Cloud-edge collaborative processing and trusted transmission method for high-position geological disaster chain monitoring data

Abstract

This invention discloses a cloud-edge collaborative processing and reliable transmission method for monitoring data of high-altitude geological disaster chains. The method includes: constructing a monitoring data processing system based on a three-tier architecture of edge layer, data fusion layer, and cloud; edge gateways collecting disaster monitoring data to obtain data samples, and setting real-time frequencies for collecting disaster monitoring data based on different risk level indicators; edge gateways collecting monitoring data according to the real-time frequency and generating feature vectors; encrypting the data and sending it to the data fusion layer, which then checks the data integrity; fusing the decrypted preprocessed monitoring data and feature vectors, and introducing a basic probability allocation based on a recognition framework to obtain a fused feature vector containing risk probability information and data feature information; and the data fusion layer inputting the fused feature vectors of each edge gateway into the cloud. This method is suitable for geological disaster monitoring in high-altitude mountainous areas and remote areas without public network access nationwide.

Description

Technical Field

[0001] This invention relates to the field of geological disaster monitoring, specifically to a cloud-edge collaborative processing and reliable transmission method for monitoring data of high-altitude geological disaster chains. Background Technology

[0002] For high-altitude, cold regions with an average elevation of 2700m–4200m, which are areas of active plate tectonics and complex disaster-prone conditions, chain disasters involving ice and rockfalls, debris flows, mudslides, and river damming are frequent. Existing geological disaster field monitoring systems have deployed a large number of monitoring devices, including ground motion, microseismic, UW-FBG distributed fiber optic DAS, infrasound, ground acoustic, radar video, and slurry water level and velocity monitoring equipment. However, in long-term on-site operation and maintenance and data processing, existing traditional monitoring data processing, transmission architectures, and processing methods suffer from key technical defects such as rigid sampling mechanisms, unreasonable cloud-edge processing architectures, difficulties in processing multi-source data, and unreliable data transmission, hindering the realization of second-level early warning for high-altitude disasters.

[0003] Traditional fixed-period sampling techniques are ill-suited to the "long-term incubation, instantaneous outbreak" evolution of high-altitude disasters. The contradiction between energy consumption and the loss of critical information is prominent. Field engineering measurements in Sedongpu and Zelongnong revealed the following: ① During the long-term stable incubation period (over 85% of the year), there is no rock mass fracturing or snow and ice slip activity. High-frequency, second-level sampling generates massive amounts of invalid and redundant data, causing a surge in power consumption of on-site solar power equipment. The low-temperature battery life in the field is shortened from the designed 12 months to 3-4 months, and equipment frequently experiences power outages in high-altitude and cold regions. ② During the sudden disaster phase (the instant of ice and rock collapse or debris flow initiation), fixed low-frequency sampling misses crucial characteristic data such as pre-vibration precursors and peak pulses. Existing technologies, with a limited number of adaptive sampling schemes, rely solely on switching sampling levels based on single rainfall and displacement thresholds. They lack a comprehensive risk quantification model that integrates multi-source monitoring parameters, failing to achieve three-level linkage closed-loop sampling control (second-hour-day), making it difficult to match the multi-causal coupled evolution characteristics of high-altitude chain disasters, resulting in poor engineering applicability.

[0004] The traditional architecture of "uploading all raw data to the cloud and centralized processing in the cloud" suffers from high bandwidth consumption and large early warning latency, making it impossible to achieve second-level local early warning. Due to the objective environment of the Grand Canyon, which lacks full terrestrial 4G / 5G public network coverage and relies solely on narrowband satellite low-bandwidth channels (narrowband satellite single link rate ≤128kbps), the uploading of massive amounts of raw data suffers from queuing congestion and large-scale delays in data packets. The latency of conventional cloud data analysis generally reaches 3 to 15 minutes, far exceeding the second-level early warning timeliness requirements for high-level disasters; at the same time, satellite communication costs have increased dramatically with the amount of data transmission.

[0005] Monitoring data in high-altitude and remote mountainous areas relies on Mesh self-organizing networks and a hybrid broadband and narrowband satellite channel for transmission. The transmission link is affected by rain, snow, mountain obstruction, electromagnetic interference, and unauthorized intrusion. The existing transmission scheme uses plaintext transmission mode, which has two major engineering pain points: ① Data transmission is prone to packet loss and malicious tampering. In the past years, field data verification of the Yajiang landslide dam area found that about 4.3% of the abnormal monitoring data could not be distinguished as either equipment malfunction or actual geological changes, leading to inaccurate judgment of disaster evolution patterns; ② Under the traditional centralized database storage mode, cloud database administrators can modify historical monitoring data, resulting in a lack of reliable original data support for subsequent disaster tracing and disaster mechanism inversion. Summary of the Invention

[0006] To address the aforementioned shortcomings of existing technologies, this invention provides a cloud-edge collaborative processing and reliable transmission method for high-altitude geological disaster chain monitoring data. To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows: A cloud-edge collaborative processing and reliable transmission method for monitoring data of high-altitude geological disaster chains is provided, which includes the following steps: S1: Construct a monitoring data processing system based on a three-tier architecture of edge layer, data fusion layer, and cloud, including an edge layer that deploys intelligent sensor nodes and edge gateways, a data fusion layer that consists of a wireless mesh self-organizing network composed of multiple edge gateways, and a cloud layer that deploys a geological disaster monitoring and early warning cloud platform; S2: The edge gateway collects disaster monitoring data to obtain data samples, calculates the risk level index of the monitoring location based on the information entropy of each type of disaster monitoring data, and sets the real-time frequency of collecting disaster monitoring data based on different risk level indicators; S3: The edge gateway collects monitoring data based on real-time frequency, extracts the time-domain and frequency-domain features of the monitoring data, and generates feature vectors; it generates data blocks from the pre-processed monitoring data, feature vectors, and risk level indicators, and encrypts them before sending them to the data fusion layer, which then determines the integrity of the data. S4: The data fusion layer decrypts the encrypted data block, formulates a risk status identification framework based on the decrypted risk level indicators, fuses the preprocessed monitoring data and feature vectors after decryption, and introduces a basic probability allocation based on the identification framework to obtain a fused feature vector containing risk probability information and data feature information. S5: The data fusion layer inputs the fusion feature vector of each edge gateway into the cloud, and the geological disaster monitoring and early warning cloud platform uses the fusion feature vector to assess the geological disaster risk at the monitoring location.

[0007] Further, step S2 includes: S21: The edge gateway collects disaster monitoring data to obtain data samples, and performs standardization processing on each type of disaster monitoring data to obtain the standardized value of each type of disaster monitoring data; S22: Calculate the proportion of the standardized value of each disaster monitoring data in the data sample to the sum of the standardized values ​​of all data samples, and then calculate the information entropy of each disaster monitoring data. , i Types of disaster monitoring data; S23: Based on information entropy Calculate the weight of each type of disaster monitoring data And based on the amount of disaster monitoring data collected. n Calculate the risk level index of monitoring locations for high-altitude geological hazards R ; ; in, Standardized values ​​for disaster monitoring data; S24: Based on risk level indicators R The disaster risk at the monitoring locations is divided into four levels, and the real-time frequency for collecting disaster monitoring data is set according to different disaster risk levels. .

[0008] Furthermore, real-time frequency The setup method is as follows: ; in, The sampling frequency is daily. The sampling frequency is the time level. The sampling frequency is on the order of seconds.

[0009] Further, step S3 includes: S31: The edge gateway is based on the real-time frequency. Collect disaster monitoring data and preprocess the disaster monitoring data to generate preprocessed monitoring data. D Simultaneously, the time-domain and frequency-domain features of disaster monitoring data are extracted to generate feature vectors of the disaster monitoring data. F ; S32: The edge gateway will preprocess the monitoring data. D eigenvectors F and corresponding risk level indicators R Generate data blocks And use the SHA-256 algorithm to generate data blocks. Data hash value ; S33: Based on data hash value Generate blockchain transaction hashes TxHashand data blocks Encryption is performed to generate encrypted data blocks. Send it to the data fusion layer; S34: Consensus nodes are formed through voting via edge gateways and the geological disaster monitoring and early warning cloud platform. These nodes are then divided into master nodes and slave nodes. The master node is responsible for generating new blocks, while the slave nodes are responsible for verifying blocks. The message complexity of the consensus process is... , U This refers to the number of consensus nodes. S35: The data fusion layer receives encrypted data blocks transmitted by the edge gateway. and transaction hash TxHash According to the transaction hash TxHash Query the corresponding data hash value in the cloud. And compare them; like Then determine the received encrypted data block If the integrity is trusted, proceed to step S4; otherwise, determine the encrypted data block. If the data is tampered with or lost during transmission, the edge gateway is required to retransmit it.

[0010] Further, step S4 includes: S41: Data fusion layer for encrypted data blocks Decryption is performed to obtain preprocessed monitoring data. eigenvectors and risk level indicators According to risk level indicators Develop an identification framework Recognition framework This corresponds to 5 mutually exclusive and exhaustive risk states; ; in, Indicates a safe state. ; Indicates a low-risk status. ; This indicates a medium-risk status. ; Indicates a high-risk status. ; This indicates an extremely high risk state. ; S42: Upload the preprocessed monitoring data from each edge gateway. and eigenvectors The fusion process is performed to obtain a multi-dimensional fused feature vector. Then, the first feature vector is used... j 3D fusion features Calculate the first aThe first edge gateway j 3D fusion features Risk status membership degree , a This is the number of the edge gateway. j To integrate the dimensions of features, b This is the number representing the risk status; S43: For each edge gateway, take the weighted average of the membership of all its fusion features to each risk state as the initial basic probability assignment. ; S44: Considering the reliability differences of different smart sensor nodes at different monitoring locations, a reliability factor based on historical accuracy is introduced. Initial basic probability assignment After making corrections, we obtain the corrected basic probability distribution. ; S45: Assigning basic probabilities By fusing with the multidimensional fusion feature vector, a fusion feature vector containing risk probability information and data feature information is obtained.

[0011] Further, calculate the membership degree. The method is as follows: ; in, For the first j 3D fusion features In a risky state Cluster centers below For the first j 3D fusion features In a risky state The standard deviation below.

[0012] Furthermore, the initial basic probability assignment The calculation method is as follows: ; in, For the first a The first edge gateway j 3D fusion features The weight, J For the number of edge gateways, For the first a The first edge gateway j 3D fusion features Risk status Information gain.

[0013] Furthermore, information gain The calculation method is as follows: ; in, Risk status Information entropy, information entropy The larger the value, the more discrete the risk status distribution and the higher the uncertainty. In the j 3D fusion features Risk status under certain conditions Information entropy for N A risk status was observed in one data sample. The prior probability, A risky situation has emerged. Number of data samples For the first j 3D fusion features A risk state appears in the data sample as a condition. The prior probability, For the first j 3D fusion features The number of data samples is the condition.

[0014] Furthermore, the initial basic probability assignment The method for making the correction is as follows: ; in, For edge gateways in a risky state The historical recognition accuracy rate.

[0015] The beneficial effects of this invention are as follows: This invention breaks away from the traditional industry model of fixed-period sampling and single-parameter threshold switching sampling. It quantifies the comprehensive disaster risk across the entire area in real time and uses a piecewise nonlinear function to dynamically and autonomously switch the sensor sampling frequency. In high-risk areas, a nonlinear rate-increment sampling method is used instead of a step-based fixed level, which aligns with the gradual instability and evolution characteristics of ice and rock avalanches. The edge layer takes effect immediately on-site, achieving dynamic closed-loop feedback of the sampling strategy across all sites. During low-risk normal periods, the average power consumption of equipment is significantly reduced, and the battery life of high-altitude solar batteries is increased from 3-4 months to more than 12 months, resulting in a low failure rate of outdoor equipment during winter power outages.

[0016] The computing power is divided into three levels: edge data-level preprocessing, fog network regional feature-level fusion, and cloud decision-level in-depth analysis. Noise reduction and feature extraction are completed locally at the edge gateway, and fog layer mesh networking realizes the aggregation of features from multiple sites in the area. Only lightweight feature vectors are uploaded to the cloud, reducing the amount of data transmitted from the source.

[0017] This invention relies on historical monitoring data to automatically calculate the weights of fused features across various dimensions through information gain. It combines the accuracy of on-site identification from various sensors to correct the basic probability allocation (BPA). It uses Shannon entropy and conditional entropy to calculate feature information gain and data-driven automatic weighting, which can effectively reduce the false alarm rate of high-altitude ice avalanches and debris flows.

[0018] Only the SHA256 hash, device ID, and timestamp are packaged and uploaded to the consortium blockchain, while the original massive amount of test data is stored in the local storage unit of the edge gateway, solving the problem of storage explosion when all massive monitoring data is uploaded to the blockchain. Attached Figure Description

[0019] Figure 1 A flowchart for cloud-edge collaborative processing and reliable transmission of high-altitude geological disaster chain monitoring data. Detailed Implementation

[0020] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0021] like Figure 1 As shown, the cloud-edge collaborative processing and reliable transmission method for high-altitude geological disaster chain monitoring data includes the following steps: S1: Construct a monitoring data processing system based on a three-tier architecture of edge layer - data fusion layer - cloud, including: Edge layer: Deploy smart sensor nodes and edge gateways. Smart sensor nodes are used to collect disaster monitoring data, including ground motion, microseismic, uwDAS, infrasound, ground sound, mud and water level, flow velocity, and video data. Edge gateways preprocess the collected disaster monitoring data, extract features in real time, and provide local early warnings. Data fusion layer: Composed of multiple edge gateways forming a wireless mesh self-organizing network, enabling data relay, multi-node collaborative computing, and regional-level data fusion; Cloud-based: Deploy a geological disaster monitoring and early warning cloud platform to achieve massive data storage, in-depth analysis, risk assessment, and global early warning dissemination.

[0022] S2: The edge gateway collects disaster monitoring data to obtain data samples, calculates the risk level index of the monitoring location based on the information entropy of each type of disaster monitoring data, and sets the real-time frequency of collecting disaster monitoring data based on different risk level indicators.

[0023] Step S2 specifically includes: S21: The edge gateway collects disaster monitoring data to obtain data samples, and performs standardization processing on each type of disaster monitoring data to obtain the standardized value of each type of disaster monitoring data; S22: Calculate the proportion of the standardized value of each disaster monitoring data in the data sample to the sum of the standardized values ​​of all data samples. Then, the information entropy of each type of disaster monitoring data is calculated. ; ; in, i For the types of disaster monitoring data, m For the sample size, k Number the sample; S23: Based on information entropy Calculate the weight of each type of disaster monitoring data And based on the amount of disaster monitoring data collected. n Calculate the risk level index of monitoring locations for high-altitude geological hazards R ; ; in, Standardized values ​​for disaster monitoring data; S24: Based on risk level indicators R The disaster risk at the monitoring locations is divided into four levels, and the real-time frequency for collecting disaster monitoring data is set according to different disaster risk levels. ; ; in, The sampling frequency is daily. The sampling frequency is the time level. The sampling frequency is on the order of seconds.

[0024] Daily sampling frequency This indicates that the disaster monitoring data is sampled once a day (low power mode), with an hourly sampling frequency. This indicates that the disaster monitoring data is sampled once per hour (standard mode), with a sampling frequency at the second level. This indicates that the sampling frequency for disaster monitoring data is once per second (early warning mode).

[0025] High-altitude geological hazard chains are characterized by "long incubation periods and short bursts of disaster onset," making it impossible to balance real-time performance and power consumption with a fixed sampling frequency. This invention can assess disaster risk levels in real time and dynamically adjust the sampling frequency, ensuring real-time early warning while minimizing system power consumption. For low-risk situations, a daily sampling frequency is used, reducing system power consumption by over 70% and extending equipment lifespan; for high-risk situations, the sampling frequency is automatically increased to the second level or even higher to ensure the capture of key characteristic information during the imminent geological disaster stage.

[0026] To address the issues of easily tampered and unreliable monitoring data in high-altitude and cold regions, this invention introduces consortium blockchain technology into the field of geological disaster monitoring, enabling reliable storage and traceability of monitoring data throughout its entire lifecycle.

[0027] S3: The edge gateway collects monitoring data based on real-time frequency, performs preprocessing, extracts the time-domain and frequency-domain features of the monitoring data, and generates feature vectors; it generates data blocks from the preprocessed monitoring data, feature vectors, and risk level indicators, and sends them to the data fusion layer after encryption.

[0028] Step S3 specifically includes: S31: The edge gateway is based on the real-time frequency. Collect disaster monitoring data and preprocess the disaster monitoring data to generate preprocessed monitoring data. D Simultaneously, the time-domain and frequency-domain features of disaster monitoring data are extracted to generate feature vectors of the disaster monitoring data. F ; Preprocessing of disaster monitoring data includes outlier removal, missing value imputation, and data standardization. In this embodiment, the extracted time-domain features include the mean, variance, peak value, and kurtosis of the disaster monitoring data. Frequency-domain features are extracted by obtaining the spectrum through Fast Fourier Transform and calculating the dominant frequency. With energy distribution ; ; in, j The imaginary unit, t For the time of disaster monitoring data collection, N The data sample size for disaster monitoring data. For the first time collected i Disaster monitoring data sequences; S32: The edge gateway will preprocess the monitoring data. D eigenvectors F and corresponding risk level indicators R Generate data blocks And use the SHA-256 algorithm to generate data blocks. Data hash value ; ; in, T For the timestamp of disaster monitoring data collection, Identifiers for smart sensor nodes that collect disaster monitoring data. This represents a string concatenation operation; S33: Based on data hash value Generate blockchain transaction hashes TxHash and data blocks Encryption is performed to generate encrypted data blocks. Send it to the data fusion layer; S34: Consensus nodes are formed through voting via edge gateways and the geological disaster monitoring and early warning cloud platform. These nodes are then divided into master nodes and slave nodes. The master node is responsible for generating new blocks, while the slave nodes are responsible for verifying blocks. The message complexity of the consensus process is... , U This refers to the number of consensus nodes. New block The structure is as follows: ; in, For the new block height, PrevHash The hash value of the previous block. Timestamp Generate a timestamp for the new block. DataHash The new block contains a set of hash values ​​of all disaster monitoring data. MerkleRoot Merkelgen is the hash value of the disaster monitoring data in the new block. Signature The master node performs the digital signature of the new block; S35: The data fusion layer receives encrypted data blocks transmitted by the edge gateway. and transaction hash TxHash According to the transaction hash TxHash Query the corresponding data hash value in the cloud. And compare them; like Then determine the received encrypted data block If the integrity is trusted, proceed to step S4; otherwise, determine the encrypted data block. If the data is tampered with or lost during transmission, the edge gateway is required to retransmit it.

[0029] S4: The data fusion layer decrypts the encrypted data block, formulates a risk status identification framework based on the decrypted risk level indicators, fuses the preprocessed monitoring data and feature vectors after decryption, and introduces a basic probability allocation based on the identification framework to obtain a fused feature vector containing risk probability information and data feature information.

[0030] Step S4 specifically includes: S41: Data fusion layer for encrypted data blocks Decryption is performed to obtain preprocessed monitoring data. eigenvectors and risk level indicators According to risk level indicators Develop an identification framework Recognition framework This corresponds to 5 mutually exclusive and exhaustive risk states; ; in, Indicates a safe state. ; Indicates a low-risk status. ; This indicates a medium-risk status. ; Indicates a high-risk status. ; This indicates an extremely high risk state. ; S42: Upload the preprocessed monitoring data from each edge gateway. and eigenvectors The fusion process is performed to obtain a multi-dimensional fused feature vector. Then, the first feature vector is used... j 3D fusion features Calculate the first a The first edge gateway j 3D fusion features Risk status membership degree ; ; in, a This is the number of the edge gateway. b This is the number representing the risk status. , For the first j 3D fusion features In a risky state Cluster centers below For the first j 3D fusion features In a risky state The standard deviation below; S43: For each edge gateway, take the weighted average of the membership of all its fusion features to each risk state as the initial basic probability assignment. ; ; in, For the first a The first edge gateway j 3D fusion features The weight, J For the number of edge gateways, For the first a The first edge gateway j 3D fusion features Risk status Information gain; Information gain The calculation method is as follows: ; in, Risk status Information entropy, information entropy The larger the value, the more discrete the risk status distribution and the higher the uncertainty. In the j 3D fusion features Risk status under certain conditions Information entropy for N A risk status was observed in one data sample. The prior probability, A risky situation has emerged. Number of data samples For the first j 3D fusion features A risk state appears in the data sample as a condition. The prior probability, For the first j 3D fusion features The number of data samples that are conditional; S44: Considering the reliability differences of different smart sensor nodes at different monitoring locations, a reliability factor based on historical accuracy is introduced. Initial basic probability assignment After making corrections, we obtain the corrected basic probability distribution. ; ; in, For edge gateways in a risky state Historical recognition accuracy; S45: Assigning basic probabilities By fusing with the multidimensional fusion feature vector, a fusion feature vector containing risk probability information and data feature information is obtained. .

[0031] The data-driven BPA construction method avoids the bias of subjective experience. The reliability-weighted correction of intelligent sensor node credibility fully considers the reliability differences of different monitoring methods, improving the credibility of feature fusion results. The introduction of uncertainty reasonably quantifies the incompleteness of evidence. Basic probability assignment characterizes the degree of support of different data features for the risk state.

[0032] S5: The data fusion layer fuses the feature vectors of each edge gateway. Inputting data into the cloud, a geological disaster monitoring and early warning cloud platform is deployed, utilizing fused feature vectors. Assess the geological disaster risks at the monitored locations.

[0033] This invention achieves full-link data traceability and high-precision multi-source fusion data, providing reliable data for the study of the disaster formation mechanism and dam evolution law of high-altitude disasters in the Grand Canyon; it is suitable for geological disaster monitoring in high-altitude mountainous areas and remote areas without public network across the country, and can be directly transferred to similar high-altitude disaster demonstration areas such as western Sichuan and Yunnan-Tibet.

Claims

1. A cloud-edge collaborative processing and reliable transmission method for high-altitude geological disaster chain monitoring data, characterized in that, Includes the following steps: S1: Construct a monitoring data processing system based on a three-tier architecture of edge layer, data fusion layer, and cloud, including an edge layer that deploys intelligent sensor nodes and edge gateways, a data fusion layer that consists of a wireless mesh self-organizing network composed of multiple edge gateways, and a cloud layer that deploys a geological disaster monitoring and early warning cloud platform; S2: The edge gateway collects disaster monitoring data to obtain data samples, calculates the risk level index of the monitoring location based on the information entropy of each type of disaster monitoring data, and sets the real-time frequency of collecting disaster monitoring data based on different risk level indicators; S3: The edge gateway collects monitoring data based on real-time frequency, extracts the time-domain and frequency-domain features of the monitoring data, and generates feature vectors; it generates data blocks from the pre-processed monitoring data, feature vectors, and risk level indicators, and encrypts them before sending them to the data fusion layer, which then determines the integrity of the data. S4: The data fusion layer decrypts the encrypted data block, formulates a risk status identification framework based on the decrypted risk level indicators, fuses the preprocessed monitoring data and feature vectors after decryption, and introduces a basic probability allocation based on the identification framework to obtain a fused feature vector containing risk probability information and data feature information. S5: The data fusion layer inputs the fusion feature vector of each edge gateway into the cloud, and the geological disaster monitoring and early warning cloud platform uses the fusion feature vector to assess the geological disaster risk at the monitoring location.

2. The cloud-edge collaborative processing and reliable transmission method for high-altitude geological disaster chain monitoring data according to claim 1, characterized in that, Step S2 includes: S21: The edge gateway collects disaster monitoring data to obtain data samples, and performs standardization processing on each type of disaster monitoring data to obtain the standardized value of each type of disaster monitoring data; S22: Calculate the proportion of the standardized value of each disaster monitoring data in the data sample to the sum of the standardized values ​​of all data samples, and then calculate the information entropy of each disaster monitoring data. , i Types of disaster monitoring data; S23: Based on information entropy Calculate the weight of each type of disaster monitoring data And based on the amount of disaster monitoring data collected. n Calculate the risk level index of monitoring locations for high-altitude geological hazards R ; ； in, Standardized values ​​for disaster monitoring data; S24: Based on risk level indicators R The disaster risk at the monitoring locations is divided into four levels, and the real-time frequency for collecting disaster monitoring data is set according to different disaster risk levels. .

3. The cloud-edge collaborative processing and reliable transmission method for high-altitude geological disaster chain monitoring data according to claim 2, characterized in that, The real-time frequency The setup method is as follows: ； in, The sampling frequency is daily. The sampling frequency is the time level. The sampling frequency is on the order of seconds.

4. The cloud-edge collaborative processing and reliable transmission method for high-altitude geological disaster chain monitoring data according to claim 2, characterized in that, Step S3 includes: S31: The edge gateway is based on the real-time frequency. Collect disaster monitoring data and preprocess the disaster monitoring data to generate preprocessed monitoring data. D Simultaneously, the time-domain and frequency-domain features of disaster monitoring data are extracted to generate feature vectors of the disaster monitoring data. F ; S32: The edge gateway will preprocess the monitoring data. D eigenvectors F and corresponding risk level indicators R Generate data blocks And use the SHA-256 algorithm to generate data blocks. Data hash value ; S33: Based on data hash value Generate blockchain transaction hashes TxHash and data blocks Encryption is performed to generate encrypted data blocks. Send it to the data fusion layer; S34: Consensus nodes are formed through voting via edge gateways and the geological disaster monitoring and early warning cloud platform. These nodes are then divided into master and slave nodes. The master node is responsible for generating new blocks, while the slave nodes are responsible for verifying blocks. The message complexity of the consensus process is... , U This refers to the number of consensus nodes. S35: The data fusion layer receives encrypted data blocks transmitted by the edge gateway. and transaction hash TxHash According to the transaction hash TxHash Query the corresponding data hash value in the cloud. And compare them; like Then determine the received encrypted data block If the integrity is trusted, proceed to step S4; otherwise, determine the encrypted data block. If the data is tampered with or lost during transmission, the edge gateway is required to retransmit it.

5. The cloud-edge collaborative processing and reliable transmission method for high-altitude geological disaster chain monitoring data according to claim 3, characterized in that, Step S4 includes: S41: Data fusion layer for encrypted data blocks Decryption is performed to obtain preprocessed monitoring data. eigenvectors and risk level indicators According to risk level indicators Develop an identification framework Recognition framework This corresponds to five mutually exclusive and exhaustive risk states; ； in, Indicates a safe state. ; Indicates a low-risk status. ; This indicates a medium-risk status. ; Indicates a high-risk status. ; This indicates an extremely high risk state. ; S42: Upload the preprocessed monitoring data from each edge gateway. and eigenvectors The fusion process is performed to obtain a multi-dimensional fused feature vector. Then, the first feature vector is used... j 3D fusion features Calculate the first a The first edge gateway j 3D fusion features Risk status membership degree , a This is the number of the edge gateway. j To integrate the dimensions of features, b This is the number representing the risk status; S43: For each edge gateway, take the weighted average of the membership of all its fusion features to each risk state as the initial basic probability assignment. ; S44: Considering the reliability differences of different smart sensor nodes at different monitoring locations, a reliability factor based on historical accuracy is introduced. Initial basic probability assignment After making corrections, we obtain the corrected basic probability distribution. ; S45: Assigning basic probabilities By fusing with the multidimensional fusion feature vector, a fusion feature vector containing risk probability information and data feature information is obtained.

6. The cloud-edge collaborative processing and reliable transmission method for high-altitude geological disaster chain monitoring data according to claim 5, characterized in that, The calculation of membership degree The method is as follows: ； in, For the first j 3D fusion features In a risky state Cluster centers below For the first j 3D fusion features In a risky state The standard deviation below.

7. The cloud-edge collaborative processing and reliable transmission method for high-altitude geological disaster chain monitoring data according to claim 5, characterized in that, The initial basic probability allocation The calculation method is as follows: ； in, For the first a The first edge gateway j 3D fusion features The weight, J For the number of edge gateways, For the first a The first edge gateway j 3D fusion features Risk status Information gain.

8. The cloud-edge collaborative processing and reliable transmission method for high-altitude geological disaster chain monitoring data according to claim 7, characterized in that, The information gain The calculation method is as follows: ； in, Risk status Information entropy, information entropy The larger the value, the more discrete the risk status distribution and the higher the uncertainty. In the j 3D fusion features Risk status under certain conditions Information entropy for N A risk status was observed in one data sample. The prior probability, A risky situation has emerged. Number of data samples For the first j 3D fusion features A risk state appears in the data sample as a condition. The prior probability, For the first j 3D fusion features The number of data samples is the condition.

9. The cloud-edge collaborative processing and reliable transmission method for high-altitude geological disaster chain monitoring data according to claim 5, characterized in that, The initial basic probability allocation The method for making the correction is as follows: ； in, For edge gateways in a risky state The historical recognition accuracy rate.

Images