CN-121982862-A - Intelligent monitoring system for high-level and hidden landslide blind areas based on Internet of things

CN121982862ACN 121982862 ACN121982862 ACN 121982862ACN-121982862-A

Abstract

The invention relates to the technical field of landslide detection, in particular to an intelligent monitoring system for high-level and hidden landslide blind areas based on the Internet of things. The method comprises the steps of S1, accessing multi-source heterogeneous time sequence data from an Internet of things monitoring network arranged on a landslide body, wherein the multi-source heterogeneous time sequence data comprise deep displacement data, ground surface displacement data, pore water pressure data and rainfall data, S2, respectively extracting preset sensitive characteristic parameters from the multi-source heterogeneous time sequence data, uniformly aligning all sensitive characteristic parameters to the same time standard to obtain a multi-dimensional characteristic vector, wherein the sensitive characteristic parameters comprise displacement acceleration and osmotic pressure change rate, S3, carrying out correlation analysis according to bit displacement acceleration and osmotic pressure change rate change trend, calculating a current abnormality index according to correlation analysis results, S4, comparing the abnormality index with a preset threshold value defined on the basis of a multi-parameter joint judgment rule, judging that the landslide body enters an abnormal state when the joint judgment rule is met, and generating a corresponding early warning signal.

Inventors

ZHANG BIN
CHEN CHONG
ZENG MINGSONG
XU QINGFANG
PANG HANG

Assignees

贵州省地质矿产勘查开发局一0六地质大队

Dates

Publication Date: 20260505
Application Date: 20260129

Claims (6)

1. High-order, hidden landslide blind area intelligent monitoring system based on thing networking, its characterized in that includes the following step: S1, accessing multi-source heterogeneous time sequence data from an Internet of things monitoring network arranged on a landslide body, wherein the multi-source heterogeneous time sequence data comprise deep displacement data, ground surface displacement data, pore water pressure data and rainfall data; S2, respectively extracting preset sensitive characteristic parameters from multi-source heterogeneous time sequence data, and uniformly aligning all the sensitive characteristic parameters to the same time reference to obtain a multi-dimensional characteristic vector, wherein the sensitive characteristic parameters comprise displacement acceleration and osmotic pressure change rate; S3, carrying out correlation analysis according to the bit displacement acceleration and the change trend of the osmotic pressure change rate, and calculating the current abnormality index according to the correlation analysis result; S4, comparing the abnormality index with a preset threshold value defined based on a multi-parameter joint judgment rule, and judging that the landslide body enters an abnormal state and generating a corresponding early warning signal when the joint judgment rule is met.
2. The intelligent monitoring system for the high-level and hidden landslide dead zone based on the Internet of things of claim 1, wherein the S1 comprises the following steps: S11, periodically acquiring deep displacement data, ground surface displacement data and pore water pressure data through sensor nodes which are deployed at different depths and are not the same as a landslide body, and acquiring rainfall data through weather stations deployed in a monitoring area; S12, transmitting the acquired multi-source heterogeneous time sequence data to a remote data center through a communication module; S13, the remote data center receives and caches data streams from different sensor nodes, and marks a source identifier, an acquisition time stamp and a preliminary quality identifier for each piece of data; S14, checking the received data, eliminating abnormal values, processing data loss caused by communication interruption, unifying data formats and units, and forming a standardized multi-source heterogeneous time sequence data stream.
3. The intelligent monitoring system for the high-level and hidden landslide dead zone based on the Internet of things of claim 2, wherein the S2 comprises the following steps: S21, extracting the time sequence of the deep displacement data and the ground surface displacement data processed in the S14 to obtain displacement acceleration characteristics through high-pass filtering, and extracting the time sequence of the pore water pressure data to obtain osmotic pressure change rate characteristics through calculating the change rate in unit time; s22, performing time alignment processing on the extracted displacement acceleration characteristic time sequence and the extracted osmotic pressure change rate characteristic time sequence, and unifying characteristic values inconsistent in sampling time to preset common time points at equal intervals by adopting a linear interpolation or nearest neighbor interpolation method; S23, combining the displacement acceleration characteristic values and the osmotic pressure change rate characteristic values which are subjected to time alignment processing and come from different data sources at each public time point to form a multidimensional characteristic vector.
4. The intelligent monitoring system for the high-order and hidden landslide dead zone based on the Internet of things of claim 3, wherein the S3 comprises the following steps: S31, calculating a pearson correlation coefficient between the displacement acceleration characteristic time sequence and the osmotic pressure change rate characteristic time sequence aiming at the multidimensional characteristic vector sequence in the current time window to obtain a correlation measurement value; s32, according to the relevance measurement value, combining the displacement acceleration characteristic value and the amplitude value of the osmotic pressure change rate characteristic in the current moment or the current time window, and obtaining an abnormality index through a preset weighted fusion formula.
5. The intelligent monitoring system for the high-order and hidden landslide dead zone based on the Internet of things of claim 4, wherein the S4 comprises the following steps: s41, acquiring the current abnormality index obtained by calculation in the step S32, and acquiring a relevance metric value, a displacement acceleration characteristic value and a osmotic pressure change rate characteristic value in a current or latest time window; S42, respectively comparing the current abnormality index, the relevance metric value, the displacement acceleration characteristic value and the osmotic pressure change rate characteristic value with corresponding preset thresholds according to a preset multi-parameter joint judgment rule; S43, judging that the landslide body enters an abnormal state according to the comparison result in the S42; s44, generating early warning signals containing warning grades, suggested measures and affected area information according to the judging result of the abnormal state, and triggering a warning release flow.
6. The intelligent monitoring system for the high-order and hidden landslide dead zone based on the Internet of things of claim 5, further comprising S5, wherein the S5 comprises the following steps: S51, continuously monitoring the actual evolution condition of a landslide body after each early warning signal is issued, and recording the issuing time of the early warning signal, the early warning level and the actual state change of the landslide body in a preset time period after early warning, wherein the actual state change comprises whether sliding occurs, the occurrence time of the sliding and the sliding scale; S52, calculating an evaluation index of the early warning based on the comparison of the early warning signal and the actual state change of the landslide body, wherein the evaluation index comprises an early warning hit rate, an early warning false alarm rate, early warning advance time and early warning grade accuracy; S53, judging whether key parameters in the system need to be adjusted according to the evaluation index obtained by the calculation in the S52, starting a parameter optimization algorithm when the evaluation index does not meet the preset performance requirement, and optimally adjusting the weight coefficient of the weighted fusion formula in the S32 and the threshold value of the multi-parameter joint judgment rule in the S42; And S54, verifying the optimized and adjusted weight coefficient and/or threshold on a historical data set, if the verification result shows that the evaluation index is improved, updating the optimized parameter into the system for subsequent abnormal index calculation and early warning judgment, otherwise, retaining the original parameter.

Description

Intelligent monitoring system for high-level and hidden landslide blind areas based on Internet of things Technical Field The invention relates to the technical field of landslide detection, in particular to an intelligent monitoring system for high-level and hidden landslide blind areas based on the Internet of things. Background China is a country with a plurality of geological disasters, wherein landslide disasters are one of main disaster types due to high occurrence frequency, wide distribution range and strong destructive power. Particularly, high-level landslide in high-steep side slopes and terrain complex areas and hidden landslide covered by vegetation and with insignificant signs of surface deformation are prone to serious casualties and property loss during sudden accidents due to special positions and high monitoring difficulty. With the acceleration of urbanization progress and the extension of major projects to mountainous areas, landslide disasters are increasingly threatening the infrastructure, resident safety and ecological environment. Therefore, the high-efficiency, accurate and reliable landslide monitoring and early warning technology is developed, early identification and early warning of landslide, especially high-level hidden landslide are realized, urgent requirements in the field of geological disaster prevention and control are met, and the method has important practical significance for guaranteeing life and property safety of people, maintaining social stability and promoting sustainable development. For a long time, a great deal of research and practice are carried out in the landslide monitoring field at home and abroad, and the monitoring technical means are gradually developed from the traditional manual inspection and simple observation to the automatic and informatization directions. The traditional monitoring method mainly relies on regular field investigation of geology personnel, and judges landslide deformation trend by means of measuring the width of surface cracks, setting simple displacement piles and the like, and the method is highly dependent on experience and limited by weather, topography and personnel accessibility, is difficult to realize continuous and real-time monitoring, and particularly cannot effectively cover high-level and hidden landslide blind areas. With the progress of mapping and sensing technologies, equipment such as a global navigation satellite system, a total station, an inclinometer and the like appears, so that the automatic monitoring of earth surface and shallow displacement can be realized, and the frequency and the accuracy of data acquisition are improved. In addition, the application of the remote sensing technology, in particular to the synthetic aperture radar interferometry technology, enables large-scale and non-contact surface deformation monitoring to be possible, and provides a new tool for identifying the hidden danger of the regional landslide. In the aspect of osmotic pressure and hydrologic monitoring, by embedding sensors such as pore water pressure gauges, rain gauges and the like, underground water dynamics and external rainfall input in a landslide body can be obtained, and data support is provided for analyzing landslide stability and rainfall correlation. The technologies achieve certain effects under different scenes, and form a main technical system for current landslide monitoring. However, facing the special and difficult task of high-level, blind landslide monitoring, the prior art system still presents several significant drawbacks and challenges. First, existing methods tend to focus on independent monitoring of a single or a few parameters, such as only surface displacement or only groundwater level measurement, and lack collaborative acquisition and fusion analysis of multiple physical fields data. The occurrence of landslide is the result of multi-factor long-term coupling action such as geological structure, groundwater, external rainfall, etc., and single parameter is difficult to comprehensively reflect the real state and evolution trend of landslide body, and misjudgment or missed judgment is easy to cause. Secondly, in the aspect of data acquisition, the existing monitoring equipment is distributed in the area with relatively convenient traffic and relatively gentle topography, and is difficult to deploy sensors, power supply and communication guarantee are weak to high side slopes with steep topography, dense vegetation and difficult personnel to reach, so that a monitoring blind area is formed. Even if the equipment is deployed, data is often lost and transmission is interrupted due to severe environments, and continuous and reliable time sequence data is difficult to obtain. In addition, in the data analysis and early warning layer, the existing system mostly adopts a single-index early warning model based on a fixed threshold, namely, when a certain monitoring value exceeds a preset threshold, an