Intelligent slope early warning method, system and equipment based on multi-source data fusion

Abstract

The invention relates to the technical field of geological disaster monitoring and early warning, and discloses a slope intelligent early warning method, system and device based on multi-source data fusion, and the method comprises the steps: collecting original multi-source data; storing the original multi-source data as time sequence monitoring data by using time sequence storage; data fusion is carried out on the time sequence monitoring data, the time sequence monitoring data after data fusion is used as a data source, prediction is carried out by using a machine learning model, and a prediction result comprises a slope deformation trend and a monitoring data abnormal state; according to the method, the limitation of single sensor data is broken through by collecting the multi-source data, the multi-source data is stored as the time sequence monitoring data, the multi-source data is stored in an associated mode according to the time dimension, the multi-source data fusion is utilized, the risk level is output, and risk early warning is carried out. The data internal law is deeply mined, prediction is carried out based on a machine learning model, the risk is early warned in advance, and data support is provided for slope safety management.

Description

Technical Field

[0001] This invention relates to the field of geological disaster monitoring and early warning technology, specifically to an intelligent early warning method, system and equipment for slopes based on multi-source data fusion. Background Technology

[0002] In fields such as geotechnical engineering, mine safety, transportation engineering, and water conservancy and hydropower engineering, slope deformation is influenced by multiple factors including geological conditions, hydrological environment, and meteorology. Landslides and collapses are typical geological hazards that directly threaten the safety of construction workers, surrounding residents, and passing vehicles, and damage infrastructure. Therefore, slope stability monitoring is crucial throughout the entire process of engineering design, construction, and operation, and is of great significance for risk prevention and control. Thus, how to achieve intelligent early warning for slopes has become an urgent problem to be solved. Summary of the Invention

[0003] This invention provides a method, system, and device for intelligent early warning of slopes based on multi-source data fusion, in order to solve the problem of how to achieve intelligent early warning of slopes.

[0004] In a first aspect, the present invention provides a slope intelligent early warning method based on multi-source data fusion, the method comprising: Raw multi-source data was collected, including data acquired using GNSS displacement gauges, inclinometers, crack gauges, rain gauges, water level gauges, stress and strain sensors, and video stream sensors. The collected raw multi-source data is stored as time-series monitoring data using time-series storage. The time-series monitoring data is fused, and the fused time-series monitoring data is used as the data source. Machine learning models are used for prediction, and the prediction results include slope deformation trends and abnormal states of monitoring data. By combining the time-series monitoring data and prediction results after data fusion, the risk level is output and a risk warning is issued.

[0005] This invention overcomes the limitations of single-sensor data by collecting multi-source data and storing it as time-series monitoring data. It links and stores multi-source data in a time-series manner, utilizes multi-source data fusion to deeply mine the inherent patterns in the data, and makes predictions based on machine learning models to achieve early warning of risks and provide data support for slope safety management.

[0006] In one optional implementation, the collected raw multi-source data is stored as time-series monitoring data using time-series storage, including: The raw multi-source data is written into the InfluxDB time series database to obtain time series monitoring data; or, The original multi-source data is written into a relational database, and the data stored in the relational database is extended with time series to obtain time series monitoring data.

[0007] This invention utilizes time-series storage to write raw multi-source data into an InfluxDB time-series database or a relational database, enabling high-concurrency data writing, automatic processing of massive multi-source data, optimization of storage structure, reduction of manual workload, and shortening of data analysis cycle.

[0008] In one optional implementation, the time-series monitoring data is fused, including: The weights of the time-series monitoring data are adjusted based on the sensor accuracy and the completeness of the sensor-acquired data, and the weights are then fused based on the adjusted weights. Deformation features, environmental features, and structural features are extracted from time-series monitoring data. The extracted features are then combined into a high-dimensional feature vector and fused at the feature level. Deformation features include displacement, displacement rate, and displacement acceleration. Environmental features include water level fluctuation amplitude. Structural features include crack width change rate.

[0009] This invention performs weighted fusion and feature-level fusion on time-series monitoring data, which ensures the accuracy of core indicators while giving the data multi-dimensional and structured feature attributes, so as to more accurately identify anomalies and potential risks.

[0010] In one optional implementation, using the fused time-series monitoring data as the data source, a machine learning model is used for prediction, including: The time-series monitoring data after data fusion is aligned according to the time dimension and input into the LSTM model to obtain the slope deformation trend; The time-series monitoring data after data fusion is converted into standardized feature vectors, which are then input into an isolated forest model. The abnormal status of the monitoring data is determined based on the anomaly score.

[0011] This invention utilizes an LSTM model to deeply mine the correlation patterns in time-series data, accurately predicting slope deformation in future time periods and providing decision-making data support for managers. It also uses an isolated forest model to mark abnormal data, facilitating precise control over abnormal states.

[0012] In one alternative implementation, the risk level is output, including: The time-series monitoring data and prediction results after data fusion are input into the random forest model, and the probability of each risk level is output. The risk level with the highest probability is output as the risk level.

[0013] This invention uses a random forest model to determine risk levels, improving the reliability of risk level assessment, providing data support for early intervention in risks, and enhancing the initiative in risk management.

[0014] In one alternative implementation, after acquiring the raw multi-source data, the method further includes: Outlier removal, missing value imputation, and spatiotemporal alignment are performed on the original multi-source data.

[0015] This invention improves the authenticity, integrity and consistency of raw multi-source data by preprocessing it, thereby reducing the risk of false alarms and missed alarms in early warning systems from the source.

[0016] In one alternative implementation, after issuing a risk warning, the method further includes: Generate periodic monitoring reports and event-driven reports.

[0017] This invention provides managers with an intuitive and comprehensive basis for decision-making by generating reports.

[0018] Secondly, this invention provides a slope intelligent early warning system based on multi-source data fusion, the system comprising: The acquisition module is used to acquire raw multi-source data, which includes data acquired using GNSS displacement gauges, inclinometers, crack gauges, rain gauges, water level gauges, stress and strain sensors, and video stream sensors. The storage module is used to store the collected raw multi-source data as time-series monitoring data using time-series storage. The prediction module is used to fuse time-series monitoring data. Using the fused time-series monitoring data as the data source, it uses machine learning models to make predictions. The prediction results include slope deformation trends and abnormal states of monitoring data. The early warning module is used to combine the time-series monitoring data and prediction results after data fusion to output the risk level and issue risk warnings.

[0019] Thirdly, the present invention provides an electronic device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to perform the slope intelligent early warning method based on multi-source data fusion described in the first aspect or any corresponding embodiment.

[0020] Fourthly, the present invention provides a computer-readable storage medium storing computer instructions for causing a computer to execute the slope intelligent early warning method based on multi-source data fusion described in the first aspect or any corresponding embodiment. Attached Figure Description

[0021] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0022] Figure 1 This is a flowchart illustrating the intelligent early warning method for slopes based on multi-source data fusion according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the core analysis process according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the intelligent early warning interface for slopes according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the risk level distribution of slope monitoring points according to an embodiment of the present invention; Figure 5 This is a structural block diagram of a slope intelligent early warning system based on multi-source data fusion according to an embodiment of the present invention; Figure 6 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0025] According to an embodiment of the present invention, a method for intelligent early warning of slopes based on multi-source data fusion is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0026] This embodiment provides a slope intelligent early warning method based on multi-source data fusion. Figure 1This is a flowchart of a slope intelligent early warning method based on multi-source data fusion according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps: Step S101: Collect raw multi-source data.

[0027] In this embodiment of the invention, multiple communication protocols are used to access various types of sensors, including TCP / IP, Modbus, MQTT, HTTP API, etc. The sensors include GNSS displacement gauges, inclinometers, crack gauges, rain gauges, water level gauges, stress-strain sensors, and video stream sensors. The multi-source data collected includes data collected using GNSS displacement gauges, inclinometers, crack gauges, rain gauges, water level gauges, stress-strain sensors, and video stream sensors.

[0028] Step S102: Use time-series storage to store the collected raw multi-source data as time-series monitoring data.

[0029] In this embodiment of the invention, time-series storage is used to efficiently store a large amount of raw multi-source data, and the raw multi-source data is stored as time-series monitoring data. This adapts to the characteristics of time-series data, which is centered on timestamps, has high-frequency writing, massive storage, and time-series querying, thereby achieving orderly storage of multi-source data.

[0030] Step S103: The time series monitoring data is fused, and the fused time series monitoring data is used as the data source to make predictions using a machine learning model.

[0031] In this embodiment of the invention, considering that slope stability is affected by multiple factors, data collected from various sensors are fused to reveal connections that cannot be shown by single data points, achieving unified access to multi-source heterogeneous monitoring data. Machine learning models such as LSTM (Long Short-Term Memory), GRU (Gate Recurrent Unit), and time series Transformer are used to predict slope deformation trends and monitor abnormal states in future time periods, providing data support for slope safety early warning.

[0032] Step S104: Combine the time-series monitoring data and prediction results after data fusion to output the risk level and issue a risk warning.

[0033] In this embodiment of the invention, the time-series monitoring data and prediction results after data fusion are incorporated into a multi-indicator evaluation system to output a risk level, which includes normal, attention, alert, danger, etc., triggering a risk warning and realizing a risk warning. The risk warning methods include, but are not limited to, pop-up windows in the software interface, sound and light alarms, SMS, email, and social software. The risk warning information includes, but is not limited to, the location of the measurement point, the risk level, the triggering indicator, and the suggested measures. At the same time, a warning log is generated to record all warning events and their processing status.

[0034] Furthermore, risk warnings are implemented using dynamic threshold calculation rules. These rules employ a logic that combines a base threshold with adjustments based on underlying factors. For example, the displacement rate threshold is the base threshold plus the sum of accumulated rainfall over a past period. This dynamic threshold is not a fixed threshold but can be dynamically adjusted based on historical baselines, current environmental conditions, and forecast results.

[0035] The slope intelligent early warning method based on multi-source data fusion provided in this embodiment breaks through the limitations of single sensor data by collecting multi-source data and storing multi-source data as time-series monitoring data. The multi-source data is stored in association according to the time dimension. By using multi-source data fusion, the inherent patterns of the data are deeply mined, and prediction is made based on machine learning models to achieve early warning of risks and provide data support for slope safety management.

[0036] This embodiment provides a slope intelligent early warning method based on multi-source data fusion, the process of which includes the following steps: Step S201: Collect raw multi-source data.

[0037] Please see details Figure 1 Step S101 of the illustrated embodiment will not be described again here.

[0038] In some alternative implementations, the method further includes: Step S202 involves performing outlier removal, missing value imputation, and spatiotemporal alignment on the original multi-source data.

[0039] In embodiments of the present invention, such as Figure 2 As shown, the original multi-source data is preprocessed, including outlier removal, missing value imputation, and spatiotemporal alignment.

[0040] Set a threshold range, identify data that exceeds the threshold range as outliers, and remove outlier data.

[0041] For short-term missing data, if the data at the time point preceding the missing point is valid, forward imputation is used. That is, the data at time t is filled with the data at time t-1, and then the data at time t+1 is obtained and linear interpolation is used to correct the data to ensure a smooth transition. For long-term missing data, multiple linear regression imputation is used to construct a regression model, calculate the missing values, and fill in the missing values.

[0042] To address the issues of inconsistent temporal frequencies and dispersed spatial locations in raw multi-source data, time alignment is achieved by setting a time benchmark and unifying the time series frequencies. Spatial alignment is achieved by converting the coordinates of monitoring points to a unified coordinate system.

[0043] By preprocessing the raw multi-source data, the authenticity, integrity, and consistency of the data can be improved, thereby reducing the risk of false alarms and missed alarms in early warnings from the source.

[0044] Step S203: Use time-series storage to store the collected raw multi-source data as time-series monitoring data.

[0045] Specifically, step S203 includes: Step S203a: Write the raw multi-source data into the InfluxDB time series database to obtain time series monitoring data; or, Step S203b: Write the original multi-source data into a relational database, and perform time-series extension on the data stored in the relational database to obtain time-series monitoring data.

[0046] In this embodiment of the invention, the original multi-source data is stored as time-series monitoring data by directly writing to an InfluxDB time-series database or a relational database combined with time-series extension.

[0047] Directly writing to the InfluxDB time series database requires adaptation to the InfluxDB data model. The original multi-source data is organized and stored according to the structure of "measurement indicator label-field-time stamp". Then, the original multi-source data is written to the InfluxDB time series database in batches through the API or SDK interface provided by the InfluxDB time series database to obtain time series monitoring data.

[0048] The approach of combining relational databases with time-series extensions involves first creating a base table in a relational database (such as MySQL) to store static data such as sensor metadata (location, type, parameters, etc.) and raw data tables containing timestamps and monitoring values. Then, a time-series extension scheme is used to optimize the storage. This scheme includes adding partitioned tables for time-based sharding, creating time indexes, and integrating time-series data engine plugins. The raw data tables are split according to the time dimension, thereby improving the efficiency of time-series data queries and forming time-series monitoring data.

[0049] It should be noted that both InfluxDB time-series databases and relational databases provide interfaces for creating, reading, updating, deleting, and querying data, as well as access control interfaces. They also support data backup and have recovery mechanisms.

[0050] By using time-series storage, raw multi-source data can be written to InfluxDB time-series databases or relational databases, enabling high-concurrency data writing, automatic processing of massive amounts of multi-source data, optimization of storage structure, reduction of manual workload, and shortening of data analysis cycle.

[0051] Step S204: The time-series monitoring data is fused, and the fused time-series monitoring data is used as the data source to make predictions using a machine learning model.

[0052] Specifically, the data fusion of time-series monitoring data in step S204 above includes: Step S2041: Adjust the weights of the time-series monitoring data according to the sensor accuracy and the integrity of the sensor-acquired data, and perform weight fusion based on the adjusted weights.

[0053] Step S2042: Extract deformation features, environmental features, and structural features from the time-series monitoring data, and combine the extracted features into a high-dimensional feature vector for feature-level fusion.

[0054] In this embodiment of the invention, weight fusion is performed, assigning different weights based on sensor accuracy (higher accuracy, higher weight) and historical sensor performance (higher completeness of sensor data, higher weight). The adjusted weights are then combined for weight fusion.

[0055] Feature-level fusion is performed to extract features from time-series monitoring data, including deformation features such as displacement, displacement rate, and displacement acceleration, environmental features such as water level fluctuation amplitude, and structural features such as crack width change rate. The extracted features are then combined into a high-dimensional feature vector.

[0056] By performing weighted fusion and feature-level fusion on time-series monitoring data, the accuracy of core indicators is ensured while the data acquires multi-dimensional and structured features, enabling more accurate identification of anomalies and potential risks.

[0057] Specifically, step S204 above, which uses the fused time-series monitoring data as the data source and employs a machine learning model for prediction, includes: Step S2043: Align the time-series monitoring data after data fusion with the time dimension, and input it into the LSTM model to obtain the slope deformation trend.

[0058] Step S2044: Convert the time-series monitoring data after data fusion into a standardized feature vector, input the standardized feature vector into the isolated forest model, and determine the abnormal status of the monitoring data based on the anomaly score.

[0059] In this embodiment of the invention, the time-series monitoring data after data fusion is aligned in time dimension using linear interpolation. For example, GNSS data within a five-minute interval and rainfall data within an hour are interpolated to a ten-minute time-series data interval, ensuring that all indicators correspond one-to-one at the same timestamp. The time-aligned time-series monitoring data is then input into a pre-loaded LSTM model. The LSTM model learns the temporal patterns and outputs the slope deformation trend for future time periods.

[0060] Statistical and trend features are extracted from the fused time-series monitoring data to form feature vectors. These feature vectors are then standardized to eliminate dimensional differences between various indicators. A pre-trained Isolation Forest model is then loaded. The core parameters of the Isolation Forest model include the number of decision trees, the number of samples per tree, and the anomaly detection threshold. The standardized feature vectors are input into the Isolation Forest model, which performs isolated evaluation on each sample and calculates its anomaly score. For example, with an anomaly score threshold of 0.5, a score less than 0.5 indicates that the data is an anomaly.

[0061] It is understandable that anomaly detection can also automatically identify abnormal data points or sequences that deviate from the normal pattern through algorithms such as One-Class SVM and Autoencoder. This is only an example and is not intended to be a limitation.

[0062] In addition, such as Figure 3 As shown, time-series data is extracted within a fixed time window, and a smooth trend curve is generated by sliding the window sequentially. Combined with linear or nonlinear regression fitting, long-term deformation trends are identified. Displacement, settlement rate, and rate of change at each measuring point are calculated in real time, achieving real-time dynamic quantification. The spatial correlation of deformation at different measuring points is analyzed, such as generating displacement cloud maps and contour maps. The temporal correlation between slope deformation and environmental inducing factors such as precipitation, water level, and construction activities is analyzed, including correlation analysis and lag effect analysis, to achieve spatiotemporal correlation analysis.

[0063] By leveraging LSTM models to deeply mine the correlation patterns in time-series data, we can accurately predict slope deformation in future time periods, providing decision-making data support for managers. We also use isolated forest models to mark abnormal data, enabling precise control of abnormal conditions.

[0064] Step S205: Combining the time-series monitoring data and prediction results after data fusion, output the risk level and issue a risk warning.

[0065] Specifically, the risk level output in step S205 above includes: Step S2051: Input the time-series monitoring data and prediction results after data fusion into the random forest model and output the probability of each risk level.

[0066] Step S2052: The risk level with the highest probability is output as the risk level.

[0067] In this embodiment of the invention, a random forest model is used as the classification model. A pre-trained random forest model is loaded, and historical time-series monitoring data is proportionally divided into training, validation, and test sets. Using the fused time-series monitoring results and prediction results as input, the random forest model outputs the probability of each risk level, which includes four levels: normal, attention, alert, and dangerous. Figure 4 As shown, different colors are used to mark each risk level, such as green for normal, yellow for caution, orange for alert, and red for red, and response procedures are set for different risk levels.

[0068] Among them, no abnormal indicators are considered normal, a single indicator approaching a temporary value is considered a warning, multiple indicators exceeding the threshold is considered a warning, and multiple indicators exceeding the threshold and lasting for a duration exceeding the preset duration is considered dangerous. The risk level with the highest probability is taken as the final output risk level.

[0069] It is understandable that classification models can also use XGBoost, Support Vector Machine (SVM), etc., but these are just examples and not intended to be limiting.

[0070] By using random forest models to determine risk levels, the reliability of risk level assessment can be improved, providing data support for early intervention in risks and enhancing the initiative in risk management.

[0071] In some alternative implementations, the method further includes: Step S206: Generate periodic monitoring reports and event-driven reports.

[0072] In this embodiment of the invention, in order to provide a richer and more intuitive way of displaying data, a customized report template is provided, key data is filled in, and periodic reports in the form of daily, weekly, and monthly reports are automatically generated, as well as event-driven reports such as early warning reports and analysis reports. The report supports exporting to PDF, Word and other formats.

[0073] Using maximum displacement, highest risk level, and real-time rainfall as key indicators, the system displays data for each key indicator on a single screen. It shows the change curves of displacement, water level, and rainfall over time, and supports overlay comparison of multiple measuring points and multiple indicators. A two-dimensional map displays the location of measuring points, real-time data, historical displacement vectors, risk level heatmaps, displacement contour lines, and cloud maps. Users can click on high-risk measuring points to access nearby surveillance video footage, and the warning information panel provides a real-time scrolling display of current warning information.

[0074] The slope intelligent early warning method based on multi-source data fusion provided in this embodiment generates reports to provide managers with an intuitive and comprehensive basis for decision-making.

[0075] This embodiment also provides a slope intelligent early warning system based on multi-source data fusion. This system is used to implement the above embodiments and preferred embodiments, and details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the system described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0076] This embodiment provides a slope intelligent early warning system based on multi-source data fusion, such as... Figure 5 As shown, it includes: The acquisition module 501 is used to acquire raw multi-source data, which includes data acquired using GNSS displacement gauges, inclinometers, crack gauges, rain gauges, water level gauges, stress and strain sensors, and video stream sensors. Storage module 502 is used to store the collected raw multi-source data as time-series monitoring data using time-series storage; The prediction module 503 is used to fuse time-series monitoring data, and uses the fused time-series monitoring data as the data source to make predictions using a machine learning model. The prediction results include slope deformation trends and abnormal states of monitoring data. The early warning module 504 is used to combine the time-series monitoring data and prediction results after data fusion to output the risk level and issue a risk warning.

[0077] In some alternative implementations, the storage module 502 includes: The first storage unit is used to write raw multi-source data into the InfluxDB time series database to obtain time series monitoring data. The second storage unit is used to write the original multi-source data into a relational database, and to perform time-series extension on the data stored in the relational database to obtain time-series monitoring data.

[0078] In some alternative implementations, the prediction module 503 includes: The weight fusion unit is used to adjust the weights of time-series monitoring data based on sensor accuracy and the completeness of sensor-acquired data, and to perform weight fusion based on the adjusted weights. The feature-level fusion unit is used to extract deformation features, environmental features, and structural features from time-series monitoring data, and to combine the extracted features into a high-dimensional feature vector for feature-level fusion. Deformation features include displacement, displacement rate, and displacement acceleration; environmental features include water level fluctuation amplitude; and structural features include crack width change rate.

[0079] In some optional implementations, the prediction module 503 further includes: The deformation trend prediction unit is used to align the time-series monitoring data after data fusion in the time dimension and input it into the LSTM model to obtain the slope deformation trend. The anomaly monitoring unit is used to convert the time-series monitoring data after data fusion into standardized feature vectors, input the standardized feature vectors into the isolated forest model, and determine the anomaly status of the monitoring data based on the anomaly score.

[0080] In some alternative implementations, the warning module 504 includes: The input unit is used to input the fused time-series monitoring data and prediction results into the random forest model and output the probability of each risk level. The output unit is used to output the risk level with the highest probability.

[0081] In some alternative embodiments, the device further includes: The preprocessing module is used to perform outlier removal, missing value imputation, and spatiotemporal alignment on the raw multi-source data.

[0082] In some alternative embodiments, the device further includes: The report generation module is used to generate periodic monitoring reports and event-driven reports.

[0083] The slope intelligent early warning system based on multi-source data fusion provided in this invention can execute the slope intelligent early warning method based on multi-source data fusion provided in any embodiment of this invention, and has the corresponding functional modules and beneficial effects of the method. Further functional descriptions of the above modules and units are the same as in the corresponding embodiments described above, and will not be repeated here.

[0084] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0085] The following is a detailed reference. Figure 6This diagram illustrates a suitable structural design for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 601, which can perform various appropriate actions and processes based on a program stored in read-only memory (ROM) 602 or a program loaded from memory 608 into random access memory (RAM) 603. RAM 603 also stores various programs and data required for the operation of the electronic device. The processor 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0086] Typically, the following devices can be connected to I / O interface 605: input devices 606 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 607 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 608 including, for example, magnetic tapes, hard disks, etc.; and communication devices 609. Communication device 609 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 6 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0087] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 609, or installed from a memory 608, or installed from a ROM 602. When the computer program is executed by the processor 601, it performs the functions defined in the slope intelligent early warning method based on multi-source data fusion according to embodiments of the present invention.

[0088] Figure 6 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0089] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the slope intelligent early warning method based on multi-source data fusion shown in the above embodiments is implemented.

[0090] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0091] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and all such modifications and variations fall within the scope defined by the appended invention.

Claims

1. A slope intelligent early warning method based on multi-source data fusion, characterized in that, The method includes: Collect raw multi-source data, which includes data collected using GNSS displacement gauges, inclinometers, crack gauges, rain gauges, water level gauges, stress and strain sensors, and video stream sensors; The collected raw multi-source data is stored as time-series monitoring data using time-series storage. The time-series monitoring data is fused, and the fused time-series monitoring data is used as the data source. A machine learning model is used to make predictions, and the prediction results include slope deformation trends and abnormal states of monitoring data. By combining the time-series monitoring data and prediction results after data fusion, the risk level is output and a risk warning is issued.

2. The method according to claim 1, characterized in that, The process of storing the collected raw multi-source data into time-series monitoring data using time-series storage includes: The raw multi-source data is written into the InfluxDB time series database to obtain time series monitoring data; or, The original multi-source data is written into a relational database, and the data stored in the relational database is extended with time series to obtain time series monitoring data.

3. The method according to claim 1, characterized in that, The process of fusing the time-series monitoring data includes: The weights of the time-series monitoring data are adjusted based on the sensor accuracy and the completeness of the sensor-acquired data, and the weights are then fused based on the adjusted weights. Deformation features, environmental features, and structural features are extracted from time-series monitoring data, and the extracted features are combined into a high-dimensional feature vector for feature-level fusion. The deformation features include displacement, displacement rate, and displacement acceleration; the environmental features include water level fluctuation amplitude; and the structural features include crack width change rate.

4. The method according to claim 1, characterized in that, The method of using time-series monitoring data after data fusion as the data source and employing machine learning models for prediction includes: The time-series monitoring data after data fusion is aligned according to the time dimension and input into the LSTM model to obtain the slope deformation trend; The time-series monitoring data after data fusion is converted into standardized feature vectors, which are then input into an isolated forest model. The abnormal status of the monitoring data is determined based on the anomaly score.

5. The method according to claim 1, characterized in that, The output risk level includes: The time-series monitoring data and prediction results after data fusion are input into the random forest model, and the probability of each risk level is output. The risk level with the highest probability is output as the risk level.

6. The method according to claim 5, characterized in that, After collecting the raw multi-source data, the method further includes: The original multi-source data is subjected to outlier removal, missing value imputation, and spatiotemporal alignment.

7. The method according to claim 1, characterized in that, After issuing a risk warning, the method further includes: Generate periodic monitoring reports and event-driven reports.

8. A slope intelligent early warning system based on multi-source data fusion, characterized in that, The system includes: The acquisition module is used to acquire raw multi-source data, which includes data acquired using GNSS displacement gauges, inclinometers, crack gauges, rain gauges, water level gauges, stress and strain sensors, and video stream sensors. The storage module is used to store the collected raw multi-source data as time-series monitoring data using time-series storage. The prediction module is used to fuse the time-series monitoring data, use the fused time-series monitoring data as the data source, and use a machine learning model to make predictions. The prediction results include slope deformation trends and abnormal states of monitoring data. The early warning module is used to combine the time-series monitoring data and prediction results after data fusion to output the risk level and issue risk warnings.

9. An electronic device, characterized in that, include: The system includes a memory and a processor, which are interconnected. The memory stores computer instructions, and the processor executes the computer instructions to perform the slope intelligent early warning method based on multi-source data fusion as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the slope intelligent early warning method based on multi-source data fusion as described in any one of claims 1 to 7.

Images