A water conservancy and water regime monitoring and management system based on Beidou short message communication

Abstract

The present invention discloses a water conservancy and water condition monitoring and management system based on Beidou short message communication, which relates to the field of water condition monitoring technology. The present invention includes scheduling, data compression, data preprocessing and early warning modules. The scheduling module determines the transmission queue by marking the water level status, calculating multiple scores and correcting parameters, and implements the preemption mechanism and silent period setting to optimize transmission; the data compression module classifies and processes data, adjusts the compression rate according to the channel signal, and adopts forward error correction coding to ensure data accuracy; the data preprocessing module filters the original data, uses spatiotemporal correlation analysis to detect anomalies and fill in missing values; the present invention effectively realizes water conservancy and water condition monitoring and management, has significant advantages in data transmission, processing and early warning, improves the accuracy and timeliness of water condition monitoring, and provides strong support for water conservancy management.

Description

Technical Field

[0001] The present invention relates to the technical field of water regime monitoring, and in particular to a water conservancy and water regime monitoring and management system based on Beidou short message communication. Background Art

[0002] Water conservancy and water regime monitoring is of great significance for flood control and drought relief, water resources management, and the construction and management of water conservancy projects. Traditional water conservancy and water regime monitoring systems primarily rely on wired or short-range wireless communication technologies, such as fiber optic communications and GPRS. However, these communication methods have numerous limitations in remote areas or complex terrain. Wired communications require laying large amounts of cables, which is costly and susceptible to damage from natural disasters. Wireless communications, such as GPRS, are unusable in areas without signal coverage, and base stations can be damaged during extreme weather or natural disasters, leading to communication interruptions. Furthermore, traditional monitoring systems lack the real-time, reliable, and bandwidth-efficient data transmission, making them difficult to meet the demand for rapid and accurate access to water regime information.

[0003] Existing water conservancy and water regime monitoring systems also have some problems with data processing and early warning. On the one hand, monitoring data is often affected by noise, data quality is uneven, and data loss or errors are prone to occur during the data transmission process. Traditional data processing methods have difficulty effectively removing noise and filling missing data, resulting in inaccurate monitoring results. On the other hand, the early warning mechanisms of existing systems are mostly based on fixed thresholds and cannot be dynamically adjusted according to real-time environmental changes and historical data. The accuracy and timeliness of early warnings are insufficient. In addition, when faced with sudden water conditions, the transmission and push methods of early warning information are relatively simple, which cannot ensure that information is conveyed to relevant personnel in a timely and accurate manner, affecting the efficiency of emergency response.

[0004] In order to solve the above-mentioned defects, a technical solution is now provided. Summary of the Invention

[0005] The purpose of the present invention is to solve the problems that the current water conservancy and water regime monitoring system based on Beidou short message communication still needs to be further improved in terms of data transmission optimization, compression and error correction, anomaly detection and early warning, and propose a water conservancy and water regime monitoring and management system based on Beidou short message communication.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A water conservancy and water regime monitoring and management system based on Beidou short message communication, comprising:

[0008] The scheduling module is used to dynamically calculate the comprehensive score of nodes, optimize the transmission queue, and support preemption and resumption of transmission to ensure the real-time transmission of high-priority data;

[0009] The data compression module is used to perform differential compression based on data type, dynamically adjust the compression ratio and fault tolerance redundancy based on channel quality, and improve bandwidth utilization and reliability;

[0010] Data preprocessing module, used for filtering and denoising, spatiotemporal anomaly detection, and missing value interpolation to ensure data quality and consistency;

[0011] The early warning module is used to integrate multi-source data to predict trends, dynamically adjust thresholds and trigger early warnings in a graded manner, and accurately push Beidou short messages.

[0012] Furthermore, the execution process of the scheduling module is as follows:

[0013] When a monitoring node generates a data packet, it marks the current water level status, including normal, warning, and over-limit, and calculates the data urgency score. The over-limit status score is 10, the warning score is 5, and the normal score is 1;

[0014] The node calculates its own historical transmission frequency score. If the number of transmissions in the past hour is greater than 3 times, the score is 0.5; otherwise, it is 1.

[0015] Introduced modified parameters to optimize the rationality of transmission queues;

[0016] The silent period is set from 0:00 to 6:00 every day, allowing only over-limit nodes to transmit, reducing channel occupancy.

[0017] Furthermore, the specific operation steps of introducing the modified parameters to optimize the transmission queue in the scheduling module are as follows:

[0018] Correction parameters include:

[0019] Node remaining power E: The monitoring node reports the remaining power percentage in real time;

[0020] Geographical risk level coefficient G: Predefined node geographical risk level:

[0021] For high-risk areas, G=1.5; for medium-risk areas, G=1.2; and for low-risk areas, G=1.0;

[0022] Data timeliness attenuation factor T: Calculate the difference between the data generation time and the current time, and define the attenuation factor formula: , is the difference between the data generation time and the current time, if >50 minutes, T=0.5; otherwise linear decay;

[0023] Comprehensive score = urgency score × frequency score × E × G × T. The comprehensive scores are sorted from high to low to generate a transmission queue;

[0024] The BeiDou dispatch center allocates transmission time slots to the first N nodes in the queue, where N = channel capacity / single packet length, and the remaining nodes enter the waiting pool.

[0025] If the water level of a node is upgraded, the preemption mechanism is triggered immediately, interrupting the transmission of low-priority nodes. Specifically:

[0026] When preemption is triggered, the node detects whether more than 50% of the data packet currently being transmitted has been sent. If more than 50% has been completed, the node is allowed to complete the current packet transmission. If less than 50%, the transmission is immediately terminated and the breakpoint position is recorded.

[0027] Add a breakpoint identifier to the interrupted node and store it at the head of the waiting pool queue. In the next scheduling cycle, the time slot is allocated to the interrupted node first, and the remaining data is continued from the breakpoint position.

[0028] Furthermore, the specific steps for adjusting the node water level status according to the dynamic threshold are as follows:

[0029] When the node status is upgraded, an additional check is performed to see if the water level change rate exceeds the threshold for 3 minutes. If so, preemption is considered valid. If not, preemption is not triggered and the node is marked as pending.

[0030] Real-time monitoring of current channel occupancy. When the load is high, only the upgrade from over-limit to critical status will trigger preemption; when the load is low, the upgrade from normal to warning will trigger preemption.

[0031] When the status of a node is upgraded, the dispatch center sends a verification request to the three adjacent nodes. If at least two adjacent nodes report similar water level change trends, the preemption is confirmed to be valid. Before triggering the preemption, the node needs to upload the original water level data of the last 10 minutes, and the dispatch center verifies the authenticity of the data through anomaly detection algorithm.

[0032] Furthermore, the specific operation steps of the data compression module are as follows:

[0033] Data classification: classify sensor data into water level, flow rate, and rainfall;

[0034] Select a compression algorithm:

[0035] Water level data: using differential encoding, only the difference between the current value and the previous moment is stored;

[0036] Velocity data: Apply wavelet transform compression to retain frequency components;

[0037] Rainfall data: Huffman coding is used to perform entropy compression on the accumulated values;

[0038] Dynamically adjust the compression rate: monitor the Beidou channel signal strength RSSI. If RSSI < -100dBm, enable aggressive compression mode, which means the compression rate is ≥ 70%; if RSSI ≥ -90dBm, enable lossless compression mode, which means the compression rate is ≤ 30%;

[0039] Encapsulate the compressed data packet, add data type identifier and decompression key; and add a fault tolerance mechanism before encapsulating the compressed packet.

[0040] Furthermore, the specific operation steps of the data compression module to add a fault tolerance mechanism before encapsulating the compressed package are as follows:

[0041] Forward Error Correction (FEC) integration: Reed-Solomon codes are used, with a set error correction capability to recover 5% of lost or erroneous bytes in a data packet. The original data is segmented into fixed-length blocks before being encapsulated and compressed. Reed-Solomon error correction code is added to each block of data, with a redundancy ratio of 10%. The Beidou channel bit error rate (BER) is monitored. If the BER is greater than 1e-3, the redundancy ratio is increased to 15%; if the BER is ≤ 1e-4, the redundancy ratio is reduced to 5%.

[0042] Block checksum and marking: The compressed data packet is divided into 512-byte blocks. A CRC-32 checksum is calculated for each block of data and appended to the end of the block. A unique identifier is added to each block in the format of packet number + block number. The data block structure is: [Block ID (2B) | Data (512B) | CRC-32 (4B) | Reed-Solomon redundancy (proportional)]; Block ID: The unique identifier of the data block in the format of packet number + block number; Data: The compressed data block; CRC-32 is a 32-bit cyclic redundancy check code used to verify data integrity; The redundant data of the Reed-Solomon error correction code of Reed-Solomon is dynamically adjusted according to the bit error rate.

[0043] Retransmission strategy: The receiving end decapsulates each block and first verifies the CRC-32 checksum. If the check passes, the data is extracted and a confirmation signal is fed back. If the check fails, the failed block ID is recorded and a negative signal is fed back. After receiving the negative signal, the dispatch center retransmits only the specific failed block, rather than the entire data packet. The retransmitted data is marked with a priority tag to ensure that it is sent first in the next transmission cycle. The maximum number of retransmissions for a single block is 2. If the retransmission still fails, the error correction code recovery mechanism is activated and the data is repaired using the RS code.

[0044] Furthermore, the specific operation steps of the data preprocessing module are as follows:

[0045] Perform sliding window mean filtering on the raw sensor data;

[0046] Detect abnormal values: If the water level change rate at a certain moment is greater than 10cm / min, it is marked as suspicious data and redundant sensor verification is initiated;

[0047] Filling missing data: Use the ARIMA model to predict the value of the missing period and cross-validate with the adjacent node data;

[0048] Introduce spatiotemporal correlation analysis in anomaly detection and missing value filling.

[0049] Furthermore, the specific operation steps of the data preprocessing module for introducing spatiotemporal correlation analysis in anomaly detection and missing value filling are as follows:

[0050] Modeling of spatiotemporal correlation features: Generate an adjacency matrix based on the geographic location of the monitoring nodes and define the distance threshold; maintain a dynamic neighbor list for each node and update it in real time;

[0051] Calculate the water level change rate of the current node and its change trend in the past hour, synchronously obtain water level data of adjacent nodes, and calculate the spatial gradient;

[0052] Multi-condition fusion judgment for anomaly detection: If the water level change rate of a single node is greater than 10 cm / min, but the spatial gradient of adjacent nodes is less than 2 cm, it is determined to be a local anomaly and triggers redundancy check. If the water level change rate of the current node and more than 50% of adjacent nodes is greater than 8 cm / min, it is determined to be a regional anomaly event, directly marked as a valid anomaly, and an early warning is initiated.

[0053] Define the scoring calculation logic: ,in is the spatiotemporal confidence score, which is used to determine whether it is a real anomaly. The time change rate is the water level change rate of the current node, the threshold is the preset anomaly threshold, the spatial gradient is the average water level difference between the current node and the adjacent nodes, and the baseline value is the reference value of the spatial gradient under normal conditions. , When it is determined to be abnormal, the threshold is dynamically relaxed or tightened according to the score;

[0054] Missing data filling with spatiotemporal coordination: When a node's data is missing, the optimal interpolation weight is calculated based on the spatiotemporal covariance model, combining historical data with real-time data from adjacent nodes; interpolation calculation logic: ,in is the spatial weight, is the time decay factor, is the interpolated water level of the current node at time t, is the water level value of the i-th adjacent node at time t;

[0055] Training the GNN model: Using node location, historical water level, and adjacent node data as input, predict missing values at the current moment; Online inference: If the missing duration is greater than 30 minutes, call GNN to generate predicted values, and weightedly fuse them with the ARIMA results. The weights are dynamically adjusted based on the prediction error.

[0056] Furthermore, the specific operation steps of the early warning module are as follows:

[0057] Receive the decompressed data and input it into the LSTM neural network to predict the water level trend in the next 2 hours;

[0058] If the predicted water level exceeds the threshold, an early warning message is generated and the impact range is calculated;

[0059] The thresholds are dynamically adjusted based on real-time environmental data and historical patterns;

[0060] When the warning is triggered, it is triggered by progressive measurement, and the process is as follows:

[0061] Warning levels and triggering conditions:

[0062] Observation level (blue): triggered when the predicted water level reaches 80% of the dynamic threshold; village inspectors are notified to strengthen monitoring. Beidou short message content: BDS-GEO: longitude, latitude, BLUE; monitoring recommendations;

[0063] Preparation level (yellow): triggered when the predicted water level reaches 95% of the dynamic threshold or the water level at the upstream node exceeds the threshold; township emergency material preparation is initiated, message content: BDS-GEO: longitude, latitude, YELLOW; material inspection;

[0064] Action level (red): triggered when the predicted water level exceeds the dynamic threshold or the river storage capacity is less than 10%; the city-level command center is linked to evacuate personnel. The message content includes: BDS-GEO: longitude, latitude, RED; evacuation instructions; village list;

[0065] Automatic upgrade and downgrade mechanism: If the water level is predicted to rise continuously for three consecutive times, the warning level will be automatically upgraded; if the water level falls below the threshold and remains stable for 30 minutes, the warning level will be automatically downgraded or lifted;

[0066] Beidou short message intelligent compression and multi-channel distribution: using key-value abbreviations and predefined codes: Latitude and longitude: LAT=31.23, LON=121.47 → L=3123,12147; Village list: pre-assigned numbers, message content is VL=01,02; Command type: Evacuation command → CMD=EVAC;

[0067] Multi-channel redundant transmission: When a red alert is issued, the alert is sent simultaneously via Beidou short messages, backup radio frequency bands, and satellite phones; receiving-end confirmation mechanism: the responsible person must reply with a confirmation code within 5 minutes, otherwise automatic retransmission will be triggered.

[0068] Furthermore, the process of dynamically adjusting the threshold in the warning module based on real-time environmental data and historical patterns is as follows:

[0069] Dynamic threshold generation model: Multi-factor weight calculation, including: real-time rainfall intensity: receiving the current regional rainfall through Beidou short messages, with a weight of 30%; upstream water level change rate: obtaining upstream node water level data and calculating the average change rate, with a weight of 25%; historical water level peak: retrieving the maximum water level in the same period of the past five years, with a weight of 20%; river water storage capacity: calculating the current river remaining water storage space based on DEM data, with a weight of 25%;

[0070] Dynamic threshold calculation logic: ,in is the basic threshold, , For real-time rainfall, The highest rainfall in history. is the water level change rate of the upstream node, is the critical rate of change.

[0071] Compared with the prior art, the present invention has the following beneficial effects:

[0072] (1) The present invention, through the optimized design of the scheduling module, realizes the dynamic comprehensive scoring of monitoring nodes and the reasonable optimization of transmission queues, and supports the preemptive retransmission mechanism. This mechanism can ensure the real-time transmission of high-priority data, especially in emergency situations, such as when the water level exceeds the limit, and can give priority to the allocation of transmission resources to ensure the rapid transmission of critical data. At the same time, the data compression module adopts differentiated compression algorithms according to different data types, and dynamically adjusts the compression rate and fault-tolerant redundancy in combination with the channel quality, effectively improving the bandwidth utilization and data transmission reliability. Through forward error correction coding and retransmission strategy, the anti-interference ability and integrity of data in complex communication environments are further enhanced, and the risk of data loss and errors is reduced;

[0073] (2) In the present invention, the data preprocessing module significantly improves the quality and consistency of monitoring data through technical means such as filtering denoising, spatiotemporal correlation anomaly detection, and missing value interpolation. Sliding window mean filtering can effectively remove noise interference in sensor data, while spatiotemporal correlation analysis can combine data from multiple nodes for anomaly detection, avoiding misjudgment due to single-point failure or local anomalies. In addition, the use of ARIMA model and graph neural network (GNN) to collaboratively fill in missing data can more accurately predict and fill in missing values, further ensuring the integrity and accuracy of the data. The application of these technologies enables the monitoring system to provide higher-quality data support, providing a reliable basis for subsequent water situation analysis and decision-making;

[0074] (3) In this invention, the early warning module integrates multi-source data, predicts water level trends through the LSTM neural network, and dynamically adjusts the early warning threshold according to real-time environmental data and historical patterns, thus achieving accurate graded early warning. This dynamic adjustment mechanism can timely adjust the early warning level according to actual water conditions, avoiding false alarms or missed alarms caused by fixed thresholds. At the same time, early warning information is accurately pushed through Beidou short messages, and intelligent compression and multi-channel distribution technology are used to ensure that the early warning information can be accurately delivered to relevant personnel in the first place. In addition, the automatic upgrade and downgrade mechanism and the receiving end confirmation mechanism further enhance the intelligence level and emergency response capability of the early warning system, effectively ensuring the timeliness and accuracy of water conservancy and water conditions monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] In order to facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings;

[0076] Figure 1 This is the overall system block diagram of the present invention. DETAILED DESCRIPTION

[0077] The following will clearly and completely describe the technical solutions of the present invention in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0078] It should be understood that the terms “include” and “comprising” used in the specification and claims of the present disclosure indicate the presence of described features, integers, steps, operations, elements and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0079] It should also be understood that the terminology used in this disclosure is for the purpose of describing specific embodiments only and is not intended to limit the disclosure. As used in this disclosure and the claims, the singular forms "a," "an," and "the" are intended to include the plural forms unless the context clearly indicates otherwise. It should be further understood that the term "and / or" as used in this disclosure and the claims refers to any and all possible combinations of one or more of the associated listed items, including and including these combinations.

[0080] like Figure 1 As shown, a water conservancy and water regime monitoring and management system based on Beidou short message communication includes a scheduling module, a data compression module, a data preprocessing module and an early warning module;

[0081] The scheduling module dynamically calculates the node's comprehensive score (urgency, power consumption, geographical risk, etc.), optimizes the transmission queue, and supports preemption and resumption of transmission to ensure real-time transmission of high-priority data;

[0082] When a monitoring node generates a data packet, it marks the current water level status (normal / warning / overrun) and calculates the data urgency score (overrun status score = 10, warning = 5, normal = 1). The node also calculates its own historical transmission frequency score (if the number of transmissions in the past hour is > 3 times, score = 0.5; otherwise = 1).

[0083] Modification parameters are introduced to optimize the rationality of the transmission queue. The modification parameters include:

[0084] Node remaining power E: The percentage of remaining power reported by the monitoring node in real time (e.g., 80% is recorded as 0.8); Geographical risk level coefficient G: Predefined node geographical risk level: high-risk area (G=1.5), medium-risk area (G=1.2), low-risk area (G=1.0); Data timeliness attenuation factor T: Calculate the difference between the data generation time and the current time, and define the attenuation factor formula: , is the difference between the data generation time and the current time, if >50 minutes, T=0.5; otherwise linear decay;

[0085] Comprehensive score = urgency score × frequency score × E × G × T. The comprehensive scores are sorted from high to low to generate a transmission queue. The Beidou dispatch center allocates transmission time slots to the first N nodes in the queue (N = channel capacity / single packet length), and the remaining nodes enter the waiting pool. If the water level status of a node escalates, the preemption mechanism is immediately triggered, interrupting the transmission of low-priority nodes. Specifically:

[0086] When preemption is triggered, the interrupted node is checked to see if it has sent more than 50% of the data packet it is currently transmitting. If so, it is allowed to complete the current packet transmission (the remaining portion continues after preemption). If less than 50%, the transmission is terminated immediately and the breakpoint position is recorded.

[0087] A breakpoint identifier (such as "Offset=256B") is added to the interrupted node and stored at the head of the waiting pool queue. In the next scheduling cycle, time slots are preferentially allocated to the interrupted node, and the remaining data is transmitted from the breakpoint. The node water level status is adjusted according to the dynamic threshold. The specific process is as follows:

[0088] When the node status is upgraded, an additional check is performed to see if the water level change rate exceeds a threshold (e.g., ≥5 cm / min) for three minutes. If so, preemption is considered valid. If not, preemption is not triggered and the node is marked as pending. The current channel occupancy rate is monitored in real time (e.g., 80% or higher indicates high load). Under high load conditions, preemption is triggered only when the status is upgraded from over-limit to critical. Under low load conditions, preemption is triggered when the status is upgraded from normal to warning.

[0089] When the status of a node is upgraded, the dispatch center sends a verification request to the three adjacent nodes. If at least two adjacent nodes report similar water level change trends, the preemption is confirmed to be effective. Before triggering the preemption, the node needs to upload the raw water level data (unfiltered) of the last 10 minutes. The dispatch center verifies the authenticity of the data through anomaly detection algorithms (such as isolation forest).

[0090] The "silent period" is set from 0:00 to 6:00 every day, allowing only over-limit nodes to transmit, reducing channel occupancy.

[0091] The data compression module performs differential compression (differential / wavelet / Huffman coding) based on data type, dynamically adjusts compression ratio and fault tolerance based on channel quality, and improves bandwidth utilization and reliability.

[0092] Data classification: sensor data is divided into water level (low-frequency slow change), flow velocity (high-frequency fluctuation), and rainfall (sudden increment). Compression algorithm selection: water level data: differential encoding (only the difference between the current value and the previous moment is stored); flow velocity data: wavelet transform compression is applied to retain the main frequency components; rainfall data: Huffman coding is used to perform entropy compression on the accumulated value.

[0093] Dynamically adjust the compression rate: monitor the Beidou channel signal strength (RSSI). If RSSI < -100dBm, enable aggressive compression mode (compression rate ≥ 70%); if RSSI ≥ -90dBm, enable lossless compression mode (compression rate ≤ 30%).

[0094] Encapsulate the compressed data packet, add the data type identifier and decompression key; add a fault tolerance mechanism before encapsulating the compressed packet. The specific process is as follows:

[0095] Forward Error Correction (FEC) integration: Using Reed-Solomon (RS) code, the error correction capability is set to recover 5% of lost or erroneous bytes in a data packet. Before encapsulating and compressing the data packet, the original data is segmented into fixed-length blocks (e.g., 240 bytes per block). RS error correction code is added to each block of data, with a redundancy ratio of 10% (i.e., 24 bytes of error correction code are added for every 240 bytes of data, for a total block length of 264 bytes). The Beidou channel bit error rate (BER) is monitored. If the BER is greater than 1e-3 (a high-error environment), the redundancy ratio is increased to 15%; if the BER is ≤ 1e-4 (a low-error environment), the redundancy ratio is reduced to 5%.

[0096] Block checksum and marking: The compressed data packet is divided into 512-byte blocks (for example, if the compressed data is 1024 bytes, it is divided into 2 blocks). A CRC-32 checksum (4 bytes) is calculated for each block of data and appended to the end of the block. A unique identifier (Block ID, 2 bytes) is added to each block in the format of "packet number + block number" (for example, 0x0001 represents the first block of the first packet). The data block structure is: [Block ID (2 bytes) | Data (512 bytes) | CRC-32 (4 bytes) | Reed-Solomon redundancy (proportional)]; Block ID: Data block unique identifier (2 bytes) in the format of "packet number + block number"; Data: Compressed data block (512 bytes); CRC-32: 32-bit cyclic redundancy check code (4 bytes) used to verify data integrity; Reed-Solomon redundancy: Redundant data of the Reed-Solomon error correction code, the proportion of which is dynamically adjusted based on the bit error rate.

[0097] Retransmission strategy: The receiving end decapsulates each block, first verifying the CRC-32 checksum. If the check passes, the data is extracted and an acknowledgment signal (ACK) is fed back. If the check fails, the failed block ID is recorded and a negative acknowledgement signal (NACK) is fed back. Upon receiving the NACK, the dispatch center retransmits only the specific failed block (e.g., Block ID = 0x0003), rather than the entire data packet. The retransmitted data is marked with a priority tag to ensure that it is sent first in the next transmission cycle. The maximum number of retransmissions for a single block is two. If the retransmission still fails, the error correction code recovery mechanism is activated, using the RS code to repair the data.

[0098] The data preprocessing module is used for filtering and denoising, spatiotemporal correlation anomaly detection, and missing value interpolation (ARIMA / GNN collaboration) to ensure data quality and consistency;

[0099] The raw sensor data is filtered using a sliding window mean filter (window size = 5 minutes). Outliers are detected: if the water level change rate at a certain moment is >10 cm / min, it is marked as suspicious data and redundant sensor verification is initiated. Missing data is filled by using the ARIMA model to predict the value of the missing period and cross-validate it with adjacent node data. Spatiotemporal correlation analysis is introduced in anomaly detection and missing value filling. The specific process is as follows:

[0100] Modeling spatiotemporal correlation features: Based on the geographic location (latitude and longitude) of the monitoring node, an adjacency matrix is generated, and distance thresholds are defined (e.g., nodes within 5 kilometers are considered neighbors). A dynamic neighbor list is maintained for each node and updated in real time (e.g., nodes are removed when they fail due to flooding). The water level change rate (e.g., ΔH / Δt) of the current node and its trend (first-order derivative) over the past hour are calculated. Water level data of adjacent nodes are simultaneously obtained, and spatial gradients (e.g., the mean difference between the water levels of the current node and its neighbors) are calculated.

[0101] Multi-condition fusion judgment for anomaly detection: If the water level change rate of a single node is greater than 10 cm / min, but the spatial gradient of adjacent nodes is less than 2 cm (i.e., high spatial consistency), it is determined to be a local anomaly (possibly a sensor failure), triggering redundancy check. If the water level change rate of the current node and more than 50% of adjacent nodes is greater than 8 cm / min, it is determined to be a regional anomaly event (such as a flood peak caused by heavy rain), directly marked as a valid anomaly, and an early warning is initiated.

[0102] Define the scoring calculation logic: ,in is the spatiotemporal confidence score used to determine whether it is a true anomaly. The time change rate is the water level change rate of the current node (unit: cm / minute), the threshold is the preset anomaly threshold, the spatial gradient is the average water level difference between the current node and the adjacent nodes, and the baseline value is the reference value of the spatial gradient under normal conditions. , When it is judged as abnormal, the threshold is dynamically relaxed or tightened according to the score (such as When the threshold value of the rate of change is lowered to 8 cm / min).

[0103] Missing data filling with spatiotemporal coordination: When a node's data is missing, the optimal interpolation weight is calculated based on the spatiotemporal covariance model, combining historical data with real-time data from adjacent nodes; interpolation calculation logic: ,in is the spatial weight, is the time decay factor, is the interpolated water level of the current node at time t, is the water level value of the i-th adjacent node at time t; train the GNN model: use the node location, historical water level, and adjacent node data as input to predict the missing value at the current moment; online inference: if the missing time is > 30 minutes, call GNN to generate the predicted value and weightedly fuse it with the ARIMA result (the weight is dynamically adjusted according to the prediction error).

[0104] The early warning module integrates multi-source data (water level / rainfall / topography) to predict trends, dynamically adjusts thresholds, and triggers warnings in different levels (blue / yellow / red), accurately pushing Beidou short messages;

[0105] The decompressed data is received and fed into an LSTM neural network to predict the water level trend over the next two hours. If the predicted water level exceeds a threshold, an alert is generated and the impact range (villages within 10 kilometers downstream) is calculated. The threshold is dynamically adjusted based on real-time environmental data and historical patterns. The process is as follows:

[0106] Dynamic threshold generation model: Multi-factor weight calculation, including: real-time rainfall intensity: receiving the current regional rainfall (unit: mm / h) through Beidou short messages, with a weight of 30%; upstream water level change rate: obtaining upstream node water level data and calculating the average change rate (ΔH / Δt), with a weight of 25%; historical water level peak: retrieving the maximum water level in the same period of the past five years, with a weight of 20%; river water storage capacity: calculating the current river remaining water storage space based on DEM data, with a weight of 25%;

[0107] Dynamic threshold calculation logic: ,in is the basic threshold (set by the water conservancy department), , For real-time rainfall, The highest rainfall in history. is the water level change rate of the upstream node, is the critical rate of change.

[0108] When the warning is triggered, it is triggered by progressive measurement, and the process is as follows:

[0109] Warning levels and trigger conditions: Observation level (blue): Triggered when the predicted water level reaches 80% of the dynamic threshold; notify village inspectors to strengthen monitoring, Beidou short message content: BDS-GEO: longitude, latitude, BLUE; monitoring recommendations; Preparation level (yellow): Triggered when the predicted water level reaches 95% of the dynamic threshold or the upstream node water level exceeds the threshold; initiate township emergency material preparation, message content: BDS-GEO: longitude, latitude, YELLOW; material inspection; Action level (red): Triggered when the predicted water level exceeds the dynamic threshold or the river water storage capacity is <10%; coordinate with the municipal command center to evacuate personnel, message content: BDS-GEO: longitude, latitude, RED; evacuation instructions; village list.

[0110] Automatic upgrade and downgrade mechanism: If the water level is predicted to rise continuously for three consecutive times, the warning level is automatically upgraded (e.g., yellow to red). If the water level falls below the threshold and remains stable for 30 minutes, the warning is automatically downgraded or lifted. Beidou short message intelligent compression and multi-channel distribution: Using key-value abbreviations and predefined encoding: Latitude and longitude: LAT=31.23, LON=121.47 → L=3123,12147; Village list: pre-assigned numbers (e.g., Village A=01, Village B=02), message content: VL=01,02; Command type: Evacuation command → CMD=EVAC; Final message example: BDS-GEO: L=3123,12147; CLS=RED; CMD=EVAC; VL=01,02 (40% length compression);

[0111] Multi-channel redundant transmission: When a red alert is issued, the alert is sent simultaneously via Beidou short messages, backup radio frequency bands, and satellite phones; receiving-end confirmation mechanism: the responsible person must reply with a confirmation code (such as ACK=RED123) within 5 minutes, otherwise automatic retransmission will be triggered.

[0112] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A water conservancy and water regime monitoring and management system based on Beidou short message communication, characterized in that: include: The scheduling module is used to dynamically calculate the comprehensive score of nodes, optimize the transmission queue, and support preemption and resumption of transmission to ensure the real-time transmission of high-priority data; The data compression module is used to perform differential compression based on data type, dynamically adjust the compression ratio and fault tolerance redundancy based on channel quality, and improve bandwidth utilization and reliability; Data preprocessing module, used for filtering and denoising, spatiotemporal anomaly detection, and missing value interpolation to ensure data quality and consistency; The early warning module is used to integrate multi-source data to predict trends, dynamically adjust thresholds, trigger early warnings in different levels, and accurately push Beidou short messages; The execution process of the scheduling module is as follows: When a monitoring node generates a data packet, it marks the current water level status, including normal, warning, and over-limit, and calculates the data urgency score. The over-limit status score is 10, the warning score is 5, and the normal score is 1; The node calculates its own historical transmission frequency score. If the number of transmissions in the past hour is greater than 3 times, the score is 0.5; otherwise, it is 1. Introduced modified parameters to optimize the rationality of transmission queues; A silent period is set from 0:00 to 6:00 every day, allowing only over-limit nodes to transmit, reducing channel occupancy; The specific operation steps of introducing the modified parameters to optimize the transmission queue in the scheduling module are as follows: Correction parameters include: Node remaining power E: The monitoring node reports the remaining power percentage in real time; Geographical risk level coefficient G: Predefined node geographical risk level: For high-risk areas, G=1.5; for medium-risk areas, G=1.2; and for low-risk areas, G=1.0; Data timeliness attenuation factor T: Calculate the difference between the data generation time and the current time, and define the attenuation factor formula: , is the difference between the data generation time and the current time, if >50 minutes, T=0.5; otherwise linear decay; Comprehensive score = urgency score × frequency score × E × G × T. The comprehensive scores are sorted from high to low to generate a transmission queue; The BeiDou dispatch center allocates transmission time slots to the first N nodes in the queue, where N = channel capacity / single packet length, and the remaining nodes enter the waiting pool. If the water level of a node is upgraded, the preemption mechanism is triggered immediately, interrupting the transmission of low-priority nodes. Specifically: When preemption is triggered, the node detects whether more than 50% of the data packet currently being transmitted has been sent. If more than 50% has been completed, the node is allowed to complete the current packet transmission. If less than 50%, the transmission is immediately terminated and the breakpoint position is recorded. Add a breakpoint identifier to the interrupted node and store it at the head of the waiting pool queue. In the next scheduling cycle, the time slot is allocated to the interrupted node first, and the remaining data is continued from the breakpoint position.

2. A water conservancy and water regime monitoring and management system based on Beidou short message communication according to claim 1, characterized in that: The specific steps for adjusting the node water level status according to the dynamic threshold are as follows: When the node status is upgraded, an additional check is performed to see if the water level change rate exceeds the threshold for 3 minutes. If so, preemption is considered valid. If not, preemption is not triggered and the node is marked as pending. Real-time monitoring of current channel occupancy. When the load is high, only the upgrade from over-limit to critical status will trigger preemption; when the load is low, the upgrade from normal to warning will trigger preemption. When the status of a node is upgraded, the dispatch center sends a verification request to the three adjacent nodes. If at least two adjacent nodes report similar water level change trends, the preemption is confirmed to be valid. Before triggering the preemption, the node needs to upload the original water level data of the last 10 minutes, and the dispatch center verifies the authenticity of the data through anomaly detection algorithm.

3. The water conservancy and water regime monitoring and management system based on Beidou short message communication according to claim 1 is characterized in that: The specific operation steps of the data compression module are as follows: Data classification: classify sensor data into water level, flow rate, and rainfall; Select a compression algorithm: Water level data: using differential encoding, only the difference between the current value and the previous moment is stored; Velocity data: Apply wavelet transform compression to retain frequency components; Rainfall data: Huffman coding is used to perform entropy compression on the accumulated values; Dynamically adjust the compression rate: monitor the Beidou channel signal strength RSSI. If RSSI < -100dBm, enable aggressive compression mode, which means the compression rate is ≥ 70%; if RSSI ≥ -90dBm, enable lossless compression mode, which means the compression rate is ≤ 30%; Encapsulate the compressed data packet, add data type identifier and decompression key; and add a fault tolerance mechanism before encapsulating the compressed packet.

4. A water conservancy and water regime monitoring and management system based on Beidou short message communication according to claim 3, characterized in that: The specific operation steps of the data compression module to add a fault tolerance mechanism before encapsulating the compressed packet are as follows: Forward Error Correction (FEC) integration: Reed-Solomon codes are used, with a set error correction capability to recover 5% of lost or erroneous bytes in a data packet. The original data is segmented into fixed-length blocks before being encapsulated and compressed. Reed-Solomon error correction code is added to each block of data, with a redundancy ratio of 10%. The Beidou channel bit error rate (BER) is monitored. If the BER is greater than 1e-3, the redundancy ratio is increased to 15%; if the BER is ≤ 1e-4, the redundancy ratio is reduced to 5%. Block checksum and marking: The compressed data packet is divided into 512-byte blocks. A CRC-32 checksum is calculated for each block of data and appended to the end of the block. A unique identifier is added to each block in the format of packet number + block number. The data block structure is: [BlockID (2B) | Data (512B) | CRC-32 (4B) | RS redundancy (proportional)]; Block ID: unique identifier of data block, the format is packet number + block number; data is the compressed data block; CRC-32 is a 32-bit cyclic redundancy check code used to verify data integrity; RS redundant Reed-Solomon error correction code redundant data, the proportion of which is dynamically adjusted according to the bit error rate; Retransmission strategy: The receiving end decapsulates each block and first verifies the CRC-32 checksum. If the check passes, the data is extracted and a confirmation signal is fed back. If the check fails, the failed block ID is recorded and a negative signal is fed back. After receiving the negative signal, the dispatch center retransmits only the specific failed block, rather than the entire data packet. The retransmitted data is marked with a priority tag to ensure that it is sent first in the next transmission cycle. The maximum number of retransmissions for a single block is 2. If the retransmission still fails, the error correction code recovery mechanism is activated and the data is repaired using the RS code.

5. The water conservancy and water regime monitoring and management system based on Beidou short message communication according to claim 1 is characterized in that: The specific operation steps of the data preprocessing module are as follows: Perform sliding window mean filtering on the raw sensor data; Detect abnormal values: If the water level change rate at a certain moment is greater than 10cm / min, it is marked as suspicious data and redundant sensor verification is initiated; Filling missing data: Use the ARIMA model to predict the value of the missing period and cross-validate with the adjacent node data; Introduce spatiotemporal correlation analysis in anomaly detection and missing value filling.

6. A water conservancy and water regime monitoring and management system based on Beidou short message communication according to claim 5, characterized in that: The specific steps of introducing spatiotemporal correlation analysis in the data preprocessing module for anomaly detection and missing value filling are as follows: Modeling of spatiotemporal correlation features: Generate an adjacency matrix based on the geographic location of the monitoring nodes and define the distance threshold; maintain a dynamic neighbor list for each node and update it in real time; Calculate the water level change rate of the current node and its change trend in the past hour, synchronously obtain water level data of adjacent nodes, and calculate the spatial gradient; Multi-condition fusion judgment for anomaly detection: If the water level change rate of a single node is greater than 10 cm / min, but the spatial gradient of adjacent nodes is less than 2 cm, it is determined to be a local anomaly and triggers redundancy check. If the water level change rate of the current node and more than 50% of adjacent nodes is greater than 8 cm / min, it is determined to be a regional anomaly event, directly marked as a valid anomaly, and an early warning is initiated. Define the scoring calculation logic: ,in is the spatiotemporal confidence score, which is used to determine whether it is a real anomaly. The time change rate is the water level change rate of the current node, the threshold is the preset anomaly threshold, the spatial gradient is the average water level difference between the current node and the adjacent nodes, and the baseline value is the reference value of the spatial gradient under normal conditions. , When it is determined to be abnormal, the threshold is dynamically relaxed or tightened according to the score; Missing data filling with spatiotemporal coordination: When a node's data is missing, the optimal interpolation weight is calculated based on the spatiotemporal covariance model, combining historical data with real-time data from adjacent nodes; interpolation calculation logic: ,in is the spatial weight, is the time decay factor, is the interpolated water level of the current node at time t, is the water level value of the i-th adjacent node at time t; Training the GNN model: Using node location, historical water level, and adjacent node data as input, predict missing values at the current moment; Online inference: If the missing duration is greater than 30 minutes, call GNN to generate predicted values, and weightedly fuse them with the ARIMA results. The weights are dynamically adjusted based on the prediction error.

7. The water conservancy and water regime monitoring and management system based on Beidou short message communication according to claim 1 is characterized in that: The specific operation steps of the early warning module are as follows: Receive the decompressed data and input it into the LSTM neural network to predict the water level trend in the next 2 hours; If the predicted water level exceeds the threshold, an early warning message is generated and the impact range is calculated; The thresholds are dynamically adjusted based on real-time environmental data and historical patterns; When the warning is triggered, it is triggered by progressive measurement, and the process is as follows: Warning levels and triggering conditions: Observation level (blue): triggered when the predicted water level reaches 80% of the dynamic threshold; village inspectors are notified to strengthen monitoring. Beidou short message content: BDS-GEO: longitude, latitude, BLUE; monitoring recommendations; Preparation level (yellow): triggered when the predicted water level reaches 95% of the dynamic threshold or the water level at the upstream node exceeds the threshold; township emergency material preparation is initiated, message content: BDS-GEO: longitude, latitude, YELLOW; material inspection; Action level (red): triggered when the predicted water level exceeds the dynamic threshold or the river storage capacity is less than 10%; the city-level command center is linked to evacuate personnel. The message content includes: BDS-GEO: longitude, latitude, RED; evacuation instructions; village list; Automatic upgrade and downgrade mechanism: If the water level is predicted to rise continuously for three consecutive times, the warning level will be automatically upgraded; if the water level falls below the threshold and remains stable for 30 minutes, the warning level will be automatically downgraded or lifted; Beidou short message intelligent compression and multi-channel distribution: using key-value abbreviations and predefined codes: Latitude and longitude: LAT=31.23, LON=121.47 → L=3123,12147; Village list: pre-assigned numbers, message content is VL=01,02; Command type: Evacuation command → CMD=EVAC; Multi-channel redundant transmission: When a red alert is issued, the alert is sent simultaneously via Beidou short messages, backup radio frequency bands, and satellite phones; receiving-end confirmation mechanism: the responsible person must reply with a confirmation code within 5 minutes, otherwise automatic retransmission will be triggered.

8. A water conservancy and water regime monitoring and management system based on Beidou short message communication according to claim 7, characterized in that: The process of dynamically adjusting the threshold in the early warning module based on real-time environmental data and historical patterns is as follows: Dynamic threshold generation model: Multi-factor weight calculation, including: real-time rainfall intensity: receiving the current regional rainfall through Beidou short messages, with a weight of 30%; upstream water level change rate: obtaining upstream node water level data and calculating the average change rate, with a weight of 25%; historical water level peak: retrieving the maximum water level in the same period of the past five years, with a weight of 20%; river water storage capacity: calculating the current river remaining water storage space based on DEM data, with a weight of 25%; Dynamic threshold calculation logic: ,in is the basic threshold, , For real-time rainfall, The highest rainfall in history. is the water level change rate of the upstream node, is the critical rate of change.

Images