Risk early warning method and device for underground pipe network, terminal equipment and storage medium

Abstract

The invention relates to the technical field of risk early warning, and discloses a risk early warning method for an underground pipe network, and the method comprises the steps: obtaining the historical track data and working data of an excavator, and the node data of the underground pipe network; the historical track data and the working data of the excavator and the underground pipe network node data are fused to obtain fused data; according to the fused data and a self-adaptive anomaly detection algorithm, obtaining an anomaly score used for judging whether the excavator is close to the underground pipe network or not; and pushing early warning information under the condition that the abnormal score exceeds a set threshold value. According to the scheme, complex nonlinear characteristics in the excavator trajectory data can be processed more effectively, the timeliness and accuracy of underground pipe network risk prediction are remarkably improved, and safe and stable operation of an urban underground pipe network is ensured. The invention further discloses a risk early warning device for the underground pipe network, terminal equipment and a storage medium.

Description

Technical Field

[0001] The present application relates to the field of risk warning technology, for example, to a risk warning method, device, terminal equipment and storage medium for underground pipe networks. Background Art

[0002] With the acceleration of urbanization, underground pipelines play an increasingly important role in urban infrastructure. As the complexity and density of underground pipelines continue to increase, the protection of excavators during construction has become a key link in ensuring the safety of urban operations. As an indispensable equipment for urban construction and maintenance, any negligence during the operation of excavators may pose a potential threat to underground pipelines, thereby causing construction safety issues. Therefore, in the modern construction environment, how to effectively protect underground pipelines and avoid accidental injuries during excavator construction has become an important issue that needs to be solved urgently.

[0003] The related technology discloses a risk warning method for underground pipelines, including: constructing a prediction model to identify the potential connection between excavator operation and damage to underground pipelines based on the time sequence relationship in historical data, thereby predicting the potential risks that the underground pipelines may face during the construction process.

[0004] In the process of implementing the embodiments of the present disclosure, it is found that there are at least the following problems in the related technology:

[0005] Although the relevant technology can identify the potential connection between excavator operation and damage to underground pipelines, this method is mainly applicable to linear relationships and it is difficult to effectively handle the complex nonlinear characteristics in the excavator trajectory data, which leads to low prediction accuracy and inability to provide timely and accurate risk warnings for underground pipelines.

[0006] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present application, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0007] In order to provide a basic understanding of some aspects of the disclosed embodiments, a brief summary is given below. The summary is not an extensive review, nor is it intended to identify key / critical components or delineate the scope of protection of these embodiments, but rather serves as a prelude to the detailed description that follows.

[0008] The embodiments of the present disclosure provide a risk warning method, apparatus, terminal device and storage medium for underground pipeline networks, which can more effectively process the complex nonlinear characteristics in excavator trajectory data and improve the accuracy and timeliness of underground pipeline network risk warnings.

[0009] In some embodiments, the risk warning method for underground pipelines includes: obtaining historical trajectory data, working data and underground pipeline node data of the excavator; fusing the historical trajectory data, working data and underground pipeline node data of the excavator to obtain fused data; obtaining an anomaly score for judging whether the excavator is close to the underground pipeline based on the fused data and an adaptive anomaly detection algorithm; and pushing a warning message when the anomaly score exceeds a set threshold.

[0010] In some embodiments, the risk warning method for underground pipelines includes: performing a wavelet transform on the fused data to obtain multiple wavelet coefficients; extracting high-dimensional features based on the multiple wavelet coefficients; and analyzing the high-dimensional features using an adaptive anomaly detection algorithm to obtain an anomaly score for determining whether an excavator is close to an underground pipeline.

[0011] In some embodiments, the risk warning method for underground pipeline networks includes: constructing a feature matrix based on multiple wavelet coefficients; inputting the adjacency matrix and the feature matrix into a graph convolutional network model, and performing feature extraction using a graph convolution operation to obtain high-dimensional features.

[0012] In some embodiments, the risk warning method for underground pipelines includes: inputting high-dimensional features into a variational autoencoder to obtain a potential variable distribution; inputting the potential variable distribution into a long short-term memory network model to obtain an abnormality score output by the long short-term memory network model for determining whether an excavator is close to an underground pipeline.

[0013] In some embodiments, the risk warning method for underground pipeline networks includes: inputting high-dimensional features into an encoder model to map the high-dimensional features to a latent space using the encoder model; and obtaining a latent variable distribution based on a mean vector and a standard deviation vector of the latent space output by the encoder model.

[0014] In some embodiments, the risk warning method for underground pipeline networks includes: determining latent variables based on the mean vector, standard deviation vector and noise vector of the standard normal distribution of the latent space; inputting the latent variables into the decoder model to use the decoder model to decode the latent variables back to the high-dimensional feature space to obtain the reconstructed data distribution output by the decoder model; calculating the loss value based on the latent variable distribution, the reconstructed data distribution and the loss function; and updating and optimizing the parameters of the decoder using the back-propagation algorithm and the loss value to improve the performance of the variational autoencoder.

[0015] In some embodiments, the risk warning method for underground pipelines includes: an acquisition module, configured to acquire historical trajectory data, working data and underground pipeline node data of an excavator; a first acquisition module, configured to fuse the historical trajectory data, working data and underground pipeline node data of the excavator to obtain fused data; a second acquisition module, configured to obtain an anomaly score for determining whether the excavator is close to an underground pipeline based on the fused data and an adaptive anomaly detection algorithm; and a push module, configured to push warning information when the anomaly score exceeds a set threshold.

[0016] In some embodiments, the risk warning method for underground pipe networks includes: a processor and a memory storing program instructions, and the processor is configured to execute the aforementioned risk warning method for underground pipe networks when running the program instructions.

[0017] In some embodiments, the terminal device includes: a terminal device body; and the aforementioned device for risk warning of underground pipe networks, which is installed on the terminal device body.

[0018] In some embodiments, the storage medium stores program instructions, and when the program instructions are run, the aforementioned risk warning method for underground pipe networks is executed.

[0019] The risk warning method, device, terminal device and storage medium for underground pipe network provided by the embodiments of the present disclosure can achieve the following technical effects:

[0020] This solution can fuse the excavator's historical trajectory data, working parameters, and underground pipe network node data, and through an adaptive anomaly detection algorithm, it can flexibly respond to the nonlinear characteristics of the fused data, accurately capture anomalies in the excavator operation process, and calculate the anomaly score. Therefore, when the anomaly score reaches or exceeds the preset safety threshold, the early warning mechanism is triggered to send real-time and targeted early warning information to relevant managers. With this solution, the complex nonlinear characteristics in the excavator trajectory data can be more effectively processed, significantly improving the timeliness and accuracy of underground pipe network risk prediction, and ensuring the safe and stable operation of the urban underground pipe network.

[0021] The above general description and the following description are exemplary and explanatory only and are not intended to limit the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] One or more embodiments are exemplarily described by corresponding drawings, which do not limit the embodiments. Elements with the same reference numerals in the drawings are shown as similar elements, and the drawings do not constitute a scale limitation, and wherein:

[0023] Figure 1is a schematic diagram of a risk warning method for an underground pipe network provided by an embodiment of the present disclosure;

[0024] Figure 2 is a schematic diagram of a method for obtaining anomaly scores provided by an embodiment of the present disclosure;

[0025] Figure 3 is a schematic diagram of a method for extracting high-dimensional features provided by an embodiment of the present disclosure;

[0026] Figure 4 is a schematic diagram of another method for obtaining anomaly scores provided by an embodiment of the present disclosure;

[0027] Figure 5 is a schematic diagram of a risk warning device for an underground pipe network provided by an embodiment of the present disclosure;

[0028] Figure 6 It is a schematic diagram of another risk warning device for underground pipe networks provided by an embodiment of the present disclosure. DETAILED DESCRIPTION

[0029] In order to be able to understand the features and technical contents of the embodiments of the present disclosure in more detail, the implementation of the embodiments of the present disclosure is described in detail below in conjunction with the accompanying drawings. The attached drawings are for reference only and are not used to limit the embodiments of the present disclosure. In the following technical description, for the convenience of explanation, a full understanding of the disclosed embodiments is provided through multiple details. However, one or more embodiments can still be implemented without these details. In other cases, to simplify the drawings, well-known structures and devices can be simplified for display.

[0030] The terms "first", "second", etc. in the specification and claims of the embodiments of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the terms used in this way can be interchanged where appropriate, so that the embodiments of the embodiments of the present disclosure described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions.

[0031] Unless otherwise stated, the term "plurality" means two or more.

[0032] In the embodiment of the present disclosure, the character " / " indicates that the preceding and following objects are in an "or" relationship. For example, A / B indicates: A or B.

[0033] The term "and / or" is a description of the association relationship between objects, indicating that three relationships can exist. For example, A and / or B means: A or B, or, A and B.

[0034] The term "correspondence" may refer to an association relationship or a binding relationship. The correspondence between A and B means that there is an association relationship or a binding relationship between A and B.

[0035] Figure 1 is a schematic diagram of a risk warning method for underground pipe networks provided by an embodiment of the present disclosure; Figure 1 As shown, a risk warning method for an underground pipe network provided by an embodiment of the present disclosure includes:

[0036] S11, the terminal device obtains the historical trajectory data, working data and underground pipe network node data of the excavator.

[0037] S12, the terminal device fuses the historical trajectory data, working data and underground pipe network node data of the excavator to obtain fused data.

[0038] S13, the terminal device obtains an anomaly score for determining whether the excavator is close to the underground pipe network based on the fused data and the adaptive anomaly detection algorithm.

[0039] S14: When the abnormality score exceeds the set threshold, the terminal device pushes a warning message.

[0040] In the embodiments of the present disclosure, a terminal device refers to an electronic device with a wireless connection function. In some embodiments, the terminal device is, for example, a mobile device, a computer, or any combination thereof. The mobile device may include, for example, a mobile phone, a wearable device, a smart mobile device, a virtual reality device, etc., or any combination thereof, wherein the wearable device may include, for example, a smart watch, a smart bracelet, a pedometer, etc.

[0041] In this solution, the historical trajectory data of the excavator includes time series position data, speed and acceleration data, attitude angle data, and trajectory feature data.

[0042] Among them, time series position data refers to data that can reflect the movement trajectory of the excavator, specifically, real-time collection of the position data of the excavator at different time points to form a continuous position data sequence, that is, the position data at different time points are the longitude, latitude and altitude data at different time points.

[0043] Speed ​​and acceleration data refer to data that can reflect changes in motion state, specifically, the excavator's moving speed, linear acceleration, and angular acceleration calculated in combination with position data.

[0044] Attitude angle data refers to data that can reflect the dynamic attitude of the excavator, including the tilt angle, roll angle and yaw angle of the excavator.

[0045] The trajectory feature data is used to describe the motion trajectory features of the excavator, including turning radius, moving direction, driving distance, dwell time, etc.

[0046] Optionally, a Beidou positioning module can be installed on the excavator to obtain the location data of the excavator. An inertial measurement unit can be installed on the excavator to collect attitude angle data in real time. With this solution, by installing a Beidou positioning module and an inertial measurement unit on the excavator, the historical trajectory data of the excavator can be fully and accurately obtained, providing a solid foundation for subsequent data analysis and mining.

[0047] In this solution, the working data of the excavator includes working status data, working time data, energy consumption data, and load and pressure data.

[0048] Among them, working status data refers to the state changes of the excavator such as starting, stopping, accelerating, decelerating, and switching of working modes.

[0049] Working time data refers to the working time of the excavator in different working modes. Energy consumption data refers to the data that monitors the fuel consumption or electricity consumption of the excavator and reflects its energy usage.

[0050] Load and pressure data refers to data that reflects the working intensity and efficiency of the excavator by monitoring parameters such as the position of the working arm, load, and pressure of the hydraulic system in real time through sensors installed on the excavator.

[0051] Optionally, the working status and energy consumption data can be obtained through the internal control system of the excavator. Various sensors installed on the excavator can be used to monitor the working parameters in real time. With this solution, the working data of the excavator can be fully and accurately obtained through the internal control system and various sensors of the excavator.

[0052] In this solution, the underground pipe network node data includes node location information, node attribute information and connection relationship data.

[0053] The node location information refers to the information reflecting the layout and direction of the pipe network, specifically the latitude and longitude coordinates and buried depth of each node in the underground pipe network.

[0054] Node attribute information is used to describe the physical characteristics of the node, including node type, material, size and other parameter information.

[0055] The connection relationship data includes the connection mode between nodes and the pipeline direction data.

[0056] Alternatively, the location information and attribute information of underground pipe network nodes can be obtained using a geographic information system. Usually, this information will be marked and recorded by a professional organization when the pipe network is laid. For pipe network nodes that are not marked or have incomplete information, accurate location information and attribute information can be obtained through on-site surveys and measurements. With this solution, accurate acquisition of underground pipe network node data can be achieved.

[0057] Optionally, when acquiring the excavator's historical trajectory data, working data and underground pipeline node data, time synchronization technology can be used to ensure that the data collected by the Beidou positioning module, the data collected by the inertial measurement unit, the working data and the pipeline node data are collected and recorded under the same time reference for subsequent data analysis and processing.

[0058] In an optimized solution, after obtaining the historical trajectory data, working data and underground pipe network node data of the excavator, the following steps are also included: removing noise, processing outliers and completing data from the collected data. In this way, the accuracy and integrity of the data can be effectively guaranteed.

[0059] Optionally, after obtaining the historical trajectory data, working data and underground pipe network node data of the excavator, the terminal device can fuse the historical trajectory data, working data and underground pipe network node data of the excavator. Specifically, data fusion can be performed in a variety of ways such as data mining technology and machine learning algorithms. In one way, the terminal device can extract information with potential value for decision-making from the historical trajectory data, working data and underground pipe network node data of the excavator by applying data mining methods such as association rule mining, cluster analysis, and classification prediction, thereby realizing deep integration of multi-source data. In another way, the terminal device can also train and learn the historical trajectory data, working data and underground pipe network node data through machine learning algorithms to identify patterns, correlations and trends in the data, thereby realizing intelligent integration of data. Here, machine learning algorithms include but are not limited to decision trees, support vector machines, random forests and other algorithms.

[0060] In an optimized example, a multimodal data fusion strategy can also be used for data fusion. Specifically, a comprehensive feature representation can be pre-constructed, which can fully reflect the motion state of the excavator, the working environment, and the status information of the underground pipe network. In the process of constructing the feature representation, a joint training method is adopted, that is, multiple data sources are used for training and learning at the same time to fully explore the advantages of each data source and improve the accuracy and comprehensiveness of feature extraction. Specifically, the fusion formula is as follows:

[0061]

[0062] Among them, Ffus i on is the comprehensive feature representation, i.e. the fused data, F k represents the feature representation of the kth data source, a k It represents the weight corresponding to the kth data source.

[0063] In this way, the feature representations of multiple data sources are fused by weighted summation to form a comprehensive feature representation with higher information content and stronger representativeness. It should be noted that the selection of weights can be dynamically adjusted according to factors such as the reliability and importance of the data source to ensure the accuracy and reliability of the fusion result.

[0064] In an optimized solution, an attention mechanism can also be introduced in the process of feature fusion. Here, the attention mechanism focuses on key features that have a potential impact on the safety of underground pipeline networks by calculating the importance of each feature for the prediction task.

[0065] Among them, the attention score can be calculated as follows:

[0066] e ij =a T tanh(W[h i ;h j ])

[0067] Among them, e ij is the attention score, a and W are trainable parameters, and h i and h j It is the node feature representation.

[0068] Furthermore, the attention score can be used to calculate the attention weight, including:

[0069]

[0070] Among them, e ij is the attention score, a ij is the attention weight. In this way, by introducing the attention mechanism, the terminal device can automatically identify and enhance the most important features for the prediction task. In this way, the key features will be given higher weights in the subsequent analysis and prediction process, thereby improving the accuracy and robustness of the model. In this solution, through the introduction of multimodal data fusion and attention mechanism, the enhanced processing of key features is achieved. This method significantly improves the prediction performance and provides strong technical support for the safety monitoring and early warning of underground pipeline networks.

[0071] Furthermore, the terminal device obtains an anomaly score for judging whether the excavator is close to the underground pipe network based on the fused data and the adaptive anomaly detection algorithm, including: the terminal device performs a wavelet transform on the fused data to obtain multiple wavelet coefficients. The terminal device extracts high-dimensional features based on the multiple wavelet coefficients. The terminal device analyzes the high-dimensional features using the adaptive anomaly detection algorithm to obtain an anomaly score for judging whether the excavator is close to the underground pipe network. In this way, an anomaly score reflecting the risk status of whether the excavator is close to the underground pipe network can be accurately output, providing an accurate data basis for risk warning of the underground pipe network.

[0072] Furthermore, when the anomaly score exceeds a set threshold, a warning message is pushed. In one example, the set threshold may be 0.8. Thus, when the anomaly score exceeds 0.8, it indicates that the excavator is close to the underground pipe network, and the terminal device may push a warning message.

[0073] The risk warning method for underground pipe networks provided by the embodiment of the present disclosure can fuse the historical trajectory data, working parameters and underground pipe network node data of the excavator, and flexibly respond to the nonlinear characteristics of the fused data through an adaptive anomaly detection algorithm, accurately capture anomalies in the excavator operation process and calculate the anomaly score. Therefore, when the anomaly score reaches or exceeds the preset safety threshold, the warning mechanism is triggered to send real-time and targeted warning information to relevant managers. With this solution, the complex nonlinear characteristics in the excavator trajectory data can be handled more effectively, significantly improving the timeliness and accuracy of underground pipe network risk prediction, and ensuring the safe and stable operation of the urban underground pipe network.

[0074] Figure 2 is a schematic diagram of a method for obtaining anomaly scores provided by an embodiment of the present disclosure; Figure 2 As shown, optionally, in S13, the terminal device obtains an anomaly score for determining whether the excavator is close to the underground pipe network according to the fused data and the adaptive anomaly detection algorithm, including:

[0075] S21, the terminal device performs wavelet transform on the fused data to obtain a plurality of wavelet coefficients.

[0076] S22, the terminal device extracts high-dimensional features based on multiple wavelet coefficients.

[0077] S23, the terminal device uses an adaptive anomaly detection algorithm to analyze the high-dimensional features to obtain an anomaly score for determining whether the excavator is close to the underground pipeline network.

[0078] In this solution, the terminal device can perform wavelet transform on the fused data to obtain multiple wavelet coefficients. Specifically, the terminal device can use discrete wavelet transform to decompose the fused data X(t) into components of different frequencies and time scales. This process can be expressed by the following formula:

[0079]

[0080] Among them, X(t) is the fused data, W i (t) is a number of wavelet coefficients, is a plurality of wavelet basis functions. In other ways, the terminal device may also adopt continuous wavelet transform or any other suitable wavelet transform method to select the optimal transform method according to the characteristics of the data. In this way, the fused data can be decomposed into multiple frequency and time scale components, which respectively reflect the characteristics of the data at different scales.

[0081] Furthermore, the terminal device extracts high-dimensional features based on multiple wavelet coefficients, including: the terminal device constructs a feature matrix based on multiple wavelet coefficients. The terminal device inputs the adjacency matrix and the feature matrix into the graph convolution network model, and uses the graph convolution operation to extract features to obtain high-dimensional features. In this scheme, the terminal device processes multiple wavelet coefficients to construct a feature matrix, and then inputs this feature matrix and the adjacency matrix into the graph convolution network model. By using the graph convolution operation to perform deep feature extraction on the input data, the terminal device can efficiently obtain high-dimensional features, thereby improving the accuracy and efficiency of data analysis and processing.

[0082] Furthermore, the terminal device can use an adaptive anomaly detection algorithm to analyze high-dimensional features and obtain an anomaly score for determining whether the excavator is close to the underground pipe network. With this solution, through the adaptive mechanism, it is possible to accurately identify potential dangerous actions in the operation of the excavator so that an early warning signal can be issued in a timely manner.

[0083] Figure 3 is a schematic diagram of a method for extracting high-dimensional features provided by an embodiment of the present disclosure; Figure 3 As shown, optionally, in S22, the terminal device extracts high-dimensional features based on multiple wavelet coefficients, including:

[0084] S31, the terminal device constructs a feature matrix based on multiple wavelet coefficients.

[0085] S32, the terminal device inputs the adjacency matrix and the feature matrix into the graph convolution network model, and uses the graph convolution operation to extract features to obtain high-dimensional features.

[0086] In this solution, the terminal device can give priority to defining nodes. Specifically, the wavelet coefficients can be regarded as nodes in the graph, and each pipe network node is naturally regarded as a node in the graph. At the same time, the trajectory points of the excavator are also regarded as other nodes in the graph. In this way, other nodes and wavelet coefficient nodes together constitute the nodes of the graph. It can be understood that each node represents a wavelet coefficient at a specific position and scale, or a specific pipe network node / excavator trajectory point.

[0087] Furthermore, optionally, a plurality of wavelet coefficients can be used to construct a feature matrix. The feature matrix is ​​usually composed of wavelet coefficient values. The feature vector of each node is its corresponding wavelet coefficient, which reflects the characteristics of the signal at different scales and positions. With this solution, the feature matrix can be accurately constructed.

[0088] In this solution, the terminal device can model the nodes and their connection relationships as a graph structure. Here, the graph structure is defined as G = (V, E), where V is the node set and E is the edge set. Node v∈V represents a trajectory point or a pipe network location, and edge e∈E represents the relationship between adjacent points.

[0089] Furthermore, the terminal device can construct an adjacency matrix A according to the graph structure G, where A[i][j] represents the connection relationship between node i and node j. For example, if an adjacency matrix A is constructed, node i is connected to node j, then A[i][j] is a non-zero value, otherwise it is 0. With this solution, the adjacency matrix A can be accurately constructed. In this way, the terminal device inputs the adjacency matrix and the feature matrix into the graph convolution network model, and uses the graph convolution operation to extract features to obtain high-dimensional features.

[0090] Optionally, the terminal device may normalize the adjacency matrix to eliminate the influence of nodes with different degrees on feature aggregation.

[0091] In an optimized solution, the adjacency matrix A and the identity matrix I may be added first and then normalized.

[0092] Optionally, the normalized adjacency matrix A' can be expressed as

[0093] Accordingly, the terminal device can use the normalized adjacency matrix A′ and feature matrix X as the input of the graph convolutional network model, so that in each layer of the graph convolutional network model, the node features can be updated using the graph convolution operation. Specifically, the graph convolution formula includes:

[0094] H l+1 =σ(D -1 / 2 AD -1 / 2 H (l) W (l) )

[0095] Among them, H l is the node feature matrix of the lth layer, A is the adjacency matrix, D is the degree matrix, W (l) is the weight matrix of the lth layer, and σ is the activation function. In this way, through multi-layer graph convolution operations, the features of the nodes will gradually become high-dimensional. After the last layer of graph convolution operations, the high-dimensional feature matrix H L , here, contains the high-dimensional feature representation of each node. In this way, the graph convolutional network model can output high-dimensional features H L In this way, the terminal device can obtain accurate high-dimensional features.

[0096] Figure 4 is another schematic diagram of a method for obtaining anomaly scores provided by an embodiment of the present disclosure; Figure 4 As shown, optionally, in S23, the terminal device uses an adaptive anomaly detection algorithm to analyze the high-dimensional features to obtain an anomaly score for determining whether the excavator is close to the underground pipe network, including:

[0097] S41, the terminal device inputs the high-dimensional features into the variational autoencoder to obtain the potential variable distribution.

[0098] S42, the terminal device inputs the potential variable distribution into the long short-term memory network model to obtain the abnormal score output by the long short-term memory network model for judging whether the excavator is close to the underground pipeline network.

[0099] In this solution, the terminal device can input high-dimensional features into the variational autoencoder to obtain the potential variable distribution. Specifically, the terminal device inputs the high-dimensional features into the encoder model to use the encoder model to map the high-dimensional features to the latent space. The terminal device obtains the potential variable distribution based on the mean vector and standard deviation vector of the latent space output by the encoder model. With this solution, complex high-dimensional feature data can be effectively input into the encoder model, and the high-dimensional features can be mapped to a more expressive latent space through the powerful learning ability of the model, which not only simplifies the complexity of the data, but also retains the key information in the data. In addition, the terminal device uses the mean vector and standard deviation vector of the latent space output by the encoder model to accurately characterize the distribution of latent variables, providing a reliable foundation for subsequent data analysis, pattern recognition or data generation tasks, greatly improving processing efficiency and accuracy.

[0100] Furthermore, the terminal device can input the potential variable distribution into the long short-term memory network model to obtain the abnormal score output by the long short-term memory network model for judging whether the excavator is close to the underground pipe network. Specifically, before inputting the potential variable distribution into the long short-term memory network model, it also includes converting it into a time series format to meet the input requirements of the long short-term memory network model. Here, the feature data can be organized into the form of sliding windows, each window containing the operation data of the excavator in a specific time period, such as position, speed, posture, etc. In order to use the converted time series data as the input of the long short-term memory network model, it is helpful to capture the timing characteristics of the excavator operation and provide rich information for subsequent anomaly detection.

[0101] Optionally, the terminal device also needs to pre-build a long short-term memory network model. This model is used to process time series data and can capture long-term dependencies in the data. In this model, the input layer is responsible for receiving the potential variable distribution from the output of the variational autoencoder as input data. The long short-term memory layer is the core part of the model, which contains multiple long short-term memory units for processing the input time series data. These long short-term memory units learn the temporal relationship in the data through the memory and forgetting mechanism, and extract features that are useful for subsequent prediction and anomaly detection. In this way, the fully connected layer maps the output of the long short-term memory layer to the final prediction result, that is, the anomaly score for judging whether the excavator is close to the underground pipeline network. This score is obtained by learning the normal operation mode and trajectory of the excavator. When the operation or trajectory of the excavator deviates significantly from the normal mode, the anomaly score will increase accordingly. In actual operation, the terminal device will obtain the potential variable distribution based on the high-dimensional multimodal data collected from the excavator in real time, and input it into the trained long short-term memory network model for prediction. The long short-term memory network model can output anomaly scores for detecting whether the excavator is close to the underground pipeline. With this solution, by combining the variational autoencoder with the long short-term memory network model and using the adaptive anomaly detection algorithm to perform in-depth analysis of high-dimensional features, the anomaly score of the excavator approaching the underground pipeline network can be accurately calculated, achieving efficient anomaly detection and early warning.

[0102] Optionally, the variational autoencoder includes an encoder model. S41, the terminal device inputs the high-dimensional features into the variational autoencoder to obtain a potential variable distribution, including:

[0103] The terminal device inputs the high-dimensional features into the encoder model to map the high-dimensional features to the latent space using the encoder model.

[0104] The terminal device obtains the potential variable distribution according to the mean vector and standard deviation vector of the latent space output by the encoder model.

[0105] In this solution, the terminal device can pre-build an encoder model, and the model input is the high-dimensional features output by the graph convolutional network model. In this way, after the terminal device inputs the high-dimensional features into the encoder model, the encoder model calculates the mean vector and standard deviation vector of the latent space based on the input high-dimensional features. And using the mean vector and standard deviation vector, the potential variable distribution can be constructed. Among them, the potential variable distribution is usually a normal distribution. With this solution, by inputting high-dimensional features into the encoder model, these complex features can be mapped to a more expressive latent space, and the distribution of the latent variables can be accurately obtained based on the mean vector and standard deviation vector output by the encoder, which provides strong data support and accurate variable description for subsequent data analysis and anomaly detection tasks.

[0106] Optionally, the variational autoencoder further includes a decoder model, and the method further includes:

[0107] The terminal device determines the latent variables according to the mean vector, standard deviation vector and noise vector of the standard normal distribution in the latent space.

[0108] The terminal device inputs the latent variables into the decoder model to use the decoder model to decode the latent variables back to the high-dimensional feature space to obtain the reconstructed data distribution output by the decoder model.

[0109] The terminal device calculates the loss value based on the potential variable distribution, the reconstructed data distribution, and the loss function.

[0110] The terminal device uses the back-propagation algorithm and the loss value to update and optimize the parameters of the decoder to improve the performance of the variational autoencoder.

[0111] In this scheme, the latent variables can be obtained according to the reparameterization technique based on the mean vector and standard deviation vector of the latent space and the noise vector sampled from the standard normal distribution. In this way, the latent variables are determined by applying the reparameterization technique, including:

[0112] z=σ×∈+μ

[0113] Among them, the mean vector of the latent space is μ, the standard deviation vector is σ, and ∈ is the noise vector sampled from the standard normal distribution. With this scheme, the latent variables can be accurately determined.

[0114] Furthermore, the terminal device can input the latent variables into the decoder model, and the decoder model uses its own neural network structure to decode the latent variables back to the high-dimensional feature space to obtain the reconstructed data distribution. In this way, the reconstruction of the data distribution can be achieved.

[0115] Furthermore, the terminal device can calculate the loss value by combining the potential variable distribution, the reconstructed data distribution and the loss function. The formula of the loss function is as follows:

[0116] L VAE =E q(z|X) [logp(X|z)]-KL(q(z|X)||p(z))

[0117] Among them, q(z|X) is the potential variable distribution output by the encoder, p(X|z) is the reconstructed data distribution output by the decoder, and KL is the Kulbak-Leibler divergence. The Kulbak-Leibler divergence is used to measure the difference between the potential variable distribution and the standard normal distribution to encourage the potential variable distribution to be close to the standard normal distribution.

[0118] Furthermore, the terminal device can update and optimize the parameters of the decoder model based on the back-propagation algorithm and the calculated loss value to minimize the total loss value. With this solution, the overall performance of the variational autoencoder can be significantly improved, enabling it to more accurately reconstruct input data and effectively identify abnormal data.

[0119] Figure 5 is a schematic diagram of a risk warning device for an underground pipe network provided by an embodiment of the present disclosure; Figure 5 As shown, the embodiment of the present disclosure provides a risk warning device 200 for underground pipe network, including an acquisition module 51, a first acquisition module 52, a second acquisition module 53 and a push module 54. The acquisition module 51 is configured to acquire the historical trajectory data, working data and underground pipe network node data of the excavator; the first acquisition module 52 is configured to fuse the historical trajectory data, working data and underground pipe network node data of the excavator to obtain fused data; the second acquisition module 53 is configured to obtain an abnormality score for judging whether the excavator is close to the underground pipe network according to the fused data and the adaptive abnormality detection algorithm; the push module 54 is configured to push the warning information when the abnormality score exceeds the set threshold.

[0120] The risk warning device 200 for underground pipe networks provided by the embodiment of the present disclosure can fuse the historical trajectory data, working parameters and underground pipe network node data of the excavator, and flexibly respond to the nonlinear characteristics of the fused data through an adaptive anomaly detection algorithm, accurately capture anomalies in the operation of the excavator and calculate the anomaly score. Therefore, when the anomaly score reaches or exceeds the preset safety threshold, the warning mechanism is triggered to send real-time and targeted warning information to relevant managers. With this solution, the complex nonlinear characteristics in the excavator trajectory data can be handled more effectively, significantly improving the timeliness and accuracy of the underground pipe network risk prediction, and ensuring the safe and stable operation of the urban underground pipe network.

[0121] Figure 6 is a schematic diagram of a risk warning device for an underground pipe network provided by an embodiment of the present disclosure; Figure 6 As shown, the embodiment of the present disclosure provides a risk warning device 300 for an underground pipe network, including a processor 301 and a memory 302. Optionally, the device 300 may also include a communication interface 303 and a bus 304. The processor 301, the communication interface 303, and the memory 302 may communicate with each other through the bus 304. The communication interface 303 may be used for information transmission. The processor 301 may call the logic instructions in the memory 302 to execute the risk warning method for an underground pipe network of the above embodiment.

[0122] In addition, the logic instructions in the memory 302 described above may be implemented in the form of software functional units and when sold or used as independent products, may be stored in a computer-readable storage medium.

[0123] The memory 302 is a computer-readable storage medium that can be used to store software programs and computer executable programs, such as program instructions / modules corresponding to the method in the embodiment of the present disclosure. The processor 301 executes the function application and data processing by running the program instructions / modules stored in the memory 302, that is, the risk warning method for underground pipe networks in the above embodiment is implemented.

[0124] The memory 302 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application required for at least one function; the data storage area may store data created according to the use of the terminal device, etc. In addition, the memory 302 may include a high-speed random access memory and may also include a non-volatile memory.

[0125] The embodiments of the present disclosure provide a terminal device, including: a terminal device body, and the above-mentioned risk warning device 200 (300) for underground pipe networks. The risk warning device 200 (300) for underground pipe networks is installed on the terminal device body. The installation relationship described here is not limited to placement inside the terminal device body, but also includes installation connections with other components of the terminal device, including but not limited to physical connections, electrical connections or signal transmission connections. It can be understood by those skilled in the art that the risk warning device 200 (300) for underground pipe networks can be adapted to a feasible terminal device body, thereby realizing other feasible embodiments.

[0126] An embodiment of the present disclosure provides a computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions are configured to execute the above-mentioned risk warning method for an underground pipe network.

[0127] The technical solution of the embodiment of the present disclosure can be embodied in the form of a software product, which is stored in a storage medium and includes one or more instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the embodiment of the present disclosure. The aforementioned storage medium may be a non-transient storage medium, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, and other media that can store program codes.

[0128] The above description and the accompanying drawings fully illustrate the embodiments of the present disclosure so that those skilled in the art can practice them. Other embodiments may include structural, logical, electrical, process and other changes. The embodiments represent only possible changes. Unless explicitly required, separate components and functions are optional, and the order of operation may vary. The parts and features of some embodiments may be included in or replace the parts and features of other embodiments. Moreover, the words used in this application are only used to describe the embodiments and are not used to limit the claims. As used in the description of the embodiments and the claims, unless the context clearly indicates, the singular forms of "a", "an" and "the" are intended to include plural forms as well. Similarly, the term "and / or" as used in this application refers to any and all possible combinations of listings containing one or more associated ones. In addition, when used in the present application, the term "comprise" and its variants "comprises" and / or comprising refer to the presence of stated features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or groups thereof. In the absence of further restrictions, the elements defined by the sentence "comprising a ..." do not exclude the presence of other identical elements in the process, method or device comprising the elements. In this article, each embodiment may focus on the differences from other embodiments, and the same and similar parts between the various embodiments may refer to each other. For the methods, products, etc. disclosed in the embodiments, if they correspond to the method part disclosed in the embodiments, then the relevant parts can refer to the description of the method part.

[0129] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software may depend on the specific application and design constraints of the technical solution. The technicians may use different methods for each specific application to implement the described functions, but such implementations should not be considered to exceed the scope of the embodiments of the present disclosure. The technicians may clearly understand that, for the convenience and simplicity of description, the specific working processes of the systems, devices and units described above may refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here.

[0130] In the embodiments disclosed herein, the disclosed methods and products (including but not limited to devices, equipment, etc.) can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units can be only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between each other shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms. The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to implement this embodiment. In addition, each functional unit in the embodiment of the present disclosure may be integrated in a processing unit, or each unit may exist physically alone, or two or more units may be integrated in one unit.

[0131] The flowchart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to the embodiment of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of the code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. In some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, which can depend on the functions involved. In the description corresponding to the flowchart and the block diagram in the accompanying drawings, the operations or steps corresponding to different boxes can also occur in a different order from the order disclosed in the description, and sometimes there is no specific order between different operations or steps. For example, two consecutive operations or steps can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, which can depend on the functions involved. Each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented by a dedicated hardware-based system that performs the specified functions or actions, or may be implemented by a combination of dedicated hardware and computer instructions.

Claims

1. A risk early warning method for underground pipe networks, characterized in that: include: Obtain the excavator's historical trajectory data, work data, and underground pipe network node data; The historical trajectory data, working data and underground pipe network node data of the excavator are integrated to obtain the integrated data; Based on the fused data and the adaptive anomaly detection algorithm, an anomaly score is obtained to determine whether the excavator is close to the underground pipe network; When the anomaly score exceeds the set threshold, a warning message is pushed.

2. The method according to claim 1, characterized in that Based on the fused data and the adaptive anomaly detection algorithm, an anomaly score is obtained to determine whether the excavator is close to the underground pipe network, including: Performing wavelet transform on the fused data to obtain multiple wavelet coefficients; Extract high-dimensional features based on multiple wavelet coefficients; The high-dimensional features are analyzed using an adaptive anomaly detection algorithm to obtain an anomaly score that is used to determine whether the excavator is close to the underground pipeline network.

3. The method according to claim 2, characterized in that Based on multiple wavelet coefficients, high-dimensional features are extracted, including: Based on multiple wavelet coefficients, construct a feature matrix; The adjacency matrix and feature matrix are input into the graph convolutional network model, and the graph convolution operation is used to extract features to obtain high-dimensional features.

4. The method according to claim 2, characterized in that: The adaptive anomaly detection algorithm is used to analyze high-dimensional features and obtain anomaly scores for judging whether the excavator is close to the underground pipe network, including: Input high-dimensional features into the variational autoencoder to obtain the potential variable distribution; The potential variable distribution is input into the long short-term memory network model to obtain the abnormal score output by the long short-term memory network model for judging whether the excavator is close to the underground pipeline network.

5. The method according to claim 4, characterized in that The variational autoencoder includes an encoder model that inputs high-dimensional features into the variational autoencoder to obtain the potential variable distribution, including: Inputting the high-dimensional features into the encoder model to map the high-dimensional features into the latent space using the encoder model; According to the mean vector and standard deviation vector of the latent space output by the encoder model, the latent variable distribution is obtained.

6. The method according to claim 5, characterized in that The variational autoencoder also includes a decoder model, and the method further includes: Determine the latent variables according to the mean vector, standard deviation vector and standard normal distribution noise vector of the latent space; Inputting the latent variables into the decoder model to decode the latent variables back to the high-dimensional feature space using the decoder model to obtain the reconstructed data distribution output by the decoder model; Calculate the loss value based on the potential variable distribution, the reconstructed data distribution, and the loss function; The back-propagation algorithm and loss value are used to update and optimize the parameters of the decoder to improve the performance of the variational autoencoder.

7. A risk warning device for underground pipe network, characterized in that: include: An acquisition module, configured to acquire historical trajectory data, working data, and underground pipe network node data of the excavator; A first obtaining module is configured to fuse historical trajectory data, working data and underground pipe network node data of the excavator to obtain fused data; A second obtaining module is configured to obtain an anomaly score for determining whether the excavator is close to the underground pipe network based on the fused data and the adaptive anomaly detection algorithm; The push module is configured to push warning information when the anomaly score exceeds a set threshold.

8. A risk warning device for underground pipe network, comprising a processor and a memory storing program instructions, characterized in that: The processor is configured to execute the risk warning method for underground pipe networks as described in any one of claims 1 to 6 when running the program instructions.

9. A terminal device, characterized in that: include: Terminal device body; The device for risk warning of underground pipe networks as described in claim 7 or 8 is installed on the terminal device body.

10. A storage medium storing program instructions, characterized in that: When the program instructions are run, the risk warning method for underground pipe networks as described in any one of claims 1 to 6 is executed.

Images