Mountain area railway foreign matter invasion monitoring method and system

Abstract

The invention belongs to the technical field of rail traffic safety protection, and particularly relates to a mountain railway foreign matter invasion monitoring method and system. Comprising the steps that S1, structure sensing data of the passive protective net are collected, whether foreign matter invasion exists outside a boundary or not is judged according to the structure sensing data, and the structure sensing data comprise vibration and inclination angle data of stand columns of the passive protective net; s2, upper track sensing data of the track area are collected, whether foreign matter invasion exists in the boundary or not is judged according to the upper track sensing data, and the upper track sensing data comprise video image information in the boundary and detection information of a three-dimensional laser radar; and S3, processing the structure sensing data and the previous sensing data, and generating alarm information. The invention further provides a mountainous area railway foreign matter invasion monitoring system which comprises a structure sensing subsystem, an upper railway sensing subsystem, a data center and a monitoring terminal and is used for achieving comprehensive monitoring of foreign matter invasion outside and inside a boundary, and reliable guarantee is provided for safe operation of the mountainous area railway.

Description

Technical Field

[0001] The present invention belongs to the technical field of rail transit safety protection, and in particular relates to a method and system for monitoring the intrusion of foreign objects on mountain railways. Background Art

[0002] Railways have become an important part of my country's basic transportation facilities. my country has a vast territory and complex and diverse terrain, and there are many safety hazards in railway operations. Especially in the southwestern mountainous areas, railways are extremely susceptible to the intrusion of foreign objects caused by geological disasters such as mudslides, rock collapses and falling rocks. Passive protection nets are usually arranged in the above-mentioned areas to intercept falling rocks and mudslides to prevent them from invading the railway. If the passive protection net successfully intercepts but is not discovered in time, it will be a huge source of risk for the safe operation of the railway; if the falling rocks have a large energy or a large bounce height, they will penetrate or jump over the passive protection net and invade the track, endangering driving safety.

[0003] In addition to geological disasters, pedestrians or animals are also considered foreign objects. Existing foreign object intrusion monitoring methods mainly rely on video monitoring or three-dimensional laser scanning, and have achieved certain results, but each has certain defects:

[0004] Video surveillance captures images within the boundaries in real time and uses deep learning image recognition to monitor whether there are any foreign objects invading the images. This method has a high rate of false alarms and missed alarms in rainy and foggy weather and in scenes where lighting conditions are prone to change.

[0005] 3D laser scanning obtains 3D information of the target through 3D point cloud data, and can accurately obtain physical space environment information to determine whether there is foreign object intrusion. However, 3D point cloud does not contain semantic features. This method cannot classify objects. Therefore, normal train stops, maintenance operations by maintenance personnel, etc. cannot be distinguished from real foreign object intrusion incidents, and the false alarm rate is high.

[0006] The above two existing foreign body intrusion monitoring methods cannot timely discover and confirm the passive protection net intercepting the falling rocks, and cannot timely eliminate the risk sources outside the boundary. Summary of the invention

[0007] The purpose of the present invention is to provide a method and system for monitoring the intrusion of foreign objects on mountain railways in order to solve the problem that the existing technology cannot timely discover and confirm the rockfall intercepted by the passive protection net and cannot timely eliminate the risk source outside the limit. The present invention can timely discover and confirm the risk source of rockfall intercepted by the passive protection net outside the limit and accurately judge the intrusion of foreign objects on the road within the limit, so as to achieve the goal of "no foreign objects on the road and no collision on the road".

[0008] In order to achieve the above-mentioned object of the invention, the present invention provides the following technical solutions:

[0009] A method for monitoring foreign body intrusion on mountain railways, characterized in that it comprises the following steps:

[0010] S1. Collecting the structure perception data of the passive protection net, and judging whether there is foreign matter invading the limit outside the limit according to the structure perception data; the structure perception data includes the vibration and inclination data of the passive protection net column;

[0011] S2. Collecting the perception data of the track area, and judging whether there is foreign matter intrusion within the limit according to the perception data of the track area; the perception data of the track area includes video image information within the limit and detection information of the three-dimensional laser radar;

[0012] S3. Process the structure perception data and the road perception data to generate alarm information.

[0013] Preferably, step S1 comprises the following steps:

[0014] S11. Collect structural perception data of the passive protection net;

[0015] S12. The structure sensing data is transmitted to the gateway via Lora wireless communication for collection, and the gateway transmits the collected structure sensing data to the data center;

[0016] S13. The data center determines in real time whether the structural perception data exceeds a preset threshold; if it exceeds the threshold, the video image acquisition device is linked to capture images of the corresponding area of ​​the passive protection net and generate an alarm message; if it does not exceed the threshold, it returns to step S11 to continue acquisition.

[0017] This optimal solution refines the collection, transmission and processing of structure perception data, and achieves low-power, long-distance data transmission through Lora wireless communication without complex wiring, significantly reducing deployment costs. Through the linkage design of threshold judgment in the data center and video image acquisition equipment, alarm information can be quickly generated, improving response efficiency and system intelligence.

[0018] Preferably, step S2 comprises the following steps:

[0019] S21. Video image information is collected within the boundaries through a video image acquisition device;

[0020] S22. The edge computing terminal analyzes the video image information based on motion detection technology to determine whether there is a target within the boundary;

[0021] S23. If there is a target in S22, the target is detected by a three-dimensional laser radar, and the stagnation time of the target within the boundary is tracked; if there is no target in S22, the process returns to step S21 to continue collecting;

[0022] S24. When the stagnation time exceeds a preset time threshold, the size and three-dimensional coordinate information of the target are obtained by a three-dimensional laser radar;

[0023] S25. Convert the three-dimensional coordinate information into the plane coordinates of the video image acquisition device, link the video image acquisition device to capture the target image information, and transmit it to the data center via wireless communication.

[0024] Through the analysis of video image data by edge computing terminals and the linkage with three-dimensional lidar, accurate detection and classification of targets in the track area are achieved, and whether to generate alarm information is determined based on the stagnation time threshold, effectively improving the system's monitoring accuracy and response speed to potential threats.

[0025] Preferably, the data center classifies the image information of the target through image recognition and classification technology, and generates alarm information of different levels according to the classification results; the image recognition and classification technology adopts YOLOv4 deep learning technology. By classifying the target image information through the deep learning technology of the data center, intelligent recognition and classification processing of foreign objects are realized. The use of the YOLOv4 algorithm not only improves the classification accuracy, but also reduces the probability of missed reports and false reports, thereby ensuring the reliability of the alarm information.

[0026] A mountain railway foreign body intrusion monitoring system is also provided, comprising:

[0027] The structural perception subsystem is used to collect structural perception data of the passive protection net, and judge whether there are foreign objects invading the limit outside the boundary according to the structural perception data; the on-track perception subsystem is used to collect on-track perception data in the track area, and judge whether there are foreign objects invading the limit within the boundary according to the data; the on-track perception data includes video image information and detection information of three-dimensional laser radar; the data center is used to process the structural perception data and the on-track perception data, generate alarm information and send the alarm information to the monitoring terminal; the monitoring terminal is used to receive and display the alarm information, and the monitoring terminal includes a PC terminal and a mobile terminal.

[0028] The system framework provided by the present invention integrates the structural perception subsystem and the on-track perception subsystem, which can simultaneously monitor the intrusion of foreign objects both inside and outside the boundaries, and uniformly process the alarm information through the data center and push the alarm information to the terminal, thus building an efficient and intelligent railway safety monitoring system.

[0029] Preferably, the structural perception subsystem includes: a vibration and inclination integrated machine, which is used to collect vibration data and inclination data of the passive protection net in real time; a gateway, which is used to receive the collected data from multiple vibration and inclination integrated machines and transmit it to the data center via wireless communication. Through the cooperation between the vibration and inclination integrated machine and the gateway, efficient collection and centralized transmission of structural perception data are achieved. The application of wireless communication technology simplifies the deployment process, greatly reduces the demand for external power and wiring, and improves the applicability and deployment efficiency of the system.

[0030] Preferably, the vibration and tilting integrated machine is operated by solar power supply and adopts a vibration trigger mode. The vibration and tilting integrated machine solves the problem of inconvenient power supply in mountainous areas by adopting solar power supply and vibration trigger mode, prolongs the service life of the equipment, reduces daily maintenance costs, and significantly improves the adaptability of the system in remote areas.

[0031] Preferably, the on-road perception subsystem includes: a camera ball, which is used to collect video image information within the boundary in real time; a three-dimensional laser radar, which is used to detect and track targets within the boundary and obtain the three-dimensional coordinate information of the target; an edge computing terminal, which is used to analyze the video image information and the detection information of the three-dimensional laser radar to determine whether there is a foreign object invading the boundary; wherein the camera ball is linked with the three-dimensional laser radar, and the three-dimensional laser radar is only activated when the edge computing terminal determines that there is a target; the edge computing terminal is further used to calculate the stagnation time of the target within the boundary, and if the stagnation time exceeds a preset threshold, an alarm message is generated. Through the linkage design of the video acquisition device and the three-dimensional laser radar, it is ensured that the laser radar is only activated when necessary, reducing energy consumption and extending the life of the equipment. The stagnation time analysis function of the edge computing terminal further improves the system's recognition accuracy and monitoring efficiency of potential threats.

[0032] Preferably, the data center includes: an alarm information generation module for generating alarm information based on structural perception data and upper road perception data; an image recognition module for classifying target image information, wherein the image recognition module performs target classification based on the deep learning technology of YOLOv4; and a data push module for pushing alarm information to the monitoring terminal through the network. The modularly designed data center realizes the automatic processing of target classification and alarm generation in combination with deep learning technology. This design significantly improves the intelligence level of the system, and at the same time ensures the timeliness of information transmission through the function of pushing alarm information through the network.

[0033] Preferably, the on-track sensing subsystem further includes a fill light device, which is automatically turned on at night or in low light conditions. By adding the fill light device, the video monitoring capability of the system at night or in low light conditions is significantly improved, ensuring the reliability of all-weather monitoring. This design further expands the application scope of the system and provides technical support for railway foreign body monitoring in complex environments.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] The present invention provides a method for monitoring foreign object intrusion on mountain railways. Through the combination of a structural sensing subsystem and an on-track sensing subsystem, comprehensive monitoring of foreign object intrusion both inside and outside the boundary is achieved. Vibration and inclination data acquisition devices are used to monitor the status of the passive protection net in real time, which can quickly detect potential risks of falling rocks or mud and rock flow intrusion. At the same time, combined with a video image acquisition device and a three-dimensional laser radar, foreign objects in the track area can be accurately located to achieve accurate monitoring of foreign objects within the boundary. The method integrates and processes two types of data through a data center, generates alarm information, and pushes it to the monitoring terminal in a timely manner, providing efficient decision-making support for railway operators. Its innovation lies in the integration of multiple monitoring technologies, which solves the problems of high false alarm rate, high missed alarm rate, and insufficient processing timeliness of a single technology in existing methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A method flow chart of a method for monitoring foreign object intrusion on mountain railways according to Embodiment 1 of the present invention;

[0037] Figure 2 This is a structural diagram of a method and system for monitoring foreign object intrusion on mountain railways according to Embodiment 3 of the present invention;

[0038] Figure 3 This is a flow chart of the monitoring method of Example 3 of the present invention;

[0039] Figure 4 This is a structural diagram of a mountain railway foreign object intrusion monitoring system according to Example 4 of the present invention. DETAILED DESCRIPTION

[0040] The following is a further detailed description of a method and system for monitoring foreign object intrusion on mountain railways provided by the present invention in conjunction with the accompanying drawings and specific embodiments. However, this should not be understood as the scope of the above-mentioned subject matter of the present invention being limited to the following embodiments. All technologies implemented based on the content of the present invention belong to the scope of the present invention. Combined with the following description, the advantages and features of the present invention will become clearer. It should be noted that the drawings are all in a very simplified form and use non-precise proportions, which are only used to conveniently and clearly assist in explaining the purpose of the embodiments of the present invention.

[0041] Example 1

[0042] This embodiment provides a specific implementation method of a method for monitoring foreign object intrusion on mountain railways. Figure 1 As shown, the following steps are included:

[0043] S1. Collecting the structure perception data of the passive protection net, and judging whether there is foreign matter invading the limit outside the limit according to the structure perception data; the structure perception data includes the vibration and inclination data of the passive protection net column;

[0044] S2. Collecting the perception data of the track area, and judging whether there is foreign matter intrusion within the limit according to the perception data of the track area; the perception data of the track area includes video image information within the limit and detection information of the three-dimensional laser radar;

[0045] S3. Process the structure perception data and the road perception data to generate alarm information.

[0046] Preferably, step S1 comprises the following steps:

[0047] S11. Collect structural perception data of the passive protection net;

[0048] S12. The structure sensing data is transmitted to the gateway via Lora wireless communication for collection, and the gateway transmits the collected structure sensing data to the data center;

[0049] S13. The data center determines in real time whether the structural perception data exceeds a preset threshold; if it exceeds the threshold, the video image acquisition device is linked to capture images of the corresponding area of ​​the passive protection net and generate an alarm message; if it does not exceed the threshold, it returns to step S11 to continue acquisition.

[0050] This optimal solution refines the collection, transmission and processing of structure perception data, and achieves low-power, long-distance data transmission through Lora wireless communication without complex wiring, significantly reducing deployment costs. Through the linkage design of threshold judgment in the data center and video image acquisition equipment, alarm information can be quickly generated, improving response efficiency and system intelligence.

[0051] Preferably, step S2 comprises the following steps:

[0052] S21. Video image information is collected within the boundaries through a video image acquisition device;

[0053] S22. The edge computing terminal analyzes the video image information based on motion detection technology to determine whether there is a target within the boundary;

[0054] S23. If there is a target in S22, the target is detected by a three-dimensional laser radar, and the stagnation time of the target within the boundary is tracked; if there is no target in S22, the process returns to step S21 to continue collecting;

[0055] S24. When the stagnation time exceeds a preset time threshold, the size and three-dimensional coordinate information of the target are obtained by a three-dimensional laser radar;

[0056] S25. Convert the three-dimensional coordinate information into the plane coordinates of the video image acquisition device, link the video image acquisition device to capture the target image information, and transmit it to the data center via wireless communication.

[0057] Through the analysis of video image data by edge computing terminals and the linkage with three-dimensional lidar, accurate detection and classification of targets in the track area are achieved, and whether to generate alarm information is determined based on the stagnation time threshold, effectively improving the system's monitoring accuracy and response speed to potential threats.

[0058] Preferably, the data center classifies the image information of the target through image recognition and classification technology, and generates alarm information of different levels according to the classification results; the image recognition and classification technology adopts YOLOv4 deep learning technology. By classifying the target image information through the deep learning technology of the data center, intelligent recognition and classification processing of foreign objects are realized. The use of the YOLOv4 algorithm not only improves the classification accuracy, but also reduces the probability of missed reports and false reports, thereby ensuring the reliability of the alarm information.

[0059] Example 2

[0060] As an optimization of the above-mentioned embodiment, embodiment 2 provides a specific implementation method of a method for monitoring foreign object intrusion on mountain railways, including:

[0061] For the structure-aware part:

[0062] Step 1: The vibration and inclination data collection device collects the vibration and inclination data of the passive protection net column in real time. The vibration and inclination data collection device is an integrated device that can be directly charged by solar energy without any wiring;

[0063] Step 2: The vibration and inclination data of multiple passive protection net columns are transmitted to the gateway in the form of wireless communication for collection;

[0064] Step 3: The gateway transmits the collected vibration and inclination data of the passive protection net columns to the data center in the form of wireless communication;

[0065] Step 4: The data center will determine in real time whether the obtained passive protection net column vibration or inclination data exceeds the set threshold. If it exceeds the threshold, it will go to step 5, otherwise repeat step 1;

[0066] Step 5: The data center will link the corresponding video image acquisition device to capture a clear image of the corresponding position of the protective net. The data center will issue an alarm message of the corresponding level to promptly discover and confirm that the passive protective net has intercepted the falling rocks.

[0067] Furthermore, the vibration and inclination data acquisition device in step one adopts a vibration-triggered acquisition working mode. The device will be in a dormant state for a long time to reduce device power consumption, but will self-wake up once every 12 hours to collect and send data to determine whether the device is in an effective working state.

[0068] For the road perception part:

[0069] Step 1: The video image data acquisition device acquires video image information within the boundary in real time;

[0070] Step 2: The edge computing terminal device uses motion detection technology to determine whether there is an object within the boundary of the collected video image information. If there is an object, proceed to step 3, otherwise repeat step 1;

[0071] Step 3: The edge computing terminal activates the 3D laser radar to detect and track targets within the boundary;

[0072] Step 4: The edge computing terminal calculates the time that the 3D laser radar detects the target and stays within the boundary. If the stay time is greater than 15 seconds, go to step 5, otherwise repeat step 1.

[0073] Step 5: The edge computing terminal obtains target information, such as target size, coordinates, etc., through the 3D laser radar;

[0074] Step 6: The edge computing terminal converts the three-dimensional coordinate information of the target obtained by the three-dimensional laser radar into the plane coordinates of the video image data acquisition device, links the video image data acquisition device to capture the image information of the target, and transmits the image information to the data center in the form of wireless communication;

[0075] Step 7: The data center classifies the target image through image recognition and classification technology, determines the alarm level based on the classification results, and pushes the alarm information.

[0076] Example 3

[0077] As an optimization of the above-mentioned embodiment, embodiment 3 provides a specific implementation method of a method for monitoring foreign object intrusion on mountain railways, such as Figure 2 As shown, a method for monitoring foreign body intrusion on mountain railways can be divided into two parts: one is a structural sensing part, and the other is an on-track sensing part. The two parts work simultaneously and in parallel. Figure 3 As shown, the following steps are included:

[0078] For the structure-aware part:

[0079] Step 1: The vibration and inclination integrated machine collects the vibration and inclination data of the passive protection net column in real time, and the vibration and inclination integrated machine can be directly charged by solar energy without any wiring;

[0080] Step 2: The vibration and inclination data of multiple passive protection net columns are transmitted to the gateway in the form of Lora wireless communication for collection;

[0081] Step 3: The gateway transmits the collected vibration and inclination data of the passive protection net columns to the data center through the operator network or the railway private network;

[0082] Step 4: The data center will determine in real time whether the obtained passive protection net column vibration or inclination data exceeds the set threshold. If it exceeds the threshold, it will go to step 5, otherwise repeat step 1;

[0083] Step 5: The data center will link the corresponding camera to capture a clear image of the corresponding position of the protective net. The data center will issue an alarm message of the corresponding level to promptly discover and confirm that the passive protective net has intercepted the falling rocks.

[0084] Furthermore, the vibration and inclination integrated machine in step one adopts a vibration-triggered collection working mode. The device will be in a dormant state for a long time to reduce device power consumption, but will wake up the device every 12 hours to collect and send data to determine whether the device is in an effective working state.

[0085] For the road perception part:

[0086] Step 1: The camera ball machine collects video image information within the boundary in real time;

[0087] Step 2: The edge computing terminal device uses motion detection technology to determine whether there is an object within the boundary of the collected video image information. If there is an object, proceed to step 3, otherwise repeat step 1;

[0088] Step 3: The edge computing terminal activates the 3D laser radar to detect and track targets within the boundary;

[0089] Step 4: The edge computing terminal calculates the time that the 3D laser radar detects the target and stays within the boundary. If the stay time is greater than 15 seconds, go to step 5, otherwise repeat step 1.

[0090] Step 5: The edge computing terminal obtains target information, such as target size, coordinates, etc., through the 3D laser radar;

[0091] Step 6: The edge computing terminal converts the 3D coordinate information of the target obtained by the 3D laser radar into the plane coordinates of the camera ball machine, links the camera ball machine to capture the image information of the target, and transmits the image information to the data center through the operator network or the railway private network;

[0092] Step 7: The data center classifies the target image through YOLOv4-based deep learning technology, determines the alarm level based on the classification results, and pushes the alarm information.

[0093] like Figure 4 As shown, the data center obtains the monitoring data of the structural perception subsystem and the upper road perception subsystem in the monitoring system in real time, processes and analyzes the monitoring data, and determines whether it is necessary to send alarm information to the monitoring terminal. At the same time, the monitoring terminal can also view the monitoring data of the data center through the network at any time;

[0094] The monitoring terminal includes a PC terminal and a mobile terminal;

[0095] The structural sensing subsystem collects and uploads the vibration and inclination information of the passive protection net outside the boundary. When foreign objects such as falling rocks hit the passive protection net, an alarm is generated immediately, supplemented by video image confirmation, including a vibration and inclination integrated machine, a camera dome and a gateway;

[0096] The on-track sensing subsystem monitors the track area within the boundary in real time, collects video images and three-dimensional point cloud data within the boundary, and can accurately issue an alarm message in the first time when falling rocks penetrate or jump over the passive protection net and invade the track due to large energy or large bounce height, endangering driving safety. At the same time, it can effectively filter out other interference information, such as normally stopped trains, maintenance personnel, etc., including camera domes, three-dimensional lidar, edge computing terminals, and industrial switches.

[0097] The on-road sensing subsystem is also provided with a fill light device, which will automatically turn on at night.

[0098] Example 4

[0099] This embodiment provides a mountain railway foreign object intrusion monitoring system, including a structure sensing subsystem, an on-track sensing subsystem, a data center and a monitoring terminal.

[0100] 1. System functions and workflow

[0101] Structural sensing subsystem: used to collect vibration data and inclination data of the passive protection net, and monitor the intrusion of foreign objects outside the boundary in real time. When foreign objects such as falling rocks hit the protection net, the system can generate an alarm immediately and confirm the risk information through video imaging equipment.

[0102] On-track sensing subsystem: Real-time monitoring of foreign objects in the track area within the boundary, collecting video images and 3D laser radar detection data. When foreign objects are found in the track, the system combines the stagnation time analysis to determine the threat level, generates alarm information, and filters other interference information.

[0103] Data center: Receives and processes data from the two subsystems, and generates alarm information through fusion analysis. The data center includes an alarm information generation module, an image recognition module, and a data push module. Among them, the image recognition module classifies targets based on YOLOv4 deep learning technology to ensure the accuracy of alarm information.

[0104] Monitoring terminal: including PC terminal and mobile terminal, used to receive and display alarm information in real time, supporting users to view monitoring data through the network at any time.

[0105] 2. Optimal design of the structure perception subsystem

[0106] Vibration and tilt integrated machine: It is powered by solar energy and operates in vibration trigger mode, which solves the problem of inconvenient power supply in mountainous areas, while reducing equipment energy consumption and daily maintenance costs, and significantly improving adaptability and service life.

[0107] Gateway: Receives data from multiple vibration and tilt integrated machines and transmits it to the data center through wireless communication technology, realizing efficient centralized processing of data collection. This design simplifies the deployment process, reduces wiring requirements, and improves the applicability of the system in remote areas.

[0108] 3. Optimal design of the on-road sensing subsystem

[0109] Composition: It includes a camera dome, a 3D laser radar and an edge computing terminal. The camera dome collects video images within the boundary in real time, and the 3D laser radar is used to detect and track the target and obtain its 3D coordinate information. The edge computing terminal determines whether there is any foreign object invading the boundary by comprehensively analyzing the video and radar data.

[0110] Linkage design: The camera ball works with the 3D laser radar, and the laser radar is activated only when it detects a target, which reduces the energy consumption of the equipment and prolongs its service life.

[0111] Dead time analysis: The edge computing terminal further improves the accuracy of identifying potential threats by analyzing the dead time of the target and generates an alarm message when it exceeds the preset threshold.

[0112] 4. Modular design of data center

[0113] The data center integrates alarm generation, image recognition and data push functions, and realizes efficient automation of data processing through modular design. Using YOLOv4 deep learning technology, the system's accuracy in target classification is significantly improved. The network push module ensures that alarm information can be quickly delivered to the monitoring terminal, improving the timeliness of information transmission and monitoring efficiency.

[0114] 5. Application of fill light device

[0115] The on-track sensing subsystem further includes a fill light device, which automatically turns on at night or in low-light conditions, significantly improving the system's all-weather monitoring capabilities and providing strong technical support for railway foreign object monitoring in complex environments.

[0116] 6. Overall advantages of the system

[0117] The system comprehensively covers the monitoring needs inside and outside the boundary through the organic combination of the structural perception subsystem and the upper track perception subsystem. The unified processing of the data center and the real-time display of the monitoring terminal have built an efficient and intelligent railway safety monitoring system. The overall design takes into account the adaptability, intelligence and economy of the system, providing reliable safety protection for mountain railways.

[0118] This embodiment can timely discover and confirm the risk source of falling rocks intercepted by the passive protection net outside the boundary, and eliminate the risk source outside the boundary; the integration of three-dimensional laser radar and video technology, using their own advantages, can solve the problem of high video monitoring underreporting and high three-dimensional laser radar false alarms, improve the accuracy of the monitoring system and reduce the false alarm rate; the life of three-dimensional laser radar is short, such as the MTTF index of semi-solid laser radar is generally 12,000 hours (about 1.37 years), which is difficult to meet the design requirements of the railway foreign body intrusion monitoring system, and the equipment maintenance within the railway boundary is inconvenient, which is not conducive to the promotion and application of the system. This embodiment uses a video image acquisition device to obtain images for mobile detection, and when an object is found, the laser radar is started to detect and track the target, which reduces the startup time of the radar and extends the life of the radar, that is, reduces the need for equipment maintenance and reduces the cost of the monitoring system.

[0119] Those skilled in the art will appreciate that the embodiments of the present invention may be provided as methods, systems or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware.

[0120] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0121] The above-mentioned embodiments only express several implementation methods of the present invention, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A method for monitoring foreign body intrusion on mountain railways, characterized in that: The following steps are involved: S1. Collecting the structure perception data of the passive protection net, and judging whether there is foreign matter invading the boundary according to the structure perception data; the structure perception data includes the vibration and inclination data of the passive protection net column; S2. Collecting the perception data of the track area, and judging whether there is foreign matter intrusion within the limit according to the perception data of the track area; the perception data of the track area includes video image information within the limit and detection information of the three-dimensional laser radar; S3. Process the structure perception data and the road perception data to generate alarm information.

2. A method for monitoring foreign matter intrusion on mountain railways according to claim 1, characterized in that: Step S1 includes the following steps: S11. Collect structural perception data of the passive protection net; S12. The structure sensing data is transmitted to the gateway via Lora wireless communication for collection, and the gateway transmits the collected structure sensing data to the data center; S13. The data center determines in real time whether the structural perception data exceeds a preset threshold; if it exceeds the threshold, the video image acquisition device is linked to capture images of the corresponding area of ​​the passive protection net and generate an alarm message; if it does not exceed the threshold, it returns to step S11 to continue acquisition.

3. A method for monitoring foreign matter intrusion on mountain railways according to claim 2, characterized in that: Step S2 includes the following steps: S21. Video image information is collected within the boundaries through a video image acquisition device; S22. The edge computing terminal analyzes the video image information based on motion detection technology to determine whether there is a target within the boundary; S23. If there is a target in S22, the target is detected by a three-dimensional laser radar, and the stagnation time of the target within the boundary is tracked; if there is no target in S22, the process returns to step S21 to continue collecting; S24. When the stagnation time exceeds a preset time threshold, the size and three-dimensional coordinate information of the target are obtained by a three-dimensional laser radar; S25. Convert the three-dimensional coordinate information into the plane coordinates of the video image acquisition device, link the video image acquisition device to capture the target image information, and transmit it to the data center via wireless communication.

4. A method for monitoring foreign matter intrusion on mountain railways according to claim 3, characterized in that: The data center classifies the image information of the target through image recognition and classification technology, and generates alarm information of different levels according to the classification results; the image recognition and classification technology adopts YOLOv4 deep learning technology.

5. A mountain railway foreign object intrusion monitoring system, characterized in that: include: The structural perception subsystem is used to collect structural perception data of the passive protection net and determine whether there is foreign matter invading the boundary based on the structural perception data; The on-track sensing subsystem is used to collect on-track sensing data in the track area and determine whether there is foreign matter intruding within the boundary based on the data; the on-track sensing data includes video image information and detection information of three-dimensional laser radar; A data center, used for processing the structure sensing data and the upper road sensing data, generating alarm information and sending the alarm information to a monitoring terminal; The monitoring terminal is used to receive and display alarm information, and the monitoring terminal includes a PC terminal and a mobile terminal.

6. A mountain railway foreign object intrusion monitoring system according to claim 5, characterized in that: The structure perception subsystem includes: Vibration and inclination integrated machine, used to collect vibration data and inclination data of passive protection net in real time; The gateway is used to receive the collected data from multiple vibration and inclination integrated machines and transmit it to the data center through wireless communication.

7. A mountain railway foreign object intrusion monitoring system according to claim 6, characterized in that: The vibration and tilting integrated machine is operated by solar power supply and adopts a vibration triggering mode.

8. A mountain railway foreign object intrusion monitoring system according to claim 7, characterized in that: The on-road perception subsystem includes: The camera ball machine is used to collect video image information within the boundary in real time; 3D laser radar, used to detect and track targets within the boundary and obtain the 3D coordinate information of the target; An edge computing terminal is used to analyze the video image information and the detection information of the three-dimensional laser radar to determine whether there is any foreign object intrusion; Among them, the camera ball is linked with a three-dimensional laser radar, and the three-dimensional laser radar is only started when the edge computing terminal determines that there is a target; the edge computing terminal is further used to calculate the stagnation time of the target within the boundary, and if the stagnation time exceeds a preset threshold, an alarm message is generated.

9. A mountain railway foreign object intrusion monitoring system according to claim 8, characterized in that: The data center includes: An alarm information generation module, used to generate alarm information based on the structure perception data and the upper road perception data; An image recognition module is used to classify target image information, wherein the image recognition module performs target classification based on the deep learning technology of YOLOv4; The data push module is used to push alarm information to the monitoring terminal through the network.

10. A mountain railway foreign object intrusion monitoring system according to claim 9, characterized in that: The on-road sensing subsystem further includes a fill light device, which is automatically turned on at night or in a low-light environment.

Images