GIS dynamic environment monitoring and early warning system based on multi-source sensor network

Abstract

The application discloses a GIS dynamic environment monitoring and early warning system based on a multi-source sensing network and belongs to the technical field of satellite navigation and positioning. The system comprises a data acquisition module, a real-time access preset satellite positioning signal and a multi-source sensing network generated sensor environment data stream; a model generation module, which generates an environment event trigger mark according to abnormal parameters in the sensor environment data stream; an evolution analysis module, which divides the triggered environment event into four stages based on the positioning point coordinates in the environment event trigger mark and the associated abnormal parameters; an early warning module, which generates a first early warning result and a second early warning result if the triggered environment event is in a stage; and a closed-loop feedback module, which dynamically adjusts the early warning module according to the first early warning result and the second early warning result. The application realizes early warning of the migration track of pollutants by generating a predicted path coordinate set, effectively overcoming the hysteresis of the passive marking point of the traditional system in the peak stage.

Description

Technical Field

[0001] This invention relates to the field of satellite navigation and positioning technology, and in particular to a GIS dynamic environment monitoring and early warning system based on multi-source sensor networks. Background Technology

[0002] As a comprehensive technology platform integrating spatial positioning, environmental perception and data analysis, GIS (Geographic Information System) collects geographic coordinates in real time through satellite navigation and positioning technologies (such as Beidou / GPS) and integrates multi-source sensor networks to monitor changes in environmental parameters, providing dynamic decision support for scenarios such as disaster early warning and pollution source tracing.

[0003] The working principle of the existing GIS dynamic environmental monitoring and early warning system is as follows: the geographic coordinates of the target area are collected in real time through satellite navigation and positioning technology (such as Beidou / GPS), and environmental parameter data streams uploaded by multi-source sensor networks (such as temperature, humidity, water quality, and air quality monitoring equipment) are received simultaneously. The sensor data and positioning coordinates are statically correlated and superimposed on the GIS platform. When the environmental parameters at a certain point exceed the preset fixed threshold, an early warning sign is triggered. Manual intervention is used to analyze whether the parameter change is caused by a real environmental event or equipment failure. Finally, a static early warning sign is output on the GIS map.

[0004] The existing technology has the following drawbacks: First, it is slow to respond and cannot predict the spread path of environmental events (such as the migration trajectory of pollutants). It can only passively mark the locations where the events have occurred at the peak stage of the environmental event, thus missing the opportunity to intervene in the development stage. Second, the timing of downgrading the environmental event depends on human experience, which may lead to premature lifting of the warning or prolonging the ineffective response period.

[0005] Therefore, there is an urgent need to provide a GIS dynamic environmental monitoring and early warning system based on multi-source sensor networks to solve the above problems. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to overcome the following shortcomings of the prior art: first, the response is delayed and it is impossible to predict the diffusion path of environmental events (such as the migration trajectory of pollutants). It can only passively mark the points where the events have occurred at the peak stage, thus losing the opportunity to intervene in the development stage; second, the decline stage of environmental events relies on human experience to determine the timing of downgrading, which is prone to prematurely lifting the warning or prolonging the ineffective response period. The present invention provides a GIS dynamic environmental monitoring and early warning system based on multi-source sensor networks.

[0007] To solve the above-mentioned technical problems, one technical solution adopted by the present invention is: to provide a GIS dynamic environment monitoring and early warning system based on multi-source sensor network, including a data acquisition module, a model generation module, an evolution analysis module, an early warning module and a closed-loop feedback module;

[0008] The data acquisition module accesses the preset satellite positioning signal and the sensor environment data stream generated by the multi-source sensor network in real time, and synchronously records the positioning point coordinates of the satellite positioning signal;

[0009] The model generation module filters and associates the positioning point coordinates in the satellite positioning signal with the abnormal parameters in the sensor environmental data stream, and inputs the abnormal parameters and the positioning point coordinates into a preset spatiotemporal association model to generate environmental event trigger markers.

[0010] The evolution analysis module, based on the location point coordinates and associated abnormal parameters in the environmental event trigger markers, divides the triggered environmental events into the following four stages: the initial stage, the development stage, the peak stage, and the decline stage.

[0011] The early warning module generates a predicted path coordinate set based on the location point coordinates and the abnormal parameters if the triggered environmental event is in the initial or development stage, and uses it as the first early warning result; if the triggered environmental event is in the peak stage, the first early warning result is continued; if the triggered environmental event is in the receding stage, a downgrade marker is generated based on the decrease rate of the abnormal parameters and the density reduction trend of the location point coordinates of the satellite positioning signal, and used as the second early warning result.

[0012] The closed-loop feedback module dynamically adjusts the early warning module based on the first early warning result, the second early warning result, and the stage information of the current environmental event determined by the evolution analysis module.

[0013] The present invention is further configured such that: the satellite positioning signal in the data acquisition module originates from the BeiDou / GPS / Galileo satellite navigation system signal, and the satellite positioning signal includes the positioning point coordinates and a first timestamp;

[0014] The sensor environmental data stream is generated by a multi-source sensor network consisting of deployed air quality monitoring stations, water quality floating sensors, and temperature and humidity remote sensing terminals. After compression and encryption, it generates data packets for transmission back. The data packets embed the unique identifier of the acquisition device in the multi-source sensor network and a second timestamp, which is aligned with the first timestamp in the satellite positioning signal.

[0015] The present invention is further configured such that: the method for generating the spatiotemporal correlation model in the model generation module is as follows:

[0016] S1. Select the historical first environmental data collected by the air quality monitoring station, the historical second environmental data collected by the water quality float sensor, and the historical third environmental data collected by the temperature and humidity remote sensing terminal from the sensor environmental data stream as preliminary training samples, input them into the preset training model for training, and extract the training parameters in the training model to generate historical abnormal fluctuation patterns.

[0017] S2. Select the real-time first environmental data collected by the air quality monitoring station, the real-time second environmental data collected by the water quality floating sensor, and the real-time third environmental data collected by the temperature and humidity remote sensing terminal from the sensor environmental data stream as optimized training samples, and input them into the training model trained in step S1 to calculate the real-time dynamic deviation.

[0018] S3. Based on the matching degree between the real-time dynamic deviation and the historical abnormal fluctuation pattern, filter the positioning point coordinates and abnormal parameters corresponding to the real-time dynamic deviation with a matching degree higher than a preset threshold. Combine the training parameters in the historical abnormal fluctuation pattern, dynamically adjust the association weights of the preliminary training samples and the optimized training samples to the training model, and generate a spatiotemporal correlation model.

[0019] The present invention is further configured such that the generation steps of the historical abnormal fluctuation pattern in step S1 are as follows:

[0020] S11. Based on the historical first environment data, historical second environment data and historical third environment data in the preliminary training samples, calculate the average fluctuation amplitude and maximum deviation of the fluctuation amplitude within the preset historical period, and input them into the training model to establish the initial fluctuation baseline.

[0021] S12. Based on the fluctuation frequency of the historical first environment data, historical second environment data and historical third environment data in the preliminary training sample within the historical period, and combined with the initial fluctuation baseline, dynamically adjust the weights of the average fluctuation amplitude and the maximum deviation within the historical period to generate historical abnormal fluctuation patterns.

[0022] The present invention is further configured such that: the specific content of the stage division of the triggered environmental event in the evolution analysis module is as follows: if the abnormal parameter associated with the initial location point coordinates in the environmental event triggering marker exceeds the maximum deviation of the abnormal parameter predicted by the spatiotemporal correlation model based on the abnormal parameter for the first time, it is determined to be the initial stage; if the number of abnormal parameters in the environmental event triggering marker increases continuously and the number of location point coordinates increases, it is determined to be the development stage; if the fluctuation range of the abnormal parameter predicted by the spatiotemporal correlation model based on the abnormal parameter in the environmental event triggering marker narrows and the change in the predicted abnormal parameter dynamic deviation approaches the preset tolerance, it is determined to be the peak stage; if the number of abnormal parameters in the environmental event triggering marker decreases continuously and the density of location point coordinates decreases, it is determined to be the decline stage.

[0023] The present invention is further configured such that the generation steps of the first warning result in the warning module are as follows:

[0024] Q1. Extract the data subset of the positioning point coordinates and associated abnormal parameters that exceed the maximum deviation of the abnormal parameters in the initial or development stage, and construct the diffusion initial point set;

[0025] Q2. Calculate the cluster center migration vector of the coordinates of the positioning points in the initial diffusion point set, and generate spatial diffusion weights by associating the rate of increase of the number of abnormal parameters. Sort the coordinates of the positioning points based on the spatial diffusion weights to construct a spatial diffusion model.

[0026] Q3. Based on the spatial diffusion model, simulate the diffusion trajectory of the coordinates of the positioning points in the initial diffusion point set, and predict the set of positioning point coordinates in future periods to form a predicted path coordinate set.

[0027] Q4. Match the predicted path coordinate set with the geographic information in the preset GIS platform to generate a visualized early warning map as the first early warning result.

[0028] The present invention is further configured such that the step of generating the second early warning result in the early warning module is as follows:

[0029] Q101. Screen out the abnormal parameters and their associated location coordinates that continuously decrease in environmental events that are in the decline phase and conform to the historical abnormal fluctuation pattern, and construct a decline analysis point set;

[0030] Q102. Calculate the slope of the spatial density change of the coordinates of the location points in the set of fading analysis points, and generate a comprehensive fading factor by associating it with the rate of decrease of the corresponding abnormal parameters.

[0031] Q103. Dynamically classify the set of fading analysis points according to the comprehensive fading factor: when the comprehensive fading factor exceeds a set range, generate an accelerated degradation marker based on the slope of the spatial density change of the positioning point coordinates; when the comprehensive fading factor is within a set range, generate a gradual degradation marker based on the rate of decrease of the abnormal parameters; when the comprehensive fading factor is less than a set range, generate a stable degradation marker based on the density reduction trend of the positioning point coordinates; and downgrade the first warning result according to the accelerated degradation marker, the gradual degradation marker, and the stable degradation marker to generate a second warning result.

[0032] The present invention is further configured such that: the specific content of the dynamic adjustment of the early warning module in the closed-loop feedback module is as follows: the closed-loop feedback module receives the first early warning result, the second early warning result, and the stage information of the current environmental event determined by the evolution analysis module, and performs the following analysis: the predicted path coordinate set in the first early warning result within a preset first period is fused with the abnormal point diffusion trajectory in the satellite positioning signal associated with the actual abnormal parameters to generate a path offset rate; based on the degradation marker in the second early warning result within a preset second period, the actual abnormal parameter change slope of the sensor environmental data stream and the positioning point coordinate distribution density of the associated satellite positioning signal are compared to generate a prediction deviation; and the first early warning result and the second early warning result are dynamically adjusted according to the path offset rate or the prediction deviation.

[0033] The beneficial effects of this invention are as follows:

[0034] 1. This invention generates a predicted path coordinate set by constructing a set of initial diffusion points and a spatial diffusion model based on the initial and development stages, thereby enabling early warning of pollutant migration trajectories. It can identify diffusion trends and trigger intervention measures in the initial stage, effectively overcoming the lag of passively marking points in the peak stage of traditional systems and seizing the critical window period for emergency response.

[0035] 2. This invention generates a comprehensive decay factor by associating the slope of spatial density change during the decay phase with the rate of decrease of abnormal parameters. This factor drives the automatic generation and dynamic adjustment of accelerated / gradual / stable degradation markers, significantly reducing the probability of prematurely lifting warnings or prolonging ineffective responses due to human experience-based misjudgments, and optimizing the efficiency of emergency response resource scheduling. Attached Figure Description

[0036] Figure 1 This is a system flowchart of the present invention;

[0037] Figure 2 This is a flowchart of the method for generating the spatiotemporal correlation model of the present invention;

[0038] Figure 3 This is a flowchart of the steps for generating the first early warning result of the present invention. Detailed Implementation

[0039] The preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, so that the advantages and features of the present invention can be more easily understood by those skilled in the art, thereby providing a clearer and more explicit definition of the scope of protection of the present invention.

[0040] Please see Figure 1 - Figure 3A GIS dynamic environment monitoring and early warning system based on multi-source sensor networks includes a data acquisition module, a model generation module, an evolution analysis module, an early warning module, and a closed-loop feedback module.

[0041] The data acquisition module accesses the preset satellite positioning signal and the sensor environment data stream generated by the multi-source sensor network in real time, and synchronously records the positioning point coordinates of the satellite positioning signal. This system uses the satellite positioning signal as the core of dynamic analysis, and the GIS platform is only used for the spatial visualization of the final result.

[0042] The model generation module filters and associates the coordinates of positioning points in satellite positioning signals based on abnormal parameters in the sensor environmental data stream, inputs the abnormal parameters and positioning point coordinates into a preset spatiotemporal association model, and generates environmental event trigger markers.

[0043] The evolution analysis module, based on the location point coordinates and associated abnormal parameters in the environmental event trigger markers, divides the triggered environmental events into the following four stages: the initial stage, the development stage, the peak stage, and the decline stage.

[0044] The early warning module, if the triggered environmental event is in its initial or developing stage, generates a predicted path coordinate set based on the location point coordinates and abnormal parameters, and uses this as the first early warning result (the predicted path coordinate set generated by the early warning module can be visualized and overlaid through a GIS platform (e.g., overlaid onto an electronic map)); if the triggered environmental event is in its peak stage, the first early warning result is executed; if the triggered environmental event is in its receding stage, a downgrade marker is generated based on the rate of decrease of abnormal parameters and the density reduction trend of the location point coordinates of the satellite positioning signal, and used as the second early warning result. The first and second early warning results are then uploaded to a preset GIS platform to provide early warning of environmental events.

[0045] The closed-loop feedback module dynamically adjusts the early warning module based on the first early warning result, the second early warning result, and the stage information of the current environmental event determined by the evolution analysis module.

[0046] By dynamically fusing multi-source sensor data and satellite positioning signals, the system accurately identifies the full-cycle characteristics of environmental events from their onset to their decline, enabling early warning based on spatial trajectory prediction and intelligent degradation response during the decline phase. Furthermore, by combining the closed-loop feedback mechanism of warning results and event evolution, the system continuously optimizes the first and second warning results, significantly improving the timeliness of environmental monitoring, the accuracy of warnings, and the system's adaptive capabilities.

[0047] One embodiment of the present invention is as follows: the satellite positioning signal in the data acquisition module originates from the BeiDou / GPS / Galileo satellite navigation system signal after multi-band dynamic weighted anti-interference processing, and the satellite positioning signal includes the positioning point coordinates and the first timestamp;

[0048] The sensor environmental data stream is generated by a multi-source sensor network consisting of deployed air quality monitoring stations, water quality floating sensors, and temperature and humidity remote sensing terminals. After being compressed and encrypted by edge computing nodes, it generates data packets for transmission back. The data packets embed the unique identifier of the acquisition device in the multi-source sensor network and the second timestamp, which is aligned with the first timestamp in the satellite positioning signal.

[0049] The specific steps for edge computing node compression and encryption are as follows: First, the raw environmental data streams from air quality monitoring stations, water quality floating sensors, and temperature and humidity remote sensing terminals are preprocessed by data cleaning and normalization to remove outliers caused by equipment transient failures or communication interference, and the units of measurement are standardized to standard units of measurement. Then, feature extraction and data compression are performed. Key feature parameters (such as extreme values ​​of temperature and humidity fluctuations and changes in pollutant concentration gradients) in the sensor environmental data streams are extracted using a sliding window mechanism. A lossy-lossless hybrid compression strategy is adopted: for low-frequency steady-state parameters (such as extreme values ​​of temperature and humidity fluctuations), a data compression based on... The lossless compression using Huffman coding achieves lossy compression for high-frequency transient parameters (such as sudden peak increases in pollutant concentration) using discrete cosine transform. After data compression, encryption and encapsulation are used. A dynamic key is generated by combining the unique identifier of the acquisition device in the multi-source sensor network with the AES-256 algorithm to encrypt the compressed data block and embed a second timestamp. Finally, the encrypted data packet interacts with the timestamp alignment module to synchronize the first timestamp of the satellite positioning signal through the NTP protocol, ensuring that the timestamp in the data packet is strictly aligned with the satellite positioning signal, forming a retrievable encrypted data stream.

[0050] Specifically, the method for generating the spatiotemporal correlation model in the model generation module is as follows:

[0051] S1. Select the historical first environmental data collected by the air quality monitoring station, the historical second environmental data collected by the water quality floating sensor, and the historical third environmental data collected by the temperature and humidity remote sensing terminal from the sensor environmental data stream as preliminary training samples, input them into the preset training model for training, and extract the training parameters in the training model to generate historical abnormal fluctuation patterns.

[0052] The preset training model is a fusion architecture based on graph convolutional neural networks and temporal convolutional networks, which jointly analyzes the spatial distribution correlation and temporal evolution of multi-source sensor environmental data streams.

[0053] Specifically, the steps for generating the historical abnormal fluctuation pattern in step S1 are as follows:

[0054] S11. Based on the historical first environment data, historical second environment data and historical third environment data in the preliminary training samples, calculate the average fluctuation amplitude and maximum deviation of the fluctuation amplitude within the preset historical period, and input them into the training model to establish the initial fluctuation baseline.

[0055] The method for calculating the average fluctuation amplitude is as follows: Based on the historical first environmental data (air quality), historical second environmental data (water quality), and historical third environmental data (temperature and humidity) in the preliminary training samples, calculate the absolute fluctuation value (the absolute value of the difference between the value of a single sampling point and the previous point) for each sampling point of each type of environmental data within the preset historical period. Then, calculate the arithmetic mean of the absolute fluctuation values ​​of all sampling points according to the time series to generate the average fluctuation amplitude of this type of environmental data.

[0056] The maximum deviation is calculated as follows: For each type of environmental data in the initial training samples, extract the numerical sequence of all sampling points within a preset historical period. First, calculate the long-term trend line of the numerical sequence (e.g., by fitting it through linear regression). Then, calculate the vertical distance deviation of the numerical sequence of the sampling points to the long-term trend line point by point. Select the maximum value among all vertical distance deviations as the maximum deviation of this type of environmental data.

[0057] S12. Based on the fluctuation frequency of historical first environmental data, historical second environmental data and historical third environmental data in the initial training sample within the historical period, and combined with the initial fluctuation baseline, dynamically adjust the weights of the average fluctuation amplitude and the maximum deviation within the historical period to generate historical abnormal fluctuation patterns. Historical abnormal fluctuation patterns can reflect the abnormal fluctuation characteristics of environmental data at different time scales, and provide accurate reference for subsequent environmental event triggering.

[0058] S2. Select the real-time first environmental data collected by the air quality monitoring station, the real-time second environmental data collected by the water quality floating sensor, and the real-time third environmental data collected by the temperature and humidity remote sensing terminal from the sensor environmental data stream as optimization training samples, and input them into the training model trained in step S1 to calculate the real-time dynamic deviation.

[0059] The calculation method for real-time dynamic deviation is as follows: real-time first environmental data collected by air quality monitoring station, real-time second environmental data collected by water quality floating sensor, and real-time third environmental data collected by temperature and humidity remote sensing terminal are used as optimization training samples and input into the training model trained in step S1. The time reference is synchronized by the first timestamp of satellite positioning signal and spatial coordinates are associated with the unique identifier of the acquisition device. Based on the model, the spatiotemporal correlation features (such as pollutant concentration gradient threshold and periodic peak of temperature and humidity fluctuation) in the historical abnormal fluctuation pattern generated in step S12 are extracted. The distribution difference between the real-time abnormal parameters (filtered output of sensor environmental data stream) and the historical abnormal fluctuation pattern in the same spatiotemporal grid is compared. The offset and spatial propagation trend of the current environmental parameters relative to the maximum deviation of abnormal parameters in the stage division (quantification benchmark of historical fluctuation threshold) are quantified to generate real-time dynamic deviation.

[0060] S3. Based on the matching degree between real-time dynamic deviation and historical abnormal fluctuation patterns, filter the positioning point coordinates and abnormal parameters corresponding to real-time dynamic deviations with matching degrees higher than a preset threshold. Combine the training parameters in the historical abnormal fluctuation patterns, dynamically adjust the association weights of the initial training samples and optimized training samples to the training model, and generate a spatiotemporal correlation model.

[0061] By dynamically fusing multi-band satellite positioning signals with sensor environmental data streams in the spatiotemporal space, the deviations and spatial propagation trends of abnormal environmental parameters are quantified based on a spatiotemporal correlation model. Combined with historical abnormal fluctuation patterns, real-time dynamic deviation calculation and weight optimization are driven to achieve closed-loop coordination of full-cycle monitoring and early warning response for environmental events, significantly improving the accuracy and reliability of system analysis.

[0062] One embodiment of the present invention is as follows: the specific content of the stage division of triggered environmental events in the evolution analysis module is as follows: if the abnormal parameters associated with the initial location point coordinates in the environmental event triggering marker exceed the maximum deviation of the abnormal parameters predicted by the spatiotemporal correlation model based on the abnormal parameters for the first time, it is determined to be the initial stage; if the number of abnormal parameters in the environmental event triggering marker increases continuously and the number of location point coordinates increases, it is determined to be the development stage; if the fluctuation range of the abnormal parameters predicted by the spatiotemporal correlation model based on the abnormal parameters in the environmental event triggering marker narrows and the change in the dynamic deviation of the predicted abnormal parameters approaches the preset tolerance, it is determined to be the peak stage; if the number of abnormal parameters in the environmental event triggering marker continues to decrease and the density of location point coordinates decreases (which must conform to the historical abnormal fluctuation pattern), it is determined to be the decline stage.

[0063] By dynamically associating the spatiotemporal variation characteristics of abnormal parameters with the coordinates of the location points, and combining the deviation and fluctuation thresholds predicted by the model, the system accurately divides environmental events into four stages: onset, development, peak, and decline, significantly improving the ability to predict event evolution trends and the timeliness of emergency response.

[0064] One embodiment of the present invention is as follows: the steps for generating the first early warning result in the early warning module are as follows:

[0065] Q1. Extract the data subset of positioning point coordinates and associated abnormal parameters that exceed the maximum deviation of abnormal parameters in the initial or development stage, and construct the diffusion initial point set;

[0066] The method for constructing the initial diffusion point set is as follows: Extract the coordinates of localization points and the associated abnormal parameters that exceed the maximum deviation of the abnormal parameters in the initial or development stage. Construct the initial diffusion point set through the following steps: First, filter the set of localization point coordinates in the initial stage (first exceeding the maximum deviation of the abnormal parameters) or development stage (the number of abnormal parameters increases continuously and the number of localization point coordinates increases) in the stage division; Second, based on the real-time first environment data, second environment data, and third environment data in the optimized training samples of step S2, extract their associated abnormal parameters (defined by the model generation module), and filter the data points that exceed the maximum deviation of the abnormal parameters; Finally, integrate all localization point coordinates and their abnormal parameter values ​​that meet the conditions to form an initial diffusion point set containing spatiotemporal location and parameter offset, which serves as the input basis for the subsequent spatial diffusion model;

[0067] Q2. Calculate the cluster center migration vector of the coordinates of the location points in the initial diffusion point set, and generate spatial diffusion weights by associating the rate of increase of the number of abnormal parameters. Sort the coordinates of the location points based on the spatial diffusion weights and construct a spatial diffusion model.

[0068] The spatial diffusion model is constructed as follows: Based on the initial diffusion point set, the cluster center migration vector (spatial coordinate reference of satellite positioning signal) of its location point coordinates is calculated, and the core migration direction is identified by density clustering algorithm (existing technology); the rate of increase of the number of abnormal parameters in the associated development stage (the number of newly added abnormal parameters per unit time) is combined with the spatial propagation trend characteristics in the historical abnormal fluctuation pattern generated in step S12 to dynamically calculate the spatial diffusion weight (characterizing the influence intensity of different location points on the diffusion path); the location point coordinates in the initial diffusion point set are sorted based on the spatial diffusion weight, and the coordinates with high weights are selected as diffusion source points; the Gaussian plume model (defined in the diffusion simulation and prediction module) is used to simulate the pollutant migration path, and the diffusion direction is adjusted by the association weight optimization mechanism in step S3 to generate a spatial diffusion model that integrates spatiotemporal evolution law;

[0069] Q3. Based on the spatial diffusion model, simulate the diffusion trajectory of the coordinates of the positioning points in the initial diffusion point set, and predict the set of positioning point coordinates in future periods to form a predicted path coordinate set.

[0070] Q4. Match the predicted path coordinate set with the geographic information in the preset GIS platform to generate a visualized early warning map, which intuitively shows the spread trend of environmental events and potential impact areas, as the first early warning result.

[0071] Specifically, the steps for generating the second early warning result in the early warning module are as follows:

[0072] Q101. Screen out the abnormal parameters and their associated location coordinates that continuously decrease and conform to the historical abnormal fluctuation pattern in environmental events that are in the decline phase, and construct a set of decline analysis points;

[0073] Q102. Calculate the slope of the spatial density change of the coordinates of the localized points in the set of fading analysis points, and generate a comprehensive fading factor by associating it with the rate of decrease of the corresponding abnormal parameters.

[0074] Q103. Dynamically classify the set of fading analysis points based on the comprehensive fading factor: When the comprehensive fading factor exceeds the set range (0.5-0.8), an accelerated degradation marker is generated based on the slope of the spatial density change of the location point coordinates (the dominant exponential decay function); when the comprehensive fading factor is within the set range, a gradual degradation marker is generated based on the rate of decrease of the abnormal parameters (the dominant linear decay function); when the comprehensive fading factor is less than the set range, a stable degradation marker is generated based on the density reduction trend of the location point coordinates. Based on the accelerated degradation marker, the gradual degradation marker, and the stable degradation marker, the first warning result is downgraded to generate the second warning result.

[0075] By constructing an initial diffusion point set and a spatial diffusion model, the migration trajectory of pollutants is accurately predicted, and a visualized early warning map is generated to achieve the first early warning result. Combined with the dynamic grading and downgrading of comprehensive decline factors during the decline phase, a second early warning result is generated, forming a closed loop of dual mechanisms: quantitative early warning of diffusion status and dynamic downgrading of the decline process. This significantly improves the accuracy of environmental event early warning and the timeliness of emergency response.

[0076] One embodiment of the present invention is as follows: The specific content of the dynamic adjustment of the early warning module in the closed-loop feedback module is as follows: The closed-loop feedback module receives the first early warning result, the second early warning result, and the stage information of the current environmental event determined by the evolution analysis module, and performs the following analysis: The predicted path coordinate set in the first early warning result within a preset first period is fused with the anomaly point diffusion trajectory in the satellite positioning signal associated with the actual anomaly parameters to generate a path offset rate; Based on the degradation marker in the second early warning result within a preset second period, the slope of the actual anomaly parameter change in the sensor data stream and the positioning point coordinate distribution density of the associated satellite positioning signal are compared to generate a prediction deviation; The first early warning result and the second early warning result are dynamically adjusted according to the path offset rate or the prediction deviation.

[0077] Specifically, the steps for generating the path offset rate are as follows:

[0078] W1. Based on the predicted path coordinate set in the first early warning result, extract the coordinates of the positioning points whose abnormal parameters exceed the maximum deviation of the abnormal parameters to form the first key node sequence; simultaneously, from the abnormal point diffusion trajectory of the actual satellite positioning signal, select the nodes whose Euclidean distance from the first key sequence is greater than the preset tolerance to generate the second key node sequence.

[0079] W2. Calculate the average cluster radius of the nodes inside the first key sequence (reflecting the theoretical diffusion concentration) and the offset vector angle of the second key sequence node relative to the first key sequence node (indicating the actual diffusion direction deviation).

[0080] W3. Using the second key node sequence as anchor points, the anomaly propagation trajectory is divided into an early stage (from the initial stage to the peak stage) and a late stage (from the peak stage to the decline stage). The offset weight for the early stage is calculated as: average cluster radius × comprehensive decline factor; the offset weight for the late stage is calculated as: the sine of the offset vector angle of the second sequence. In the early stage, the Euclidean distance between the first key node sequence and the nodes in the anomaly propagation trajectory of the actual satellite positioning signal is calculated point-by-point and recorded as the first single-point offset distance. The arithmetic mean of all first single-point offset distances in the early stage is taken to generate the early average offset distance. In the late stage, the Euclidean distance between the second key node sequence and the nodes in the anomaly propagation trajectory of the actual satellite positioning signal is calculated point-by-point and recorded as the second single-point offset distance. The arithmetic mean of all second single-point offset distances in the late stage is taken to generate the late average offset distance.

[0081] W4. Path offset rate is generated by merging the two-stage weights: Path offset rate = offset weight of the early stage × average offset distance of the early stage + offset weight of the late stage × average offset distance of the late stage.

[0082] Example 1: Suppose that in a certain pollutant diffusion event, the theoretically predicted path coordinate set is [(0,0), (2,2), (4,3)], and the actual satellite positioning signal trajectory is [(0.1,0.2), (2.1,1.8), (3.8,3.2)], with a preset tolerance of 0.3;

[0083] W1: First critical node sequence [(0,0), (4,3)]; Second critical node sequence [(2.1,1.8)] (Euclidean distance over-tolerance);

[0084] W2: Average cluster radius = 0.85 (theoretical concentration), offset vector angle = 30° (direction deviation);

[0085] W3: Early offset weight = 0.85 × 0.7 = 0.595 (overall decay factor is 0.7), early average offset distance = 0.15; late offset weight = sin(30°) = 0.5, late average offset distance = 0.25;

[0086] W4: Path offset rate = 0.595 × 0.15 + 0.5 × 0.25 = 0.214;

[0087] By integrating the phased weighted calculation of theoretical diffusion concentration and directional deviation, the actual deviation of pollutant diffusion paths is quantified, significantly improving the spatiotemporal accuracy and dynamic response adaptability of environmental event early warning.

[0088] Specifically, the steps for generating prediction bias are as follows:

[0089] H1. Extract the degradation markers from the second early warning results, and associate them with the actual abnormal parameter change slope of the sensor data stream and the location point coordinate distribution density of the associated satellite positioning signal. Construct a comparison framework between the theoretical degradation trend and the actual degradation trend within the preset second period.

[0090] H2. Within the comparison framework, calculate the difference between the predicted degradation rate of the degradation marker in the theoretical degradation trend and the actual degradation rate of the degradation marker in the actual degradation trend to generate a preliminary prediction bias.

[0091] H3. Based on the fluctuation characteristics of the slope of the actual abnormal parameter changes in the sensor data stream and the changing pattern of the positioning point coordinate distribution density of the satellite positioning signal, the weight of the preliminary prediction deviation is dynamically adjusted to generate the prediction deviation. The comprehensive prediction deviation can reflect the accuracy of the early warning module in the degradation process and provide a key basis for the subsequent dynamic adjustment of the early warning module.

[0092] Example 2: Suppose that in a pollutant receding event in a certain water area, the theoretical degradation trend prediction rate is 5% per day (based on accelerated degradation markers), the actual monitored slope of the abnormal parameter change rate is 3% per day, and the reduction rate of the location point coordinate distribution density is 0.1 / km².

[0093] H2: Preliminary prediction deviation = Theoretical degradation rate - Actual degradation rate = 5% - 3% = 2%;

[0094] H3: Based on the actual abnormal parameter fluctuation characteristics (fluctuation coefficient 0.3) and the location point density change pattern (density weight 0.7), the weight is dynamically adjusted to 0.3×0.7+0.1×0.3=0.24, and the final prediction deviation = 2%×0.24=0.48%;

[0095] By dynamically adjusting the deviation weights by integrating actual parameter fluctuations and spatial density changes, the accuracy of response during the early warning and downgrade phase is quantified, significantly improving the timeliness and reliability of resource scheduling decisions during the environmental event's decline phase.

[0096] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A GIS dynamic environment monitoring and early warning system based on multi-source sensor networks, characterized in that: It includes a data acquisition module, a model generation module, an evolution analysis module, an early warning module, and a closed-loop feedback module; The data acquisition module accesses the preset satellite positioning signal and the sensor environment data stream generated by the multi-source sensor network in real time, and synchronously records the positioning point coordinates of the satellite positioning signal; The model generation module filters and associates the positioning point coordinates in the satellite positioning signal with the abnormal parameters in the sensor environmental data stream, and inputs the abnormal parameters and the positioning point coordinates into a preset spatiotemporal association model to generate environmental event trigger markers. The evolution analysis module, based on the location point coordinates and associated abnormal parameters in the environmental event trigger markers, divides the triggered environmental events into the following four stages: the initial stage, the development stage, the peak stage, and the decline stage. The early warning module generates a predicted path coordinate set based on the location point coordinates and the abnormal parameters if the triggered environmental event is in the initial or development stage, and uses it as the first early warning result; if the triggered environmental event is in the peak stage, the first early warning result is continued; if the triggered environmental event is in the receding stage, a downgrade marker is generated based on the decrease rate of the abnormal parameters and the density reduction trend of the location point coordinates of the satellite positioning signal, and used as the second early warning result. The steps for generating the first early warning result in the early warning module are as follows: Q1. Extract the data subset of the positioning point coordinates and associated abnormal parameters that exceed the maximum deviation of the abnormal parameters in the initial or development stage, and construct the diffusion initial point set; Q2. Calculate the cluster center migration vector of the coordinates of the positioning points in the initial diffusion point set, and generate spatial diffusion weights by associating the rate of increase of the number of abnormal parameters. Sort the coordinates of the positioning points based on the spatial diffusion weights to construct a spatial diffusion model. Q3. Based on the spatial diffusion model, simulate the diffusion trajectory of the coordinates of the positioning points in the initial diffusion point set, and predict the set of positioning point coordinates in future periods to form a predicted path coordinate set. Q4. Match the predicted path coordinate set with the geographic information in the preset GIS platform to generate a visualized early warning map as the first early warning result. The steps for generating the second early warning result in the early warning module are as follows: Q101. Screen out the abnormal parameters and their associated location coordinates that continuously decrease and conform to the historical abnormal fluctuation pattern in environmental events that are in the decline phase, and construct a set of decline analysis points; Q102. Calculate the slope of the spatial density change of the coordinates of the location points in the set of fading analysis points, and generate a comprehensive fading factor by associating it with the rate of decrease of the corresponding abnormal parameters. Q103. Dynamically classify the set of fading analysis points according to the comprehensive fading factor: when the comprehensive fading factor exceeds the set range, generate an accelerated fading marker based on the slope of the spatial density change of the positioning point coordinates; when the comprehensive fading factor is within the set range, generate a gradual fading marker based on the rate of decrease of the abnormal parameters. When the comprehensive decay factor is less than the set range, a stable degradation marker is generated based on the density reduction trend of the positioning point coordinates. According to the accelerated degradation marker, the gradual degradation marker and the stable degradation marker, the first warning result is downgraded to generate a second warning result. The closed-loop feedback module dynamically adjusts the early warning module based on the first early warning result, the second early warning result, and the stage information of the current environmental event determined by the evolution analysis module.

2. The GIS dynamic environment monitoring and early warning system based on multi-source sensor networks according to claim 1, characterized in that: The satellite positioning signal in the data acquisition module originates from the BeiDou / GPS / Galileo satellite navigation system signal, and the satellite positioning signal includes the positioning point coordinates and the first timestamp; The sensor environmental data stream is generated by a multi-source sensor network consisting of deployed air quality monitoring stations, water quality floating sensors, and temperature and humidity remote sensing terminals. After compression and encryption, it generates data packets for transmission back. The data packets embed the unique identifier of the acquisition device in the multi-source sensor network and a second timestamp, which is aligned with the first timestamp in the satellite positioning signal.

3. The GIS dynamic environment monitoring and early warning system based on multi-source sensor networks according to claim 2, characterized in that: The method for generating the spatiotemporal correlation model in the model generation module is as follows: S1. Select the historical first environmental data collected by the air quality monitoring station, the historical second environmental data collected by the water quality float sensor, and the historical third environmental data collected by the temperature and humidity remote sensing terminal from the sensor environmental data stream as preliminary training samples, input them into the preset training model for training, and extract the training parameters in the training model to generate historical abnormal fluctuation patterns. S2. Select the real-time first environmental data collected by the air quality monitoring station, the real-time second environmental data collected by the water quality floating sensor, and the real-time third environmental data collected by the temperature and humidity remote sensing terminal from the sensor environmental data stream as optimized training samples, and input them into the training model trained in step S1 to calculate the real-time dynamic deviation. S3. Based on the matching degree between the real-time dynamic deviation and the historical abnormal fluctuation pattern, filter the positioning point coordinates and abnormal parameters corresponding to the real-time dynamic deviation with a matching degree higher than a preset threshold. Combine the training parameters in the historical abnormal fluctuation pattern, dynamically adjust the association weights of the preliminary training samples and the optimized training samples to the training model, and generate a spatiotemporal correlation model.

4. The GIS dynamic environment monitoring and early warning system based on multi-source sensor networks according to claim 3, characterized in that: The steps for generating the historical abnormal fluctuation pattern in step S1 are as follows: S11. Based on the historical first environment data, historical second environment data and historical third environment data in the preliminary training samples, calculate the average fluctuation amplitude and maximum deviation of the fluctuation amplitude within the preset historical period, and input them into the training model to establish the initial fluctuation baseline. S12. Based on the fluctuation frequency of the historical first environment data, historical second environment data and historical third environment data in the preliminary training sample within the historical period, and combined with the initial fluctuation baseline, dynamically adjust the weights of the average fluctuation amplitude and the maximum deviation within the historical period to generate historical abnormal fluctuation patterns.

5. The GIS dynamic environment monitoring and early warning system based on multi-source sensor networks according to claim 4, characterized in that: The evolution analysis module divides triggered environmental events into stages as follows: if the abnormal parameters associated with the initial location point coordinates in the environmental event triggering marker exceed the maximum deviation of the abnormal parameters predicted by the spatiotemporal correlation model based on the abnormal parameters for the first time, it is determined to be the initial stage; if the number of abnormal parameters in the environmental event triggering marker increases continuously and the number of location point coordinates increases, it is determined to be the development stage; if the fluctuation range of the abnormal parameters predicted by the spatiotemporal correlation model based on the abnormal parameters in the environmental event triggering marker narrows and the change in the predicted dynamic deviation of the abnormal parameters approaches the preset tolerance, it is determined to be the peak stage; if the number of abnormal parameters in the environmental event triggering marker decreases continuously and the density of location point coordinates decreases, it is determined to be the decline stage.

6. The GIS dynamic environment monitoring and early warning system based on multi-source sensor networks according to claim 5, characterized in that: The specific content of the dynamic adjustment of the early warning module in the closed-loop feedback module is as follows: The closed-loop feedback module receives the first early warning result, the second early warning result, and the stage information of the current environmental event determined by the evolution analysis module, and performs the following analysis: the predicted path coordinate set in the first early warning result within a preset first period is fused with the anomaly point diffusion trajectory in the satellite positioning signal associated with the actual anomaly parameters to generate a path offset rate; based on the degradation marker in the second early warning result within a preset second period, the actual anomaly parameter change slope of the sensor environmental data stream and the positioning point coordinate distribution density of the associated satellite positioning signal are compared to generate a prediction deviation; and the first early warning result and the second early warning result are dynamically adjusted according to the path offset rate or the prediction deviation.

Images