Landslide mass displacement monitoring method based on Internet of Things

Abstract

The invention relates to a landslide mass displacement monitoring method based on the Internet of Things, and the method comprises the steps: collecting multi-source data from a monitoring point through a multi-hop self-organizing network, carrying out the preprocessing of the multi-source data, and obtaining a stable data sequence; according to the transmission path of the stable data sequence in the complex environment, obtaining signal strength and delay indexes between network nodes; optimizing a transmission path according to the signal strength and the delay index, and obtaining a transmission link; acquiring a data stream of each monitoring point position in the transmission link, extracting space coordinate information, performing grouping association on the displacement information and the rainfall value through a clustering algorithm, and determining a clustered space association group to form an integrated data set of space matching; and carrying out feature fusion on the integrated data set subjected to space matching, calculating a comprehensive risk value of each group by adopting a weighted average method, and obtaining a risk assessment result. According to the method, the accuracy and the real-time performance of disaster early warning are remarkably improved, and reliable support is provided for disaster prevention and control in a complex environment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of landslide disaster monitoring, and in particular to a landslide displacement monitoring method based on the Internet of Things. BACKGROUND

[0002] Landslide disaster monitoring is an important field of geological disaster prevention, which is related to people's life and property safety and regional stable development. Its importance is self-evident. With the change of natural environment and the intensification of human activities, landslide disasters occur frequently, which puts forward higher requirements for the accuracy and timeliness of monitoring technology. Effective monitoring means is not only the key to disaster warning, but also an important basis for scientific decision-making.

[0003] However, many current landslide monitoring methods have significant shortcomings in practical application, especially in data collection and processing capability in complex environments. Existing technologies often fail to adapt to changing geological conditions and dynamic disaster evolution processes, especially in data transmission and multi-source information integration, which is prone to information island phenomenon, resulting in limitations in the comprehensiveness and real-time of monitoring. This limitation makes the accuracy of disaster warning greatly discounted, which is difficult to meet the actual demand.

[0004] Under this background, the landslide monitoring field faces many technical difficulties, the most core of which is how to realize stable transmission of data in a multi-hop self-organizing network environment. Complex mountainous terrain and harsh weather conditions often lead to unstable wireless network signals, and data transmission is prone to interruption or loss, which further affects the integrity of monitoring information. This unstable transmission further exacerbates another key problem, which is the difficulty of matching multi-source data in time and space. The displacement, inclination angle and rainfall information collected by different monitoring points need to be synchronized in time and correlated in space, but due to transmission delay or data loss, it is difficult to form a unified analysis basis. For example, after a heavy rain, the rainfall data of a certain monitoring point fails to be transmitted in time, while the displacement data shows abnormality, which ultimately leads to the inability to accurately determine whether the region has entered a dangerous state.

[0005] Therefore, how to ensure stable transmission of multi-source data in complex environments and realize effective matching in time and space has become a key problem in landslide disaster monitoring and warning. SUMMARY

[0006] The purpose of the present application is to provide a landslide displacement monitoring method based on the Internet of Things, which significantly improves the accuracy and real-time of disaster warning and provides reliable support for disaster prevention and control in complex environments.

[0007] To achieve the above purpose, the present application provides the following scheme:

[0008] A landslide displacement monitoring method based on the Internet of Things, comprising:

[0009] Collecting multi-source data from monitoring points through a multi-hop self-organizing network, pre-processing the multi-source data, and obtaining a stable data sequence, wherein the multi-source data includes displacement information, an inclination angle, and rainfall values of a target landslide body;

[0010] According to a transmission path of the stable data sequence in a complex environment, obtaining signal strength and delay indicators between network nodes;

[0011] According to the signal strength and delay indicators, optimizing the transmission path, and obtaining a transmission link;

[0012] Obtaining data streams of each monitoring point in the transmission link, extracting spatial coordinate information, grouping and associating the displacement information and rainfall values through a clustering algorithm, determining a spatially associated group after clustering to form an integrated data set matched in space;

[0013] Performing feature fusion on the integrated data set matched in space, calculating a comprehensive risk value of each group using a weighted average method, and obtaining a risk assessment result.

[0014] Optionally, the pre-processing of the multi-source data to obtain a stable data sequence includes:

[0015] Performing a noise filtering operation on the multi-source data using a particle filtering algorithm, eliminating abnormal interference in the data, completing preliminary state estimation, and outputting an intermediate data sequence after processing;

[0016] Extracting independent data streams of the displacement information, the inclination angle, and the rainfall values from the intermediate data sequence, and performing time series alignment processing on each type of data stream;

[0017] According to the aligned data stream, if a data point of displacement information or an inclination angle exceeds a preset threshold range, the data point is marked, and a correlation analysis is performed in combination with a historical record of rainfall values to determine whether an abnormal state exists;

[0018] Obtaining the marked data points and the correlation analysis result, performing data interpolation repair on the data points in an abnormal state, filling in missing or abnormal parts using a linear interpolation method, and obtaining a repaired data sequence;

[0019] Through the repaired data sequence, performing smoothing processing on the change trend of the displacement information, the inclination angle, and the rainfall values, and obtaining the stable data sequence.

[0020] Optionally, the time series alignment processing on each type of data stream includes:

[0021] For the data stream of each monitoring point, recording the corresponding time stamp, and obtaining an initial multi-source data set;

[0022] For the initial multi-source data set, a synchronous calibration method is used to process the differences in timestamps. By comparing the timestamps of each data stream, the offset is calculated, and the calibrated timestamp dataset is obtained.

[0023] Based on the calibrated timestamp dataset, alignment processing is performed to arrange each data stream according to a unified time axis, obtain the aligned data sequence, calculate the synchronization error between each data stream, and determine whether the synchronization requirements are met by comparing it with a preset range, and obtain the error judgment result.

[0024] If the error judgment result shows that the synchronization error exceeds the preset range, the synchronization calibration method is re-executed on the data stream that exceeds the range, the timestamp offset is adjusted, and the updated aligned data sequence is obtained.

[0025] Using the updated aligned data sequence, synchronization error data within a preset range are filtered out, a time-matched subset of data is determined, and the aligned data stream is obtained.

[0026] Optionally, optimizing the transmission path based on the signal strength and delay metrics, and obtaining the transmission link includes:

[0027] If the signal strength or delay index does not meet the preset threshold, an alternative path is selected based on historical data.

[0028] The alternative paths are comprehensively evaluated using the signal strength and delay metrics to obtain the transmission link status after the switch.

[0029] Based on the state of the switched transmission link, the continuity of the data stream is detected. If a data stream interruption is detected, the path is adjusted to obtain a stable transmission link.

[0030] Through the stable transmission link, the performance of network nodes in complex environments is continuously monitored, and changes in signal strength and latency are dynamically recorded to determine the direction of transmission path optimization.

[0031] Based on the dynamic recording of performance changes and the filtering results of the data sequence, the final transmission link is obtained.

[0032] Optionally, the displacement information and rainfall values ​​are grouped and correlated using a clustering algorithm to determine the spatially correlated groups after clustering, forming a spatially matched integrated dataset, including:

[0033] Based on the spatial coordinate information, the displacement information and rainfall value are grouped using the K-means clustering method to determine the grouped category labels;

[0034] For the grouped category labels, obtain spatial coordinates and related index data within each category, and determine spatial correlation strength between categories;

[0035] According to the correlation strength, determine the correlation mode of spatial distribution, and obtain the integrated data set of spatial matching.

[0036] Optionally, according to the correlation strength, determining the correlation mode of spatial distribution and obtaining the integrated data set of spatial matching comprises:

[0037] If the spatial correlation strength exceeds a preset threshold, it is classified as a strong correlation group, and a strong correlation spatial grouping result is obtained;

[0038] From the strong correlation spatial grouping result, extract the distribution characteristics of displacement information and rainfall value, construct a mapping table of spatial distribution, and determine the correlation mode of spatial distribution;

[0039] According to the correlation mode of spatial distribution, integrate various index information in the data grouping, form a data structure of the matching result, and obtain an intermediate data set of spatial matching;

[0040] Through the intermediate data set of spatial matching, analyze the corresponding relationship between spatial coordinates and index analysis, and obtain the integrated data set of spatial matching.

[0041] Optionally, the feature fusion is performed on the integrated data set of spatial matching, and a weighted average method is used to calculate the comprehensive risk value of each group, comprising:

[0042] From the integrated data set of spatial matching, obtain spatial matching related information, classify and arrange data points in different regions, and obtain a preliminary classified regional data set;

[0043] According to the preliminary classified regional data set, a multi-dimensional attribute merging is performed on data points in each region by using a feature fusion method, and a fused feature set is obtained;

[0044] For the fused feature set, a weighted average method is used to comprehensively calculate multiple attribute values of each region, and a comprehensive risk value of each region is obtained.

[0045] Optionally, obtaining the risk assessment result comprises:

[0046] If the comprehensive risk value of each region exceeds a preset threshold, the corresponding region is marked as high risk, and the distribution of high risk regions is determined;

[0047] According to the distribution of high risk regions, combined with the standard of regional division, the boundary of the marked region is adjusted, and the final risk region division result is obtained;

[0048] Through the final risk area division result, corresponding risk assessment data is generated to judge the risk level distribution of each area.

[0049] The present application has the advantages that: the present application solves the problem of inaccurate risk assessment caused by unstable data transmission, large time synchronization error and insufficient spatial correlation in complex environment by collecting, processing and analyzing multi-source data of monitoring points. The present application first uses the particle filter algorithm to filter noise and estimate the state of original data such as displacement, tilt angle and rainfall, to obtain a stable data sequence; in the case of insufficient signal strength in the transmission path, the data transmission path is dynamically adjusted to ensure data flow continuity; K-means clustering algorithm is used for spatial correlation grouping to form an integrated data set; finally, the comprehensive risk value is calculated by feature fusion and weighted average, and the high-risk area is marked and the warning signal is broadcast. The present application significantly improves the accuracy and real-time performance of disaster warning through comprehensive technical means of data processing, time alignment and spatial matching, and provides reliable support for disaster prevention and control in complex environment. BRIEF DESCRIPTION OF DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0051] Figure 1 A flowchart of a landslide displacement monitoring method based on the Internet of Things according to an embodiment of the present application. DETAILED DESCRIPTION

[0052] The technical solutions in the embodiments of the present application will be described clearly and completely with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0053] In order to make the above-mentioned purposes, features and advantages of the present application more apparent and easy to understand, the present application will be further described in detail with reference to the drawings and specific embodiments.

[0054] As shown in Figure 1 The present embodiment proposes a landslide displacement monitoring method based on the Internet of Things, which includes:

[0055] The multi-source data is collected from the monitoring points through the multi-hop self-organizing network, and the multi-source data is preprocessed to obtain a stable data sequence, wherein the multi-source data includes displacement information, an inclination angle and rainfall values of a target landslide body;

[0056] According to a transmission path of the stable data sequence in a complex environment, signal strength and delay indicators between network nodes are obtained;

[0057] According to the signal strength and delay indicators, the transmission path is optimized to obtain a transmission link;

[0058] Data streams of each monitoring point in the transmission link are obtained, spatial coordinate information is extracted, displacement information and rainfall values are grouped and associated through a clustering algorithm, and a spatially associated group after clustering is determined to form a spatially matched integrated data set;

[0059] The spatially matched integrated data set is subjected to feature fusion, a weighted average method is used to calculate a comprehensive risk value of each group, and a risk assessment result is obtained.

[0060] Further, the preprocessing of the multi-source data to obtain the stable data sequence includes:

[0061] A particle filtering algorithm is used to perform a noise filtering operation on the multi-source data, abnormal interference in the data is removed, preliminary state estimation is completed, and an intermediate data sequence after processing is output;

[0062] Independent data streams of displacement information, inclination angles and rainfall values are extracted from the intermediate data sequence, and time sequence alignment processing is performed on each type of data stream;

[0063] According to the aligned data stream, if the data points of the displacement information or the inclination angle exceed a preset threshold range, the data points are marked, and associated analysis is performed in combination with historical records of the rainfall values to determine whether there is an abnormal state;

[0064] The marked data points and the associated analysis result are obtained, data interpolation repair is performed on the data points in the abnormal state, a linear interpolation method is used to fill in the missing or abnormal parts, and a repaired data sequence is obtained;

[0065] Through the repaired data sequence, the change trend of the displacement information, the inclination angle and the rainfall values is smoothed to obtain the stable data sequence.

[0066] Specifically, in the process of collecting multi-source data through multi-hop network and self-organizing network cooperation mechanism, a landslide monitoring scene can be imagined. The multi-hop network realizes data relay transmission through multiple sensor nodes, and the self-organizing network dynamically adjusts the connection between nodes to ensure that data is transmitted from remote monitoring points to the center node. Assuming that there are three monitoring points, respectively collecting displacement information, inclination angle and rainfall value, the initial data record has a displacement value of 5.2 cm, an inclination angle of 3.5 degrees, and a rainfall of 20 mm. These data form the original data set, but may contain noise interference. In one possible implementation, the particle filter algorithm is used for noise filtering for the original data set. Particle filtering estimates the true value by simulating multiple particle states to remove abnormal interference. For example, if a point in the displacement data suddenly jumps to 10 cm, which is obviously deviated from the trend, the particle filter will correct it to a reasonable value close to 5.3 cm according to the historical data and probability distribution. Such processing can effectively improve the reliability of the data and output the intermediate data sequence to lay the foundation for subsequent analysis. For example, time series alignment processing of the intermediate data sequence can ensure that the displacement, inclination angle and rainfall value correspond to the same time point through timestamp matching. Assuming that the displacement data is missing at a certain time point, the alignment processing can find that the corresponding inclination angle is 3.8 degrees and the rainfall is 25 mm, thereby providing consistency for subsequent analysis. Time alignment can avoid misjudgment caused by misplaced data. In one possible implementation, if the displacement or inclination angle exceeds the threshold value, for example, the displacement threshold is set to 6 cm, and the data of a certain point is 6.5 cm, it is marked as an abnormal point, and is analyzed in combination with the historical record of rainfall. If the recent rainfall is continuously higher than 30 mm, it can be inferred that the abnormality is related to rainfall. Such correlation analysis helps to determine whether the abnormality is a natural phenomenon and improves the accuracy of monitoring. For example, when performing linear interpolation repair on the abnormal data point, assuming that the displacement data is 6.5 cm at a certain time point, and the data of the previous and next time points is 5.4 cm and 5.6 cm respectively, the interpolation can be corrected to 5.5 cm. The repaired data sequence is smoother, avoiding the interference of abnormal points on trend analysis, and enhancing the usability of data. In one possible implementation, a stable sequence is constructed through the repaired data sequence, and the displacement, inclination angle and rainfall value are smoothed.

[0067] Further, the time series alignment processing for each type of data stream includes:

[0068] For the data stream of each monitoring point, record the corresponding time stamp, and obtain the initial multi-source data set;

[0069] For the initial multi-source data set, use a synchronous calibration method to process the difference of the time stamp, calculate the offset by comparing the time stamps of each data stream, and obtain the calibrated time stamp data set;

[0070] According to the calibrated timestamp data set, alignment processing is performed to arrange each data stream according to a unified time axis, obtain an aligned data sequence, calculate the synchronization error between each data stream, compare with a preset range to determine whether the synchronization requirement is met, and obtain an error judgment result;

[0071] If the error judgment result shows that the synchronization error exceeds the preset range, the synchronization calibration method is re-executed for the data stream exceeding the range, the timestamp offset is adjusted, and an updated aligned data sequence is obtained;

[0072] Through the updated aligned data sequence, the synchronization error data within the preset range is filtered out, the time-matched data subset is determined, and the aligned data stream is obtained.

[0073] Further, the transmission path is optimized according to the signal strength and delay index, and the transmission link is obtained:

[0074] If the signal strength or delay index does not meet the preset threshold, a backup path is selected according to historical data;

[0075] The backup path is comprehensively evaluated through the signal strength and delay index, and the state of the transmission link after switching is obtained;

[0076] According to the state of the transmission link after switching, the continuity of the data stream is detected, and if the data stream is interrupted, the path is adjusted to obtain a stable transmission link;

[0077] Through the stable transmission link, the network node performance in a complex environment is continuously monitored, the changes of signal strength and delay index are dynamically recorded, and the optimization direction of the transmission path is determined;

[0078] According to the dynamically recorded performance changes, combined with the filtering processing result of the data sequence, the final transmission link is obtained.

[0079] Specifically, when monitoring the transmission path of each node in real time, the signal strength and delay changes can be continuously recorded by the sensor nodes deployed near the monitoring points. Assuming that the preset signal strength threshold is -70 decibels, if the signal strength of a certain node decreases to -80 decibels, it indicates that the current path may be affected by environmental interference, such as mountain obstruction or weather factors. At this time, it is necessary to trigger the backup path selection mechanism. The selection of backup path can be based on historical data, and the path with signal strength stable above -65 decibels is preferred. For example, in the feasibility analysis of backup path, the comprehensive evaluation of signal strength and delay indicators is the key. Assuming that the signal strength of path A is -68 decibels and the delay is 180 milliseconds, while the signal strength of path B is -72 decibels and the delay is 150 milliseconds. The path switching technology will tend to select the path A with better comprehensive performance, and route the data packets to this path. After switching, the transmission link state needs to be monitored to ensure the continuity of data flow is not affected. For example, if data flow interruption is detected, such as a node after switching has a data packet loss rate of 10%, the data packet routing strategy needs to be adjusted. Data packets can be dynamically allocated to another backup path to ensure the stability of the transmission link. Assuming that the loss rate decreases to 1% after adjustment, it indicates that the new link can support continuous transmission. For example, when continuously monitoring the performance of network nodes in complex environments, it is particularly important to dynamically record the changes in signal strength and delay indicators. Assuming that the signal strength of a certain node fluctuates from -70 decibels to -85 decibels within a day, and the delay increases from 150 milliseconds to 300 milliseconds, recording these changes helps to determine the optimization direction of the transmission path, such as avoiding areas with unstable signals. Periodic detection of network node running state is indispensable for the final transmission link configuration. Assuming that it is detected once an hour, if it is found that the delay of a certain path continuously exceeds 250 milliseconds, it is judged that it does not meet the performance requirements, and the routing strategy needs to be optimized again. This way can ensure the timeliness and integrity of the monitoring data, and provide stable support for mountain landslide warning.

[0080] Further, the displacement information and rainfall values are grouped and associated by a clustering algorithm to determine the spatial correlation groups after clustering to form a spatially matched integrated data set, including:

[0081] According to the spatial coordinate information, the K-means clustering method is used to group the displacement information and rainfall values, and the class labels after grouping are determined;

[0082] For the class labels after grouping, the spatial coordinates and related index data in each class are obtained, and the spatial correlation strength between classes is judged;

[0083] According to the correlation strength, the correlation mode of spatial distribution is determined, and the spatially matched integrated data set is obtained.

[0084] Further, the acquiring the spatially matched integrated dataset according to the spatial distribution correlation mode comprises:

[0085] If the spatial correlation strength exceeds a preset threshold, it is classified as a strong correlation group, and a strong correlation spatial grouping result is obtained.

[0086] From the strong correlation spatial grouping result, the distribution characteristics of displacement information and rainfall values are extracted, a spatial distribution mapping table is constructed, and a spatial distribution correlation mode is determined.

[0087] According to the spatial distribution correlation mode, the various types of index information in the data grouping are integrated, a data structure of the matching result is formed, and a spatially matched intermediate dataset is obtained.

[0088] Through the spatially matched intermediate dataset, the corresponding relationship between the spatial coordinates and the index analysis is analyzed, and the spatially matched integrated dataset is obtained.

[0089] Specifically, assuming that there are multiple sensor points in a monitoring area, each point records latitude and longitude information, such as the coordinates of point A are 116.5 degrees east longitude and 39.8 degrees north latitude, and the coordinates of point B are 116.6 degrees east longitude and 39.9 degrees north latitude. These data constitute a spatial position distribution record. By sorting these coordinates, a preliminary spatial distribution table is formed to lay the foundation for subsequent analysis. For example, for the spatial position distribution record, the K-means clustering method is used to group the displacement index and rainfall value. The displacement index and rainfall value can be used as features, and it is assumed that the displacement of point A in a certain area is 5.2 cm and the rainfall is 30 mm, and the displacement of point B is 5.5 cm and the rainfall is 32 mm. Through clustering analysis, points with similar characteristics are classified into the same category, and each category is assigned a label, such as "high-risk group" or "low-risk group", so as to clearly define the risk characteristics of different regions. For example, after obtaining the category label, the spatial coordinates and index data in each category are analyzed to determine the spatial correlation strength between categories. Assuming that the points in the "high-risk group" are mostly concentrated in a certain slope area, the spatial distance is close, and the index data such as displacement and rainfall show a high consistency trend, the correlation strength can be calculated. If the preset threshold is 0.8 and the calculation result is 0.85, it is classified as a strong correlation group, and a spatial grouping result is formed to provide a basis for subsequent risk assessment. For example, from the strong correlation spatial grouping result, the distribution characteristics of displacement index and rainfall value are extracted, and a spatial distribution mapping table is constructed. Assuming that in the strong correlation group, the points with large displacement are mostly distributed in the lower part of the slope, and the points with high rainfall are also concentrated in this area, a correlation mode can be determined, i.e. the positive correlation distribution of rainfall and displacement in space. This mapping table helps to reveal the potential landslide inducement.

[0090] Based on the spatial distribution of the correlation mode, integrate various index information to form the data structure of the matching result. Assuming that the displacement, rainfall and corresponding coordinates in a certain category are integrated into a data unit, such as the data unit of point A containing coordinates, displacement 5.2 cm, rainfall 30 mm and other information, the intermediate data set of spatial matching is generated. This structured processing facilitates subsequent analysis of the relationship between different indicators. Through the intermediate data set of spatial matching, the corresponding relationship between spatial coordinates and indicators is analyzed, and the final integrated data is constructed. Assuming that it is found that the displacement of the lower part of the hillside is generally higher than that of the upper part, and the rainfall distribution also presents a similar law, then these relationships can be recorded as part of the complete data set. This integration method helps to fully understand the spatial characteristics of the monitoring area and provides data support for landslide warning.

[0091] Further, the integrated data set of spatial matching is subjected to feature fusion, and a weighted average method is used to calculate the comprehensive risk value of each group, including:

[0092] From the integrated data set of spatial matching, obtain the relevant information of spatial matching, classify and arrange the data points of different regions, and obtain the region data set after preliminary classification;

[0093] According to the region data set after preliminary classification, a multi-dimensional attribute merging method is used to merge the data points of each region, and a fused feature set is obtained;

[0094] For the fused feature set, the weighted average method is used to calculate the comprehensive risk value of each region.

[0095] Specifically, assuming that there are multiple sensor points in a monitoring area, the points are distributed at different heights of the hillside, and the points on the upper, middle and lower parts of the hillside can be classified respectively to form three preliminary region data sets, providing clear regional division for subsequent analysis. For the region data set after preliminary classification, the feature fusion method can be used to integrate multi-dimensional attributes. Assuming that each point contains displacement, rainfall and soil humidity attributes, these attributes can be merged into a comprehensive feature vector according to certain rules during the fusion process. In calculating the comprehensive risk value, the weighted average method can be used to process the fused feature set. Assuming that the weight of displacement is 0.4, the weight of rainfall is 0.3, and the weight of soil humidity is 0.3, the weighted calculation is performed on the feature set of the lower part of the hillside to obtain the comprehensive risk value. If the preset threshold is 0.7 and the calculation result is 0.75, the region is marked as a high-risk region. This method can fully reflect the potential threat of the region.

[0096] Further, the risk assessment result is obtained, including:

[0097] If the comprehensive risk value of each region exceeds a preset threshold, the corresponding region is marked as high risk, and the distribution of the high-risk region is determined.

[0098] According to the distribution of the high-risk region, the boundary of the marked region is adjusted in combination with the standard of region division, and finally, the risk region division result is obtained.

[0099] Through the final risk region division result, corresponding risk assessment data is generated, and the risk level distribution of each region is determined.

[0100] The above-described embodiments are only descriptions of the preferred modes of the present application and do not limit the scope of the present application. Without departing from the design spirit of the present application, various modifications and improvements to the technical solutions of the present application made by those skilled in the art shall fall within the protection scope determined by the claims of the present application.

Claims

1. A landslide displacement monitoring method based on the Internet of Things, characterized in that, include: Multi-source data is collected from monitoring points through a multi-hop self-organizing network. The multi-source data is preprocessed to obtain a stable data sequence. The multi-source data includes displacement information, tilt angle and rainfall value of the target landslide body. Based on the transmission path of the stable data sequence in a complex environment, obtain the signal strength and delay indicators between network nodes; Optimize the transmission path and obtain the transmission link based on the signal strength and delay indicators; Data streams from each monitoring point in the transmission link are acquired, and spatial coordinate information is extracted. The displacement information and rainfall values ​​are grouped and associated using a clustering algorithm to determine the spatial association groups after clustering, thereby forming a spatially matched integrated dataset. Feature fusion is performed on the integrated dataset of spatial matching, and a weighted average method is used to calculate the comprehensive risk value of each group to obtain the risk assessment result.

2. The landslide displacement monitoring method based on the Internet of Things according to claim 1, characterized in that, Preprocessing the multi-source data to obtain a stable data sequence includes: The particle filter algorithm is used to perform noise filtering on the multi-source data to remove abnormal interference in the data, complete the preliminary state estimation, and output the processed intermediate data sequence. Independent data streams containing displacement information, tilt angle, and rainfall values ​​are extracted from the intermediate data sequence, and time series alignment processing is performed on each type of data stream; Based on the aligned data stream, if the displacement information or tilt angle data points exceed the preset threshold range, the data points are marked, and correlation analysis is performed in conjunction with the historical rainfall data to determine whether there is an abnormal state. The labeled data points and the results of the correlation analysis are obtained. Data points in abnormal states are repaired by data interpolation. Linear interpolation is used to fill in missing or abnormal parts and obtain the repaired data sequence. The stable data sequence is obtained by smoothing the trends of displacement information, tilt angle and rainfall value using the repaired data sequence.

3. The landslide displacement monitoring method based on the Internet of Things according to claim 2, characterized in that, Time series alignment for each type of data stream includes: For the data stream at each monitoring point, record the corresponding timestamp to obtain the initial multi-source data set; For the initial multi-source data set, a synchronous calibration method is used to process the differences in timestamps. By comparing the timestamps of each data stream, the offset is calculated, and the calibrated timestamp dataset is obtained. Based on the calibrated timestamp dataset, alignment processing is performed to arrange each data stream according to a unified time axis, obtain the aligned data sequence, calculate the synchronization error between each data stream, and determine whether the synchronization requirements are met by comparing it with a preset range, and obtain the error judgment result. If the error judgment result shows that the synchronization error exceeds the preset range, the synchronization calibration method is re-executed on the data stream that exceeds the range, the timestamp offset is adjusted, and the updated aligned data sequence is obtained. Using the updated aligned data sequence, synchronization error data within a preset range are filtered out, a time-matched subset of data is determined, and the aligned data stream is obtained.

4. The landslide displacement monitoring method based on the Internet of Things according to claim 1, characterized in that, Optimizing the transmission path based on the signal strength and delay metrics, and obtaining the transmission link includes: If the signal strength or delay index does not meet the preset threshold, an alternative path is selected based on historical data. The alternative paths are comprehensively evaluated using the signal strength and delay metrics to obtain the transmission link status after the switch. Based on the state of the switched transmission link, the continuity of the data stream is detected. If a data stream interruption is detected, the path is adjusted to obtain a stable transmission link. Through the stable transmission link, the performance of network nodes in complex environments is continuously monitored, and changes in signal strength and latency are dynamically recorded to determine the direction of transmission path optimization. Based on the dynamic recording of performance changes and the filtering results of the data sequence, the final transmission link is obtained.

5. The landslide displacement monitoring method based on the Internet of Things according to claim 1, characterized in that, The displacement information and rainfall values ​​are grouped and correlated using a clustering algorithm to determine the spatially correlated groups after clustering, forming a spatially matched integrated dataset, including: Based on the spatial coordinate information, the displacement information and rainfall value are grouped using the K-means clustering method to determine the grouped category labels; For the grouped category labels, obtain the spatial coordinates and related indicator data within each category, and determine the strength of the spatial association between categories; Based on the association strength, the spatial distribution association pattern is determined, and the integrated dataset of spatial matching is obtained.

6. The landslide displacement monitoring method based on the Internet of Things according to claim 5, characterized in that, Determining the spatial distribution association pattern based on the association strength, and obtaining the integrated dataset for spatial matching includes: If the spatial correlation strength exceeds a preset threshold, it is classified as a strong correlation group, and the spatial grouping result of strong correlation is obtained. From the strongly correlated spatial grouping results, the distribution characteristics of displacement information and rainfall values ​​are extracted, a spatial distribution mapping table is constructed, and the spatial distribution association pattern is determined. Based on the spatial distribution association pattern, various indicator information within the data group is integrated to form the data structure of the matching results, thus obtaining the intermediate dataset for spatial matching. By analyzing the correspondence between spatial coordinates and index analysis using the intermediate dataset of spatial matching, an integrated dataset of spatial matching is obtained.

7. The landslide displacement monitoring method based on the Internet of Things according to claim 1, characterized in that, Feature fusion is performed on the integrated dataset of spatial matching, and the comprehensive risk value of each group is calculated using a weighted average method, including: Relevant information on spatial matching is obtained from the integrated dataset of spatial matching, and data points in different regions are classified and organized to obtain a preliminary classified regional dataset. Based on the pre-classified regional dataset, a feature fusion method is used to merge the multi-dimensional attributes of the data points in each region to obtain the fused feature set. The combined risk value of each region is obtained by comprehensively calculating the multiple attribute values ​​of each region using a weighted average method on the fused feature set.

8. The landslide displacement monitoring method based on the Internet of Things according to claim 1, characterized in that, Obtaining risk assessment results includes: If the overall risk value of each region exceeds the preset threshold, the corresponding region is marked as high-risk to determine the distribution of high-risk regions. Based on the distribution of the high-risk areas and the criteria for area division, the boundaries of the marked areas are adjusted to obtain the final risk area division results. Based on the final risk area division results, corresponding risk assessment data is generated to determine the risk level distribution of each area.

Images