Urban rail transit fire rescue and inspection vehicle

Abstract

The application relates to the field of inspection vehicles and discloses a city rail transit fire-fighting rescue inspection vehicle, which comprises a vehicle compartment, second electrically-driven holders are fixedly connected to the four corners of the vehicle compartment, a high-definition camera is arranged at one end of the second electrically-driven holder, a smoke sensor is fixedly connected to one side of the top of the vehicle compartment, a mounting plate is fixedly connected to the other side of the top of the vehicle compartment, a pantograph is mounted on the top of the mounting plate, connecting boxes are fixedly connected to the bottom of the vehicle compartment in a symmetrical mode, audio acquisition equipment is arranged on the side away from the connecting boxes, and ultrasonic radars and single-line laser radar sensors are uniformly arranged on the bottom of the vehicle compartment. The smoke sensor is used for detecting the smoke concentration in the vehicle compartment and the surrounding environment in real time, and timely alarm is sent; the first electrically-driven holders, the fire-fighting water gun, the fire-fighting water belt, the electromagnetic valve and the water storage tank arranged on the top of the vehicle compartment in a symmetrical mode form a fire-fighting system; and the audible and light signals are sent by the audible and light alarm when the abnormal situation is detected.

Description

Technical Field

[0001] This invention relates to the field of inspection vehicle technology, specifically to an urban rail transit fire and rescue inspection vehicle. Background Technology

[0002] An inspection vehicle is a vehicle specifically designed to perform inspection tasks. It integrates a variety of advanced technologies and equipment and can conduct patrol inspections in different environments according to set routes and requirements. In urban rail transit, subway operation focuses on safety and punctuality, and is gradually becoming the first choice for urban residents' travel.

[0003] Urban subway lines are mostly designed as bidirectional single-track lines, with the majority of the lines located underground. Due to limited construction space and conditions, as well as the influence of factors such as urban planning and construction investment, they cannot be designed and laid out with multiple distribution lines or auxiliary lines like surface railways. Therefore, after an accident occurs on a subway train, it is not possible to organize other normal trains to overtake or detour as in the railway operation organization, which leads to congestion on the operating line, inability to provide timely rescue, and causes huge economic losses and negative social impacts. Therefore, a fire rescue and inspection vehicle for urban rail transit is proposed to improve the above situation. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an urban rail transit fire rescue and inspection vehicle, which solves the problem of being unable to provide timely rescue after a rail train accident.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a fire rescue and inspection vehicle for urban rail transit, comprising a carriage, with a second electric pan-tilt unit fixedly connected to each of the four corners of the carriage. A high-definition camera is installed at one end of each second electric pan-tilt unit. A smoke sensor is fixedly connected to one side of the top of the carriage, and a mounting plate is fixedly connected to the other side of the top of the carriage. A pantograph is installed on the top of the mounting plate. A connection box is symmetrically fixedly connected to the bottom of the carriage, and an audio acquisition device is installed on the side of the connection box away from the other side. Ultrasonic radar and single-line laser radar sensors are evenly arranged on the bottom of the carriage. A first electric pan-tilt unit is symmetrically fixedly connected to the top of the carriage. A fire nozzle is installed at one end of the first electric pan-tilt unit, and a fire hose is fixedly connected to one end of the fire nozzle. A solenoid valve is installed on the top of the fire hose, and a water storage tank is fixedly connected to the end of the fire hose away from the fire nozzle. An audible and visual alarm is fixedly installed on the top of the carriage.

[0006] Preferably, a control panel is provided at the bottom of the inner wall of the carriage, seats are fixedly connected to both sides of the inner wall of the carriage, a fire extinguisher is connected to one side of the bottom of the inner wall of the carriage, and lighting lamps are symmetrically fixedly connected to the top of the inner wall of the carriage.

[0007] Preferably, the bottom of the water tank is fixedly connected to the other side of the bottom of the inner wall of the carriage, the top of the water tank is fixedly connected to a connecting plate, a medical bed is fixedly connected to one side of the top of the connecting plate, a defibrillator is connected to the other side of the top of the connecting plate, a ventilator is connected to the top side of the connecting plate away from the defibrillator, and a folding stretcher is provided on the top side of the connecting plate away from the medical bed.

[0008] Preferably, a drive motor is fixedly installed on one side of the inner wall of the connecting box, a first pulley is fixedly installed at the output end of the drive motor, a belt is installed on the outer wall of the first pulley, a second pulley is connected to the inner wall of the belt, an axle is fixedly connected to the middle of the second pulley, and track wheels are evenly fixedly connected to both ends of the axle.

[0009] Preferably, the interior of the vehicle compartment is equipped with a battery compartment containing a spare battery. Doors are installed on both the front and rear sides of the vehicle compartment. Observation windows are evenly distributed on the outer walls of the vehicle compartment. Headlights are installed on both the left and right sides of the vehicle compartment.

[0010] A method for track flaw detection of an urban rail transit fire rescue inspection vehicle includes the following steps:

[0011] S1. Acoustic Flaw Detection: The ultrasonic radar on the inspection vehicle emits ultrasonic signals to the track and receives the reflected signals. By analyzing the amplitude, speed of sound, main frequency, and waveform changes of the reflected signals, the nature and size of internal track defects are determined. Combined with B-mode images, the type and location of track damage are identified. Simultaneously, normal track structure waveform data is collected to train a YOLO neural network model. The real-time track waveform data is input into the trained model to filter out the complete set of damaged waveforms. Image processing algorithms are then used to process the complete set of damaged waveforms to obtain the mileage and depth information of suspected defects.

[0012] S2. Audio Flaw Detection: The audio acquisition equipment on the inspection vehicle is used to collect wheel-rail vibration and noise data. The audio data is enhanced by high-frequency transformation. The short-time average energy, short-time average zero-crossing rate, 1 / 3 octave spectrum sound pressure and MFCC feature vector of the enhanced audio data are extracted. The vectors are input into the pre-established and trained XGboost model. The defects in the track are judged based on the model output.

[0013] S3. Assisted Flaw Detection: The single-line lidar sensor on the inspection vehicle is activated to collect track point cloud data. Based on the track's occlusion characteristics and height jumps, key points on the track top are detected. A track cross-section model is constructed and matched to determine the final track position. Kalman filtering is used to correlate the track top detection points. Simultaneously, track images are acquired using a camera. A deep learning object detection algorithm is used to identify track surface information in the images. Information obtained from an infrared rangefinder and temperature and humidity sensors is combined with audio flaw detection results and acoustic flaw detection results for data fusion. A decision fusion model is used to determine the type and location of faults and potential faults.

[0014] Preferably, in step S1, the step of processing the entire set of damaged waveforms using an image processing algorithm to obtain the mileage and depth information of suspected defects includes an edge detection algorithm, gray-level co-occurrence matrix calculation, and Fourier transform. The edge detection algorithm is used to extract the edge contour of the damaged area, the gray-level co-occurrence matrix calculation is used to obtain the texture features of the damaged area, and the Fourier transform is used to analyze the frequency domain features to determine the depth of damage. The depth of damage d satisfies the relationship d = v × t with the propagation speed v of the ultrasonic wave in the track and the time difference t between the ultrasonic wave and the reception of the ultrasonic wave in the damaged area.

[0015] Preferably, in step S2, where an audio acquisition device is used to collect wheel-rail vibration noise data, and the audio data is enhanced using a treble shift method, the treble shift method is expressed by the formula y(n) = x(n) × a. n The process is performed, where x(n) is the original audio signal, y(n) is the transformed audio signal, a is the pitch transformation factor, and n is the sampling point number.

[0016] Preferably, in step S2, where the vector is input into a pre-established and trained XGboost model, and the objective function of the XGboost model is determined based on the model output to identify any defects in the track, the objective function of the XGboost model is... in The loss function measures the predicted value. Compared with the true value y i The difference between them, Ω(f) k ) represents the regularization term, and K represents the number of decision trees.

[0017] Preferably, in step S3, where a camera is used to acquire track images and a deep learning target detection algorithm is used to identify track surface information in the images, the output of the deep learning target detection algorithm is O(x, y, c, p), where x and y are the coordinates of the detected target in the image, c is the category of the detected target, and p is the confidence level of the detected target.

[0018] Working principle: The drive motor rotates the first pulley and belt, which in turn drives the second pulley, causing the track wheels at both ends of the axle to rotate. This allows the inspection vehicle to move along the track. A control panel provides operators with a platform to control vehicle and equipment operation. Ultrasonic radar emits ultrasonic signals to the track and receives reflected signals. Audio acquisition equipment collects wheel-rail vibration noise data, and a single-line lidar sensor collects track point cloud data. The track vehicle can then perform comprehensive flaw detection on the track. A second electric pan-tilt unit and a high-definition camera further enhance the inspection capabilities. It can capture images of the surrounding environment of the carriage from all angles, thereby monitoring the track and surrounding conditions in real time. The smoke sensor detects the smoke concentration in the carriage and surrounding environment in real time and issues an alarm in a timely manner. The first electric pan-tilt unit, symmetrically set on the top of the carriage, together with the fire gun, fire hose, solenoid valve and water tank, constitutes a fire protection system. At the same time, the sound and light alarm will emit sound and light signals when abnormalities are detected. The medical rescue area is formed by setting up a medical bed, defibrillator, ventilator and folding stretcher on the connecting plate. Thus, the inspection vehicle integrates fire rescue, medical rescue and track flaw detection, which enhances its practicality.

[0019] This invention provides an urban rail transit fire rescue and inspection vehicle. It has the following beneficial effects:

[0020] 1. This invention uses a smoke sensor to detect the smoke concentration in the carriage and surrounding environment in real time and issue an alarm in a timely manner. The first electric pan-tilt unit, symmetrically arranged on the top of the carriage, together with fire hoses, fire nozzles, solenoid valves, and water tanks, constitutes a fire protection system. At the same time, an audible and visual alarm will emit audible and visual signals when an abnormality is detected, which improves the speed of emergency response. A medical bed, defibrillator, ventilator, and folding stretcher are set up on the connecting plate to form a medical rescue area, which solves the problem of not being able to provide timely rescue after a rail train accident.

[0021] 2. This invention transmits ultrasonic signals to the track using an ultrasonic radar and receives reflected signals. It also collects wheel-rail vibration noise data using an audio acquisition device and collects track point cloud data using a single-line lidar sensor. This allows the track vehicle to perform comprehensive flaw detection on the track, thereby enhancing the practicality of the invention. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the front right side structure of the present invention;

[0023] Figure 2 This is a schematic diagram of the front bottom side structure of the present invention;

[0024] Figure 3 This is a schematic diagram of the internal structure of the carriage of the present invention;

[0025] Figure 4This is a partial structural diagram of the control panel of the present invention;

[0026] Figure 5 This is a partial structural diagram of the drive motor of the present invention;

[0027] Figure 6 This is a flowchart of a method for track flaw detection of an urban rail transit fire rescue inspection vehicle proposed in this invention.

[0028] The components include: 1. Carriage; 2. Track wheels; 3. Observation window; 4. Car door; 5. Mounting plate; 6. Pantograph; 7. First electric pan-tilt unit; 8. Fire nozzle; 9. Fire hose; 10. Solenoid valve; 11. Audible and visual alarm; 12. Smoke sensor; 13. High-definition camera; 14. Second electric pan-tilt unit; 15. Headlights; 16. Connecting box; 17. Audio acquisition equipment; 18. Seat; 19. Control panel; 20. Fire extinguisher; 21. Backup battery; 22. Lighting; 23. Single-line lidar sensor; 24. Ultrasonic radar; 25. Water tank; 26. Connecting plate; 27. Medical bed; 28. Defibrillator; 29. ​​Ventilator; 30. Folding stretcher; 31. Battery compartment; 32. Drive motor; 33. First pulley; 34. Belt; 35. Second pulley; 36. Axle. Detailed Implementation

[0029] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0030] Example:

[0031] Please see the appendix Figure 1 - Appendix Figure 5This invention provides a fire rescue and inspection vehicle for urban rail transit, including a carriage 1. Each of the four corners of the carriage 1 is fixedly connected to a second electric pan-tilt unit 14. A high-definition camera 13 is installed at one end of each second electric pan-tilt unit 14. A smoke sensor 12 is fixedly connected to one side of the top of the carriage 1, and a mounting plate 5 is fixedly connected to the other side of the top of the carriage 1. A pantograph 6 is installed on the top of the mounting plate 5. A connecting box 16 is symmetrically fixedly connected to the bottom of the carriage 1. An audio acquisition device 17 is installed on the side of the connecting box 16 furthest from the bottom. Ultrasonic radar 24 and single-line laser radar sensors 23 are evenly arranged on the bottom of the carriage 1. A first electric pan-tilt unit 7 is symmetrically fixedly connected to the top of the carriage 1. A fire nozzle 8 is installed at one end of the first electric pan-tilt unit 7. A fire hose 9 is fixedly connected to one end of the fire nozzle 8. A solenoid valve 10 is installed on the top of the fire hose 9. A water storage tank 25 is fixedly connected to the end of the fire hose 9 furthest from the fire nozzle 8. An audible and visual alarm 11 is fixedly installed on the top of the carriage 1.

[0032] Specifically, a second electric gimbal 14 and a high-definition camera 13 are installed at the four corners of the carriage 1. Through the flexible rotation of the second electric gimbal 14, the high-definition camera 13 can capture the surrounding environment of the carriage from all directions and multiple angles, thereby monitoring the track and surrounding conditions in real time, and promptly detecting potential safety hazards, such as foreign objects on the track or personnel intrusion, thus realizing the visual monitoring of the track environment.

[0033] The smoke sensor 12 can detect the smoke concentration in the carriage and surrounding environment in real time, thereby quickly sensing smoke signals in the early stage of a fire and issuing an alarm in a timely manner, buying valuable time for personnel evacuation and fire fighting. The first electric pan-tilt unit 7, symmetrically arranged on the top of the carriage 1, together with the fire nozzle 8, fire hose 9, solenoid valve 10 and water storage tank 25, constitute a fire protection system. The first electric pan-tilt unit 7 adjusts the spray angle of the fire nozzle 8, and combined with the water source provided by the water storage tank 25, water can be quickly delivered to the fire source for fire fighting when a fire occurs. The solenoid valve 10 can control the water flow, thereby achieving precise fire fighting and effectively responding to track fires. At the same time, the audible and visual alarm 11 emits audible and visual signals when abnormal situations (such as fire, track faults, etc.) are detected, thereby alerting the surrounding personnel to safety and assisting personnel in quickly locating the accident scene. This achieves effective reminders to the surrounding personnel, improves the speed of emergency response, and solves the problem of not being able to provide timely rescue after a track train accident.

[0034] The ultrasonic radar 24 emits ultrasonic signals to the track and receives reflected signals, the audio acquisition device 17 collects wheel-rail vibration noise data, and the single-line lidar sensor 23 collects track point cloud data, thereby detecting flaws in the track and enhancing the practicality of the inspection vehicle.

[0035] Please see the appendix Figure 1 - Appendix Figure 5 The bottom of the inner wall of the carriage 1 is equipped with a control panel 19. Seats 18 are fixedly connected to both sides of the inner wall of the carriage 1. A fire extinguisher 20 is connected to one side of the bottom of the inner wall of the carriage 1. Lighting lamps 22 are symmetrically fixedly connected to the top of the inner wall of the carriage 1. The bottom of the water tank 25 is fixedly connected to the other side of the bottom of the inner wall of the carriage 1. A connecting plate 26 is fixedly connected to the top of the water tank 25. A medical bed 27 is fixedly connected to one side of the top of the connecting plate 26. A defibrillator 28 is connected to the other side of the top of the connecting plate 26. A ventilator 29 is connected to the top side of the connecting plate 26 away from the defibrillator 28. A folding stretcher 30 is provided on the top side of the connecting plate 26 away from the medical bed 27.

[0036] Specifically, the control panel 19 is located at the bottom of the inner wall of the carriage 1, providing a platform for operators to control vehicle operation and equipment operation. The fire extinguisher 20 is connected to one side of the bottom of the inner wall of the carriage, which can be quickly used to extinguish fires in the early stage of a fire or in a small-scale fire. The medical rescue area is formed by the medical bed 27, defibrillator 28, ventilator 29 and folding stretcher 30 set on the connecting plate 26. When an accident occurs and people are injured, these devices can be used to provide on-site treatment and emergency care to the injured, such as using a defibrillator to provide first aid to patients with cardiac arrest and using a ventilator to assist breathing, thereby improving the survival rate of the injured.

[0037] Please see the appendix Figure 1 - Appendix Figure 5 A drive motor 32 is fixedly installed on one side of the inner wall of the connecting box 16. A first pulley 33 is fixedly installed at the output end of the drive motor 32. A belt 34 is installed on the outer wall of the first pulley 33. A second pulley 35 is connected to the inner wall of the belt 34. An axle 36 is fixedly connected to the middle of the second pulley 35. Track wheels 2 are evenly fixedly connected to both ends of the axle 36. A battery compartment 31 is installed inside the carriage 1. A spare battery 21 is installed in the battery compartment 31 inside the carriage 1. Doors 4 are installed on both the front and rear sides of the carriage 1. Observation windows 3 are evenly arranged on the outer wall of the carriage 1. Headlights 15 are installed on both the left and right sides of the carriage 1.

[0038] Specifically, the operation of the drive motor 32 drives the first pulley 33 and belt 34, and at the same time drives the second pulley 35 to rotate, thereby causing the track wheels 2 at both ends of the wheel axle 36 to rotate, thus enabling the inspection vehicle to move on the track and travel along the predetermined route. The backup battery 21 in the battery compartment 31 provides emergency power for the vehicle when there is no external power supply or the power supply is abnormal, thereby ensuring the normal operation of the vehicle's key equipment (such as lighting, communication, and some detection equipment).

[0039] Please see the appendix Figure 6 A method for track flaw detection in urban rail transit fire rescue inspection vehicles includes the following steps:

[0040] S1. Acoustic Flaw Detection: The ultrasonic radar 24 on the inspection vehicle emits ultrasonic signals to the track and receives reflected signals. By analyzing the amplitude, speed of sound, main frequency, and waveform changes of the reflected signals, the nature and size of internal defects in the track are determined. Combined with B-mode images, the type and location of track damage are identified. At the same time, normal track structure waveform data are collected to train a YOLO neural network model. The real-time track waveform data is input into the trained model to filter out the complete set of damaged waveforms. Image processing algorithms are then used to process the complete set of damaged waveforms to obtain the mileage and depth information of suspected defects.

[0041] Specifically, the ultrasonic radar 24 transmits ultrasonic signals to the track and receives reflected signals. By utilizing the differences in the propagation characteristics of ultrasonic waves in different media, and based on the amplitude, speed, dominant frequency, and waveform changes of the reflected signals, analysts can determine the nature and size of defects inside the track. Combined with B-mode images, they can intuitively identify the type and location of track damage, thereby achieving high-precision detection of the internal structure of the track.

[0042] S2. Audio Flaw Detection: The audio acquisition device 17 on the inspection vehicle is used to collect wheel-rail vibration noise data. The audio data is enhanced by high-frequency transformation. The short-time average energy, short-time average zero-crossing rate, 1 / 3 octave spectrum sound pressure and MFCC feature vector of the enhanced audio data are extracted. The vectors are input into the pre-established and trained XGboost model. The defects in the track are judged based on the model output.

[0043] Specifically, the audio acquisition device 17 is used to collect wheel-rail vibration noise data. Since the vibration noise characteristics of the track are different under different conditions (normal or defective), the audio data is enhanced by high-frequency transformation, which can expand the data sample size and improve the generalization ability of the model. The short-time average energy, short-time average zero-crossing rate, 1 / 3 octave spectrum sound pressure and MFCC feature vector of the enhanced audio data are extracted. These feature vectors are input into the pre-established and trained XGboost model. Based on the model output, the defects of the track are judged, thereby realizing the rapid and accurate judgment of the track status using audio features.

[0044] S3. Assisted Flaw Detection: The single-line lidar sensor 23 on the inspection vehicle is activated to collect track point cloud data. Based on the track's occlusion characteristics and height jumps, key points on the track top are detected. A track cross-section model is constructed and matched to determine the final track position. Kalman filtering is used to associate the track top detection points. At the same time, track images are collected by a camera. A deep learning target detection algorithm is used to identify track surface information in the images. Information obtained by an infrared rangefinder and temperature and humidity sensor is combined with audio flaw detection results and acoustic flaw detection results for data fusion. The type and location of faults and potential faults are determined through a decision fusion model.

[0045] Specifically, track point cloud data is collected by single-line lidar sensor 23. Based on the occlusion characteristics of the track to the surrounding environment and the detection of key points on the track top according to height jumps, a track cross-section model is constructed and matched to determine the final track position. Kalman filtering is used to associate the detection points on the track top, thereby accurately obtaining the geometric shape and position information of the track, and realizing high-precision modeling and real-time monitoring of the track structure.

[0046] In step S1, the image processing algorithm is used to process the entire set of damaged waveforms to obtain the mileage and depth information of suspected defects. The image processing algorithm includes an edge detection algorithm, gray-level co-occurrence matrix calculation, and Fourier transform. The edge detection algorithm is used to extract the edge contour of the damaged area, the gray-level co-occurrence matrix calculation is used to obtain the texture features of the damaged area, and the Fourier transform is used to analyze the frequency domain features to determine the damage depth. The damage depth d satisfies the relationship d = v × t with the propagation speed v of the ultrasonic wave in the track and the time difference t of the ultrasonic wave from transmission to reception through the damaged area.

[0047] In step S2, audio acquisition equipment is used to collect wheel-rail vibration noise data. In the step of enhancing the audio data using a treble shift method, the treble shift method is calculated using the formula y(n) = x(n) × a. n The process is performed, where x(n) is the original audio signal, y(n) is the transformed audio signal, a is the pitch transformation factor, and n is the sampling point number.

[0048] In step S2, where the vector is input into the pre-built and trained XGboost model, and the objective function of the XGboost model is determined based on the model output to identify any defects in the orbit, the objective function of the XGboost model is... in The loss function measures the predicted value. Compared with the true value y i The difference between them, Ω(f) k ) represents the regularization term, and K represents the number of decision trees.

[0049] In step S3, which involves acquiring track images using a camera and identifying track surface information in the images using a deep learning object detection algorithm, the output of the deep learning object detection algorithm is O(x, y, c, p), where x and y are the coordinates of the detected target in the image, c is the category of the detected target, and p is the confidence level of the detected target.

[0050] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A fire and rescue inspection vehicle for urban rail transit, comprising a carriage (1), characterized in that: Each of the four corners of the carriage (1) is fixedly connected to a second electric pan-tilt unit (14). A high-definition camera (13) is installed at one end of each second electric pan-tilt unit (14). A smoke sensor (12) is fixedly connected to one side of the top of the carriage (1). A mounting plate (5) is fixedly connected to the other side of the top of the carriage (1). A pantograph (6) is installed on the top of the mounting plate (5). A connecting box (16) is symmetrically fixedly connected to the bottom of the carriage (1). An audio acquisition device (17) is installed on the side of the connecting box (16) furthest from the other side. The bottom is uniformly equipped with ultrasonic radar (24) and single-line laser radar sensor (23). The top of the carriage (1) is symmetrically and fixedly connected with a first electric pan-tilt unit (7). A fire gun (8) is installed at one end of the first electric pan-tilt unit (7). A fire hose (9) is fixedly connected at one end of the fire gun (8). A solenoid valve (10) is installed on the top of the fire hose (9). A water storage tank (25) is fixedly connected at the end of the fire hose (9) away from the fire gun (8). An audible and visual alarm (11) is fixedly installed on the top of the carriage (1). The bottom of the water tank (25) is fixedly connected to the other side of the bottom of the inner wall of the carriage (1). A connecting plate (26) is fixedly connected to the top of the water tank (25). A medical bed (27) is fixedly connected to one side of the top of the connecting plate (26). A defibrillator (28) is connected to the other side of the top of the connecting plate (26). A ventilator (29) is connected to the top side of the connecting plate (26) away from the defibrillator (28). A folding stretcher (30) is provided on the top side of the connecting plate (26) away from the medical bed (27). A drive motor (32) is fixedly installed on one side of the inner wall of the connecting box (16). A first pulley (33) is fixedly installed at the output end of the drive motor (32). A belt (34) is installed on the outer wall of the first pulley (33). A second pulley (35) is connected to the inner wall of the belt (34). An axle (36) is fixedly connected to the middle of the second pulley (35). Track wheels (2) are evenly fixedly connected to both ends of the axle (36).

2. The urban rail transit fire rescue inspection vehicle according to claim 1, characterized in that: The bottom of the inner wall of the carriage (1) is provided with a control panel (19), and seats (18) are fixedly connected to both sides of the inner wall of the carriage (1). A fire extinguisher (20) is connected to one side of the bottom of the inner wall of the carriage (1), and lighting lamps (22) are symmetrically fixedly connected to the top of the inner wall of the carriage (1).

3. The urban rail transit fire rescue inspection vehicle according to claim 1, characterized in that: The interior of the carriage (1) is equipped with a battery compartment (31), and a spare battery (21) is installed in the battery compartment (31) inside the carriage (1). Doors (4) are installed on both the front and rear sides of the carriage (1). Observation windows (3) are evenly arranged on the outer wall of the carriage (1). Headlights (15) are installed on both the left and right sides of the carriage (1).

4. A method for track flaw detection in urban rail transit fire rescue inspection vehicles, characterized in that: The method applied to the urban rail transit fire rescue inspection vehicle according to any one of claims 1-3 includes the following steps: S1. Acoustic flaw detection: The ultrasonic radar (24) on the inspection vehicle is used to transmit ultrasonic signals to the track and receive reflected signals. By analyzing the amplitude, speed of sound, main frequency and waveform changes of the reflected signals, the nature and size of internal defects in the track are determined. Combined with B-display images, the type and location of track damage are determined. At the same time, the normal structure waveform data of the track is collected to train the YOLO neural network model. The real-time track waveform data is input into the trained model to filter out the complete set of damaged waveforms. The image processing algorithm is used to process the complete set of damaged waveforms to obtain the mileage and depth information of suspected defects. S2, Audio Flaw Detection: Use the audio acquisition device (17) on the inspection vehicle to collect wheel-rail vibration noise data, enhance the audio data through high-frequency transformation, extract the short-time average energy, short-time average zero-crossing rate, 1 / 3 octave spectrum sound pressure and MFCC feature vector of the enhanced audio data, input the vector into the pre-established and trained XGboost model, and judge the defects in the track based on the model output. S3, Auxiliary Flaw Detection: The single-line lidar sensor (23) on the inspection vehicle is activated to collect track point cloud data. Based on the occlusion characteristics of the track to the surrounding environment and the height jump detection of key points on the track top, a track cross-section model is constructed and matched to determine the final track position. Kalman filtering is used to associate the track top detection points. At the same time, the camera is used to collect track images. A deep learning target detection algorithm is used to identify track surface information in the images. The information obtained by the infrared rangefinder and temperature and humidity sensor is combined with the audio flaw detection results and the acoustic flaw detection results to perform data fusion. The type and location of faults and potential faults are determined through the decision fusion model. In step S1, the image processing algorithm is used to process the entire set of damaged waveforms to obtain the mileage and depth information of suspected defects. The image processing algorithm includes an edge detection algorithm, a gray-level co-occurrence matrix calculation, and a Fourier transform. The edge detection algorithm is used to extract the edge contour of the damaged area, the gray-level co-occurrence matrix calculation is used to obtain the texture features of the damaged area, and the Fourier transform is used to analyze the frequency domain features to determine the damage depth. The damage depth d satisfies the relationship d = v × t with the propagation speed v of the ultrasonic wave in the track and the time difference t between the ultrasonic wave transmission and reception in the damaged area. In step S2, where audio acquisition equipment is used to collect wheel-rail vibration noise data, and audio data enhancement is performed using a treble shift method, the treble shift method is expressed by the formula y(n) = x(n) × a. n The process is performed, where x(n) is the original audio signal, y(n) is the transformed audio signal, a is the pitch transformation factor, and n is the sampling point number.

5. A method for track flaw detection of an urban rail transit fire rescue inspection vehicle according to claim 4, characterized in that: In step S2, where the vector is input into a pre-built and trained XGboost model, and the objective function of the XGboost model is determined based on the model output to identify any defects in the orbit, the objective function of the XGboost model is... in The loss function measures the predicted value. Compared with the true value y i The difference between them, Ω(f) k ) represents the regularization term, and K represents the number of decision trees.

6. A method for track flaw detection of an urban rail transit fire rescue inspection vehicle according to claim 4, characterized in that: In step S3, where a camera is used to acquire track images and a deep learning target detection algorithm is used to identify track surface information in the images, the output of the deep learning target detection algorithm is O(x, y, c, p), where x and y are the coordinates of the detected target in the image, c is the category of the detected target, and p is the confidence level of the detected target.

Images