sends a signal to the controller in real time.

[0014] A measurement method for an automatic measurement system for urban rail transit includes the following steps:

[0015] Step S1: The system starts up, the controller initializes and establishes communication connections with the third sensor, the second sensor, the fourth sensor, the first camera and the second camera;

[0016] Step S2: The system moves along the track, the second camera captures the view in front of the track in real time and transmits the video data to the controller, and the first camera simultaneously captures the view of the outer surface of the track head and transmits it to the controller;

[0017] Step S3: During the movement, the third sensor monitors the distance change from the gas spring housing to the swing block in real time, the fourth sensor monitors the distance change from the connecting frame to the side of the measuring wheel in real time, and the second sensor monitors the distance change between the ends of each pair of top blocks in real time. Each sensor transmits the monitoring data to the controller in real time.

[0018] Step S4: The controller processes the video data transmitted by the second camera using a preset algorithm to determine whether there are obstacles in front of the track; it also processes the image transmitted by the first camera using a preset algorithm to analyze whether the outer surface and shape of the track head have changed.

[0019] Step S5: The controller compares the analysis results from the first camera with the monitoring data from the third, second, and fourth sensors to determine whether the track has generated an abnormal signal that requires maintenance.

[0020] Step S6: The controller transmits real-time monitoring data and analysis results to the cloud, and at the same time determines whether the monitoring data of each sensor exceeds the preset threshold. If it does, an alarm operation is executed.

[0021] Further, the specific process of determining whether there is an obstacle in front of the track in step S4 is as follows: After the second camera transmits real-time video data to the controller, the controller extracts features from the video frames based on a multi-feature weighted matching obstacle recognition algorithm, identifies the contours of abnormal objects in the image, compares the contour parameters with a preset obstacle feature library, and determines that an obstacle exists when the matching degree exceeds a preset threshold; wherein the matching degree calculation formula is:

[0022]

[0023] In the formula, M is the contour matching degree, n is the number of feature parameters, and w i Let Sim(P) be the weight of the i-th feature parameter (e.g., contour area weight 0.3, perimeter weight 0.2, vertex angle weight 0.5). i Q i ) represents the similarity between the i-th parameter of the contour to be identified and the corresponding parameter of the standard obstacle in the feature library, with a value range of [0,1]; when M≥0.7, it is determined that an obstacle exists.

[0024] Further, the specific process of analyzing whether the outer surface and shape of the rail head have changed in step S4 is as follows: After the first camera captures an image of the outer surface of the rail head, the controller performs edge detection and shape modeling on the image using a rail head deformation detection algorithm based on shape deviation analysis. The modeling result is then compared with a preset standard rail head shape model to calculate the deviation. When the deviation value exceeds a preset range, it is determined that the shape of the rail head has changed. The formula for calculating the deviation value is:

[0025]

[0026] In the formula, D is the shape deviation value (unit: mm), m is the number of sampling points, and S... j S represents the coordinates of the j-th sampling point after modeling the actual railhead. 0j These are the coordinates of the sampling points corresponding to the standard model; when D > 1 mm, it is determined that the rail head shape has changed.

[0027] Further, the specific process of comprehensive comparison in step S5 is as follows: the controller fuses the rail head shape change data analyzed by the first camera, the distance change data monitored by the third sensor, the top block distance change data monitored by the second sensor, and the lateral distance change data monitored by the fourth sensor. A comprehensive anomaly index is calculated using a multi-source data fusion comprehensive anomaly assessment algorithm. When the comprehensive anomaly index exceeds a preset threshold, a signal requiring maintenance is generated. The formula for calculating the comprehensive anomaly index is:

[0028] In the formula, I is the comprehensive anomaly index, a, b, c, and d are weighting coefficients (satisfying a+b+c+d=1, such as taking 0.3, 0.2, 0.3, and 0.2 respectively), and D is the railhead shape deviation value. max =1mm is the rail head deviation threshold, ΔL3 is the change in distance monitored by the third sensor, ΔL 3max =1mm is its threshold, ΔL2 is the change in the distance monitored by the second sensor, ΔL 2max =1mm is its threshold, ΔL4 is the change in distance monitored by the fourth sensor, ΔL 4max A preset threshold is set for it; a maintenance signal is generated when I ≥ 0.6.

[0029] Furthermore, the specific process of performing the alarm operation in step S6 is as follows: after the controller determines that the monitoring data exceeds the preset threshold, it controls the alarm to issue an audible and visual alarm signal on the one hand, and records the track position information where the abnormality occurred, the corresponding sensor monitoring data and the camera image on the other hand, and uploads the abnormal information to the cloud storage synchronously after marking it.

[0030] Compared with the prior art, the advantages and positive effects of the present invention are:

[0031] The automatic measurement system utilizes the third sensor (laser displacement sensor) and the fourth sensor (infrared distance sensor) in the track testing assembly, as well as the second sensor (rod-type displacement sensor) in the auxiliary testing assembly, to monitor multiple key components of the track in real time. The laser displacement sensor accurately measures the distance change between the wheel and the connecting frame, the infrared distance sensor acquires real-time distance data from the connecting frame to the wheel's side, and the rod-type displacement sensor monitors the distance change between the ends of each pair of top blocks. These sensors work together to achieve comprehensive acquisition of multi-dimensional track data. Compared to traditional single-point acquisition methods, the acquired data is more continuous and comprehensive, avoiding measurement blind spots and significantly improving the efficiency and quality of data acquisition.

[0032] The system is equipped with a first camera and a second camera, which respectively use a rail head deformation detection algorithm based on shape deviation analysis and an obstacle recognition algorithm based on contour feature matching to accurately judge changes in the outer surface and shape of the rail head and obstacles in front of the track. Simultaneously, the controller comprehensively compares the camera analysis results with the monitoring data from various sensors using a multi-source data fusion algorithm for anomaly assessment. When the monitoring data exceeds a preset threshold, it can promptly trigger an alarm. This multi-source data fusion fault early warning mechanism greatly improves the accuracy and reliability of fault judgment, enabling early detection of potential track safety hazards and providing strong support for timely maintenance measures.

[0033] The controller transmits real-time monitoring data and analysis results to the cloud, facilitating long-term storage and in-depth analysis of historical track data. Simultaneously, the system can determine whether track maintenance is required based on a comprehensive anomaly index, providing a scientific basis for track maintenance. For example, a comprehensive anomaly index is calculated by integrating data such as rail head shape deviation and changes in monitoring distances from various sensors. When the index exceeds a preset threshold, a maintenance signal is generated, facilitating track inspection by staff. Attached Figure Description

[0034] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. Obviously, the drawings described below are merely some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without creative effort. In the drawings:

[0035] Figure 1 The three-dimensional representation of the present invention Figure 1 ;

[0036] Figure 2 This is a partial cross-sectional view of the rail of the present invention;

[0037] Figure 3 The three-dimensional representation of the present invention Figure 2 ;

[0038] Figure 4 The three-dimensional representation of the present invention Figure 3 ;

[0039] In the diagram: 1. Trolley; 2. First camera; 3. Connecting plate; 4. Lighting lamp; 5. Gas spring housing; 6. Motor; 7. Measuring wheel; 8. Wheel hub; 9. Connecting frame; 10. Moving block; 11. Top block; 12. Support rod; 13. First connecting column; 14. Second connecting column; 15. Guide block; 16. Knob; 17. Screw; 18. Fixing frame; 191. Rail head tread; 192. Rail head side; 193. Rail bottom upper surface; 194. Rail head lower jaw; 195. Rail bottom upper jaw; 20. Second camera; 21. Fixing seat; 22. Fixing block; 23. Swing block; 24. Third sensor; 25. Gas spring telescopic rod; 26. Fourth sensor. Detailed Implementation

[0040] The present invention will be further described below with reference to the accompanying drawings and embodiments:

[0041] An automatic measurement system for urban rail transit includes a trolley 1. A set of evenly arranged track testing components are fixedly mounted on both sides of the lower part of the trolley 1. Each track testing component includes a third sensor 24 and a fourth sensor 26, which test the track. The track testing components also include a fixed base 21, which is inverted U-shaped and fixed to the lower side of the trolley 1. Symmetrical connecting rods are rotatably connected to one end of each pair of connecting rods on both sides of the fixed base 21. The other end of each pair of connecting rods is rotatably connected to both ends of one side of a rotating frame. The center of the rotating frame is rotatably connected to the central axis of a measuring wheel 7. One end of the central axis of the measuring wheel 7... The output shaft of the fixedly connected motor 6 is fixed inside the rotating frame. The upper middle part of the rotating frame rotates to connect to the lower end of the gas spring housing 5. The gas spring housing 5 is matched with the gas spring telescopic rod 25. The upper end of the gas spring telescopic rod 25 is fixedly connected to the swing block 23. The swing block 23 is rotatably connected to the fixed block 22. The fixed block 22 is fixedly connected to the fixed seat 21. The third sensor 24 is fixed on the side of the gas spring housing 5 corresponding to the fixed block 22 and is used to monitor the distance between the gas spring housing 5 and the swing block 23. The fourth sensor 26 is fixedly connected to the side of the connecting frame 9 corresponding to the measuring wheel 7. The fourth sensor 26 is used to monitor the distance between the connecting frame 9 and the side of the measuring wheel 7.

[0042] It also includes an auxiliary testing component, which includes a fixed frame 18, which is fixed to one end of the trolley 1. Guide blocks 15 are fixedly connected to both ends of the lower side of the trolley 1. Guide blocks 15 are screwed to screw rods 17. One end of screw rod 17 is fixedly connected to both ends of knob 16. The other end of screw rod 17 is rotatably connected to moving blocks 10. Moving blocks 10 are slidably connected to the fixed frame 18. The outer end of one side of moving blocks 10 is rotatably connected to the center of two support rods 12. The two support rods 12 are arranged in an X shape. The outer ends of support rods 12 are fixedly connected to top blocks 11. The surface of top blocks 11 is arc-shaped. The side of one end of moving blocks 10 corresponding to guide blocks 15 is fixedly connected to a second connecting post 14. The other end of support rods 12 is fixedly connected to a first connecting post 13. The first connecting post 13 and the second connecting post 14 are connected by an elastic connector.

[0043] The trolley 1 has two sets of independently suspended measuring wheels 7 on each side. Each set of measuring wheels 7 consists of a high-stiffness preloaded roller and a low-stiffness floating roller. The high-stiffness roller records the wheel-rail normal force in real time through a thin-film pressure sensor array (range 0–50kN, resolution 0.05%FS). The low-stiffness floating roller can float freely in the vertical ±10mm and the lateral ±5mm under the constraint of the suspension formed by the gas spring and connecting rod. Its displacement is captured by a laser displacement sensor (resolution 0.5μm).

[0044] Track testing components are installed on both sides of the lower part of the trolley. The third sensor 24 and the fourth sensor 26 in the track testing components are used to test the track. The fixed base 21 is inverted U-shaped and connected to the bracket. It is connected to the rotating frame through the connecting rod. The measuring wheel 7 and the motor 6 are installed in the center of the rotating frame. The gas spring assembly is used for adjustment to form a suspension system for the measuring wheel 7. The third sensor 24 monitors the distance from the gas spring housing 5 to the swing block 23, and the fourth sensor 26 monitors the distance from the connecting frame 9 to the side of the wheel. The fixed frame 18 of the auxiliary testing component is connected to the bracket to realize multi-directional testing of the track. The track-related data are acquired in real time through the third sensor 24 and the fourth sensor 26 to provide a basis for judging the track status. The auxiliary testing component helps to detect the track from other angles and enhances the comprehensiveness of the detection. The combination of the measuring wheel 7 and the motor 6 allows the device to move along the track, while the gas spring assembly ensures that a suitable contact state is maintained during the test.

[0045] Furthermore, the elastic connector is a spring, with the first connecting post 13 and the second connecting post 14 fixedly connected to both ends of the spring, respectively.

[0046] The track includes a rail head tread 191, a rail head side 192, and a rail bottom upper surface 193. The rail bottom upper surface 193 can abut against the measuring wheel 7, the rail head tread 191 abuts against the central roller of the measuring wheel 7, and the rail head side 192 corresponds to the side of the measuring wheel 7.

[0047] The elastic connector is a spring, with its two ends connected to the first connecting post 13 and the second connecting post 14, respectively. It provides elastic cushioning within the auxiliary testing assembly, ensuring that the connections between components maintain their relative positions during operation. Furthermore, it allows for adaptive adjustment through spring extension and contraction when encountering changes in external force, preventing damage from hard collisions and facilitating more accurate perception of subtle changes in track-related parts.

[0048] Furthermore, the track also includes a lower rail head jaw 194 and a lower rail bottom jaw 195. The lower rail head jaw 194 can abut against the lower top block, and the upper rail bottom jaw 195 can abut against the upper top block. A second sensor is fixedly installed at a corresponding position on each pair of support rods.

[0049] Furthermore, it also includes a controller, which is installed inside the trolley 1. The other end of the trolley 1 is fixedly connected to a connecting plate 3. A second camera 20 is fixedly connected to the lower side of the connecting plate 3. The second camera 20 is used to eliminate obstacles. When an obstacle is encountered, the controller processes the video data through an obstacle recognition algorithm based on contour feature matching to determine whether it is an obstacle and controls the motion motor 6 to stop moving forward in advance. A first camera 2 is fixedly connected to the rear side of the connecting plate 3. The first camera 2 is used to capture the track and uses a rail head deformation detection algorithm based on shape deviation analysis to determine whether the outer surface and shape of the rail head have changed. A comprehensive anomaly assessment algorithm based on multi-source data fusion is used to comprehensively determine whether a signal requiring maintenance is generated by combining the captured content and the sensor-sensed content.

[0050] Among them, the obstacle recognition algorithm based on contour feature matching extracts feature parameters such as contour area, perimeter, and vertex angle of abnormal objects in video frames, and performs weighted matching calculation with a preset obstacle feature library to achieve obstacle determination; the rail head deformation detection algorithm based on shape deviation analysis models the rail head edge contour, calculates the root mean square deviation between the actual shape and the standard model, and analyzes the degree of rail head deformation; the comprehensive anomaly assessment algorithm based on multi-source data fusion calculates a comprehensive anomaly index by fusing image analysis results with multi-sensor monitoring data to achieve accurate generation of maintenance signals.

[0051] The controller is installed inside the trolley 1, which is connected to a second camera 20 and a first camera 2. The second camera 20 is used to eliminate obstacles, using an obstacle recognition algorithm based on contour feature matching to process video data and determine if there are obstacles. The first camera 2 is used to photograph the track, using a rail head deformation detection algorithm based on shape deviation analysis to determine changes in the outer surface and shape of the rail head. Through a comprehensive anomaly assessment algorithm that integrates the camera footage and sensor data, a comprehensive judgment is made on whether a maintenance signal is generated, achieving all-round intelligent monitoring of the track, early detection of obstacles ahead of the track to prevent collisions, and timely detection of problems such as rail head deformation. By integrating multi-source data, the accuracy of determining whether the track needs maintenance is improved, ensuring the safe operation of urban rail transit.

[0052] Furthermore, the controller transmits data to the third sensor 24, the second sensor, and the fourth sensor 26. The second camera 20 and the first camera 2 transmit data to the controller. An alarm is also provided. The controller controls the alarm, which is fixed on the vehicle. The first camera 2 and the second camera 20 transmit the monitoring images to the controller in real time. The third sensor 24, the second sensor, and the fourth sensor 26 perform real-time monitoring and transmit the monitoring data to the controller. The controller transmits the monitoring data to the cloud. The controller determines whether the monitoring data exceeds the set threshold. If it exceeds the set threshold, it controls the alarm to sound.

[0053] The second sensor is a displacement sensor, specifically a lever-type displacement sensor. Its lever end is connected to one of the top blocks, and its housing end is connected to the other top block. It detects the distance change between the ends of each pair of top blocks in real time by sensing the extension and retraction of the lever, and transmits the data to the controller in real time. When the distance change exceeds the preset range, such as 1mm, the controller sends a warning signal. The third sensor 24 is a displacement sensor, specifically a laser displacement sensor. It emits a laser beam to the surface of the swing block 23 and receives the reflected signal to detect the distance change between the wheel and the connecting frame 9 in real time, and transmits the data to the controller in real time. When the distance change exceeds the preset range, such as 1mm, the central processing unit sends a warning signal. The fourth sensor 26 is a distance sensor, specifically an infrared distance sensor. Its transmitting end is fixed to the connecting frame 9, and its receiving end faces the side of the measuring wheel. It calculates the distance by the time difference of the infrared beam reflection. When lateral movement or swing occurs, the sensor sends a signal to the controller in real time.

[0054] The controller transmits data with the third sensor 24, the second sensor, the fourth sensor 26, the second camera 20, and the first camera 2. It also connects to an alarm. Each sensor monitors and transmits data in real time. The controller determines if the data exceeds a threshold; if so, it activates the alarm and transmits the data to the cloud. Specific types of the second, third, and fourth sensors are described: the second sensor is a lever-type displacement sensor that detects changes in the distance to the top block end through lever extension; the third sensor 24 is a laser displacement sensor that detects changes in the distance between the wheel and the connecting frame through laser reflection; and the fourth sensor 26 is an infrared distance sensor that calculates distance through the time difference of infrared beam reflection. The sensors provide accurate data in real time, which the controller analyzes and makes decisions about. When an anomaly occurs in the track, an alarm is triggered promptly to alert staff. The data is also uploaded to the cloud for subsequent analysis and management. Different types of sensors address different monitoring needs, improving the accuracy and reliability of the system.

[0055] A measurement method for an automatic measurement system for urban rail transit includes the following steps:

[0056] Step S1: The system starts up, the controller initializes and establishes communication connections with the third sensor 24, the second sensor, the fourth sensor 26, the first camera 2 and the second camera 20;

[0057] Step S2: The system moves along the track, the second camera 20 captures the image in front of the track in real time and transmits the video data to the controller, and the first camera 2 simultaneously captures the image of the outer surface of the track head and transmits it to the controller;

[0058] Step S3: During the movement, the third sensor 24 monitors the distance change from the gas spring housing 5 to the swing block 23 in real time, the fourth sensor 26 monitors the distance change from the connecting frame to the side of the measuring wheel 7 in real time, and the second sensor monitors the distance change between the ends of each pair of top blocks in real time. Each sensor transmits the monitoring data to the controller in real time.

[0059] Step S4: The controller processes the video data transmitted by the second camera 20 using a preset algorithm to determine whether there are obstacles in front of the track; it also processes the image transmitted by the first camera 2 using a preset algorithm to analyze whether the outer surface and shape of the track head have changed.

[0060] Step S5: The controller compares the analysis results from the first camera 2 with the monitoring data from the third sensor 24, the second sensor, and the fourth sensor 26 to determine whether the track has generated an abnormal signal that requires maintenance.

[0061] Step S6: The controller transmits real-time monitoring data and analysis results to the cloud, and at the same time determines whether the monitoring data of each sensor exceeds the preset threshold. If it does, an alarm operation is executed.

[0062] Further, the specific process of determining whether there is an obstacle in front of the track in step S4 is as follows: After the second camera 20 transmits real-time video data to the controller, the controller extracts features from the video frames based on a multi-feature weighted matching obstacle recognition algorithm, identifies the contours of abnormal objects in the image, compares the contour parameters with a preset obstacle feature library, and determines that an obstacle exists when the matching degree exceeds a preset threshold; wherein the matching degree calculation formula is:

[0063]

[0064] Specifically, after the second camera 20 transmits video data to the controller, the controller uses an algorithm to extract features from the video frames, identify the outlines of abnormal objects, and perform weighted matching calculations between the outline parameters and a preset obstacle feature library. When the matching degree exceeds a preset threshold, it is determined that there is an obstacle. The controller extracts and weighted compares the key feature parameters of the target outline to achieve accurate identification of obstacles in front of the track. The specific process is as follows: Feature extraction stage: After the controller preprocesses the video frames transmitted by the second camera 20, it extracts the outline features of abnormal objects in the picture through an edge detection algorithm, including but not limited to the key parameters of outline area, perimeter, and vertex angle.

[0065] Feature comparison stage: The extracted actual contour feature parameters are compared with the standard parameters in the preset obstacle feature library one by one to calculate the similarity. The similarity is normalized to ensure that the value range is [0,1].

[0066] Weighted matching calculation: Different weights are assigned to each feature parameter based on its importance to obstacle recognition (e.g., contour area weight 0.3, perimeter weight 0.2, vertex angle weight 0.5), and the result is calculated using the formula...

[0067]

[0068] Calculate the overall matching degree M; when the overall matching degree M≥0.7, the controller determines that there is an obstacle in front of the track. This algorithm improves the accuracy of obstacle recognition and anti-interference ability through multi-feature fusion and weighted calculation.

[0069] Further, the specific process of analyzing whether the outer surface and shape of the rail head have changed in step S4 is as follows: After the first camera 2 captures the image of the outer surface of the rail head, the controller performs edge detection and shape modeling on the image using a rail head deformation detection algorithm based on shape deviation analysis. The modeling result is then compared with the preset standard rail head shape model to calculate the deviation. When the deviation value exceeds the preset range, it is determined that the shape of the rail head has changed. The formula for calculating the deviation value is:

[0070]

[0071] Specifically, after the first camera 2 captures an image of the rail head, the controller uses algorithms to perform edge detection and shape modeling on the image. The modeling result is then compared to a preset standard rail head shape model to calculate the deviation. When the deviation exceeds a preset range, the rail head shape is determined to have changed. After the first camera 2 captures an image of the rail head's outer surface, the controller uses an edge detection algorithm (such as the Canny operator) to extract the rail head's edge contour. Then, contour fitting technology is used to establish the actual rail head's shape model, obtaining the coordinate parameters of multiple sampling points (i.e., S in the formula). j ); The coordinates of the sampling points obtained from the actual modeling are compared with the coordinates of the corresponding sampling points in the preset standard shape model of the rail head (i.e., S in the formula). 0j The comparison was performed using the root mean square deviation formula.

[0072]

[0073] Calculate the shape deviation value D (unit: mm); when the deviation value D>1mm, it is determined that the rail head shape has changed significantly. This algorithm achieves accurate detection of rail head deformation by quantifying the shape deviation.

[0074] Further, the specific process of comprehensive comparison in step S5 is as follows: The controller fuses the rail head shape change data analyzed by the first camera 2, the distance change data monitored by the third sensor 24, the top block distance change data monitored by the second sensor, and the lateral distance change data monitored by the fourth sensor 26. A comprehensive anomaly index is calculated using a multi-source data fusion comprehensive anomaly assessment algorithm. When the comprehensive anomaly index exceeds a preset threshold, a signal requiring maintenance is generated. The formula for calculating the comprehensive anomaly index is:

[0075]

[0076] Specifically, the controller fuses the rail head shape change data analyzed by the first camera 2, the distance change data monitored by the third sensor 24, the top block distance change data monitored by the second sensor, and the lateral distance change data monitored by the fourth sensor 26. It then calculates a comprehensive anomaly index using an algorithm. When the comprehensive anomaly index exceeds a preset threshold, a maintenance signal is generated. The controller standardizes the rail head shape deviation value, the distance change value ΔL3 from the third sensor 24, the distance change value ΔL2 from the second sensor, and the distance change value ΔL4 from the fourth sensor 26, by dividing each parameter value by its corresponding preset threshold (e.g., D). max =1mm, ΔL 3max =1mm, etc., to obtain the standardized anomaly ratio of each parameter; assign weight coefficients to each parameter according to its importance to the track condition assessment (e.g., a=0.3, b=0.2, c=0.3, d=0.2, and satisfying a+b+c+d=1), and calculate the comprehensive anomaly index I through the formula; when the comprehensive anomaly index I≥0.6, the controller generates a signal that maintenance is required. This algorithm avoids misjudgment of a single parameter through multi-source data fusion, and improves the reliability of track anomaly judgment.

[0077] The specific process of performing the alarm operation in step S6 is as follows: After the controller determines that the monitoring data exceeds the preset threshold, it controls the alarm to issue an audible and visual alarm signal on the one hand, and records the track position information where the abnormality occurred, the corresponding sensor monitoring data and the camera image on the other hand, and uploads the abnormal information to the cloud storage synchronously after marking it.

[0078] It also includes a lighting lamp 4, which is fixed to the lower side of the connecting plate and is used for lighting in dim environments to provide illumination for the first camera 2 and the second camera 20.

[0079] The cloud can transmit data between electronic products such as mobile phones and computers, controllers and the cloud through existing transmission modes (including but not limited to wireless transmission, CAN bus transmission, etc.).

[0080] The method of using this invention is as follows: Upon use, the system starts, the controller initializes and establishes communication connections with the third sensor 24, the second sensor, the fourth sensor 26, the first camera 2, and the second camera 20; the measuring wheel 7 abuts against the rail, the motor 6 is started, and the system moves along the rail. The second camera 20 captures the image in front of the rail in real time and transmits the video data to the controller. The first camera 2 simultaneously captures the image of the outer surface of the rail head and transmits it to the controller. During the movement, the third sensor 24 monitors the distance change from the gas spring housing 5 to the swing block 23 in real time, the fourth sensor 26 monitors the distance change from the connecting frame 9 to the side of the measuring wheel 7 in real time, and the second sensor monitors the distance between the ends of each pair of top blocks in real time. As the distance between the sensors changes, each sensor transmits its monitoring data to the controller in real time. The controller processes the video data transmitted by the second camera 20 using a preset algorithm to determine if there are any obstacles in front of the track. It also processes the image transmitted by the first camera 2 using a preset algorithm to analyze whether the outer surface and shape of the track head have changed. The controller compares the analysis results from the first camera 2 with the monitoring data from the third sensor 24, the second sensor, and the fourth sensor 26 to determine if the track has generated any abnormal signals requiring maintenance. The controller transmits the real-time monitoring data and analysis results to the cloud and simultaneously determines whether the monitoring data from each sensor exceeds a preset threshold. If it does, an alarm is triggered.

[0081] The effectiveness of this invention was verified by applying it to the track of the Wujiabao section of Jinan Metro Line 2.

[0082] The Wujiabao section of Jinan Metro Line 2 is located in western Jinan City. The line traverses urban-rural fringe areas and some industrial zones. The tracks are constantly affected by train loads, geological subsidence, and the surrounding environment, posing potential risks such as rail head wear, lateral track displacement, and localized deformation. Traditional manual inspection methods are insufficient to meet the high-frequency, high-precision monitoring requirements of this section. Applying this urban rail transit automatic measurement system to this section has demonstrated significant effects in data acquisition, fault early warning, and operation and maintenance decision-making. Specific verification is as follows:

[0083] The Wujiabao section of track is approximately 5.2 kilometers long, including multiple curves and turnout areas. Traditional manual inspection requires three workers, with a daily inspection efficiency of only 1.5 kilometers per day. Furthermore, it suffers from oversights in monitoring hidden areas such as rail head side wear and rail base jaw deformation. After implementing this system, the system automatically moves along the track. The laser displacement sensor in the track testing component monitors the real-time distance change between the wheel and the connecting frame (sampling frequency up to 10Hz), while the infrared distance sensor records the lateral distance data between the connecting frame and the wheel side every 0.5 seconds. The lever-type displacement sensor in the auxiliary testing component continuously collects the distance change at the end of the top block.

[0084] During 30 consecutive days of monitoring, the system collected a total of 126,000 sets of rail head tread data, 98,000 sets of lateral displacement data, and 83,000 sets of top block distance change data, covering all key points along the entire track with no monitoring blind spots. Compared to traditional manual single-point sampling (selecting 6 points every 250 meters), the data collection efficiency has been improved by 8 times, and the system has achieved a transformation from "discrete point monitoring" to "continuous line monitoring," fully capturing detailed characteristics such as the accelerated wear rate of rail heads on curved sections and lateral displacement fluctuations in the turnout area.

[0085] Due to prolonged lateral force, a section of a curve in the Wujiabao area experienced slight deformation of the rail head's lower jaw. Traditional manual inspections, using visual inspection and simple measuring tools, failed to detect the abnormality. This system uses a first camera (2) to capture images of the rail head. An algorithm based on shape deviation analysis calculates the rail head shape deviation value in this area as D = 1.2 mm (exceeding the preset threshold of 1 mm). Simultaneously, a third sensor (24) detects a distance change ΔL3 = 1.1 mm between the gas spring housing 5 and the swing block 23, and a second sensor detects a distance change ΔL2 = 0.9 mm between the top block and the gas spring housing 5. The controller calculates a comprehensive anomaly index using a multi-source data fusion algorithm.

[0086] If the threshold of 0.6 is exceeded, the alarm will be activated immediately and the abnormal location will be marked.

[0087] Maintenance personnel verified the warning information on-site and confirmed that the lower jaw of the rail head 194 had localized plastic deformation. They promptly ground and repaired it, preventing increased train vibration and noise, as well as accelerated wheel-rail wear, caused by further deformation. Furthermore, the system successfully identified a rock (contour matching degree M = 0.82) that had fallen due to heavy rain beside the track using the obstacle recognition algorithm of the second camera 20. This allowed the system to stop moving forward in advance, avoiding the risk of equipment collision.

[0088] The system uploaded 30 days of monitoring data from the Wujiabao section to the cloud. Analysis of historical data revealed that the rail head wear rate (0.03 mm / day) in the K3+200-K3+500 section was higher than in other areas (average 0.015 mm / day), and lateral displacement fluctuations were positively correlated with train frequency. Based on the trend of the comprehensive anomaly index, the system generated maintenance recommendations: conduct a special inspection every 15 days in the K3+200-K3+500 section, prioritize rail head grinding, and strengthen the foundation reinforcement of the turnout area to reduce lateral displacement fluctuations. After adopting these recommendations, the average rail head lifespan of this section increased from 18 months under the traditional maintenance model to 24 months, annual maintenance costs decreased by 32%, and the train running stability score (based on vibration acceleration monitoring) improved from 85 to 92 points, significantly improving passenger comfort. This data-driven maintenance model achieved "on-demand maintenance" rather than "scheduled maintenance," greatly improving the scientific and economical nature of track maintenance in this section.

[0089] The above-disclosed embodiments are merely specific examples of the present invention. However, the present invention is not limited thereto, and any variations that can be conceived by those skilled in the art should fall within the protection scope of the present invention.

Claims

1. An automatic measurement system for urban rail transit, comprising a trolley (1) applied on a track, characterized in that: A set of evenly arranged track testing components are fixedly installed on both sides of the lower part of the trolley (1). The track testing components include a third sensor (24) and a fourth sensor (26). The track is tested by the third sensor (24) and the fourth sensor (26). The track testing components also include a fixed seat (21). The fixed seat (21) is U-shaped and is fixed to the lower side of the trolley (1). The two sides of the fixed seat (21) are rotatably connected to one end of a symmetrical connecting rod. The other end of each pair of connecting rods is rotatably connected to one end of a rotating frame. The center of the rotating frame is rotatably connected to the central shaft of the measuring wheel (7). One end of the central shaft of the measuring wheel (7) is fixedly connected to the output shaft of the motor (6). The motor (6) is fixed to the rotating frame. Inside the frame, the upper middle part of the rotating frame rotates to the lower end of the gas spring housing (5). The gas spring housing (5) matches the gas spring telescopic rod (25). The upper end of the gas spring telescopic rod (25) is fixedly connected to the swing block (23). The swing block (23) is rotatably connected to the fixed block (22). The fixed block (22) is fixedly connected to the fixed seat (21). The third sensor (24) is fixed on the side of the gas spring housing (5) corresponding to the fixed block (22) and is used to monitor the distance between the gas spring housing (5) and the swing block (23). The fourth sensor (26) is fixedly connected on the side of the connecting frame (9) corresponding to the measuring wheel (7). The fourth sensor (26) is used to monitor the distance between the connecting frame (9) and the side of the measuring wheel (7).

2. The automatic measurement system for urban rail transit according to claim 1, characterized in that: It also includes an auxiliary testing component, which includes a fixing frame (18) fixed to one end of the trolley (1). Guide blocks (15) are fixedly connected to both ends of the lower side of the trolley (1). The guide blocks (15) are screwed to screw rods (17). One end of the screw rod (17) is fixedly connected to both ends of a knob (16). The other end of the screw rod (17) is rotatably connected to a moving block (10). The moving block (10) is slidably connected to the fixing frame (18). 0) The outer ends of the two support rods (12) are rotatably connected to the center of the two support rods (12). The two support rods (12) are arranged in an X shape. The outer ends of the support rods (12) are fixedly connected to the top block (11). The surface of the top block (11) is arc-shaped. The moving block (10) is fixedly connected to the second connecting column (14) on one side of the guide block (15). The other end of the support rod (12) is fixedly connected to the first connecting column (13). The first connecting column (13) and the second connecting column (14) are connected by an elastic connector.

3. The automatic measurement system for urban rail transit according to claim 2, characterized in that: The elastic connector is a spring, and the two ends of the spring are respectively fixedly connected to the first connecting post (13) and the second connecting post (14); The track includes a rail head tread (191), a rail head side (192), and a rail bottom upper surface (193). The rail bottom upper surface (193) can abut against the measuring wheel (7), the rail head tread (191) abuts against the center roller of the measuring wheel (7), and the rail head side (192) corresponds to the side of the measuring wheel (7). The track also includes a lower jaw (194) and an upper jaw (195). The lower jaw (194) can abut against the lower top block (11), and the upper jaw (195) can abut against the upper top block (11). A second sensor is fixedly installed at a corresponding position on each pair of support rods (12).

4. The automatic measurement system for urban rail transit according to claim 3, characterized in that: It also includes a controller, which is installed inside the trolley (1). The other end of the trolley (1) is fixedly connected to a connecting plate (3). The lower side of the connecting plate (3) is fixedly connected to a second camera (20). The second camera (20) is used to eliminate obstacles. When an obstacle is encountered, the controller processes video data through an obstacle recognition algorithm based on contour feature matching to determine whether it is an obstacle and controls the motion motor (6) to stop moving forward in advance. The rear side of the connecting plate (3) is fixedly connected to a first camera (2). The first camera (2) is used to shoot the track. The rail head deformation detection algorithm based on shape deviation analysis is used to determine whether the outer surface and shape of the rail head have changed. The comprehensive anomaly evaluation algorithm based on multi-source data fusion is used to comprehensively determine whether a signal requiring maintenance is generated by combining the shooting content and the sensor perception content. Specifically, the obstacle recognition algorithm based on contour feature matching extracts feature parameters such as the contour area, perimeter, and vertex angle of abnormal objects in video frames and performs weighted matching calculations with a preset obstacle feature library to determine obstacles; the rail head deformation detection algorithm based on shape deviation analysis models the rail head edge contour, calculates the root mean square deviation between the actual shape and the standard model, and analyzes the degree of rail head deformation; the comprehensive anomaly assessment algorithm based on multi-source data fusion calculates a comprehensive anomaly index by fusing image analysis results with multi-sensor monitoring data to achieve accurate generation of maintenance signals.

5. The automatic measurement system for urban rail transit according to claim 4, characterized in that: The controller transmits data to the third sensor (24), the second sensor, and the fourth sensor (26). The second camera (20) and the first camera (2) transmit data to the controller. An alarm is also provided. The controller controls the alarm. The alarm is fixed on the car (1). The first camera (2) and the second camera (20) transmit the monitoring images to the controller in real time. The third sensor (24), the second sensor, and the fourth sensor (26) perform real-time monitoring and transmit the monitoring data to the controller. The controller transmits the monitoring data to the cloud. The controller determines whether the monitoring data exceeds the set threshold. If it exceeds the set threshold, it controls the alarm to sound. The second sensor is a displacement sensor, specifically a lever-type displacement sensor. Its lever end is connected to one of the top blocks (11), and its housing end is connected to the other top block (11). It detects the distance change between the ends of each pair of top blocks (11) in real time by sensing the extension and retraction of the lever, and transmits the data to the controller in real time. When the distance change exceeds a preset range, such as 1mm, the controller sends a warning signal. The third sensor (24) is a displacement sensor, specifically a laser displacement sensor. It emits a laser beam to the surface of the swing block (23) and receives the reflected signal, detecting the distance change between the wheel and the connecting frame (9) in real time, and transmitting the data to the controller in real time. When the distance change exceeds a preset range, such as 1mm, the central processing unit sends a warning signal. The fourth sensor (26) is a distance sensor, specifically an infrared distance sensor. Its transmitting end is fixed to the connecting frame (9), and its receiving end faces the side of the measuring wheel (7). It calculates the distance using the time difference of infrared beam reflection. When lateral movement or swinging occurs, the sensor sends a signal to the controller in real time.

6. The measurement method of an automatic measurement system for urban rail transit according to claim 5, characterized in that, Includes the following steps: Step S1: The system starts up, the controller initializes and establishes communication connections with the third sensor (24), the second sensor, the fourth sensor (26), the first camera (2) and the second camera (20); Step S2: The system moves along the track, the second camera (20) captures the front of the track in real time and transmits the video data to the controller, and the first camera (2) captures the outer surface of the track head in sync and transmits it to the controller; Step S3: During the movement, the third sensor (24) monitors the distance change from the gas spring housing to the swing block (23) in real time, the fourth sensor (26) monitors the distance change from the connecting frame to the side of the measuring wheel in real time, and the second sensor monitors the distance change between the ends of each pair of top blocks in real time. Each sensor transmits the monitoring data to the controller in real time. Step S4: The controller processes the video data transmitted by the second camera (20) using a preset algorithm to determine whether there are obstacles in front of the track; The image transmitted by the first camera (2) is processed by a preset algorithm to analyze whether the outer surface and shape of the rail head have changed. Step S5: The controller compares the analysis results of the first camera (2) with the monitoring data of the third sensor (24), the second sensor, and the fourth sensor (26) to determine whether the track has generated an abnormal signal that requires maintenance. Step S6: The controller transmits real-time monitoring data and analysis results to the cloud, and at the same time determines whether the monitoring data of each sensor exceeds the preset threshold. If it does, an alarm operation is executed.

7. The measurement method according to claim 6, characterized in that: The specific process for determining whether there is an obstacle in front of the track in step S4 is as follows: After the second camera (20) transmits the real-time video data to the controller, the controller extracts features from the video frame based on the obstacle recognition algorithm of multi-feature weighted matching, identifies the outline of abnormal objects in the picture, compares the outline parameters with the preset obstacle feature library, and determines that there is an obstacle when the matching degree exceeds the preset threshold. The formula for calculating the matching degree is: In the formula, M is the contour matching degree, n is the number of feature parameters, and w i Let Sim(P) be the weight of the i-th feature parameter (e.g., contour area weight 0.3, perimeter weight 0.2, vertex angle weight 0.5). i Q i ) represents the similarity between the i-th parameter of the contour to be identified and the corresponding parameter of the standard obstacle in the feature library, with a value range of [0,1]; when M≥0.7, it is determined that an obstacle exists.

8. The measurement method according to claim 7, characterized in that, The specific process of analyzing whether the outer surface and shape of the rail head have changed in step S4 is as follows: After the first camera (2) captures the image of the outer surface of the rail head, the controller performs edge detection and shape modeling on the image through the rail head deformation detection algorithm based on shape deviation analysis, calculates the deviation between the modeling result and the preset standard shape model of the rail head, and determines that the shape of the rail head has changed when the deviation value exceeds the preset range. The formula for calculating the deviation value is: In the formula, D is the shape deviation value (unit: mm), m is the number of sampling points, and S... j S represents the coordinates of the j-th sampling point after modeling the actual railhead. 0j These are the coordinates of the sampling points corresponding to the standard model; when D > 1 mm, it is determined that the rail head shape has changed.

9. The measurement method according to claim 6, characterized in that, The specific process of comprehensive comparison in step S5 is as follows: The controller integrates the rail head shape change data analyzed by the first camera (2), the distance change data monitored by the third sensor (24), the top block distance change data monitored by the second sensor, and the lateral distance change data monitored by the fourth sensor (26), and calculates the comprehensive anomaly index through the comprehensive anomaly evaluation algorithm of multi-source data fusion. When the comprehensive anomaly index exceeds the preset threshold, a signal that requires maintenance is generated. The formula for calculating the comprehensive anomaly index is as follows: In the formula, I is the comprehensive anomaly index, a, b, c, and d are weighting coefficients (satisfying a+b+c+d=1, such as taking 0.3, 0.2, 0.3, and 0.2 respectively), and D is the railhead shape deviation value. max =1mm is the rail head deviation threshold, ΔL3 is the distance change value monitored by the third sensor (24), ΔL 3max =1mm is its threshold, ΔL2 is the change in the distance monitored by the second sensor, ΔL 2max =1mm is its threshold, ΔL4 is the distance change value monitored by the fourth sensor (26), ΔL 4max Set a preset threshold for it; A maintenance signal is generated when I ≥ 0.

6.

10. The measurement method according to claim 6, characterized in that: The specific process of performing the alarm operation in step S6 is as follows: after the controller determines that the monitoring data exceeds the preset threshold, it controls the alarm to issue an audible and visual alarm signal on the one hand, and records the track position information where the abnormality occurred, the corresponding sensor monitoring data and the camera image on the other hand, and uploads the abnormal information to the cloud storage synchronously after marking it.

Images