A highway maintenance road construction safety management method, device, medium and equipment

Abstract

The application discloses a kind of highway maintenance road construction safety management method, device, medium and equipment, it is related to highway construction management field.The method comprises: extracting the first characteristic vector of the road surface electromagnetic wave data of different structural layers of road surface;First characteristic vector is marked with aging parameter marking with structural layer, obtains training data set, inputs deep convolutional neural network training, obtains inversion model;The to-be-maintained highway is divided into multiple sub-regions, the road surface electromagnetic wave data of different structural layers in each sub-region is acquired, and the second characteristic vector of the road surface electromagnetic wave data corresponding to different structural layers is extracted, and input into inversion model, obtains the aging parameter of four structural layers of each sub-region;According to aging parameter, determine the current aging grade of four structural layers, and the period of transformation into next aging grade, further determine the maintenance order of each sub-region.The application can improve the efficiency of maintaining highway by the above scheme.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of highway construction management, and in particular to a highway maintenance road construction safety management method, device, medium and equipment. BACKGROUND

[0002] With the improvement and development of highway traffic network, the length and quantity of highways are increasing, and in order to ensure the safety of motor vehicles on highways, it is necessary to regularly maintain highways. Highway maintenance is the maintenance of highways. Maintain highways and structures and facilities on highways to maintain highway performance as much as possible, repair damaged parts in time, ensure safe, comfortable and smooth driving, save transportation costs and time; take correct technical measures to improve engineering quality, extend the service life of highways and delay the reconstruction time.

[0003] In the prior art, the technical scheme for maintaining highways usually uses unmanned aerial vehicles, roadside cameras and other shooting devices to shoot images of highways, and then uses image recognition technology to analyze the shot images to obtain aging areas on the surface of the highway, further determines the aging areas as damaged areas of the highway to be maintained, and then performs highway maintenance work on the damaged areas.

[0004] However, due to the complexity of the road surface structure, and the possible changes in the aging degree of the aging areas on the road surface during highway maintenance, the traditional highway maintenance scheme cannot reasonably arrange the construction sequence of multiple aging areas on the highway, resulting in insufficient precision and low efficiency of highway maintenance. SUMMARY

[0005] Therefore, it is necessary to provide a highway maintenance road construction safety management method, device, medium and equipment to solve the technical problems of insufficient maintenance precision and low efficiency of existing highway maintenance technology.

[0006] The present application adopts the following technical solutions:

[0007] In a first aspect, the present application provides a highway maintenance road construction safety management method, which comprises:

[0008] Obtaining road surface electromagnetic wave data corresponding to different structural layers of the road surface, the different structural layers of the road surface including surface layer, base layer, cushion layer and bottom base layer; extracting a first feature vector of the road surface electromagnetic wave data corresponding to the different structural layers of the road surface, the first feature vector including amplitude, phase, frequency and wavelength;

[0009] The first feature vectors corresponding to different structural layers of the pavement are respectively marked with structural layer labels and aging parameter labels to obtain a training data set, the training data set is input into a deep convolutional neural network for training to obtain an inversion model;

[0010] The to-be-maintained highway is divided into multiple sub-regions, pavement electromagnetic wave data corresponding to different structural layers in each sub-region is obtained, and second feature vectors of the pavement electromagnetic wave data corresponding to different structural layers in each sub-region are extracted; the second feature vectors corresponding to different structural layers in each sub-region are input into the inversion model to obtain aging parameters respectively associated with four structural layers of the each sub-region;

[0011] According to the aging parameters, the current aging levels of the four structural layers of the each sub-region are determined respectively, and the conversion periods from the current aging levels to the next aging levels are determined;

[0012] According to the current aging levels of the four structural layers of the each sub-region and the conversion periods from the current aging levels to the next aging levels, the maintenance sequence of the each sub-region is determined.

[0013] Further, the first feature vectors corresponding to different structural layers of the pavement are respectively marked with structural layer labels and aging parameter labels, specifically including:

[0014] According to the range in which the energy value of the amplitude in each first feature vector is located, the each first feature vector is marked as one of a surface layer, a base layer, a cushion layer, and a sub-base layer;

[0015] According to the abnormal values of the phase, the abnormal values of the frequency, and the abnormal values of the wavelength contained in each first feature vector, the each first feature vector is marked with an aging parameter.

[0016] Further, the aging parameters are used to respectively determine the current aging levels of the four structural layers of the each sub-region and the conversion periods from the current aging levels to the next aging levels, specifically including:

[0017] The aging parameters are reconstructed by using a pre-constructed parameter quantization table to obtain aging values corresponding to different aging types contained in the four structural layers respectively;

[0018] According to the aging values corresponding to the different aging types and the weights of the different aging types in the structural layers in which the different aging types are located, the current aging levels respectively corresponding to the four structural layers of the each sub-region and the aging stages in which the current aging levels are located are determined;

[0019] determine a transformation period of the four structural layers of each sub-region from the current aging level to a next aging level according to the aging stage of the four structural layers of each sub-region at the current aging level;

[0020] The aging stage includes an early stage, a middle stage and a late stage.

[0021] Further, the determination of the maintenance sequence of each sub-region according to the current aging level of the four structural layers of each sub-region and the transformation period from the current aging level to the next aging level specifically includes:

[0022] comparing the aging levels of the same type of structural layers of each sub-region to obtain four aging level sequences of the same type of structural layers of each sub-region, each of the four aging level sequences containing a plurality of elements related to the aging level;

[0023] based on any of the four aging level sequences, comparing the aging stages of elements with the same aging level, and adjusting the four aging level sequences according to the comparison result;

[0024] determining the maintenance sequence of each sub-region according to the adjusted four aging level sequences.

[0025] Further, the construction of the inversion model specifically includes:

[0026] obtaining electromagnetic wave data related to the four structural layers of the highway respectively;

[0027] extracting feature vectors of the electromagnetic wave data related to the four structural layers respectively;

[0028] inputting the feature vectors of the electromagnetic wave data related to the four structural layers respectively into four deep convolutional neural networks for training to obtain the inversion model associated with the four structural layers respectively.

[0029] Further, the method further includes:

[0030] determining the amount of raw materials required for the maintenance of each sub-region according to the current aging level of the four structural layers of each sub-region and the transformation period from the current aging level to the next aging level;

[0031] maintaining each sub-region according to the amount of raw materials required for the maintenance of each sub-region.

[0032] In a second aspect, the present application provides a highway maintenance road construction safety management device, comprising:

[0033] An acquisition module is configured to acquire road surface electromagnetic wave data corresponding to different structural layers of a road surface, the different structural layers of the road surface including a surface layer, a base layer, a cushion layer and a subbase layer; and a first feature vector of the road surface electromagnetic wave data corresponding to the different structural layers of the road surface is extracted, the first feature vector including amplitude, phase, frequency and wavelength;

[0034] A training module is configured to mark the first feature vector corresponding to the different structural layers of the road surface with structural layer labels and aging parameter labels respectively to obtain a training data set, and input the training data set into a deep convolutional neural network for training to obtain an inversion model;

[0035] An inversion module is configured to divide a road to be maintained into a plurality of sub-regions, acquire road surface electromagnetic wave data corresponding to different structural layers in each sub-region, and extract a second feature vector of the road surface electromagnetic wave data corresponding to the different structural layers in each sub-region; and input the second feature vector corresponding to the different structural layers in each sub-region into the inversion model to obtain aging parameters associated with four structural layers of each sub-region respectively.

[0036] A first determination module is configured to determine, according to the aging parameters, a current aging level of the four structural layers of each sub-region and a conversion period from the current aging level to a next aging level.

[0037] A second determination module is configured to determine, according to the current aging level of the four structural layers of each sub-region and the conversion period from the current aging level to the next aging level, a maintenance sequence of each sub-region.

[0038] The present application provides a computer readable storage medium, the storage medium stores a computer program, the computer program is executed by a processor to realize the road maintenance road construction safety management method.

[0039] The present application provides a computer device, including a memory, a processor and a computer program stored on the memory and executable on the processor, the processor executes the program to realize the road maintenance road construction safety management method.

[0040] The at least one technical solution adopted by the present application can achieve the following beneficial effects:

[0041] The application first acquires road surface electromagnetic wave data corresponding to different structural layers of the road surface, the different structural layers of the road surface including a surface layer, a base layer, a cushion layer and a subbase layer, and extracts a first feature vector of the road surface electromagnetic wave data corresponding to the different structural layers, the feature vector including amplitude, phase, frequency and wavelength; then the first feature vector corresponding to the different structural layers of the road surface is respectively marked with structural layer and aging parameter to obtain a training data set, the training data set is input into a deep convolutional neural network for training to obtain an inversion model; then the road to be maintained is divided into multiple sub-regions, road surface electromagnetic wave data corresponding to different structural layers in each sub-region is acquired, and a second feature vector of the road surface electromagnetic wave data corresponding to the different structural layers in each sub-region is extracted; the second feature vector corresponding to the different structural layers in each sub-region is input into the inversion model to obtain aging parameters respectively associated with the four structural layers of each sub-region; finally, the current aging level of the four structural layers of each sub-region is determined according to the aging parameters, and the conversion period from the current aging level to the next aging level is determined; further, the maintenance order of each sub-region is determined according to the current aging level of the four structural layers of each sub-region and the conversion period from the current aging level to the next aging level.

[0042] Through the above scheme, the aging level of the multiple road surface structural layers of the road and the conversion period from the current aging level to the next aging level are combined, which can preferentially find the sub-region where the inner layer structure of the road surface appears aging, and preferentially maintain the sub-region where the inner layer structure of the road surface exists aging, and after the maintenance of the sub-region where the inner layer structure exists aging is completed, the sub-region where the outer layer structure of the road surface exists aging is maintained in the standing time after maintenance, which can reduce the waiting time in the road maintenance process; and when the aging levels of the same structural layers of different sub-regions are the same, the sub-region where the structural layer is about to convert to the next aging level is preferentially maintained, which can avoid further aging of the structural layers of each sub-region due to the lack of timely maintenance in the road maintenance process, thereby reducing the difficulty of road maintenance and reducing the cost of road maintenance. In summary, the precision of road maintenance can be improved and the efficiency of road maintenance can be improved through the application. BRIEF DESCRIPTION OF DRAWINGS

[0043] The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiments of the application and together with the description serve to explain the application. In the drawings:

[0044] Figure 1 A road maintenance road construction safety management method flowchart provided by the application;

[0045] Figure 2A schematic diagram of an application scene of a road maintenance road construction safety management method provided by the present application is provided.

[0046] Figure 3 A principle diagram for obtaining a road surface aging degree corresponding to each sub-region included in the to-be-maintained road provided by the present application is provided.

[0047] Figure 4 A schematic diagram of a road maintenance road construction safety management device provided by the present application is provided.

[0048] Figure 5 A schematic diagram of a computer device for implementing the road maintenance road construction safety management method provided by the present application is provided. DETAILED DESCRIPTION

[0049] In order to make the objectives, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be described below in detail with reference to specific embodiments of the present application and corresponding drawings. Obviously, the described embodiments are only some of the embodiments of the present application, but not 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 work fall within the scope of protection of the present application.

[0050] The server mentioned in the present application can be a server arranged in a business platform, or a device such as a desktop computer, a notebook computer, etc. capable of executing the scheme of the present application. For the convenience of description, the server will be taken as the execution subject for description below. The technical solutions provided by the embodiments of the present application will be described in detail below with reference to the drawings.

[0051] Figure 1 A schematic diagram of a road maintenance road construction safety management method flow provided by the present application is provided, which specifically includes the following steps.

[0052] S10: Obtain road surface electromagnetic wave data corresponding to different structural layers of the road surface, the different structural layers of the road surface including a surface layer, a base layer, a cushion layer and a bottom base layer; extract a first feature vector of the road surface electromagnetic wave data corresponding to the different structural layers of the road surface, the feature vector including amplitude, phase, frequency and wavelength.

[0053] In the embodiment, ground penetrating radar (GPR) is used to obtain pavement electromagnetic wave data corresponding to different structural layers of the pavement. The ground penetrating radar is a device for detecting underground medium by using high-frequency electromagnetic waves. It transmits electromagnetic waves to the underground through a transmitting antenna. When the electromagnetic waves encounter underground media or target bodies with different electrical properties, they are reflected, and the reflected waves are received by a receiving antenna. By analyzing the characteristics of the received electromagnetic wave signals, such as waveform, intensity, phase, and frequency, the spatial position, structure, and morphology of the underground medium can be inferred. The parameters of the first feature vector of the electromagnetic wave data include amplitude, frequency, phase, and wavelength. When the electromagnetic waves emitted by the ground penetrating radar encounter different structural layers of the medium, reflection and transmission occur. The intensity of the reflected signal mainly depends on the difference in the relative permittivity of the upper and lower layers of the medium. The greater the difference in permittivity, the greater the reflected electromagnetic wave energy. Therefore, when the ground penetrating radar detects different structural layers, the received electromagnetic wave signals will have different reflection characteristics due to the difference in the permittivity of the materials contained in the different structural layers, further resulting in different frequencies, phases, wavelengths, and amplitudes of the electromagnetic waves.

[0054] S20: The first feature vector corresponding to different structural layers of the pavement is marked with structural layer and aging parameter, respectively, to obtain a training data set. The training data set is input into a deep convolutional neural network for training to obtain an inversion model.

[0055] In the embodiment, the aging parameters refer to parameters corresponding to multiple damage types contained in each structural layer, for example, the aging parameters of the surface layer are: crack length 12 cm, crack depth 6 cm, and asphalt thickness 10 cm. The aging parameters of the base layer are: base layer thickness 15 cm, compaction degree 92%, crack length 10 cm, and crack depth 5 cm.

[0056] The inversion model is obtained by training the feature vectors of electromagnetic wave data related to multiple structural layers of the road surface using a deep convolutional neural network (DCNN). The deep convolutional neural network includes an input layer, a hidden layer, and an output layer. Among them, the hidden layer includes a convolution layer, a pooling layer, and a fully connected layer. It should be noted that the number of convolution kernels of the inversion model is determined according to the number of damage types possessed by different structural layers of the road to be maintained. For example, the surface layer of the road to be maintained may contain the following types of damage: crack length, crack depth, and asphalt thickness. The first layer of the convolution layer of the inversion model corresponding to the surface layer needs to have at least 3 convolution kernels, so as to use the 3 convolution kernels to extract the features of the three damage types from the input surface layer electromagnetic wave data. Furthermore, since the number of convolution kernels of the second layer should be the number of all possible combinations of the features of the first layer, and the features of each damage type in the first layer contain three states (-1, 0, 1), the second layer of the convolution layer of the inversion model needs to have at least 3 convolution kernels. 3 Convolution kernels, or 27 convolution kernels. Pooling layers are used to reduce the number of parameters and computational complexity of the inversion model. The fully connected layer acts as a "classifier" in the inversion model, mapping the learned feature representations to the sample's label space, integrating the features of the previous layer, and outputting the prediction results through an activation function.

[0057] Optionally, before extracting the characteristic vectors of the electromagnetic wave data corresponding to the four structural layers, the electromagnetic wave data corresponding to the four structural layers can be denoised using methods such as wavelet filtering to improve the accuracy of the inversion model output of the aging parameters associated with the four structural layers.

[0058] S30: Divide the highway to be maintained into multiple sub-areas, obtain pavement electromagnetic wave data corresponding to different structural layers in each sub-area, and extract the second eigenvectors of the pavement electromagnetic wave data corresponding to different structural layers in each sub-area; input the second eigenvectors corresponding to different structural layers in each sub-area into the inversion model to obtain aging parameters respectively associated with the four structural layers in each sub-area.

[0059] In this embodiment, the road to be maintained refers to a damaged road surface that needs maintenance but has not yet been maintained. Each sub-region refers to a plurality of regions obtained by dividing the road to be maintained according to its structural characteristics.

[0060] refer to Figure 2 A small car equipped with a ground-penetrating radar is emitting electromagnetic waves to the road under maintenance and receiving the reflected electromagnetic waves. Figure 2 The road to be maintained is an S-shaped road. Based on the structural characteristics of the road, multiple sets of parallel coordinates can be marked on both sides of the road. The road can be divided into multiple sub-areas according to the multiple sets of parallel coordinates, such as Figure 2Neutron region 1, sub-region 2, sub-region 3,..., sub-region n.

[0061] With reference to the foregoing Figure 2 In different sub-regions of the highway to be maintained, the pavement structure of each sub-region includes four layers: surface layer, base layer, cushion layer and sub-base layer. The surface layer is the uppermost layer of the highway to be maintained, directly contacting with vehicle load and atmosphere. Its main features include high strength, strong deformation resistance, good stability and flatness, good wear resistance and skid resistance, and water resistance. Commonly used materials include cement concrete, asphalt concrete, block stone, asphalt macadam mixture, etc. The base layer is located below the surface layer, mainly bearing the driving load transmitted from the surface layer, and diffusing and transmitting it to the cushion layer and soil foundation. The base layer features high strength, good water stability and high flatness. Commonly used materials of the base layer include various binder stabilized soil or gravel, lean concrete, natural sand and gravel, and stone, etc. The cushion layer is located between the base layer and the sub-base layer, mainly improving the humidity and temperature conditions of the soil foundation, and preventing the roadbed soil from extruding into the sub-base layer. The main functions of the sub-base layer are to improve the overall bearing capacity of the pavement, reduce the compressive stress on the top surface of the roadbed, mitigate the influence of uneven deformation of the roadbed on the surface layer, and provide waterproof and construction convenience. The materials of the sub-base layer usually include gravel, gravel, stabilized soil, etc., and sometimes industrial waste slag is also used as the material.

[0062] S40: determining the current aging level of each sub-region of the four structure layers according to the aging parameters, and the conversion period from the current aging level to the next aging level.

[0063] In this embodiment, the current aging level refers to the aging level of each structure layer at the current time, the next aging level refers to the aging level one level higher than the current aging level, and the conversion period refers to the length of time required to convert from the current aging level to the next aging level.

[0064] S50: determining the maintenance order of each sub-region according to the current aging level of each sub-region of the four structure layers, and the conversion period from the current aging level to the next aging level.

[0065] In this embodiment, the maintenance order refers to the order of maintenance of each sub-region. According to the current aging level of each sub-region of the four structure layers, and the conversion period from the current aging level to the next aging level, the maintenance order of each sub-region is determined.

[0066] Optionally, the four structure layers are sequentially lowered in priority from the inner layer to the outer layer, that is, the priority of the base layer is the highest, and the priority of the surface layer is the lowest, the sub-regions where the base layer is damaged are preferentially screened out, the aging grades of the base layer of each sub-region are sorted to obtain an aging grade sequence of the base layer, and the maintenance sequence of each sub-region is determined according to the aging grade sequence of the base layer. When the aging grades of the base layer are the same, the sub-regions with the same aging grade are sorted from small to large according to the length of the conversion period from the current aging grade to the next aging grade, and the sub-region with a short conversion period is preferentially maintained.

[0067] It should be noted that in the present embodiment, "first" and "second" only serve to distinguish, and do not represent the order.

[0068] Based on Figure 1The method for managing road construction safety in road maintenance comprises the following steps: obtaining road electromagnetic wave data corresponding to different structural layers of a road surface, the different structural layers of the road surface comprising a surface layer, a base layer, a cushion layer and a subbase layer, and extracting a first feature vector of the road electromagnetic wave data corresponding to the different structural layers of the road surface, the feature vector comprising amplitude, phase, frequency and wavelength; marking the first feature vector corresponding to the different structural layers of the road surface with structural layer labels and aging parameter labels to obtain a training data set; inputting the training data set into a deep convolutional neural network for training to obtain an inversion model; dividing a road to be maintained into a plurality of sub-regions, obtaining road electromagnetic wave data corresponding to different structural layers in each sub-region, and extracting a second feature vector of the road electromagnetic wave data corresponding to the different structural layers in each sub-region; inputting the second feature vector corresponding to the different structural layers in each sub-region into the inversion model to obtain aging parameters associated with the four structural layers of each sub-region; and finally determining the current aging level of the four structural layers of each sub-region and the conversion period from the current aging level to the next aging level according to the aging parameters. Through the embodiment, the aging level of the plurality of road surface structural layers of the road and the conversion period from the current aging level to the next aging level are combined, the sub-regions with aging in the inner layer structure of the road surface are preferentially maintained, the sub-regions with aging in the outer layer structure of the road surface are maintained during the standing time after maintenance, the waiting time in the road maintenance process is reduced, the structural layers of each sub-region are preferentially maintained when the aging levels of the same structural layers of different sub-regions are the same, the structural layers of each sub-region are preferentially maintained when the aging levels of the same structural layers of different sub-regions are the same, the difficulty of road maintenance is reduced, and the cost of road maintenance is reduced. In summary, the scheme shown in the embodiment can improve the accuracy of road maintenance and improve the efficiency of road maintenance.

[0069] In the application of the method for managing road construction safety in road maintenance, the steps can be executed in any order, and the execution order of the steps can be determined as required, and the application does not limit the execution order of the steps. Figure 1

[0070] Optionally, in the embodiment, the feature vectors of the electromagnetic wave data can be classified to obtain the feature vectors of the surface layer electromagnetic wave data, the feature vectors of the base layer electromagnetic wave data, the feature vectors of the cushion layer electromagnetic wave data and the feature vectors of the subbase layer electromagnetic wave data.

[0071] In the embodiment, the feature vectors of the electromagnetic wave data can be classified to obtain the feature vectors of the surface layer electromagnetic wave data, the feature vectors of the base layer electromagnetic wave data, the feature vectors of the cushion layer electromagnetic wave data and the feature vectors of the subbase layer electromagnetic wave data. Figure 3 ​, the electromagnetic wave data corresponding to the four structural layers are mixed before being input into the feature classification model, i.e., the electromagnetic wave data corresponding to the four structural layers are overlapped. Since the electromagnetic wave is absorbed by the medium contained in different structural layers when passing through the different structural layers, and scattering occurs in each structural layer, the electromagnetic wave will be attenuated to different degrees when passing through each structural layer. Moreover, since the conductivity of the medium contained in different structural layers is different, when the electromagnetic wave passes from one structural layer to another structural layer, the phenomenon of electromagnetic wave energy reflection occurs at the interface of the structural layers, i.e., part of the energy is reflected back to the upper structural layer from the lower structural layer according to the reflection law. Therefore, the feature classification model can process the multiple feature vectors contained in the mixed electromagnetic wave, and classify the feature vectors of the electromagnetic wave data corresponding to different structural layers according to the energy attenuation of the electromagnetic wave after passing through different structural layers.

[0072] The feature classification model is obtained by training the feature vectors corresponding to the electromagnetic wave data of different structural layers based on a self-organizing feature map network (SOM).

[0073] The input layer is configured to input the feature vectors of the electromagnetic wave data, the competitive layer is configured to cluster the elements in the feature vectors of the electromagnetic wave data according to the similarity of the element values, and the output layer is configured to output the clustered feature vectors of the electromagnetic wave data.

[0074] Optionally, after the feature vectors of the mixed electromagnetic wave returned by the road surface are classified by the self-organizing neural network, the support vector machine can be used to find the best boundary between the feature vectors of the electromagnetic wave data of different categories in the classified feature vectors of the electromagnetic wave data, so as to realize more accurate classification of the feature vectors corresponding to different types of electromagnetic wave.

[0075] After the feature vectors of the electromagnetic wave data corresponding to different structural layers are classified by the self-organizing map network, the performance of the feature classification model is evaluated using a test set, and the parameters of the feature classification model are modified to further improve the classification ability of the feature classification model for the feature vectors of complex electromagnetic wave.

[0076] Optionally, in this embodiment, the feature vectors corresponding to the surface layer electromagnetic wave data, the base layer electromagnetic wave data, the cushion layer electromagnetic wave data and the bottom base layer electromagnetic wave data are extracted, and the feature vectors corresponding to the surface layer electromagnetic wave data, the base layer electromagnetic wave data, the cushion layer electromagnetic wave data and the bottom base layer electromagnetic wave data are input into the first inversion model, the second inversion model, the third inversion model and the fourth inversion model respectively, to obtain the aging parameters associated with the surface layer, the base layer, the cushion layer and the bottom base layer respectively.

[0077] In this embodiment, reference Figure 3 The first inversion model refers to an inversion model specifically used to process the characteristic vector of the surface layer electromagnetic wave data, which can process the characteristic vector of the surface layer electromagnetic wave data to obtain the aging parameters of the surface layer; the second inversion model refers to an inversion model specifically used to process the characteristic vector of the base layer electromagnetic wave data, which can process the characteristic vector of the base layer electromagnetic wave data to obtain the aging parameters of the base layer; the third inversion model refers to an inversion model specifically used to process the characteristic vector of the cushion layer electromagnetic wave data, which can process the characteristic vector of the cushion layer electromagnetic wave data to obtain the aging parameters of the cushion layer; the fourth inversion model refers to an inversion model specifically used to process the characteristic vector of the bottom layer electromagnetic wave data, which can process the characteristic vector of the bottom layer electromagnetic wave data to obtain the aging parameters of the bottom layer.

[0078] Among them, the aging parameters common to the four structural layers include: crack length, crack depth and number of cracks; the aging parameters unique to the surface layer include but are not limited to: asphalt thickness and smoothness; the aging parameters unique to the base layer include but are not limited to: base density; the aging parameters unique to the cushion layer include but are not limited to: compaction; the aging parameters unique to the subbase layer include but are not limited to: looseness.

[0079] Optionally, the asphalt thickness is calculated as follows:

[0080]

[0081] Where X is the thickness of asphalt, I is the intensity of the received electromagnetic wave, I0 is the initial intensity of the electromagnetic wave, and K is the attenuation coefficient of the electromagnetic wave in the asphalt material.

[0082] The solution shown in this embodiment classifies the eigenvectors of electromagnetic wave data corresponding to each of the four structural layers to obtain eigenvectors for the surface layer, base layer, cushion layer, and subbase layer. These eigenvectors are then input into the first, second, third, and fourth inversion models, respectively, to obtain aging parameters associated with the surface layer, base layer, cushion layer, and subbase layer, respectively. Compared to existing techniques that directly invert the eigenvectors of electromagnetic wave data using a single inversion model, this solution utilizes strongly correlated inversion models to process the electromagnetic wave data associated with each structural layer, thereby obtaining more accurate aging parameters for each structural layer.

[0083] Furthermore, in one or more embodiments of the present invention, the current aging level of the four structural layers of each sub-region and the conversion period from the current aging level to the next aging level are determined based on the aging parameters, specifically including:

[0084] S401: Reconstructing the aging parameters by using the pre-constructed parameter quantization table to obtain the aging values corresponding to different aging types contained in the four structure layers.

[0085] In this embodiment, the parameter quantization table refers to converting the aging parameters corresponding to each aging type into specific aging values associated with each aging type. For example, for the first sub-region, the aging parameters of the surface layer structure thereof include: crack length (1) of 10 cm, crack length (2) of 12 cm, crack length (3) of 9 cm; crack depth (1) of 4 cm, crack depth (2) of 6 cm, crack depth (3) of 8 cm; asphalt thickness of 10 cm. The aging values obtained by using the parameter quantization table for the surface layer structure of the first sub-region include: crack length 31, crack depth 6, asphalt thickness 10. The numerical calculation method of each aging value is: crack length = 10 + 12 + 9, crack depth = (4 + 6 + 8) / 3, i.e. the value of crack length is the sum of the lengths of all cracks, and the dimension is removed; the value of crack depth is the average depth of all cracks, and the dimension is removed; the aging value of asphalt thickness is the average asphalt thickness of the sub-region, and the dimension is removed.

[0086] S402: According to the aging values corresponding to different aging types and the weights of different aging types in the structure layer where they are located, determine the current aging grade of each sub-region corresponding to the four structure layers, and the aging stage where the current aging grade is located.

[0087] In this embodiment, the weight of different aging types in the structure layer where they are located refers to the importance of each aging type in the structure layer where it is located, and the sum of the weights of all aging types in each structure layer is 1.

[0088] For the surface layer structure of the first sub-region, the weight of crack length is 0.1, the weight of crack depth is 0.4, and the weight of asphalt thickness is 0.5. Assuming that the aging values are: crack length 31, crack depth 6, asphalt thickness 10, the aging grade of the weighted sum of different values is: greater than 2 and less than 8, which is the first aging grade; greater than 8 and less than 15, which is the second aging grade; greater than 15, which is the third aging grade. The calculation method of the aging grade of the surface layer structure of the first sub-region is: 31 x 0.1 + 6 x 0.4 + 10 x 0.5 = 10.5, which is in the second aging grade. The division range of each aging grade can be adjusted adaptively according to different types of roads.

[0089] S403: Determine the conversion period of the four structural layers of each sub-region from the current aging level to the next aging level according to the aging stage of the four structural layers of each sub-region at the current aging level.

[0090] The aging stage includes early stage, middle stage and late stage.

[0091] In this embodiment, each aging level includes early stage, middle stage and late stage. The numerical range corresponding to each aging level is further divided into three intervals, i.e. each aging level is further divided into early stage, middle stage and late stage. For example, for the first sub-region, the range of the first aging level of the surface layer is greater than 2 and less than 8, and the division can obtain the numerical range of the early stage is greater than or equal to 2, and less than 4; the numerical range of the middle stage is greater than or equal to 4, and less than 6; the numerical range of the late stage is greater than or equal to 6, and less than 8.

[0092] Optionally, the way of determining the conversion period of the four structural layers of each sub-region from the current aging level to the next aging level in this embodiment includes but is not limited to:

[0093] First, obtain the time length data set of each structural layer of the road from different aging stages in the i-th aging level to the i+1-th aging level, i is a positive integer. Then, calculate the average time length of each structural layer of the road from different aging stages in the i-th aging level to the i+1-th aging level according to the above time length data set, and determine the average time length as the conversion period of each structural layer from different aging stages in the i-th aging level to the i+1-th aging level.

[0094] The embodiment reconstructs the aging parameters by using the pre-constructed parameter quantization table, obtains the aging values corresponding to different aging types contained in the four structure layers respectively, and determines the current aging grade of each sub-region corresponding to the four structure layers respectively according to the aging values corresponding to different aging types and the weight of different aging types in the structure layer where the aging types are located, and finally determines the conversion period of the four structure layers of each sub-region from the current aging grade to the next aging grade according to the aging stage where the four structure layers of each sub-region in the four structure layers respectively correspond to the current aging grade, which can accurately determine the conversion period of each structure layer of any sub-region of the to-be-maintained highway from the current aging grade to the next aging grade, provide priority basis for subsequent determination of the maintenance sequence of each sub-region, that is, preferentially maintain the sub-region where the structure layer corresponding to the short conversion period from the current aging grade to the next aging grade is located, which can avoid the further aging of the structure layer of each sub-region in a short time due to the failure to obtain timely maintenance in the highway maintenance process, thereby reducing the difficulty of highway maintenance and reducing the cost of highway maintenance.

[0095] In addition, in one or more embodiments of the present application, the maintenance sequence of each sub-region is determined according to the current aging grade of the four structure layers of each sub-region and the conversion period from the current aging grade to the next aging grade, and specifically includes:

[0096] S501: Compare the aging grades of the same type structure layers of each sub-region to obtain four aging grade sequences of the same type structure layers of each sub-region, and each of the four aging grade sequences contains a plurality of elements related to the aging grade.

[0097] In the embodiment, the aging grades of the surface layer, the base layer, the cushion layer and the sub-base layer of each sub-region are compared respectively to obtain the surface layer aging grade sequence, the base layer aging grade sequence, the cushion layer aging grade sequence and the sub-base layer aging grade sequence, and each aging grade sequence is arranged from high to low in aging grade. Each aging grade sequence is converted into an aging grade matrix respectively, which is convenient for subsequent data processing, for example, the surface layer aging grade sequence is a matrix The to-be-maintained highway is divided into five sub-regions, and the subscript of each element in the matrix represents the aging grade of the surface layer of the sub-region, and the superscript of each element represents the number of the sub-region. Similarly, the base layer aging grade sequence matrix B, the cushion layer aging grade sequence matrix C and the sub-base layer aging grade sequence matrix D can be obtained, which will not be described here.

[0098] S502: Based on any aging grade sequence in the four aging grade sequences, compare the aging stages of elements with the same aging grade, and adjust the four aging grade sequences according to the comparison result.

[0099] In this embodiment, for any aging level sequence, the aging stages of elements with the same aging level are compared, and any aging level sequence is adjusted according to the comparison result. For example, the surface layer aging level sequence is a matrix in The aging level of the surface layer of sub-region 5 is the first aging level. Assuming that the aging stage of the surface layer of sub-region 4 is the early aging stage of the first aging level, the aging stage of the surface layer of sub-region 2 is the late aging stage of the first aging level, the matrix A is adjusted to obtain The corresponding surface aging grade sequence has also been adjusted.

[0100] S503: Determine the maintenance order of each sub-area according to the adjusted four aging level sequences.

[0101] In this embodiment, the curing order of each sub-area is determined based on the aging grade sequence corresponding to the sub-base layer. If any number of elements in the sub-base layer aging grade sequence are at the same aging grade and aging stage, the curing order of each sub-area is then determined based on the aging grade sequence of the cushion layer. Similarly, the curing order of each sub-area is determined based on the subscripts of each element in the inner layer aging grade sequence. If any number of sub-areas in the same structural layer are at the same aging grade and aging stage, the curing order of these any number of sub-areas is determined based on the aging grade sequence of the next structural layer.

[0102] It should be noted that the "inside" in the above embodiment refers to the lower structural layer of the pavement structure layer, and the "outside" refers to the upper structural layer of the pavement structure layer. For example: the surface layer is the upper layer relative to the base layer, the base layer is the upper layer relative to the cushion layer, and the cushion layer is the upper layer relative to the subbase layer.

[0103] The present invention first compares the aging levels of similar structural layers in each sub-region, generating four aging level sequences for the same structural layers in each sub-region. Based on any of the four aging level sequences, the aging stages of elements with the same aging level are compared and the four aging level sequences are adjusted based on the comparison results. Finally, the maintenance sequence for each sub-region is determined based on the adjusted four aging level sequences. This solution prioritizes the maintenance sequence for each sub-region based on the aging level of the inner layer structure and the aging stage of each level within the highway sub-region. This allows for batch maintenance of sub-regions with damaged inner layers, minimizing damage to the pavement structure during the maintenance process. Furthermore, after the inner layer structure is cured, the next damaged structural layer can be cured during the rest period. Furthermore, when performing maintenance on any structural layer of the pavement, this solution prioritizes the structural layer that is about to transition to the next aging level, reducing the rate of aging level increases for each structural layer during the maintenance process, thereby reducing maintenance costs.

[0104] In addition, in one or more embodiments of the present invention, the construction of the inversion model specifically includes:

[0105] The pre-collected electromagnetic wave data related to the four structural layers of the highway are input into four deep convolutional neural networks for training respectively, and inversion models associated with the four structural layers are obtained.

[0106] In this embodiment, reference Figure 3 The feature vectors of the pre-collected electromagnetic wave data related to the four structural layers of the highway are input into the feature classification model for data classification, and the feature vectors of the surface layer electromagnetic wave data, the feature vectors of the base layer electromagnetic wave data, the feature vectors of the cushion layer electromagnetic wave data, and the feature vectors of the sub-base layer electromagnetic wave data are obtained. The feature vectors of the surface layer electromagnetic wave data, the feature vectors of the base layer electromagnetic wave data, the feature vectors of the cushion layer electromagnetic wave data, and the feature vectors of the sub-base layer electromagnetic wave data are then respectively input into four deep convolutional neural networks for training to obtain inversion models associated with the four structural layers. This solution can train strongly correlated inversion models for electromagnetic wave data related to different structural layers, and can more accurately process the feature vectors of electromagnetic wave data corresponding to different structural layers, thereby improving the accuracy of aging level judgment for different structural layers.

[0107] Optionally, in one or more embodiments of the present invention, the present invention further includes:

[0108] The amount of raw materials required for maintenance of each sub-area is determined based on the current aging level of the four structural layers of each sub-area and the conversion cycle from the current aging level to the next aging level.

[0109] In the embodiment, for any sub-region, the amount of raw materials required for maintenance of the sub-region is different when different structural layers of the sub-region correspond to different aging grades, and the higher the aging grade, the greater the amount of raw materials required. Moreover, the aging stages of the same aging grade are also different, and the amount of raw materials required for maintenance of the sub-region in the early stage is less than that in the late stage.

[0110] Optionally, the manner of determining the amount of raw materials required for maintenance of each sub-region includes but is not limited to: obtaining a historical raw material consumption data set for road maintenance, and correcting the raw material consumption data set according to the actual maintenance effect. Based on the corrected raw material consumption data set, a maintenance material consumption table corresponding to each aging grade and each aging grade containing different aging stages is established, and the amount of raw materials required for maintenance of each sub-region is determined according to the maintenance material consumption table.

[0111] According to the amount of raw materials required for maintenance of each sub-region, each sub-region is maintained.

[0112] The embodiment can determine the amount of raw materials required for maintenance of each sub-region according to the aging grade corresponding to each structural layer of each sub-region and the aging stage in the aging grade, can reduce the cost of maintenance of each sub-region, and can avoid excessive use of raw materials in the process of road maintenance, which can cause some trace elements to exceed the standard and reduce environmental pollution.

[0113] The above is a road maintenance road construction safety management method provided by one or more embodiments of the present application. Based on the same idea, the present application also provides a corresponding road maintenance road construction safety management device, as shown in Figure 4 .

[0114] Figure 4 A road maintenance road construction safety management device provided by the present application is shown in the figure, which includes:

[0115] The acquisition module 01 is used to acquire road surface electromagnetic wave data corresponding to different structural layers of the road surface, and the different structural layers of the road surface include the surface layer, the base layer, the cushion layer and the bottom base layer. The first feature vector of the road surface electromagnetic wave data corresponding to the different structural layers of the road surface is extracted, and the first feature vector includes amplitude, phase, frequency and wavelength.

[0116] The training module 02 is used to mark the first feature vector corresponding to the different structural layers of the road surface with structural layer marks and aging parameter marks respectively to obtain a training data set, and the training data set is input into a deep convolutional neural network for training to obtain an inversion model.

[0117] The inversion module 03 is used for dividing the road to be maintained into a plurality of sub-regions, acquiring the pavement electromagnetic wave data corresponding to different structure layers in each sub-region, and extracting the second feature vector of the pavement electromagnetic wave data corresponding to different structure layers in each sub-region; the second feature vector corresponding to different structure layers in each sub-region is input into the inversion model to obtain the aging parameters respectively associated with the four structure layers of each sub-region.

[0118] The first determination module 04 is used for determining the current aging level of the four structure layers of each sub-region according to the aging parameters, and determining the conversion period from the current aging level to the next aging level.

[0119] The second determination module 05 is used for determining the maintenance sequence of each sub-region according to the current aging level of the four structure layers of each sub-region and the conversion period from the current aging level to the next aging level.

[0120] The specific limitations of the road maintenance road construction safety management device can be referred to the limitations of the road maintenance road construction safety management method in the above, which will not be repeated here. Each module in the road maintenance road construction safety management device can be realized by software, hardware and their combination. The above modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so that the processor calls and executes the operations corresponding to each module.

[0121] The application also provides a computer readable storage medium, which stores a computer program, and the computer program can be used to execute the road maintenance road construction safety management method. Figure 1 The application also provides a computer readable storage medium, which stores a computer program, and the computer program can be used to execute the road maintenance road construction safety management method.

[0122] The application also provides a computer readable storage medium, which stores a computer program, and the computer program can be used to execute the road maintenance road construction safety management method. Figure 5 The structure diagram of the computer device is shown in the figure, which includes a processor, an internal bus, a network interface, a memory and a non-volatile memory, and can also include other hardware required by the business. Figure 5 The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs to realize the road maintenance road construction safety management method. Figure 1 The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs to realize the road maintenance road construction safety management method.

[0123] Those skilled in the art can understand that all or part of the processes in the methods of the embodiments can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the embodiments of the methods. Any reference to memory, storage, database or other medium used in the embodiments of the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0124] The technical features of the above embodiments can be combined in any manner. For the sake of brevity, not all possible combinations of the technical features in the embodiments are described, but it should be considered that any combination of the technical features is within the scope of the present application as long as there is no contradiction.

Claims

1. A highway maintenance and road construction safety management method, characterized in that: include: Acquire pavement electromagnetic wave data corresponding to different pavement structural layers, the different pavement structural layers including a surface layer, a base layer, a cushion layer, and a subbase layer; extract first eigenvectors of the pavement electromagnetic wave data corresponding to the different pavement structural layers, the first eigenvector including amplitude, phase, frequency, and wavelength; The first eigenvectors corresponding to different structural layers of the pavement are respectively labeled with structural layers and aging parameters to obtain a training data set, and the training data set is input into a deep convolutional neural network for training to obtain an inversion model; Dividing the road to be maintained into multiple sub-areas, obtaining pavement electromagnetic wave data corresponding to different structural layers in each sub-area, and extracting second eigenvectors of the pavement electromagnetic wave data corresponding to different structural layers in each sub-area; inputting the second eigenvectors corresponding to different structural layers in each sub-area into the inversion model to obtain aging parameters respectively associated with the four structural layers in each sub-area; Determining, based on the aging parameters, the current aging levels of the four structural layers of each sub-region and the conversion period from the current aging level to the next aging level; Determining a maintenance order for each sub-area according to the current aging level of the four structural layers of each sub-area and a conversion period from the current aging level to the next aging level; The step of marking the first characteristic vectors corresponding to different structural layers of the road surface with structural layer markings and aging parameter markings specifically includes: Marking each first eigenvector as one of a surface layer, a base layer, a cushion layer, and a subbase layer according to a range of an energy value of an amplitude in each first eigenvector; Marking each first eigenvector with an aging parameter according to an abnormal value of phase, an abnormal value of frequency, and an abnormal value of wavelength contained in each first eigenvector; Determining the current aging level of the four structural layers of each sub-region and the conversion period from the current aging level to the next aging level according to the aging parameters specifically includes: Reconstructing the aging parameters using a pre-constructed parameter quantization table to obtain aging values ​​corresponding to different aging types contained in the four structural layers; Determining the current aging levels corresponding to the four structural layers of each sub-region and the aging stages at the current aging levels according to the aging values ​​corresponding to the different aging types and the weights of the different aging types in the structural layers in which they are located; determining, according to the aging stages of the four structural layers of each sub-region at the current aging level, a conversion period for the four structural layers of each sub-region to convert from the current aging level to the next aging level; The aging stages include early, middle and late stages.

2. The highway maintenance and road construction safety management method according to claim 1, characterized in that: Determining the maintenance order of each sub-area according to the current aging level of the four structural layers of each sub-area and the conversion period from the current aging level to the next aging level specifically includes: Comparing the aging levels of the same type of structural layers in the respective sub-regions to obtain four aging level sequences for the same type of structural layers in the respective sub-regions, wherein the four aging level sequences respectively include a plurality of elements related to the aging level; Based on any one of the four aging level sequences, comparing aging stages of elements having the same aging level, and adjusting the four aging level sequences according to the comparison results; The maintenance sequence of each sub-area is determined according to the adjusted four aging level sequences.

3. The highway maintenance and road construction safety management method according to any one of claims 1 to 2, characterized in that: The construction of the inversion model specifically includes: Acquire electromagnetic wave data related to the four structural layers of the highway; Extracting characteristic vectors of electromagnetic wave data respectively related to the four structural layers; The feature vector groups of electromagnetic wave data respectively related to the four structural layers are respectively input into four deep convolutional neural networks for training to obtain inversion models respectively associated with the four structural layers.

4. The highway maintenance and road construction safety management method according to any one of claims 1 to 2, characterized in that: The method further comprises: Determining the amount of raw materials required for curing each sub-region based on the current aging level of the four structural layers of each sub-region and the conversion period from the current aging level to the next aging level; Each sub-region is cured according to the amount of raw materials required for curing each sub-region.

5. A highway maintenance and road construction safety management device, used to implement the highway maintenance and road construction safety management method according to any one of claims 1 to 4, characterized in that: include: an acquisition module for acquiring pavement electromagnetic wave data corresponding to different pavement structural layers, the different pavement structural layers including a surface layer, a base layer, a cushion layer, and a subbase layer; and extracting first eigenvectors of the pavement electromagnetic wave data corresponding to the different pavement structural layers, the first eigenvector including amplitude, phase, frequency, and wavelength; A training module is used to label the first eigenvectors corresponding to different structural layers of the road surface with structural layers and aging parameters to obtain a training data set, and input the training data set into a deep convolutional neural network for training to obtain an inversion model; an inversion module for dividing the road to be maintained into a plurality of sub-regions, obtaining pavement electromagnetic wave data corresponding to different structural layers in each sub-region, and extracting second eigenvectors of the pavement electromagnetic wave data corresponding to different structural layers in each sub-region; inputting the second eigenvectors corresponding to different structural layers in each sub-region into the inversion model to obtain aging parameters respectively associated with the four structural layers in each sub-region; A first determining module is configured to determine, based on the aging parameters, the current aging level of the four structural layers of each sub-region and a conversion period from the current aging level to the next aging level; A second determining module is configured to determine a maintenance order for each sub-area based on the current aging level of the four structural layers of each sub-area and a conversion period from the current aging level to the next aging level; The training module is further configured to label each first eigenvector as one of a surface layer, a base layer, a cushion layer, and a sub-base layer according to a range of an energy value of an amplitude in each first eigenvector; and to label each first eigenvector with an aging parameter according to abnormal values ​​of phase, frequency, and wavelength contained in each first eigenvector; The first determination module is further configured to reconstruct the aging parameters using a pre-constructed parameter quantization table to obtain aging values ​​corresponding to different aging types contained in the four structural layers; determine the current aging levels corresponding to the four structural layers of each sub-area and the aging stages at the current aging levels according to the aging values ​​corresponding to the different aging types and the weights of the different aging types in the structural layers in which they are located; determine the conversion period of the four structural layers of each sub-area from the current aging level to the next aging level according to the aging stages at the current aging levels of the four structural layers of each sub-area; wherein the aging stages include early, middle and late stages.

6. A computer-readable storage medium, characterized in that The storage medium stores a computer program, and when the computer program is executed by a processor, the highway maintenance and road construction safety management method according to any one of claims 1 to 4 is implemented.

7. A computer device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the program, the method for managing highway maintenance and road construction safety according to any one of claims 1 to 4 is implemented.

Images