Slope snow removal system and method

Abstract

The present application discloses a slope snow removal system and method, which relates to the technical field of acoustic wave data processing. The method includes: obtaining multi-source remote sensing data of a first area of the slope; the multi-source remote sensing data includes snow depth image data generated based on LiDAR data; using the multi-source remote sensing data and an avalanche risk assessment model to evaluate whether there is a snow accumulation risk in the first area; when there is a snow accumulation risk in the first area, obtaining first acoustic wave data of each snow removal unit; based on the first acoustic wave data of each snow removal unit, determining the snow layer distribution characteristics of the snow accumulation of each snow removal unit; based on the snow layer distribution characteristics, determining target snow removal units for performing snow removal operations from all snow removal units. The technical solution provided by the present application improves the accuracy and reliability of slope monitoring results, thereby enabling accurate and timely snow removal operations and achieving efficient and safe snow cleaning.

Description

Technical Field

[0001] This application relates to the technical field of acoustic wave data processing, and specifically to a slope snow removal system and method. Background Technique

[0002] In transportation infrastructure such as highways, railways, and airports, snow accumulation is one of the key factors affecting traffic safety. Especially on slopes such as the side slopes, when the snow slides, it is extremely easy to cause natural disasters such as avalanches, which will affect road traffic, and even damage the vehicles driving on the road, causing casualties. Therefore, snow accumulation on slopes poses a great threat to traffic safety. In the prior art, for areas with thick snow accumulation, it is often necessary to manually clean it on-site by shoveling snow or other means. Manual snow removal not only has low efficiency, but also is extremely easy to trigger avalanches during the snow removal process, and it is easy to have the situation of personnel slipping, presenting great potential safety hazards. On the other hand, in the prior art, the monitoring of snow accumulation on slopes mainly relies on ground surveys, meteorological data, and historical records. However, with the trend of irregular climate change and more abnormal climates in recent years, it has been difficult to accurately predict slope avalanches relying on historical meteorological data. At the same time, traditional slope snow accumulation monitoring methods also have problems such as limited coverage, poor real-time performance, and difficulty in comprehensive monitoring, resulting in low accuracy of avalanche prediction results, so that the snow accumulation on slopes cannot be cleared in a timely and effective manner, posing great potential safety hazards to the safe operation of transportation facilities.

[0003] Therefore, how to overcome the above-mentioned existing technical problems and defects has become a key problem to be solved. Summary of the Invention

[0004] The purpose of this application is to provide a slope snow removal system and method to solve the problems raised in the above background technique, improve the accuracy and reliability of slope monitoring results, so as to achieve accurate and timely snow removal operations, and be able to efficiently and safely clean the snow.

[0005] The technical solution of the embodiment of this application is implemented as follows:

[0006] The embodiment of this application provides a slope snow removal system, which includes a plurality of snow shock removal units, a communication unit, and a processing unit. The plurality of snow shock removal units are continuously laid in the first area of the highway slope; the snow shock removal unit includes a shock removal panel and an acoustic wave sensor, and the acoustic wave sensor is arranged on the surface of the shock removal panel;

[0007] The acoustic wave sensor is used to emit acoustic waves to the snow on the shock removal panel where it is located, receive the echo signals returned by different snow layers in the snow, generate first acoustic wave data, and send it to the processing unit;

[0008] The snow removal panel is used to remove the snow on the surface area by vibrating when receiving the control signal sent by the processing unit;

[0009] The communication unit is used to obtain multi-source remote sensing data of the first area; the multi-source remote sensing data includes snow depth image data generated based on Light Detection and Ranging (LiDAR) data;

[0010] The processing unit is used to use the multi-source remote sensing data as input data, and evaluate whether there is a snow accumulation risk in the first area by using an avalanche risk assessment model; when there is a snow accumulation risk in the first area, based on the first acoustic wave data sent by each snow removal unit in the first area, determine the structural differences of each snow layer in the snow along the vertical direction of each snow removal unit, and obtain the snow layer distribution characteristics of each snow removal unit; based on the snow layer distribution characteristics of each snow removal unit, determine the target snow removal unit for performing snow removal operations from all snow removal units, and send a control signal to the target snow removal unit.

[0011] In the above solution, the system further includes a plurality of LiDAR detection units; the plurality of LiDAR detection units are evenly arranged in the first area; the LiDAR detection unit includes a mounting rod and a phase-type laser rangefinder, the mounting rod is arranged on the slope surface, and the phase-type laser rangefinder is arranged at the upper end of the mounting rod;

[0012] The phase-type laser rangefinder is used to emit laser light to the snow, receive the reflected signals returned by the snow surface and the snow removal panel respectively, and obtain LiDAR data; and, based on the time difference between the reflected signals of the snow surface and the snow removal panel in the LiDAR data, determine the snow depth and obtain the snow depth image data.

[0013] An embodiment of the present application also provides a method for clearing snow on a slope, and the method includes:

[0014] Obtain multi-source remote sensing data of the first area of the slope; the multi-source remote sensing data includes snow depth image data generated based on LiDAR data;

[0015] Use the multi-source remote sensing data as input data, and evaluate whether there is a snow accumulation risk in the first area by using an avalanche risk assessment model;

[0016] When there is a snow accumulation risk in the first area, obtain the first acoustic wave data of each snow removal unit among the plurality of snow removal units of the slope snow removal system; the first acoustic wave data includes at least one echo signal received after sending an ultrasonic signal to the snow of the snow removal unit;

[0017] Based on the first acoustic wave data of each snow removal unit, determine the snow layer distribution characteristics of the snow in each snow removal unit; the snow layer distribution characteristics characterize the structural differences of each snow layer in the snow along the vertical direction;

[0018] Based on the snow layer distribution characteristics of each snow removal unit, determine the target snow removal unit for snow removal operation from all snow removal units, and send a control signal to the target snow removal unit; the control signal is used to control the target snow removal unit to vibrate and remove the snow on the surface area.

[0019] In the above solution, before obtaining the multi-source remote sensing data of the first area of the slope, the method further includes:

[0020] Obtain a synthetic aperture radar (SAR) image of the slope;

[0021] Extract features from the SAR image to generate feature information of the slope; the feature information characterizes the differences in echo intensities of different pixels in the SAR image; the feature information includes statistical features, texture features, and polarization features;

[0022] Based on the feature information, perform image segmentation on the SAR image to obtain a plurality of second areas; the area types of different second areas are different; the area types include snow-covered areas and non-snow-covered areas;

[0023] Use the second area with the area type of snow-covered area as the first area to be detected.

[0024] In the above solution, the determining the snow layer distribution characteristics of the snow in each snow removal unit based on the first acoustic wave data of each snow removal unit includes:

[0025] For each snow removal unit, based on the first acoustic wave data, determine the structural information of the snow in the corresponding area; the structural information includes the structural attributes of each snow layer in at least one snow layer of the snow;

[0026] Based on the structural information, determine the structural differences of the snow from top to bottom to obtain the snow layer distribution characteristics of the snow in the corresponding snow removal unit.

[0027] In the above solution, the determining the structural information of the snow in the corresponding area based on the first acoustic wave data includes:

[0028] Based on the acoustic response characteristics of each echo signal in the multiple echo signals of the first acoustic wave data, determine the snow layer type and structural parameters corresponding to each echo signal; the acoustic response characteristics include multiple characteristic factors, and the characteristic factors are echo intensity, scattering characteristics, or phase characteristics;

[0029] Generate the structural information of the snow cover in the corresponding area based on the snow layer type and structural parameters of all echo signals.

[0030] In the above solution, determining the snow layer type corresponding to each echo signal based on the acoustic response characteristics of each echo signal in the multiple echo signals based on the first acoustic wave data includes:

[0031] For each echo signal, compare each characteristic factor in the acoustic response characteristics with the corresponding preset threshold to obtain the comparison result of each characteristic factor;

[0032] Based on the comparison results of each characteristic factor, judge that the type of the snow layer is dry snow, wet snow or compacted snow to obtain at least one initial judgment result;

[0033] Based on the proportion of each snow layer type in all initial judgment results, determine the snow layer type of the corresponding snow layer.

[0034] In the above solution, determining the structural parameter corresponding to each echo signal based on the acoustic response characteristics of each echo signal in the multiple echo signals based on the first acoustic wave data includes:

[0035] Based on the snow layer type of each snow layer and the initial acoustic wave propagation speed of each snow layer, determine the first thickness of each snow layer; the initial acoustic wave propagation speed is pre-configured according to the snow layer type;

[0036] Take the snow cover thickness on the surface area of the snow removal unit as the iteration target, and use the gradient descent method to iteratively optimize the initial acoustic wave propagation speed of each snow layer to obtain the optimized acoustic wave propagation speed of each snow layer;

[0037] Based on the optimized acoustic wave propagation speed, determine the second thickness of each snow layer;

[0038] Based on the first acoustic wave data of each snow removal unit and the second thickness of each snow layer, determine the structural parameters of the snow cover in the corresponding area; where,

[0039] During the iteration process, the defined error function is used to evaluate the difference between the predicted snow cover thickness of the model and the actual measured snow cover thickness, and the gradient descent valve is used to adjust the acoustic wave propagation speed to minimize the error function; the error function is expressed as:

[0040] ;

[0041] Where, represents the error function, represents the thickness of the i-th snow layer, D represents the actual measured snow cover thickness;

[0042] The gradient descent rule is expressed as:

[0043] ;

[0044] Among them, represents the acoustic wave propagation speed of snow layer i after the (k + 1)-th iteration, represents the acoustic wave propagation speed after the k-th iteration, represents the learning rate;

[0045] Using the acoustic wave propagation speed of snow layer i after the (k + 1)-th iteration to calculate the snow layer thickness after the corresponding iteration, so as to determine the difference between the predicted snow accumulation thickness and the actual measured snow accumulation thickness; the snow layer thickness after each iteration optimization is expressed as:

[0046] ;

[0047] Among them, represents the thickness of snow layer i after the (k + 1)-th iteration, represents the echo duration of snow layer i.

[0048] In the above solution, the snow layer distribution characteristics further include the porosity and effective density of each snow layer; the method further includes:

[0049] Using the optimized acoustic wave propagation speed of each snow layer to calculate the corresponding effective density:

[0050] Using the effective density of each snow layer to calculate the porosity of the corresponding snow layer; among them,

[0051] The calculation formulas for the effective density and porosity are expressed as:

[0052] ;

[0053] ;

[0054] Among them, represents the effective density of snow layer i, represents the effective bulk modulus of the snow layer, represents the optimized acoustic wave propagation speed of the snow layer, represents the porosity of the snow layer, represents the bulk density.

[0055] In the above solution, determining the target snow removal unit for performing snow removal operations from all snow removal units based on the snow layer distribution characteristics of each snow removal unit for snow accumulation includes:

[0056] Dividing the first area into grids to obtain a plurality of first grids; the first grids correspond to the snow removal units one by one;

[0057] Based on the snow layer distribution characteristics corresponding to each first grid, determine whether there is an avalanche risk in the first grid;

[0058] When there is an avalanche risk in the first grid, use the first grid and the grids around the first grid as the second grid;

[0059] Use the snow removal units corresponding to the second grid and the radiation grids of the second grid as the target snow removal units; the radiation grids include the first grids located downstream of the second grid.

[0060] The slope snow removal system and method provided by the embodiments of the present application utilize multi-source remote sensing data to determine the surface snow coverage and snow depth in each area of the slope, so as to determine the areas with relatively thick snow, achieve comprehensive and efficient determination of risk areas, improve the monitoring efficiency, and reduce the consumption of human and material resources;

[0061] At the same time, by laying snow removal units on the slope, the surface snow in the areas with relatively thick snow can be removed in advance, triggering small avalanches in advance, thereby preventing large-scale avalanches from occurring and ensuring the safety of transportation facilities;

[0062] Furthermore, by setting acoustic wave sensors on the snow removal units, the acoustic wave sensors of each snow removal unit analyze the snow layer structure in the snow through the response of the transmitted signal, so as to determine specific avalanche risk points in the snow-covered area, and then be able to remove the snow in the avalanche risk points and their radiation areas. Since the concept that the weak interface between different snow layers is likely to cause snow layer sliding is introduced, the avalanche risk can be more accurately evaluated, thereby improving the accuracy of avalanche monitoring results and ensuring the accuracy and reliability of the snow removal effect;

[0063] Furthermore, by remotely controlling the snow removal units to vibrate and remove snow, remote snow removal on the slope is realized. Since there is no need for manual on-site snow removal, the problems of low efficiency and low safety caused by manual snow removal can be avoided, thereby improving the timeliness and safety of snow removal. Description of the Drawings

[0064] Figure 1 It is a schematic structural diagram of a slope snow removal system provided by the embodiments of the present application;

[0065] Figure 2 It is a three-dimensional structural diagram of the snow removal unit in the slope snow removal system provided by the embodiments of the present application;

[0066] Figure 3 It is a side structural diagram of the snow removal unit in the slope snow removal system provided by the embodiments of the application;

[0067] Figure 4It is a schematic structural diagram of the LiDAR detection unit in the embodiments of the present application;

[0068] Figure 5 It is a schematic flowchart of a slope snow removal method provided by the embodiments of the present application. Specific embodiments

[0069] Next, the technical solutions in the embodiments of the present application will be clearly and completely described with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0070] The stability of slope snow cover and potential avalanche risks have always been important concerns for public safety and infrastructure protection. However, traditional slope snow cover monitoring methods have the problem of low accuracy of monitoring results. At the same time, for the monitored avalanche risk areas, manual snow removal is often required, with low execution efficiency and safety. It is also possible to use external large equipment (such as snow plows) for snow removal operations. However, in this process, on the one hand, the installation of the equipment has high requirements for the terrain of the installation location, which results in many areas where snow removal equipment cannot be installed. On the other hand, the process from discovering the avalanche risk to installing the snow removal equipment takes a lot of time, resulting in a large lag in snow removal operations and low efficiency.

[0071] Based on this, in various embodiments of the present application, the slope snow removal method provided by the embodiments of the present application uses multi-source remote sensing data to determine the surface snow cover and snow depth in each area of the slope, so as to determine the areas with thick snow, achieving comprehensive and efficient determination of risk areas, improving monitoring efficiency, and reducing human and material resource consumption. At the same time, by laying snow shock removal units on the slope, the surface snow in the areas with thick snow can be removed in advance, triggering small avalanches in advance, thereby preventing large-scale avalanches and ensuring the safety of transportation facilities. Further, by setting acoustic sensors on the snow shock removal units, the acoustic sensors of each snow shock removal unit analyze the snow layer structure in the snow by transmitting signals, so as to determine specific avalanche risk points in the snow-covered area, and then the snow in the avalanche risk points and their radiation areas can be removed. Since the concept that the weak interface between different snow layers is likely to cause snow layer sliding is introduced, the avalanche risk can be more accurately evaluated, thereby improving the accuracy of avalanche monitoring results and ensuring the accuracy and reliability of the snow shock removal effect. Further, by remotely controlling the snow shock removal units to vibrate and remove snow, remote snow removal from the slope is realized. Since there is no need for manual on-site snow removal, the problems of low efficiency and low safety caused by manual snow removal can be avoided, thereby improving the timeliness and safety of snow removal.

[0072] An embodiment of the present application provides a slope snow removal system, as Figure 1 shown. The system may include a plurality of snow removal units 101, a communication unit 102, and a processing unit 103. The plurality of snow removal units 101 are continuously laid in a first area of the highway slope.

[0073] As Figure 2 and Figure 3 shown, the snow removal unit 101 includes a removal panel 201 and a sound wave sensor 202. The sound wave sensor 202 is disposed on the surface of the removal panel 201. The snow removal unit 101 may further include feet 203. The removal panel 201 is mounted on the slope through the feet 203 and is parallel to the slope surface. A power module 204 is disposed below the vibration panel 201. The power module 204 includes a vibration motor for providing vibration power to the vibration panel 201. In order to reduce the vibration impact of the vibration motor on the slope surface, a shock absorption module 205 may be disposed between the feet 203 and the vibration panel 201. The snow removal unit 101 may further include a communication module 206. The communication module 206 may be disposed on the feet 203. The communication module 206 may be a 5G communication module for sending the detection signal of the sound wave sensor 202 to the processing unit 103, and receiving the control signal sent by the processing unit 103 and sending the control signal to the power module 204.

[0074] In actual application, the communication module 206 of the snow removal unit 101 may further include a power source for powering the power module 204.

[0075] The sound wave sensor 202 is configured to emit sound waves to the snow on the corresponding removal panel 201, receive the echo signals returned by different snow layers in the snow, generate first sound wave data, and send the first sound wave data to the processing unit 103;

[0076] The removal panel 201 is configured to vibrate and remove the surface snow when receiving the control signal sent by the processing unit 103;

[0077] The communication unit 102 is configured to obtain multi-source remote sensing data of the first area. The multi-source remote sensing data includes snow depth image data generated based on LiDAR data;

[0078] The processing unit 103 is configured to use the multi-source remote sensing data as input data, and evaluate whether there is a snow accumulation risk in the first area by using an avalanche risk assessment model; when there is a snow accumulation risk in the first area, based on the first acoustic wave data sent by each snow removal unit in the first area, determine the structural differences of each snow layer in the snow of each snow removal unit in the vertical direction, and obtain the snow layer distribution characteristics of each snow removal unit; based on the snow layer distribution characteristics of each snow removal unit, determine the target snow removal unit for performing snow removal operations from all snow removal units, and send a control signal to the target snow removal unit.

[0079] In practical applications, the slope can be divided into areas according to SAR data first, and the area with a relatively high snow accumulation risk can be selected from the divided areas as the first area.

[0080] In one embodiment, the system further includes a plurality of LiDAR detection units, and the plurality of LiDAR detection units are uniformly arranged in the first area;

[0081] In practical applications, the layout density of the LiDAR detection units can be determined according to the detection range of the LiDAR detection units, so as to ensure that all snow removal units in the first area can be detected by the LiDAR detection units.

[0082] In one embodiment, as Figure 4 shown, the LiDAR detection unit includes a mounting rod 401 and a phase laser rangefinder 402. The mounting rod 401 is arranged on the slope surface, and the phase laser rangefinder 402 is arranged at the upper end of the mounting rod;

[0083] The phase laser rangefinder is configured to emit laser light to the snowfield by using the phase laser rangefinder 402, receive the echo signals returned by the snow surface and the removal panel respectively, and obtain LiDAR data; and, based on the time difference between the echo signals of the snow surface and the removal panel 201 in the LiDAR data, determine the snow depth and obtain snow depth image data.

[0084] Here, the time difference between the echo signals of the snow surface and the removal panel 201 can be understood as the time difference between the echo signal returned by the snow surface and the echo signal returned by the removal panel 201; after determining the snow depth of each snow removal unit, an image of the snow surface in the first area is generated, that is, snow depth image data.

[0085] Based on the architecture of the above slope snow removal system, an embodiment of the present application also provides a slope snow removal method, as Figure 5 shown, this method includes S501 to S505; S501 to S505 will be described in detail below with reference to specific embodiments.

[0086] S501: Obtain multi-source remote sensing data of the first area of the slope; the multi-source remote sensing data includes snow depth image data generated based on LiDAR data.

[0087] In practical applications, a LiDAR detection unit can be used to generate snow depth image data.

[0088] Based on this, in one embodiment, the method may further include:

[0089] Use the phase laser rangefinder 402 of the LiDAR detection unit to emit laser light towards the snowfield, respectively receive the echo signals returned by the snow surface and the vibration removal panel, obtain LiDAR data; and based on the time difference between the echo signals of the snow surface and the vibration removal panel in the LiDAR data, determine the snow accumulation thickness to obtain snow depth image data.

[0090] In practical applications, the first area can be understood as the area of the slope where the snow accumulation is thick and there is a risk of snow accumulation.

[0091] In practical applications, before executing S501, the slope area can be divided first to determine the area of the slope where the snow accumulation is thick and there is a risk.

[0092] Based on this, in one embodiment, before obtaining the multi-source remote sensing data of the first area of the slope, the method may further include:

[0093] Obtain the SAR image of the slope;

[0094] Extract features from the SAR image to generate feature information of the slope; the feature information characterizes the difference in echo intensity of different pixels in the SAR image; the feature information includes statistical features, texture features, and polarization features;

[0095] Based on the feature information, perform image segmentation on the SAR image to obtain multiple second areas; the area types of different second areas are different; the area types include snow accumulation areas and non-snow accumulation areas;

[0096] Take the second area with the area type of snow accumulation area as the first area to be detected.

[0097] In practical applications, the statistical feature can be statistical quantities such as the mean and variance of each pixel or region, the texture feature can be extracted by methods such as the gray-level co-occurrence matrix (GLCM) and local binary pattern (LBP), and the polarization feature can be determined by analyzing the ratio of echo intensities under different polarization combinations.

[0098] After obtaining the slope feature information, a classification algorithm is used to classify each pixel region in the SAR image, thereby performing image segmentation on the SAR image. Here, the classification algorithm can use supervised classification algorithms such as support vector machines and random forest, unsupervised classification algorithms such as clustering algorithms, or deep learning algorithms such as convolutional neural networks (CNNs). The embodiments of the present application do not limit this here.

[0099] S502: Use the multi-source remote sensing data as input data, and use the avalanche risk assessment model to evaluate whether there is a snow accumulation risk in the first region.

[0100] In practical applications, the snow depth image data can be input into the avalanche risk assessment model. The avalanche risk assessment model calculates the snow accumulation risk index based on the input data. When the snow accumulation risk index is higher than the risk threshold, there is a snow accumulation risk in the first region; correspondingly, when the snow accumulation risk index is not higher than the risk threshold, there is no snow accumulation risk in the first region.

[0101] In practical applications, the risk index can be expressed as:

[0102] ;

[0103] Among them, represents the risk coefficient of region j, represents the snow depth of region j, represents the slope of region j, represents the temperature of region j, represents the wind speed of region j, to represent weights.

[0104] Here, the risk threshold can be pre-configured according to historical data and expert experience.

[0105] In practical applications, the LiDAR and SAR historical data of the regions where avalanches have occurred can be collected to construct a training sample set. The historical data includes the snow depth, slope, temperature, wind speed, etc. of the avalanche regions, and then the avalanche risk assessment model is constructed using the training sample set.

[0106] S503: When there is a snow accumulation risk in the first region, obtain the first acoustic wave data of each snow removal unit in the slope snow removal system; the first acoustic wave data includes at least one echo signal received after sending an ultrasonic signal to the snow in the snow removal unit.

[0107] In practical applications, a snow removal unit is provided in each second area, and the snow removal system may include the snow removal units in all the second areas. Here, when there is a risk of snow accumulation in the first area, obtaining the first acoustic wave data of each snow removal unit among the multiple snow removal units of the slope snow removal system can be understood as obtaining the first acoustic wave data of each snow removal unit among the multiple snow removal units corresponding to the first area in the snow removal system, that is, obtaining the first acoustic wave data of each snow removal unit among the multiple snow removal units located in the first area of the snow removal system.

[0108] In practical applications, the snow accumulation risk index of the first area can be used for preliminary screening of avalanche risk areas. Judging the snow layer stability by using the snow layer structure of each snow removal unit can further refine the screening of avalanche risk areas, thereby improving the accuracy of the screening results and snow removal results, and ensuring the effectiveness and reliability of avalanche prevention. At the same time, since the number of snow removal units is large, processing all the ultrasonic data of the snow removal units results in a large amount of data processing, which affects the reliability and efficiency of the calculation results. By performing ultrasonic detection operations on the first area determined after preliminary screening, the amount of ultrasonic data processing can be reduced, and the reliability and efficiency of the calculation results can be improved.

[0109] In practical applications, after the acoustic wave sensor provided on the snow removal unit receives the control instruction sent by the processing unit, it emits ultrasonic waves into the snow. Since snow layers with different densities will form reflection interfaces, echo signals returned from different reflection interfaces will be received. Therefore, the structure of the snow layer in the snow accumulation can be judged according to the number and intensity of the received echo signals.

[0110] S504: Based on the first acoustic wave data of each snow removal unit, determine the snow layer distribution characteristics of the snow accumulation of each snow removal unit; the snow layer distribution characteristics characterize the structural differences of each snow layer in the snow accumulation along the vertical direction.

[0111] In one embodiment, the determining the snow layer distribution characteristics of the snow accumulation of each snow removal unit based on the first acoustic wave data of each snow removal unit includes:

[0112] For each snow removal unit, based on the first acoustic wave data, determine the structural information of the snow accumulation in the corresponding area; the structural information includes the structural attributes of each snow layer in at least one snow layer of the snow accumulation.

[0113] Based on the structural information, determine the structural differences of the snow accumulation from top to bottom to obtain the snow layer distribution characteristics of the snow accumulation of the corresponding snow removal unit.

[0114] In one embodiment, the determining the structural information of the snow accumulation in the corresponding area based on the first acoustic wave data includes:

[0115] Based on the acoustic response characteristics of each echo signal among the multiple echo signals of the first acoustic wave data, determine the snow layer type and structural parameters corresponding to each echo signal; the acoustic response characteristics include multiple characteristic factors, and the characteristic factors are echo intensity, scattering characteristics, or phase characteristics;

[0116] Based on the snow layer types and structural parameters of all echo signals, generate the structural information of the snow cover in the corresponding area.

[0117] In one embodiment, the determining the snow layer type corresponding to each echo signal based on the acoustic response characteristics of each echo signal among the multiple echo signals of the first acoustic wave data includes:

[0118] For each echo signal, compare each characteristic factor in the acoustic response characteristics with the corresponding preset threshold to obtain the comparison result of each characteristic factor;

[0119] Based on the comparison results of each characteristic factor, judge that the snow layer type of the snow layer is dry snow, wet snow, or compacted snow to obtain at least one initial judgment result;

[0120] Based on the proportion of each snow layer type in all initial judgment results of the same snow layer, determine the snow layer type of the corresponding snow layer.

[0121] In practical applications, the characteristics presented by different types of snow layers for characteristic factors are shown in Table 1:

[0122] Table 1:

[0123]

[0124] In practical applications, for the echo intensity, an echo intensity range can be pre-configured for each snow layer type, and based on the comparison result of the echo signal intensity with each snow layer intensity range, judge the snow layer type;

[0125] For the scattering characteristics, a coherence coefficient range can be pre-configured for each snow layer type, and based on the comparison result of the coherence coefficient of the echo signal with each snow layer coherence coefficient range, judge the snow layer type;

[0126] For the phase characteristics, a phase standard deviation range can be pre-configured for each snow layer type, and based on the comparison result of the phase standard deviation of the echo signal with each snow layer phase standard deviation range, judge the snow layer type.

[0127] In practical applications, a snow layer type judgment result can be obtained for each characteristic factor, and according to the proportion of the judgment results, the judgment result with the highest proportion can be used as the final snow layer type judgment result. For example, for snow layer S1, if two judgment results are for dry snow and one judgment result is for wet snow, and the proportion of dry snow is greater than that of wet snow, then the snow layer type of this snow layer is dry snow.

[0128] In one embodiment, determining the structural parameter corresponding to each echo signal based on the acoustic response characteristics of each echo signal among the multiple echo signals of the first acoustic data may include:

[0129] Determine the first thickness of each snow layer based on the snow layer type of each snow layer and the initial acoustic wave propagation speed of each snow layer; the initial acoustic wave propagation speed is pre-configured according to the snow layer type;

[0130] Taking the snow thickness on the surface area of the snow removal unit as the iteration target, and using the gradient descent method to iteratively optimize the initial acoustic wave propagation speed of each snow layer to obtain the optimized acoustic wave propagation speed of each snow layer;

[0131] Determine the second thickness of each snow layer based on the optimized acoustic wave propagation speed;

[0132] Based on the first acoustic data of each snow removal unit and the second thickness of each snow layer, determine the structural parameter of the snow accumulation in the corresponding area; where

[0133] During the iteration process, the defined error function is used to evaluate the difference between the predicted snow thickness of the model and the actual measured snow thickness, and the acoustic wave propagation speed is adjusted using the gradient descent valve to minimize the error function; the error function is expressed as:

[0134] ;

[0135] Where represents the error function, represents the thickness of the i-th snow layer, D represents the actually measured snow thickness;

[0136] The gradient descent rule is expressed as:

[0137] ;

[0138] Where represents the acoustic wave propagation speed of the i-th snow layer after the (k + 1)-th iteration, represents the acoustic wave propagation speed after the k-th iteration, represents the learning rate;

[0139] Calculate the snow layer thickness after the corresponding iteration using the acoustic wave propagation speed of the i-th snow layer after the (k + 1)-th iteration, that is, in the error function , to determine the difference between the predicted snow depth and the actually measured snow depth;

[0140] The snow layer thickness after each iterative optimization, that is, the snow layer thickness determined according to the sound wave propagation speed of the snow layer after each iteration, is expressed as:

[0141] ;

[0142] Wherein, represents the thickness of snow layer i after the (k + 1)-th iteration, represents the echo duration of snow layer i.

[0143] In one embodiment, the snow layer distribution characteristics further include the porosity and effective density of each snow layer; the method further includes:

[0144] Using the optimized sound wave propagation speed of each snow layer, calculate the corresponding effective density:

[0145] Using the effective density of each snow layer, calculate the porosity of the corresponding snow layer; wherein,

[0146] The calculation formulas for effective density and porosity are expressed as:

[0147] ;

[0148] ;

[0149] Wherein, represents the effective density of snow layer i, represents the effective bulk modulus of the snow layer, which can be obtained through experiments or literature, represents the optimized sound wave propagation speed of snow layer i, represents the porosity of snow layer i, represents the ice density.

[0150] Here, the porosity directly reflects the proportion of air in the snow layer. A high porosity means that the snow layer is relatively loose and prone to sliding. By measuring the porosity, the stability of the snow layer can be judged more accurately; while the effective density reflects the mass per unit volume of the snow layer. A high density usually means that the snow layer is more compact and stable. By measuring the effective density, the bearing capacity and stability of the snow layer can be evaluated more accurately. Therefore, by introducing porosity and effective density, a more complex multi-factor comprehensive evaluation model can be constructed, such as a linear weighted model or a non-linear model (Logistic regression, machine learning, etc.). These models can consider the physical properties of the snow layer more comprehensively and improve the accuracy of prediction.

[0151] S505: Based on the snow layer distribution characteristics of each snow removal unit, determine the target snow removal unit for performing snow removal operations from all snow removal units, and send a control signal to the target snow removal unit; the control signal is used to control the target snow removal unit to vibrate and remove the snow on the surface area.

[0152] In one embodiment, the determining the target snow removal unit for performing snow removal operations from all snow removal units based on the snow layer distribution characteristics of each snow removal unit may include:

[0153] Divide the first area into a plurality of first grids; the first grids correspond to the snow removal units one by one;

[0154] Based on the snow layer distribution characteristics corresponding to each first grid, determine whether there is an avalanche risk in the first grid;

[0155] When there is an avalanche risk in the first grid, use the first grid and the grids around the first grid as second grids;

[0156] Use the snow removal units corresponding to the second grids and the radiation grids of the second grids as the target snow removal units; the radiation grids include the first grids located downstream of the second grids.

[0157] In practical applications, the snow layer distribution characteristics of each first grid can be input into a comprehensive evaluation model, that is, the snow layer distribution characteristics of each snow removal unit are used as input data to input into the comprehensive evaluation model, and the comprehensive evaluation model is used to evaluate whether there is an avalanche risk in the first grid and output an evaluation result; the comprehensive evaluation model can adopt a linear weighted model or a non-linear model, such as Logistic regression, machine learning, etc.

[0158] In practical applications, porosity refers to the proportion of air in the snow layer and reflects the looseness of the snow layer; a high porosity means that the snow layer is relatively loose and prone to sliding; a low porosity means that the snow layer is relatively dense and has good stability; for dry snow, when the porosity is high, the snow layer is relatively loose and prone to avalanches; for wet snow, although the moisture in the wet snow increases the adhesiveness, if the porosity is still high, there is still a high avalanche risk;

[0159] The effective density refers to the mass per unit volume of the snow layer, which reflects the compactness of the snow layer. A high density usually means that the snow layer is more compact and stable, but an excessively high density may lead to local stress concentration and increase the risk of avalanches. For dry snow, when the effective density is low, the snow layer is relatively loose and prone to sliding. For wet snow, when the effective density is high, the moisture in the wet snow increases the density, but it may also form larger snow blocks, increasing the likelihood of avalanches. For compacted snow, the structure is relatively stable, but when the effective density is high, it may also trigger large-scale avalanches.

[0160] Based on this, it can be determined that there is an avalanche risk in the first grid when one of the following conditions is met:

[0161] There are wet snow layers and dry snow layers in the snow cover corresponding to the first grid, and the dry snow layer is located above the wet snow layer;

[0162] There is a wet snow layer in the first grid. When the porosity of the wet snow layer is higher than the preset wet snow porosity threshold, or the effective density of the wet snow layer is higher than the preset wet snow density threshold; for example, when the porosity of the wet snow layer is greater than 0.5, or the effective density is greater than 500 kg / m 3 there is an avalanche risk;

[0163] There is only a dry snow layer in the first grid. When the porosity of the dry snow layer is higher than the preset dry snow porosity threshold, or the effective density of the dry snow layer is lower than the preset dry snow density threshold; for example, when the porosity of the dry snow layer is greater than 0.7 or the effective density is less than 200 kg / m 3 there is an avalanche risk;

[0164] When there is compacted snow in the first grid and the effective density of the compacted snow is higher than the preset compacted snow density threshold; for example, when the effective density of the compacted snow is greater than 800 kg / m 3 there is an avalanche risk.

[0165] In practical applications, the grids around the first grid can be the grids adjacent to the four sides of the first grid, or, with the first grid as the center, according to the 3x3 window size, the first grid and the 8 first grids around it are used as the second grid.

[0166] In practical applications, the snow removal unit corresponding to the radiation grid can be the snow removal unit below the mapping point of the second grid on the slope.

[0167] In summary, the slope snow removal method provided by the embodiments of the present application determines the surface snow coverage and snow depth of each area of the slope by using multi-source remote sensing data, so as to determine the areas with relatively thick snow, achieving comprehensive and efficient determination of risk areas, improving the monitoring efficiency, and reducing the consumption of human and material resources. At the same time, by laying snow shock removal units on the slope, the snow on the surface of the areas with relatively thick snow can be removed in advance, triggering small avalanches in advance, thereby preventing large-scale avalanches from occurring and ensuring the safety of transportation facilities. Further, by setting acoustic wave sensors on the snow shock removal units, the acoustic wave sensors of each snow shock removal unit analyze the snow layer structure in the snow through the response of the transmitted signal, so as to determine specific avalanche risk points in the snow-covered area, and then the snow in the avalanche risk points and their radiation areas can be removed. Since the concept that the weak interface between different snow layers is likely to cause snow layer sliding is introduced, the avalanche risk can be more accurately evaluated, thereby improving the accuracy of avalanche monitoring results, and further ensuring the accuracy and reliability of the snow removal effect. Further, by remotely controlling the snow shock removal units to vibrate and remove snow, remote snow removal from the slope is realized. Since there is no need for manual on-site snow removal, the problems of low efficiency and low safety caused by manual snow removal can be avoided, thereby improving the timeliness and safety of snow removal.

[0168] It should be noted that when the slope snow removal system provided in the above embodiment performs slope snow removal, only the above division of each program module is used for illustration. In practical applications, the above processing can be allocated to different program modules according to needs, that is, the internal structure of the system is divided into different program modules to complete all or part of the above-described processing. In addition, the slope snow removal method provided in the above embodiment and the slope snow removal system embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0169] It should be noted that "first", "second", etc. are used to distinguish similar objects and do not necessarily need to describe a specific order or sequence.

[0170] In addition, the technical solutions described in the embodiments of the present application can be arbitrarily combined without conflict.

[0171] The above is only a preferred embodiment of the present application and is not intended to limit the protection scope of the present application.

Claims

1. A slope snow removal system, characterized in that: The system includes a plurality of snow-shock units, a communication unit and a processing unit, wherein the plurality of snow-shock units are continuously laid in a first area of ​​a road slope; the snow-shock units include a snow-shock panel and an acoustic wave sensor, wherein the acoustic wave sensor is arranged on the surface of the snow-shock panel; The acoustic wave sensor is used to emit acoustic waves to the snow on the snow removal panel, receive echo signals returned by different snow layers in the snow, generate first acoustic wave data and send it to the processing unit; The vibration-removal panel is used to vibrate and remove snow on the surface upon receiving a control signal sent by the processing unit; The communication unit is used to obtain multi-source remote sensing data of the first area; the multi-source remote sensing data includes snow depth image data generated based on LiDAR data; The processing unit is used to use the multi-source remote sensing data as input data, and use the avalanche risk assessment model to evaluate whether there is a snow accumulation risk in the first area; when there is a snow accumulation risk in the first area, based on the first acoustic wave data sent by each snow removal unit in the first area, determine the structural differences of each snow layer in the snow of each snow removal unit along the vertical direction, and obtain the snow layer distribution characteristics of each snow removal unit; the snow layer distribution characteristics characterize the structural differences of each snow layer in the snow along the vertical direction; based on the snow layer distribution characteristics of each snow removal unit, determine the target snow removal unit that performs snow removal operations from all snow removal units, and send a control signal to the target snow removal unit; wherein, Based on the first sound wave data of each snow removal unit, the snow layer distribution characteristics of the snow in each snow removal unit are determined, including: for each snow removal unit, based on the first sound wave data, the structural information of the snow in the corresponding area is determined; the structural information includes the structural attributes of each snow layer in at least one snow layer of the snow; based on the structural information, the structural difference of the snow from top to bottom is determined to obtain the snow layer distribution characteristics of the snow in the corresponding snow removal unit; The determining of the structural information of snow accumulation in the corresponding area based on the first acoustic wave data includes: determining the type and structural parameters of the snow layer corresponding to each echo signal based on the acoustic response characteristics of each echo signal in the multiple echo signals of the first acoustic wave data; generating the structural information of snow accumulation in the corresponding area based on the snow layer types and structural parameters of all echo signals; The method of determining a target snow removal unit for performing snow removal operations from all snow removal units based on the snow layer distribution characteristics of each snow removal unit includes: dividing the first area into grids to obtain multiple first grids; the first grids correspond one-to-one to the snow removal units; judging whether the first grid has an avalanche risk based on the snow layer distribution characteristics corresponding to each first grid; when the first grid has an avalanche risk, using the first grid and grids around the first grid as second grids; using the snow removal units corresponding to the second grid and the radial grids of the second grid as target snow removal units; the radial grids include a first grid located downstream of the second grid; wherein, The snow layer distribution characteristics include the porosity and effective density of each snow layer; Based on the snow layer distribution characteristics corresponding to each first grid, the first grid is judged to have avalanche risk when one of the following conditions is met: The snow accumulation corresponding to the first grid contains a wet snow layer and a dry snow layer, and the dry snow layer is located above the wet snow layer; There is a wet snow layer in the first grid, and the porosity of the wet snow layer is higher than a preset wet snow porosity threshold, or the effective density of the wet snow layer is higher than a preset wet snow density threshold; There is only a dry snow layer in the first grid, the porosity of the dry snow layer is higher than a preset dry snow porosity threshold, or the effective density of the dry snow layer is lower than a preset dry snow density threshold; When there is compacted snow in the first grid, and the effective density of the compacted snow is higher than a preset compacted snow density threshold.

2. The system according to claim 1, characterized in that The system further comprises a plurality of LiDAR detection units; the plurality of LiDAR detection units are evenly arranged in the first area; the LiDAR detection unit comprises a mounting rod and a phase laser rangefinder, the mounting rod is arranged on the slope surface, and the phase laser rangefinder is arranged on the upper end of the mounting rod; The phase laser rangefinder is used to emit laser to the snow, receive the reflected signals returned by the snow surface and the vibration removal panel respectively, and obtain LiDAR data; and determine the thickness of snow based on the time difference of the reflected signals of the snow surface and the vibration removal panel in the LiDAR data to obtain snow depth image data.

3. A method for removing snow from a slope, characterized in that: The method comprises: Acquire multi-source remote sensing data of a first area of ​​the slope; the multi-source remote sensing data includes snow depth image data generated based on LiDAR data; Using the multi-source remote sensing data as input data, and using an avalanche risk assessment model to assess whether there is a snow accumulation risk in the first area; When there is a risk of snow accumulation in the first area, obtaining first acoustic wave data of each snow-shock removal unit in the plurality of snow-shock removal units of the slope snow removal system; the first acoustic wave data includes at least one echo signal received after sending an ultrasonic signal to the snow of the snow-shock removal unit; Based on the first acoustic wave data of each snow removal unit, the snow layer distribution characteristics of the snow in each snow removal unit are determined; the snow layer distribution characteristics represent the structural differences of each snow layer in the snow along the vertical direction; Based on the snow layer distribution characteristics of each snow-vibrating unit, a target snow-vibrating unit for performing snow removal operations is determined from all snow-vibrating units, and a control signal is sent to the target snow-vibrating unit; the control signal is used to control the target snow-vibrating unit to vibrate and remove snow on the surface; wherein, The method of determining the snow layer distribution characteristics of the snow accumulated in each snow removal unit based on the first sound wave data of each snow removal unit includes: for each snow removal unit, determining the structural information of the snow accumulated in the corresponding area based on the first sound wave data; the structural information includes the structural attributes of each snow layer in at least one snow layer of the snow accumulation; based on the structural information, determining the structural difference of the snow from top to bottom, and obtaining the snow layer distribution characteristics of the snow accumulated in the corresponding snow removal unit; The determining of the structural information of snow accumulation in the corresponding area based on the first acoustic wave data includes: determining the type and structural parameters of the snow layer corresponding to each echo signal based on the acoustic response characteristics of each echo signal in the multiple echo signals of the first acoustic wave data; generating the structural information of snow accumulation in the corresponding area based on the snow layer types and structural parameters of all echo signals; The step of determining a target snow-vibrating removal unit for performing snow removal operations from all snow-vibrating removal units based on the snow layer distribution characteristics of each snow-vibrating removal unit includes: The first area is gridded to obtain a plurality of first grids; the first grids correspond to the snow removal units one by one; based on the snow layer distribution characteristics corresponding to each first grid, it is determined whether the first grid has an avalanche risk; when the first grid has an avalanche risk, the first grid and the grids around the first grid are used as second grids; the snow removal units corresponding to the second grid and the radiation grids of the second grid are used as target snow removal units; the radiation grids include the first grid located downstream of the second grid; wherein, The snow layer distribution characteristics include the porosity and effective density of each snow layer; Based on the snow layer distribution characteristics corresponding to each first grid, the first grid is judged to have avalanche risk when one of the following conditions is met: The snow accumulation corresponding to the first grid contains a wet snow layer and a dry snow layer, and the dry snow layer is located above the wet snow layer; There is a wet snow layer in the first grid, and the porosity of the wet snow layer is higher than a preset wet snow porosity threshold, or the effective density of the wet snow layer is higher than a preset wet snow density threshold; There is only a dry snow layer in the first grid, and the porosity of the dry snow layer is higher than a preset dry snow porosity threshold or the effective density of the dry snow layer is lower than a preset dry snow density threshold; When there is compacted snow in the first grid, and the effective density of the compacted snow is higher than a preset compacted snow density threshold.

4. The method according to claim 3, characterized in that Before acquiring the multi-source remote sensing data of the first area of ​​the slope, the method further includes: Acquire SAR images of the slope; Extracting features from the SAR image to generate feature information of the slope; the feature information represents the difference in echo intensity of different pixels in the SAR image; the feature information includes statistical features, texture features and polarization features; Based on the feature information, the SAR image is segmented to obtain a plurality of second regions; different second regions have different region types; the region types include snow-covered regions and non-snow-covered regions; The second area whose area type is a snow area is used as the first area to be detected.

5. The method according to claim 3, characterized in that: The acoustic response characteristics include multiple characteristic factors, and the characteristic factors are echo intensity, scattering characteristics or phase characteristics.

6. The method according to claim 3, characterized in that The step of determining the snow layer type corresponding to each echo signal based on the acoustic response characteristics of each echo signal among the plurality of echo signals of the first acoustic wave data comprises: For each echo signal, each characteristic factor in the acoustic response feature is compared with a corresponding preset threshold value to obtain a comparison result of each characteristic factor; Based on the comparison result of each characteristic factor, determining the type of the snow layer as dry snow, wet snow or compacted snow, and obtaining at least one initial determination result; Based on the proportion of each snow layer type in all initial judgment results, the snow layer type of the corresponding snow layer is determined.

7. The method according to claim 3, characterized in that The step of determining a structural parameter corresponding to each echo signal based on an acoustic response characteristic of each echo signal among a plurality of echo signals of the first acoustic wave data comprises: determining a first thickness of each snow layer based on a snow layer type of each snow layer and an initial sound wave propagation velocity of each snow layer; the initial sound wave propagation velocity being preconfigured according to the snow layer type; The snow thickness on the surface of the snow removal unit is taken as an iterative target, and the initial sound wave propagation velocity of each snow layer is iteratively optimized using a gradient descent method to obtain an optimized sound wave propagation velocity of each snow layer; determining a second thickness of each snow layer based on the optimized sound wave propagation speed; Based on the first acoustic wave data of each snow removal unit and the second thickness of each snow layer, the structural parameters of the snow in the corresponding area are determined; wherein, In the iterative process, the error function defined is used to evaluate the difference between the snow thickness predicted by the model and the actual measured snow thickness, and the sound wave propagation speed is adjusted using the gradient descent method to minimize the error function; the error function is expressed as: Where E represents the error function, d i represents the thickness of the i-th snow layer, and D represents the actual measured snow thickness; The gradient descent law is expressed as: Among them, v i (k+1) represents the sound wave propagation speed of snow layer i after the k+1th iteration, v i (k) represents the speed of sound wave propagation after the kth iteration, and η represents the learning rate; The acoustic wave propagation velocity of snow layer i after the k+1th iteration is used to calculate the thickness of the snow layer after the corresponding iteration to determine the difference between the predicted snow thickness and the actual measured snow thickness; the thickness of the snow layer after each iteration optimization is expressed as: Among them, d i (k+1) represents the thickness of snow layer i after the k+1th iteration, Δt i Indicates the echo duration of snow layer i.

8. The method according to claim 7, characterized in that The method further comprises: Using the optimized sound wave propagation speed for each snow layer, calculate the corresponding effective density: Using the effective density of each snow layer, the porosity of the corresponding snow layer is calculated; where, The calculation formula of effective density and porosity is expressed as: Among them, ρ eff,i represents the effective density of snow layer i, K eff represents the effective bulk modulus of the snow layer, v i represents the optimized sound wave propagation speed of the snow layer, n i represents the porosity of the snow layer, ρ ice Represents the density of ice.

Images