High and steep slope monitoring method and device based on Beidou positioning technology

Abstract

The invention provides a high and steep slope monitoring method and device based on a Beidou positioning technology. The method comprises the following steps: acquiring route mileage information of a to-be-monitored high and steep slope; dividing the high and steep slope to be monitored into a short-distance high and steep slope and a long-distance high and steep slope according to the route mileage information; performing fixed-point monitoring on the short-distance high and steep slope by adopting a Beidou positioning technology to obtain fixed-point monitoring data; carrying out distributed monitoring on the long-distance high and steep slope by adopting a multi-sensor fusion technology to obtain distributed monitoring data; and performing analysis based on the fixed-point monitoring data and the distributed monitoring data, and giving out an early warning when the analysis result is that the security level is exceeded. Based on the Beidou high-precision positioning technology, fixed-point monitoring is carried out on the short-distance high and steep slope with the short line mileage range, fixed-point monitoring is carried out on the long-distance high and steep slope, and therefore monitoring of the safety state of the high and steep slope, dangerous situation alarming and advanced early warning are achieved.

Description

Technical Field

[0001] The present disclosure relates to the technical field of slope monitoring, and in particular to a monitoring method and device for high and steep slope monitoring based on Beidou positioning technology. Background Art

[0002] The high and steep slopes along the heavy-load railway and above the tunnel entrances are complex and have steep terrain. Due to human construction, unauthorized entry and destruction of infrastructure, and rain erosion, natural disasters such as landslides and mud-rock flows are prone to occur, posing a serious threat to the safe operation of heavy-load railway transport vehicles.

[0003] Although a variety of safety status monitoring methods for high and steep slopes have been developed, most of these monitoring methods for high and steep slopes with short line mileage have low accuracy and make it difficult to achieve displacement monitoring and super-strong early warning for landslides, mudslides, etc. Summary of the invention

[0004] The embodiments of the present disclosure provide a monitoring method and device for high and steep slope monitoring based on Beidou positioning technology, which are used to realize displacement monitoring and advance warning of landslides, mud-rock flows, etc.

[0005] In a first aspect, the present disclosure provides a monitoring method for high and steep slope monitoring based on Beidou positioning technology, characterized in that it includes:

[0006] Obtain line mileage information of high and steep slopes to be monitored;

[0007] According to the line mileage information, the high and steep slopes to be monitored are divided into short-distance high and steep slopes and long-distance high and steep slopes;

[0008] Beidou positioning technology is used to carry out fixed-point monitoring of the short-distance high and steep slopes to obtain fixed-point monitoring data;

[0009] Multi-sensor fusion technology is used to perform distributed monitoring of the long-distance high and steep slopes to obtain distributed monitoring data;

[0010] An analysis is performed based on the fixed-point monitoring data and the distributed monitoring data, and an early warning is issued when the analysis result exceeds the safety level.

[0011] In an exemplary embodiment, the fixed-point monitoring data includes: sensor monitoring data, positioning raw data and three-dimensional coordinates of the monitoring point; a Beidou monitoring station is deployed at the monitoring point on the short-distance high and steep slope; a variety of sensors are integrated in the Beidou monitoring station; the Beidou monitoring station includes an above-ground part and an underground part; the above-ground part includes a main box; the main box includes a monitoring receiver; the Beidou positioning technology is used to perform fixed-point monitoring on the short-distance high and steep slope to obtain fixed-point monitoring data, including:

[0012] The sensor is used to monitor the short-distance high and steep slope to obtain sensor monitoring data;

[0013] The monitoring receiver is used to receive the sensor monitoring data and synchronously output the original positioning data;

[0014] The monitoring points are distinguished according to the IP address and port number of the monitoring receiver, and the three-dimensional coordinates of all monitoring points are calculated.

[0015] In an exemplary embodiment, the distributed monitoring data includes: vibration signals, alarm information and video information; positioning vibration optical fibers and laser cameras are laid on the monitoring points on the long-distance high and steep slopes; multi-sensor fusion technology is used to perform distributed monitoring on the long-distance high and steep slopes to obtain distributed monitoring data, including:

[0016] Based on the vibration signal, the type and orientation of the intruding foreign object are determined; the vibration signal is output through an interferometer connected to the positioning type vibration optical fiber when the positioning type vibration optical fiber senses vibration in the laying area;

[0017] The vibration signal is converted into an electrical signal by the photoelectric detector of the positioning type vibration optical fiber; the electrical signal is used to determine abnormal information and generate corresponding alarm information;

[0018] The laser camera responds to the alarm information to identify the position of the foreign object and collect the video information.

[0019] In an exemplary embodiment, determining the type and position of the intruding foreign object based on the vibration signal includes:

[0020] Using an optical fiber vibration signal denoising algorithm to denoise the vibration signal;

[0021] Using a positioning algorithm, the vibration event is positioned on the denoised vibration signal to obtain the type and position of the foreign object;

[0022] The vibration characteristics of the vibration signal after denoising are analyzed to determine the type of the foreign matter.

[0023] In an exemplary embodiment, performing analysis based on the fixed-point monitoring data and the distributed monitoring data, and issuing an early warning when the analysis result exceeds the safety level, includes:

[0024] Determining the change offset of the fixed-point monitoring data, and classifying the offset excess level according to the change offset;

[0025] Determine the foreign object intrusion event category according to the offset excess level, the foreign object category and the foreign object orientation;

[0026] Based on the offset exceeding level and / or the foreign object intrusion event category, determine whether it exceeds the safety level. If so, generate an alarm and notify the management personnel of the corresponding level; the safety level is divided according to the safety range of the high and steep slope to be monitored.

[0027] In an exemplary embodiment, determining the category of the foreign object intrusion event according to the offset exceeding level, the foreign object category and the foreign object orientation includes:

[0028] A decision tree model is constructed using a preset training set, and after integration and evaluation of the decision tree model, a target decision tree model is obtained; the training set is characterized by the training offset over-limit level and the training foreign body intrusion event category, and the training security level is a label;

[0029] The offset excess level, the foreign object category and the foreign object orientation are input into the target decision tree model to obtain the foreign object intrusion event category.

[0030] In a second aspect, the present disclosure provides a monitoring device for high and steep slope monitoring based on Beidou positioning technology, comprising:

[0031] An acquisition module is used to obtain the line mileage information of the high and steep slope to be monitored;

[0032] A division module, used for dividing the high and steep slopes to be monitored into short-distance high and steep slopes and long-distance high and steep slopes according to the line mileage information;

[0033] A fixed-point monitoring module is used to use Beidou positioning technology to perform fixed-point monitoring on the short-distance high and steep slope to obtain fixed-point monitoring data;

[0034] A distributed monitoring module is used to perform distributed monitoring of the long-distance high and steep slopes using multi-sensor fusion technology to obtain distributed monitoring data;

[0035] The analysis module is used to perform analysis based on the fixed-point monitoring data and the distributed monitoring data, and issue an early warning when the analysis result exceeds the safety level.

[0036] In a third aspect, the present disclosure provides an electronic device, comprising a processor and a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the steps in the method provided in the first aspect are executed.

[0037] In a fourth aspect, the present disclosure provides a storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps in the method provided in the first aspect are executed.

[0038] In a fifth aspect, the present disclosure provides a computer program product, comprising a computer program, which, when executed by a processor, runs the steps in the method provided in the first aspect above.

[0039] It can be seen from the above technical solutions that the present disclosure has the following advantages:

[0040] The present disclosure provides a monitoring method and device for high-steep slope monitoring based on Beidou positioning technology, the method comprising: obtaining line mileage information of the high-steep slope to be monitored; dividing the high-steep slope to be monitored into short-distance high-steep slopes and long-distance high-steep slopes according to the line mileage information; using Beidou positioning technology to perform fixed-point monitoring on the short-distance high-steep slope to obtain fixed-point monitoring data; using multi-sensor fusion technology to perform distributed monitoring on the long-distance high-steep slope to obtain distributed monitoring data; performing analysis based on the fixed-point monitoring data and the distributed monitoring data, and issuing an early warning when the analysis result exceeds the safety level. Based on Beidou high-precision positioning technology, fixed-point monitoring is performed on the short-distance high-steep slope with a short line mileage range, and fixed-point monitoring is performed on the long-distance high-steep slope, thereby realizing the monitoring of the safety status of the high-steep slope, emergency alarm and advance warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0042] Figure 1 The following is a flowchart showing the steps of a monitoring method for high and steep slope monitoring based on Beidou positioning technology provided in the first embodiment of the present disclosure;

[0043] Figure 2 A flowchart of the steps of a monitoring method for high and steep slope monitoring based on Beidou positioning technology provided in the second embodiment of the present disclosure is exemplarily shown;

[0044] Figure 3 The structural block diagram of a monitoring device for high and steep slope monitoring based on Beidou positioning technology provided in the third embodiment of the present disclosure is exemplarily shown. DETAILED DESCRIPTION

[0045] The embodiments of the present disclosure provide a monitoring method and device for high and steep slope monitoring based on Beidou positioning technology, which are used to realize displacement monitoring and advance warning of landslides, mud-rock flows, etc.

[0046] In order to make the purpose, features, and advantages of the invention disclosed in the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] Embodiment 1

[0048] See also Figure 1 , Figure 1 The following is a flowchart of the steps of the shore-based ship posture measurement method provided by the first embodiment of the present disclosure, and the method specifically includes:

[0049] Step S101, obtaining line mileage information of the high and steep slope to be monitored;

[0050] In the embodiment of the present application, the line mileage information of the high and steep slope to be monitored is obtained by collection or input, and the information includes the starting point, the end point and the overall length of the slope.

[0051] Step S102, dividing the high and steep slopes to be monitored into short-distance high and steep slopes and long-distance high and steep slopes according to the line mileage information;

[0052] In an embodiment of the present application, based on the acquired line mileage information, the high and steep slopes to be monitored are divided into two categories: short-distance high and steep slopes and long-distance high and steep slopes, so that different monitoring strategies can be adopted for slopes of different lengths.

[0053] Step S103, using Beidou positioning technology to perform fixed-point monitoring on the short-distance high and steep slope to obtain fixed-point monitoring data;

[0054] In the embodiment of the present application, for short-distance high-steep slopes, Beidou high-precision positioning technology is used for fixed-point monitoring to provide high-precision location information and obtain fixed-point monitoring data of short-distance slopes. The fixed-point monitoring data includes: sensor monitoring data, original positioning data and three-dimensional coordinates of monitoring points; Beidou monitoring stations are deployed at monitoring points on short-distance high-steep slopes; multiple sensors are integrated in the Beidou monitoring station; the Beidou monitoring station includes an above-ground part and an underground part; the above-ground part includes a main box; the main box includes a monitoring receiver.

[0055] Step S104, using multi-sensor fusion technology to perform distributed monitoring on the long-distance high and steep slope to obtain distributed monitoring data;

[0056] In the embodiment of the present application, for long-distance high and steep slopes, multi-sensor fusion technology is used to integrate multiple sensors (such as displacement sensors and strain sensors, etc.) to monitor at different locations on the slope to obtain distributed monitoring data. Among them, the distributed monitoring data includes: vibration signals, alarm information and video information; positioning vibration optical fibers and laser cameras are laid on the monitoring points on the long-distance high and steep slopes.

[0057] Step S105, performing analysis based on the fixed-point monitoring data and the distributed monitoring data, and issuing an early warning when the analysis result exceeds the safety level.

[0058] In the embodiment of the present application, the environmental condition of the slope is determined by analyzing the fixed-point monitoring data and the distributed monitoring data. If the analysis result shows that the environmental condition analysis result of the slope exceeds the preset safety level, an early warning is issued to remind relevant personnel to take measures.

[0059] The monitoring method for high-steep slope monitoring based on Beidou positioning technology provided by the disclosed embodiment includes: obtaining the line mileage information of the high-steep slope to be monitored; dividing the high-steep slope to be monitored into short-distance high-steep slopes and long-distance high-steep slopes according to the line mileage information; using Beidou positioning technology to perform fixed-point monitoring on the short-distance high-steep slope to obtain fixed-point monitoring data; using multi-sensor fusion technology to perform distributed monitoring on the long-distance high-steep slope to obtain distributed monitoring data; performing analysis based on the fixed-point monitoring data and the distributed monitoring data, and issuing an early warning when the analysis result exceeds the safety level. By performing fixed-point monitoring on the short-distance high-steep slopes with a short line mileage range and fixed-point monitoring on the long-distance high-steep slopes based on Beidou high-precision positioning technology, the safety status of high-steep slopes can be monitored, and emergency alarms and advance warnings can be provided.

[0060] Embodiment 2

[0061] See also Figure 2 , Figure 2 The following is a flowchart of the steps of a monitoring method for high and steep slope monitoring based on Beidou positioning technology provided in the first embodiment of the present disclosure, and the method specifically includes:

[0062] Step S201, obtaining line mileage information of the high and steep slope to be monitored;

[0063] Step S202, dividing the high and steep slopes to be monitored into short-distance high and steep slopes and long-distance high and steep slopes according to the line mileage information;

[0064] Step 203, using Beidou positioning technology to perform fixed-point monitoring on the short-distance high and steep slope to obtain fixed-point monitoring data;

[0065] It should be noted that in the process of fixed-point monitoring of short-distance high-steep slopes, a variety of equipment is configured, including base stations and Beidou monitoring stations. The base station is used to build a differential service CORS network and provide Beidou high-precision differential services. Its components include reference antennas, base stations, power supply equipment, lightning protection devices, routers and CORS solution software. The Beidou monitoring station is responsible for real-time monitoring of water and soil changes at the monitoring point. It is divided into two parts: aboveground and underground. The aboveground part includes the main box (containing monitoring receivers, communication equipment and power controllers), monitoring antennas, solar panels, observation piers, lightning rods and guardrails; the underground part includes battery boxes and a variety of sensors, such as crack displacement monitors, pressure gauges, tilt sensors, water level gauges and rain gauges.

[0066] When selecting the locations of the base station and Beidou monitoring station, it is necessary to conduct on-site surveys in combination with the locations in the monitoring point distribution table to ensure that the base station is far away from the vibration source, has no obstructions around it, and has a stable signal that is not easily damaged; the Beidou monitoring station needs to have no obstructions around it, a stable signal, and be as close to the monitoring point as possible without intruding on the line. After the on-site location is determined, the location coordinates, kilometer mark, and on-site environment need to be recorded.

[0067] In an embodiment of the present application, the fixed-point monitoring process includes: using a variety of sensors to monitor short-distance high and steep slopes to obtain sensor monitoring data; using a monitoring receiver to receive these data and synchronously output the original positioning data; distinguishing monitoring points according to the IP address and port number of the monitoring receiver, and calculating the three-dimensional coordinates of all monitoring points.

[0068] The specific implementation steps are as follows: determine the monitoring points and deploy Beidou monitoring stations at the monitoring points; deploy base stations at the monitoring points and build a differential service CORS network; integrate multiple sensors in the Beidou monitoring station; monitor through multiple sensors and obtain sensor monitoring data; receive sensor monitoring data through monitoring receivers and synchronously output raw positioning data, including carrier phase data and ephemeris data; transmit the raw positioning data to the control center through communication equipment; the control center distinguishes the monitoring points according to the IP address and port number of the monitoring receiver, solves the three-dimensional coordinates of each monitoring point and stores them in the database; finally, the sensor monitoring data, raw positioning data and three-dimensional coordinates of the monitoring points are combined to form fixed-point monitoring data.

[0069] Among them, the CORS system consists of antennas, base stations, power supply, lightning protection, routers and CORS solution software. The CORS system can receive satellite signals broadcast by GNSS all day, all day and with high reliability, and provide differential data for high-precision positioning of Beidou terminals. In addition to reusing existing base stations at monitoring points along the line, this application deploys a set of base stations at the slope in front of the tunnel, covering a range of 10 to 20 kilometers, and the signal can cover the station and surrounding line sections. This application also deploys debris flow and landslide monitoring stations at the slope in front of the tunnel to monitor natural disasters such as debris flow and landslides at the tunnel entrance. The monitoring station has the functions of real-time satellite positioning data tracking, collection, recording, equipment integrity monitoring, etc. It is divided into above-ground and underground parts. The above-ground part is mainly composed of a main box, antenna, solar panel, observation pier, lightning rod, guardrail, etc. The underground part is mainly composed of battery boxes and related sensors (such as crack meters, pressure gauges, tilt sensor tubes, etc.).

[0070] Among the various sensors, the crack meter uses a wire displacement sensor to sense the displacement changes of the cracks; the pressure gauge is used to monitor the changes in soil stress in the geotechnical structure; the tilt sensor is equipped with an inclinometer tube, which monitors the slope inclination by observing the deformation of the inclinometer tube. The inclinometer tube is usually installed in a vertical borehole that passes through the unstable soil layer to the lower stable stratum. By analyzing the data from these sensors, alarm thresholds can be defined, including first, second and third level warning values, to ensure that abnormal changes in the slope are discovered in a timely manner.

[0071] Specifically, when the soil stress in the structure under test changes, the pressure gauge sensing plate will synchronously sense the stress change, and the sensing plate will deform. The deformation is transmitted to the vibrating string and converted into a change in the stress of the vibrating string, thereby changing the vibration frequency of the vibrating string. The electromagnetic coil excites the vibrating string and measures its vibration frequency. The frequency signal is transmitted to the reading device via a cable, and the compressive stress value of the structure under test can be measured, and the temperature value of the buried point can be measured synchronously. The tilt sensor tube is usually installed in a vertical borehole that passes through an unstable soil layer to the lower stable stratum. A digital vertical active inclinometer probe, a control cable, a pulley device and a reading instrument are used to observe the deformation of the inclinometer tube. The inclinometer tube is installed in a vertical borehole that passes through an unstable soil layer to the lower stable stratum.

[0072] In addition, the definition of the alarm threshold is as follows:

[0073] Level 1 warning value: the average daily displacement rate in the horizontal direction exceeds 1mm / d for five consecutive days and the displacement direction is basically the same; the cumulative displacement in five days exceeds 6mm, and the average daily displacement rate during the period exceeds 0.5mm / d and the direction is consistent and no convergence is seen. The vertical direction is controlled at twice the value of the horizontal direction.

[0074] Level 2 warning value: The average daily displacement rate in the horizontal direction exceeds 1.5mm / d for five consecutive days and the displacement direction is basically the same; the cumulative displacement in five days exceeds 9mm, and the average daily displacement rate during the period exceeds 0.8mm / d and the direction is consistent and no convergence is seen. The vertical direction is controlled at twice the value of the horizontal direction.

[0075] Level 3 warning value: The average daily displacement rate in the horizontal direction exceeds 2mm / d for five consecutive days and the displacement direction is basically consistent; the cumulative displacement in five days exceeds 15mm, and the average daily displacement rate during the period exceeds 1mm / d and the direction is consistent and no convergence is seen. The vertical direction is controlled at twice the value of the horizontal direction.

[0076] Displacement monitoring accuracy requirements: horizontal ±3.0mm, vertical ±5.0mm. After the monitoring system runs stably, the actual monitoring data can be reported to the construction unit and the warning value can be recalculated and optimized, and can be configured in the background.

[0077] Through the above steps and equipment configuration, it is possible to achieve fixed-point monitoring of short-distance high and steep slopes, and issue early warnings based on monitoring data to ensure the safety and stability of the slopes. This fixed-point monitoring method has the advantages of high precision, strong synchronization, and comprehensive information, which can effectively improve the accuracy and reliability of monitoring.

[0078] Step S204, using multi-sensor fusion technology to perform distributed monitoring on the long-distance high and steep slope to obtain distributed monitoring data;

[0079] In an embodiment of the present application, the distributed monitoring process includes: determining the type and orientation of the invading foreign object based on the vibration signal; the vibration signal is output through an interferometer connected to the positioning vibration optical fiber when vibration occurs in the positioning vibration optical fiber sensing laying area; the vibration signal is converted into an electrical signal through the photoelectric detector of the positioning vibration optical fiber; the electrical signal is used to determine abnormal information and generate corresponding alarm information; and the orientation of the foreign object is identified and video information is collected in response to the alarm information through a laser camera.

[0080] In a specific implementation, the laser of the positioning vibration optical fiber emits a laser which passes through the optical path part and enters the on-site optical cable; the on-site optical cable senses changes in the external environment and brings back the vibration information generated by the external environment through the laser; the vibration information passes through the optical path part and enters the photoelectric detector of the positioning vibration optical fiber and is converted into an electrical signal; data collection and analysis are performed on the electrical signal and an alarm message is generated; according to the vibration orientation in the alarm message, a laser camera configured nearby is linked to perform secondary identification; the laser camera collects video information and stores it in a video database; among them, the vibration information, alarm information and video information are combined to form distributed monitoring data.

[0081] For example, during the monitoring process, the laser emitted by the laser passes through the optical path and enters the on-site optical cable. When natural disasters such as mudslides or falling rocks occur, the optical cable will feel the vibration of the external environment. Some lasers with on-site vibration information return along the optical cable, pass through the optical path, enter the photoelectric detection and convert into electrical signals, and collect and analyze the electrical signals to generate alarms. According to the alarm location, the configured laser camera nearby can be linked to collect video information of the alarm location in real time. After the collection is completed, the video data is transmitted to the background server through 4G transmission for analysis and processing. Finally, the processed alarm video information is transmitted to the general monitoring room to observe the alarm information. At the same time, the data with different environmental conditions during the collection are classified and sorted, and a video database under different times, different weather, different types and different vehicle conditions is established to accumulate samples for more detailed evaluation of perimeter intrusion and disaster alarms in the later stage.

[0082] It should be noted that the positioning vibration optical fiber is laid around the perimeter of the monitoring section. The continuous light emitted by the ultra-narrow linewidth laser is modulated into pulse light with a wavelength of 1550nm by an acousto-optic modulator (AOM) and a function generator. After the pulse light is amplified by an erbium-doped fiber amplifier (EDFA), it is injected into the single-mode sensing optical fiber by a circulator. When the optical fiber senses the vibration of the external environment, such as natural disasters such as mudslides or falling rocks, the laser in the optical fiber will generate backward Rayleigh scattered light. These scattered lights are converted into electrical signals by photodetectors, and the received signals are sampled by a data acquisition card with hardware averaging function to further suppress thermal noise and amplifier spontaneous radiation noise. Subsequently, the electrical signal is transmitted to the computer for analysis and processing. The disturbance event near the optical fiber is discovered and located by the change in the intensity of Rayleigh scattered light. When there is no vibration disturbance in the outside world, the light intensity of the scattering curves of the adjacent periods changes little; when there is a disturbance in the outside world, the light intensity of the curves of different periods at the disturbance will fluctuate violently. Therefore, the disturbance point can be determined by differentiating the curves of different periods.

[0083] The environment of high and steep slopes is complex and there is a lot of noise. Due to the high sensitivity of Rayleigh scattered light, it is very easy to be disturbed by the surrounding environment. Common external random vibrations mainly refer to meteorological changes such as wind and rain, as well as unpredictable and unavoidable traffic, mechanical vibrations and shocks caused by construction work, which make useful signals easily covered. Therefore, after collecting the vibration signal, the original waveform needs to be preprocessed to facilitate the perception and positioning of subsequent intrusion events.

[0084] After preprocessing, the vibration signal is output through an interferometer connected to the positioning vibration optical fiber and converted into an electrical signal through a photoelectric detector. Data collection and analysis are performed on the electrical signal to determine abnormal information and generate corresponding alarm information.

[0085] When the system generates an alarm message, the laser camera responds to the alarm message, identifies the location of the foreign object, and collects video information. The specific steps are as follows:

[0086] The laser camera performs secondary identification based on the vibration direction in the alarm information and collects video information of the alarm location in real time. The collected video information is transmitted to the backend server via 4G for analysis and processing, and the processed alarm video information is transmitted to the general monitoring room for observation.

[0087] The optical camera uses video detection technology for secondary category recognition, and the video detection technology is formed by the fusion of hybrid attention and bidirectional feature pyramid. Specifically, the present application proposes a long-range high and steep slope clearance safety event detection method that integrates hybrid attention and bidirectional feature pyramid. First, based on the latest YoloX framework, a detection method for perimeter intrusion, debris flow, and rockfall is designed; secondly, in order to solve the problem of similar detection with the background, a hybrid attention module is integrated into the network structure to enhance the specific target area of ​​interest while weakening the irrelevant background area, thereby improving the detection ability of similar features of the network; finally, in order to improve the detection ability of small-scale rockfall foreign objects with an image area of ​​less than 30×30 pixels, a bidirectional feature pyramid network is integrated into the network structure to strengthen the mutual communication of features at different levels.

[0088] The detection algorithm is implemented based on the improved YOLOX target detection framework with hybrid attention and bidirectional feature pyramid network. The YOLOX algorithm is a new target detection framework, which is mainly based on the improvement of the YOLOv3 algorithm. The network improvements mainly include the backbone network structure, the classification and regression decoupling head, the anchor-free mechanism, and the dynamic matching positive sample. The YOLOX model is modified based on the YOLOv3-Darknet53 baseline. The backbone network is combined with the cross-stage partial network (CSPNet) to form CSPDarknet to improve the network learning performance. The backbone network is composed of a residual network structure. The residual network includes a backbone part and a residual side part. The backbone part consists of a 1×1 and a 3×3 convolution. The residual side does not perform any operation, and then the above two parts are spliced. Since repeated gradient information in the network will have a significant impact on the network's computing performance and reasoning speed, CSPNet forms a large residual structure based on the stacking of residual blocks. The main body is divided into the above residual blocks, and the residual edges are spliced ​​with the output of the main body after a small amount of processing to form the output of CSPNet. CSPNet uses cross-stage feature fusion strategies and truncated gradient flows to enhance the variability of learning features in different layers.

[0089] The Focus module is added to the YOLOX backbone. Sub-feature maps are extracted every feature point in the horizontal and vertical directions of the feature map to obtain four sub-feature maps. These four parts are stacked, and the length and width of the final feature map are halved, and the number of channels is four times the original. This module increases the receptive field of feature points, reduces information loss, reduces the scale of network parameters, and improves the speed of network training and inference.

[0090] The YOLOX backbone network also uses the Spatial Pyramid Pooling (SPP) structure. SPP passes the input through the maximum pooling kernels of different sizes (1×1, 5×5, 9×9, 13×13) to obtain different levels of receptive field information, and then splices the input with the output of each maximum pooling to obtain the final output of SPP. This structure effectively integrates feature information at all levels. Adding the SPP module to the network expands the receptive field while having almost no effect on the network inference speed.

[0091] The clearance safety detection algorithm for steep slopes needs to run around the clock. At night, the visibility of the video decreases, and the difference between the natural disaster area and the overall background decreases. In particular, the pixel areas of scattered rocks at different positions in the picture vary greatly. In order to improve the robustness of the detection algorithm, this application adds a hybrid attention module after the last three CSP layers of the YoloX backbone network CSPDarknet to further enhance the network's feature extraction capabilities, enabling it to select more critical information for the current detection target.

[0092] 1) Channel Attention

[0093] The channel attention module performs average pooling and maximum pooling operations on the input feature map from the wide and high dimensions respectively, and then sends it to the shared network composed of multi-layer perceptrons. It then outputs the channel attention features after global average pooling and global maximum pooling, and sums them element by element to get the feature vector. After sigmoid, it obtains the weight of channel attention learning, and multiplies the weight feature by the input feature map along the channel direction to obtain the final feature map.

[0094] 2) Spatial Attention

[0095] The spatial attention module performs average pooling and maximum pooling operations on the input feature map from the channel dimension to form two feature maps with the same dimension but channel 1. Then they are merged in the channel direction and sent to the convolution with output channel 1. After sigmoid, the spatial attention with the same spatial dimension as the input feature map and channel dimension 1 is obtained. Finally, the feature map and the spatial attention are multiplied to the fine-tuned feature map.

[0096] 3) Mixed Channel Attention

[0097] The hybrid channel attention module connects channel attention and spatial attention in series. The feature map first performs attention operation in the channel direction, followed by attention operation in the width and height directions, and finally obtains the attention feature map. Hybrid attention allows the network to highlight the weight of the foreground rockfall target part during the learning process, which is equivalent to weighting the original feature map, so that the network can autonomously learn to pay more attention to the features of the foreground rockfall during the learning process.

[0098] As the depth of the network continues to increase, the features extracted by the network will be lost to a certain extent. The feature pyramid network is an effective way to solve this problem. It can fuse features at different levels. After feature extraction, YoloX's backbone network sends the features of the middle layer, the lower middle layer, and the bottom layer to the feature pyramid network, which fuses the bottom-up and top-down features. The previous feature fusion method treated all input features equally. The bidirectional feature pyramid network added lateral connections between features of the same scale and introduced a weighting strategy. The final output feature map combines the features of the current layer and the upper and lower layers, a total of three layers. The calculation method is as follows:

[0099]

[0100] Among them, Out and In represent the output and input features respectively. i Then apply the Relu activation function to ensure the weight ω i ≥0, and the learning rate ε=0.0001 is a small value to avoid numerical instability.

[0101] The YoloX feature pyramid network is partially modified into a bidirectional feature pyramid network, and the lateral connection of the intermediate feature layer is added. At the same time, in order to adapt to the backbone network structure, the BiFPN network is simplified into three layers. The features of different levels are fused and output to the three detection heads of YoloX respectively.

[0102] In an optional embodiment, determining the type and location of the intruding foreign object based on the vibration signal includes:

[0103] The optical fiber vibration signal denoising algorithm is used to denoise the vibration signal;

[0104] Wherein, the optical fiber vibration signal denoising algorithm includes wavelet threshold denoising and empirical mode decomposition;

[0105] A positioning algorithm is used to locate the vibration event that causes the vibration signal;

[0106] The positioning algorithms include direct spectrum method, time domain difference method and ratio positioning method.

[0107] It should be noted that wavelet threshold denoising has the characteristics of multi-resolution analysis, and can realize localized analysis of signals in the time domain and frequency domain. In the fiber optic vibration sensing system, the wavelet threshold shrinkage method is used to reduce the time domain noise of the signal. It is suitable for most square-integrable signals, has a good suppression effect on white noise, and can retain the peak points and smoothness characteristics of the signal while denoising. The basic idea of ​​wavelet threshold shrinkage denoising is: select a suitable wavelet to decompose the original signal into waves and obtain wavelet coefficients of each scale; then, since the wavelet coefficient of the noise is smaller than the wavelet coefficient of the signal, a threshold that meets the conditions can be selected to filter the coefficient; finally, the denoised signal is reconstructed.

[0108] Empirical mode decomposition adaptively decomposes nonlinear and non-stationary signals into the sum of several intrinsic mode functions distributed from high to low frequency through a "screening" process. EMD decomposes based on the time scale characteristics of the signal and does not require any basis function to be set in advance. Compared with the wavelet threshold denoising method, its most significant advantage is that it solves the problem that the basis function cannot be adapted.

[0109] At present, there are two methods for locating vibration events: direct spectrum method and time domain difference method. The direct spectrum method performs FFT on the signal at the same location at different times. If the location is affected by the vibration event, the spectrum will contain the characteristic frequency of the signal.

[0110] The difference positioning method is to make a difference between the optical time domain reflection curves of two adjacent measurements. Before the position point affected by the vibration event, the curve difference is small; at the optical fiber position affected by the vibration event and all subsequent positions, the difference curve fluctuates violently, and then the vibration event position is determined based on the fluctuation of multiple difference curves.

[0111] Ratio positioning method, the initial data received by the acquisition card is one-dimensional data. First, the one-dimensional data needs to be integrated according to the acquisition length of the acquisition card to obtain an M×N two-dimensional matrix, which represents the M optical time domain reflection curves collected in M ​​consecutive pulse cycles, each curve has N data points. Then the M optical time domain reflection curves are accumulated and averaged to obtain the polarization noise reduction curve. Because the change of the polarization state of the sensing fiber will continue from the vibration position to the end of the fiber, it will cause the curves of the fiber position affected by the vibration and its subsequent positions to fluctuate smoothly, and the vibration event position can be determined.

[0112] Step S205, determining the change offset of the fixed-point monitoring data, and dividing the offset exceeding level according to the change offset;

[0113] Step S206, determining the type of the foreign object intrusion event according to the offset exceeding level, the foreign object type and the foreign object orientation;

[0114] In an optional embodiment, determining the category of the foreign object intrusion event according to the offset exceeding level, the foreign object category and the foreign object orientation includes:

[0115] A decision tree model is constructed using a preset training set, and after integration and evaluation of the decision tree model, a target decision tree model is obtained; the training set is characterized by the training offset over-limit level and the training foreign body intrusion event category, and the training security level is a label;

[0116] The offset excess level, the foreign object category and the foreign object orientation are input into the target decision tree model to obtain the foreign object intrusion event category.

[0117] In an embodiment of the present application, a target decision tree model is constructed with training offset excess levels and training foreign object intrusion event categories as features and training security levels as labels. Subsequently, the offset excess levels and foreign object intrusion event categories are input into a random forest model to output the security level.

[0118] It should be noted that the random forest model is an integrated learning algorithm that is widely used in classification and regression problems. It improves the accuracy and robustness of the model by building multiple decision trees and aggregating their prediction results. When each tree is generated, samples and features are randomly selected to increase the diversity between trees, thereby reducing overfitting. In addition, random forests can evaluate the importance of each feature, which helps users understand which features have the greatest impact on the results. The model is highly adaptable, can handle high-dimensional data, and is suitable for numerical and categorical data. Since each tree can be built independently, random forests support parallel computing to improve training efficiency. Its successful application in fields such as architecture, finance, and medicine shows its importance in capturing complex data relationships and supporting decision-making. This makes random forests an effective tool in data analysis and predictive modeling, helping companies and researchers make more scientific decisions in complex environments.

[0119] Furthermore, the steps for building the random forest model are as follows:

[0120] (1) Data preparation: Prepare a data set containing features and labels. The features are the offset exceeding level and foreign object intrusion event category, and the label is the security level.

[0121] (2) Dataset division: The collected data set is divided into a training set and a test set. In this embodiment, 70% of the data samples in the data set are used as the training set, and 30% of the data samples are used as the test set. This division not only helps model learning and optimization, but also effectively evaluates the performance and generalization ability of the model. In order to ensure the randomness and representativeness of the division, a common method is random sampling to avoid deviations caused by order or other factors. When dividing, it is also necessary to pay attention to maintaining a balanced distribution of various labels in the training set and the test set to ensure that the model can obtain comprehensive learning information. Finally, after the division is completed, be sure to check the size and content of the training set and the test set to ensure that they meet the predetermined proportions and data requirements. This process is not only the basis for model training, but also provides a reliable basis for subsequent model evaluation.

[0122] (3) Random sampling: Randomly select several explanatory variables and data samples from the training set, and construct multiple different training sets with replacement.

[0123] (4) Constructing a decision tree: For each randomly sampled training set, a decision tree model is constructed. Furthermore, the decision tree model is configured as a classification and regression tree (CART model). The CART model recursively divides the data set into smaller subsets, selects the best features and split points at each node to minimize the Gini impurity (for classification) or the mean square error (for regression), and finally generates a tree for prediction. The CART model is easy to understand and highly interpretable, and can handle nonlinear relationships and high-dimensional data. In a random forest, each decision tree is built based on the CART algorithm. Each tree is an independent CART decision tree that randomly selects features for modeling. Random forests make the final decision by integrating the prediction results of multiple CART decision trees, thereby improving the accuracy and generalization ability of the overall model.

[0124] (5) Integrated decision tree: Integrate multiple decision tree models into a random forest model and use a voting mechanism to make the final prediction. It should be noted that in a random forest, each decision tree is built based on a different training data set and feature subset, and has a certain degree of randomness. The final prediction result can be determined by a voting mechanism or by taking the average. In classification problems, a majority vote can be used to select the category with the most votes as the final prediction result; in regression problems, the prediction results of multiple trees can be averaged as the final output. By integrating multiple decision trees into a random forest model, the risk of overfitting can be effectively reduced and the robustness and accuracy of the model can be improved.

[0125] (6) Model evaluation: The prediction performance of the random forest model on the test set is evaluated using evaluation indicators to determine the final model with good prediction performance. Further, the evaluation indicators use accuracy, precision, recall or F1 value. In this embodiment, accuracy is the ratio of the number of samples correctly predicted by the model to the total number of samples. It is one of the most commonly used evaluation indicators and is calculated as follows: accuracy = (TP + TN) / (TP + TN + FP + FN), where TP represents true positive examples, TN represents true negative examples, FP represents false positive examples, and FN represents false negative examples. The precision rate measures how many of the positive examples predicted by the model are true positive examples, and is an important indicator for evaluating the accuracy of model prediction. The calculation formula is: precision = TP / (TP + FP). The recall rate measures the ratio of the number of positive examples that the model can correctly predict to the actual number of positive examples, also known as the recall rate, and the calculation formula is: recall rate = TP / (TP + FN). The F1 value is the harmonic mean of precision and recall, which comprehensively considers the accuracy and comprehensiveness of the model. The higher the F1 value, the better the balance between precision and recall. The calculation formula is: F1 = 2×(precision×recall) / (precision+recall).

[0126] In addition, in order to solve the compatibility design problem of the Beidou satellite navigation system BDS and GPS, the parameter design method needs to be considered when designing the tracking loop to ensure the flexibility of the loop. The mathematical model of the Beidou high-precision positioning algorithm and its specific steps are as follows:

[0127] (1) In terms of the function model, a function model based on code division multiple access technology is used to perform double difference processing in the system. For GNSS observation data of multiple frequencies, the non-combined original double difference observation equation is used for positioning solution. The specific model is as follows:

[0128]

[0129] Among them, V C represents the residual of Beidou double difference observation value, A C is the coefficient matrix of the corresponding coordinate parameters, are the parameters to be estimated except the ambiguity parameters, N C is the corresponding mode double difference ambiguity parameter to be estimated, l C is the difference between the observed value and the calculated value of each system.

[0130] (2) In the random model, the altitude angle weighted model is used. The altitude angle of each satellite relative to the base station and the mobile station is calculated, and the satellite with the largest altitude angle is selected as the reference satellite. The unit weight variance can usually be taken as 0.003*0.003, and then calculated according to the formula:

[0131]

[0132] Among them, α is the amplification factor, which can be taken as 1.5.

[0133] (3) In order to improve the reliability of Beidou high-precision positioning algorithm, a combination of multiple cycle slip detection methods is adopted. The gross error elimination part is mainly divided into pseudorange and carrier for gross error detection, and the pseudorange observation value is adaptively weighted or eliminated according to different residual sizes. As for the carrier, it is reinitialized and then filtered to provide relatively clean data for the subsequent parameter estimation module.

[0134] (4) In order to obtain high-precision positioning results, the Beidou double-difference ambiguity is accurately and reliably fixed. This application conducts a lot of research on the ambiguity algorithm part, adopts the idea of ​​partial ambiguity fixation, and improves the overall ambiguity fixation rate while ensuring reliability. The process of rapid ambiguity search is actually selected according to the least squares criterion. The specific objective function is as follows:

[0135]

[0136] Among them, y is the matrix composed of all observations, a is the unknown ambiguity parameter matrix, b is the position coordinate parameter, and A and B are the corresponding coefficient matrices. After obtaining the accurate Beidou double difference ambiguity through search, the Beidou high-precision positioning result can be calculated.

[0137] It should be noted that based on Beidou's high-precision positioning technology, for high and steep slopes with short line mileage, Beidou monitoring stations are installed and deployed at monitoring points, and differential service CORS stations are built to provide Beidou high-precision differential services, integrating displacement / crack sensors and water level analysis technology to achieve displacement monitoring and advance warning of landslides, mudslides, etc.

[0138] The main objects of high and steep slopes near railways are landslides and slope protection. The geology and structure of slopes near railways are uneven, which is the focus of monitoring. Through monitoring, we can understand the dynamic changes of geology around heavy-duty railways and provide timely and reliable information for the subsequent assessment of slope reinforcement and protection.

[0139] Step S207, based on the offset exceeding level and / or the foreign object intrusion event category, determine whether it exceeds the safety level, if so, generate an alarm and notify the management personnel of the corresponding level; the safety level is divided according to the safety range of the high and steep slope to be monitored.

[0140] A method for measuring the attitude of a shore-based ship provided by an embodiment of the present disclosure includes: using a camera to obtain the ship side image data of the shore-based ship to be measured; the camera is installed on the dock and the installation height is higher than the ship side of the shore-based ship to be measured; the camera focal length, the ship length of the shore-based ship to be measured, and the horizontal distance between the camera and the shore-based ship to be measured are obtained; using an image feature extraction method, a plurality of ship side features are extracted from the ship side image data; based on the ship side features, the ship pitch angle of the shore-based ship to be measured is determined; based on the ship side features, the ship roll angle of the shore-based ship to be measured is determined in combination with the camera focal length, the ship length and the horizontal distance; based on the ship pitch angle and the ship roll angle, the ship attitude of the shore-based ship to be measured is determined. Using a shore-based camera as a sensor to obtain the ship side image data of the shore-based ship to be measured, and then determining the ship attitude of the shore-based ship to be measured, it is convenient for the dock to grasp the ship attitude in the first time, and avoid accidents caused by significant changes in the ship attitude during loading and unloading.

[0141] Embodiment 3

[0142] See also Figure 3 , Figure 3 The structural block diagram of a monitoring device for high and steep slope monitoring based on Beidou positioning technology provided in the third embodiment of the present disclosure is exemplarily shown, and the device includes:

[0143] An acquisition module 301 is used to acquire line mileage information of the high and steep slope to be monitored;

[0144] A division module 302 is used to divide the high and steep slopes to be monitored into short-distance high and steep slopes and long-distance high and steep slopes according to the line mileage information;

[0145] The fixed-point monitoring module 303 is used to use Beidou positioning technology to perform fixed-point monitoring on the short-distance high and steep slope to obtain fixed-point monitoring data;

[0146] A distributed monitoring module 304 is used to perform distributed monitoring on the long-distance high and steep slopes by using multi-sensor fusion technology to obtain distributed monitoring data;

[0147] The analysis module 305 is used to perform analysis based on the fixed-point monitoring data and the distributed monitoring data, and issue an early warning when the analysis result exceeds the safety level.

[0148] In an exemplary embodiment, the fixed-point monitoring data includes: sensor monitoring data, positioning raw data and three-dimensional coordinates of the monitoring point; a Beidou monitoring station is deployed at the monitoring point on the short-distance high and steep slope; a variety of sensors are integrated in the Beidou monitoring station; the Beidou monitoring station includes an above-ground part and an underground part; the above-ground part includes a main box; the main box includes a monitoring receiver; the fixed-point monitoring module 303 includes:

[0149] A sensor monitoring submodule, used for using the sensor to monitor the short-distance high and steep slope to obtain sensor monitoring data;

[0150] A raw data output submodule, used to utilize the monitoring receiver to receive the sensor monitoring data and synchronously output the raw positioning data;

[0151] The submodule for determining the three-dimensional coordinates of the monitoring points is used to distinguish the monitoring points according to the IP address and port number of the monitoring receiver and calculate the three-dimensional coordinates of all monitoring points.

[0152] In an exemplary embodiment, the distributed monitoring data includes: vibration signals, alarm information and video information; positioning vibration optical fibers and laser cameras are laid on the monitoring points on the long-distance high and steep slopes; the distributed monitoring module 304 includes:

[0153] A foreign body information determination submodule, for determining the type and orientation of the invading foreign body based on the vibration signal; the vibration signal is outputted through an interferometer connected to the positioning type vibration optical fiber when the positioning type vibration optical fiber senses vibration in the laying area;

[0154] An alarm information generating submodule is used to convert the vibration signal into an electrical signal through the photoelectric detector of the positioning type vibration optical fiber; the electrical signal is used to determine abnormal information and generate corresponding alarm information;

[0155] The video information acquisition submodule is used to identify the position of the foreign object and acquire the video information in response to the alarm information through the laser camera.

[0156] In an exemplary embodiment, determining the type and position of the intruding foreign object based on the vibration signal includes:

[0157] Using an optical fiber vibration signal denoising algorithm to denoise the vibration signal;

[0158] Using a positioning algorithm, the vibration event is positioned on the denoised vibration signal to obtain the type and position of the foreign object;

[0159] The vibration characteristics of the vibration signal after denoising are analyzed to determine the type of the foreign matter.

[0160] In an exemplary embodiment, the analysis module 305 includes:

[0161] An offset over-limit level determination submodule, used to determine the change offset of the fixed-point monitoring data, and to classify the offset over-limit level according to the change offset;

[0162] A foreign body intrusion event category determination submodule, used to determine the foreign body intrusion event category according to the offset overlimit level, the foreign body category and the foreign body orientation;

[0163] The judgment submodule is used to judge whether the safety level is exceeded based on the offset excess level and / or the foreign object intrusion event category. If so, an alarm is generated and the management personnel of the corresponding level are notified; the safety level is divided according to the safety range of the high and steep slope to be monitored.

[0164] In an exemplary embodiment, the foreign body intrusion event category determination submodule includes:

[0165] A model building unit is used to build a decision tree model using a preset training set, and obtain a target decision tree model after integrating and evaluating the decision tree model; the training set is characterized by a training offset over-limit level and a training foreign body intrusion event category, and a training security level is a label;

[0166] The intrusion event category determination unit is used to input the offset excess level, the foreign object category and the foreign object orientation into the target decision tree model to obtain the foreign object intrusion event category.

[0167] Embodiment 4: The embodiment of the present disclosure further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of a monitoring method for high and steep slope monitoring based on Beidou positioning technology in any embodiment.

[0168] Embodiment 5, the embodiment of the present disclosure further provides a computer storage medium having a computer program stored thereon, and when the computer program is executed by the processor, the steps of a monitoring method for high and steep slope monitoring based on Beidou positioning technology of any embodiment are implemented.

[0169] Embodiment 6: The embodiments of the present disclosure further provide a computer program product having a computer program stored thereon, and when the computer program is executed by the processor, the steps of a monitoring method for high and steep slope monitoring based on Beidou positioning technology of any embodiment are implemented.

[0170] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0171] In the several embodiments provided in the present application, it should be understood that the methods, devices, electronic devices and storage media disclosed in the present disclosure can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0172] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0173] In addition, each functional unit in each embodiment of the present disclosure may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0174] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present disclosure is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a readable storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present disclosure. The aforementioned readable storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0175] As described above, the above embodiments are only used to illustrate the technical solutions of the present disclosure, rather than to limit them. Although the present disclosure has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present disclosure.

Claims

1. A monitoring method for high and steep slopes based on Beidou positioning technology, characterized in that: include: Obtain line mileage information of high and steep slopes to be monitored; According to the line mileage information, the high and steep slopes to be monitored are divided into short-distance high and steep slopes and long-distance high and steep slopes; Beidou positioning technology is used to carry out fixed-point monitoring of the short-distance high and steep slope to obtain fixed-point monitoring data; Multi-sensor fusion technology is used to perform distributed monitoring of the long-distance high and steep slopes to obtain distributed monitoring data; An analysis is performed based on the fixed-point monitoring data and the distributed monitoring data, and an early warning is issued when the analysis result exceeds the safety level.

2. The method for monitoring high and steep slopes based on Beidou positioning technology according to claim 1 is characterized in that: The fixed-point monitoring data includes: sensor monitoring data, positioning raw data and three-dimensional coordinates of the monitoring point; Beidou monitoring stations are deployed at the monitoring points on the short-distance high and steep slopes; multiple sensors are integrated in the Beidou monitoring stations; the Beidou monitoring stations include an above-ground part and an underground part; the above-ground part includes a main box; the main box includes a monitoring receiver; Beidou positioning technology is used to perform fixed-point monitoring on the short-distance high and steep slopes to obtain fixed-point monitoring data, including: The sensor is used to monitor the short-distance high and steep slope to obtain sensor monitoring data; The monitoring receiver is used to receive the sensor monitoring data and synchronously output the original positioning data; The monitoring points are distinguished according to the IP address and port number of the monitoring receiver, and the three-dimensional coordinates of all monitoring points are calculated.

3. The method for monitoring high and steep slopes based on Beidou positioning technology according to claim 2 is characterized in that: The distributed monitoring data includes: vibration signals, alarm information and video information; positioning vibration optical fibers and laser cameras are laid on the monitoring points on the long-distance high and steep slopes; multi-sensor fusion technology is used to perform distributed monitoring on the long-distance high and steep slopes to obtain distributed monitoring data, including: Based on the vibration signal, the type and orientation of the intruding foreign object are determined; the vibration signal is output through an interferometer connected to the positioning type vibration optical fiber when the positioning type vibration optical fiber senses vibration in the laying area; The vibration signal is converted into an electrical signal by the photoelectric detector of the positioning type vibration optical fiber; the electrical signal is used to determine abnormal information and generate corresponding alarm information; The laser camera responds to the alarm information to identify the position of the foreign object and collect the video information.

4. The method for monitoring high and steep slopes based on Beidou positioning technology according to claim 3 is characterized in that: Based on the vibration signal, determining the type and location of the intruding foreign object includes: Using an optical fiber vibration signal denoising algorithm to denoise the vibration signal; Using a positioning algorithm, the vibration event is positioned on the denoised vibration signal to obtain the type and position of the foreign object; The vibration characteristics of the vibration signal after denoising are analyzed to determine the type of the foreign matter.

5. The method for monitoring high and steep slopes based on Beidou positioning technology according to claim 4 is characterized in that: Analyzing the fixed-point monitoring data and the distributed monitoring data, and issuing an early warning when the analysis result exceeds the safety level, including: Determining the change offset of the fixed-point monitoring data, and classifying the offset excess level according to the change offset; Determine the foreign object intrusion event category according to the offset excess level, the foreign object category and the foreign object orientation; Based on the offset exceeding level and / or the foreign object intrusion event category, determine whether it exceeds the safety level. If so, generate an alarm and notify the management personnel of the corresponding level; the safety level is divided according to the safety range of the high and steep slope to be monitored.

6. The method for monitoring high and steep slopes based on Beidou positioning technology according to claim 5 is characterized in that: Determining the type of foreign object intrusion event according to the offset over-limit level, the type of foreign object, and the orientation of the foreign object includes: A decision tree model is constructed using a preset training set, and after integration and evaluation of the decision tree model, a target decision tree model is obtained; the training set is characterized by the training offset over-limit level and the training foreign body intrusion event category, and the training security level is a label; The offset excess level, the foreign object category and the foreign object orientation are input into the target decision tree model to obtain the foreign object intrusion event category.

7. A monitoring device for high and steep slope monitoring based on Beidou positioning technology, characterized in that: include: An acquisition module is used to obtain the line mileage information of the high and steep slope to be monitored; A division module, used for dividing the high and steep slopes to be monitored into short-distance high and steep slopes and long-distance high and steep slopes according to the line mileage information; A fixed-point monitoring module is used to use Beidou positioning technology to perform fixed-point monitoring on the short-distance high and steep slope to obtain fixed-point monitoring data; A distributed monitoring module is used to perform distributed monitoring of the long-distance high and steep slopes using multi-sensor fusion technology to obtain distributed monitoring data; The analysis module is used to perform analysis based on the fixed-point monitoring data and the distributed monitoring data, and issue an early warning when the analysis result exceeds the safety level.

8. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the method according to any one of claims 1 to 6 is executed.

9. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is performed.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images