Wetland ecological security early warning method and system based on remote sensing analysis

Abstract

The present application provides a kind of wetland ecological security early warning method and system based on remote sensing analysis, first, obtain the multi-dimensional remote sensing data set of target wetland area including multi-temporal imaging, spectral radiation and microwave scattering data, and obtain the ecological process signal set by interpretation;Establish the mapping relationship model of ecological process signal and wetland ecological security state, obtain the current ecological security state description;Cross-process correlation analysis is carried out on the ecological process signal, and the abnormal conduction chain of ecological process is formed;Construct early warning trigger node system and generate early warning rules;Real-time multi-dimensional remote sensing data is obtained, and after interpretation, it is matched with early warning rules, and wetland ecological security early warning information including abnormal ecological process link, abnormal conduction path and early warning response node is generated.The present application can accurately and comprehensively early warn wetland ecological security risk, and protect the stability of wetland ecosystem.

Description

Technical Field

[0001] This invention relates to the field of wetland ecological protection and management technology, and more specifically, to a wetland ecological security early warning method and system based on remote sensing analysis. Background Technology

[0002] In the field of wetland ecological protection and management, wetland ecological security early warning is a crucial link in ensuring the stability and sustainable development of wetland ecosystems. Traditional wetland ecological security monitoring and early warning methods have many limitations. On the one hand, data acquisition methods are relatively singular and one-sided. Previously, wetland ecological information was mainly obtained through field sampling and manual observation. These methods not only consume significant manpower, resources, and time, but also, due to the vast area and complex environment of wetlands, field sampling and observation often only obtain limited information from local areas, making it difficult to comprehensively and accurately reflect the ecological status of the entire wetland. For example, for key ecological processes such as hydrological connectivity, vegetation succession dynamics, and soil-water interaction processes in large-scale wetlands, real-time, continuous, and comprehensive monitoring through field observation is difficult.

[0003] On the other hand, existing early warning methods lack in-depth analysis of the complex relationships between ecological processes. Wetland ecosystems are complex wholes, with various ecological processes interconnected and influencing each other. However, traditional early warning methods typically analyze individual ecological indicators in isolation, without considering the transmission relationships and interaction mechanisms between signals from different ecological processes. This leads to an inability to accurately identify the root causes and transmission pathways of ecological anomalies when faced with complex changes in wetland ecosystems, making it difficult to detect potential ecological security risks in advance. Consequently, effective early warning and response measures cannot be taken in a timely manner, significantly reducing the accuracy and timeliness of wetland ecological security early warnings, which is detrimental to the protection and restoration of wetland ecosystems. Summary of the Invention

[0004] In view of the aforementioned problems, and in conjunction with the first aspect of the present invention, embodiments of the present invention provide a wetland ecological security early warning method based on remote sensing analysis, the method comprising:

[0005] A multi-dimensional remote sensing data set of the target wetland area is acquired, and ecological process signals are interpreted on the multi-dimensional remote sensing data set to obtain an ecological process signal set of the target wetland area. The multi-dimensional remote sensing data set includes multi-temporal imaging data, spectral radiation data and microwave scattering data of the target wetland area. The ecological process signal set includes hydrological connectivity process signals, vegetation succession process signals and soil-water interaction process signals.

[0006] A mapping relationship model between ecological process signals and wetland ecological security status is established. The mapping relationship model is used to perform state matching on the set of ecological process signals to obtain a description of the current ecological security status of the target wetland area. The mapping relationship model is obtained by associating ecological process signals interpreted from historical remote sensing data with the results of field surveys of wetland ecological security during the same period.

[0007] Cross-process correlation analysis is performed on each signal in the ecological process signal set to identify abnormal transmission relationships between different ecological process signals, determine the initiating process signal, intermediate transmission process signal and final influencing process signal of abnormal transmission, and form an abnormal transmission chain of ecological processes;

[0008] Based on the description of the abnormal transmission chain of the ecological process and the current ecological security status, an early warning triggering node system is constructed. The early warning triggering node system is combined with the dynamic characteristics of the ecological process of the target wetland area to generate wetland ecological security early warning rules. The early warning triggering node system includes the early warning activation conditions corresponding to each abnormal transmission link.

[0009] Real-time multi-dimensional remote sensing data of the target wetland area is acquired, and ecological process signals are interpreted from the real-time multi-dimensional remote sensing data to obtain a set of real-time ecological process signals. The set of real-time ecological process signals is matched with wetland ecological security early warning rules to generate wetland ecological security early warning information for the target wetland area. The wetland ecological security early warning information includes abnormal ecological process links, abnormal transmission paths, and early warning response nodes.

[0010] In another aspect, embodiments of the present invention also provide a wetland ecological security early warning system based on remote sensing analysis, including a processor and a machine-readable storage medium connected to the processor. The machine-readable storage medium is used to store programs, instructions, or code, and the processor is used to run the programs, instructions, or code in the machine-readable storage medium to implement the above-described method.

[0011] Based on the above, this invention first acquires a multi-dimensional remote sensing data set of the target wetland area, including multi-temporal imaging data, spectral radiation data, and microwave scattering data, and interprets ecological process signals. This allows for the comprehensive and accurate acquisition of key ecological process signals such as hydrological connectivity, vegetation succession, and soil-water interaction. Then, a mapping model between ecological process signals and wetland ecological security status is established, and historical data is used for correlation training. This model scientifically and accurately transforms ecological process signals into a description of the current wetland ecological security status, improving the objectivity and accuracy of ecological security status assessment. Cross-process correlation analysis of ecological process signals identifies abnormal transmission relationships and forms anomaly transmission chains, deeply revealing the interaction mechanisms between various ecological processes in the wetland ecosystem. This accurately identifies the root causes and transmission paths of ecological anomalies. Based on the ecological process anomaly transmission chains and the current ecological security status description, an early warning trigger node system is constructed, and wetland ecological security early warning rules are generated. These rules closely integrate with the dynamic characteristics of wetland ecological processes, improving the targeting and timeliness of early warnings. Finally, by acquiring real-time multi-dimensional remote sensing data and performing signal interpretation and rule matching, wetland ecological security early warning information containing abnormal ecological processes, abnormal transmission paths, and early warning response nodes can be generated in real time. This provides timely and effective safety early warning suggestions for wetland ecological protection and management departments, and helps to take measures in advance to prevent wetland ecological security risks. Attached Figure Description

[0012] Figure 1 This is a schematic diagram of the execution flow of the wetland ecological security early warning method based on remote sensing analysis provided in the embodiments of the present invention.

[0013] Figure 2 This is a schematic diagram of the hardware architecture of a wetland ecological security early warning system based on remote sensing analysis provided in an embodiment of the present invention. Detailed Implementation

[0014] The present invention will now be described in detail with reference to the accompanying drawings. Figure 1 This is a flowchart illustrating a wetland ecological security early warning method based on remote sensing analysis, as provided in one embodiment of the present invention. The following is a detailed description of this wetland ecological security early warning method based on remote sensing analysis.

[0015] Step S110: Obtain a multi-dimensional remote sensing data set of the target wetland area, interpret the ecological process signals of the multi-dimensional remote sensing data set to obtain an ecological process signal set of the target wetland area. The multi-dimensional remote sensing data set includes multi-temporal imaging data, spectral radiation data and microwave scattering data of the target wetland area. The ecological process signal set includes hydrological connectivity process signals, vegetation succession process signals and soil-water interaction process signals.

[0016] In this embodiment, the target wetland area is a specific wetland. A multi-dimensional remote sensing data set of the wetland is acquired using a remote sensing data receiving device and a data storage system. The multi-temporal imaging data originates from data captured by optical remote sensing satellites at different times, covering imaging results of the wetland under different seasons and weather conditions. Spectral radiation data is collected by hyperspectral remote sensing equipment, containing the spectral reflectance and radiation characteristics of different land features within the wetland. Microwave scattering data is acquired through synthetic aperture radar, reflecting the scattering characteristics of soil moisture and water distribution in the wetland. After data acquisition, a remote sensing data interpretation algorithm is used to process the multi-dimensional remote sensing data set, extracting signals related to hydrological connectivity, vegetation succession, and soil-water interaction, ultimately forming a set of ecological process signals containing these three types of signals.

[0017] Step S111: Retrieve a multi-dimensional remote sensing data set of the target wetland area from the data storage system. The multi-dimensional remote sensing data set includes multi-temporal imaging data, spectral radiation data and microwave scattering data of the target wetland area.

[0018] A multi-dimensional remote sensing data set of a wetland was retrieved from a dedicated remote sensing data storage system. This system employs a distributed storage architecture, classifying and storing multi-temporal imaging data, spectral radiometric data, and microwave scattering data according to data type and acquisition time. During the retrieval process, a connection was established through a data interface protocol. After verifying data integrity, all three types of data were extracted completely to the data processing platform. The multi-temporal imaging data includes images taken over several consecutive months, with each time period's imaging data containing timestamps and geographic location information. The spectral radiometric data is stored by spectral band, covering multiple bands from visible light to near-infrared. The microwave scattering data is differentiated according to polarization and imaging resolution.

[0019] Step S112: Split the multi-dimensional remote sensing data set according to data type to separate independent multi-temporal imaging data subsets, spectral radiometric data subsets, and microwave scattering data subsets.

[0020] A data classification algorithm was used to segment the retrieved multi-dimensional remote sensing data set. First, data type identifiers were identified based on metadata: multi-temporal imaging data included an "optical imaging" identifier, spectral radiometric data included a "hyperspectral" identifier, and microwave scattering data included a "radar scattering" identifier. Based on these identifiers, the multi-dimensional remote sensing data set was divided into three independent subsets: the multi-temporal imaging data subset, the spectral radiometric data subset, and the microwave scattering data subset. Each subset retained the original data's temporal and location-related information to accurately correspond to the specific area and time period of a wetland during subsequent processing.

[0021] Step S113: Perform temporal feature extraction on the subset of multi-temporal imaging data, track the temporal change trajectory of water body boundary and vegetation coverage area in the target wetland area, and convert the temporal change trajectory of water body boundary and vegetation coverage area in the target wetland area into the original signal that can reflect the hydrological connectivity status, and obtain the hydrological connectivity original signal.

[0022] Temporal feature extraction was performed on a subset of multi-temporal imaging data of a wetland. First, a temporal analysis algorithm was used to sequentially arrange the imaging data from different time periods. Then, an edge detection algorithm was used to identify the water body boundaries and vegetation cover boundaries for each time period. For the water body boundaries, their expansion or contraction changes were tracked at different time periods, and the spatial coordinate changes of the boundaries were recorded. For the vegetation cover areas, the temporal changes in their range and distribution were similarly tracked. These changes were then processed using a signal conversion algorithm to transform the spatial and temporal variation features into continuous electrical signals, which reflect the changes in the connectivity between water bodies and vegetation within the wetland, thus obtaining the original hydrological connectivity signal.

[0023] Step S1131: Perform radiometric correction on the subset of multi-temporal imaging data, extract the water body reflection features of each time period from the radiometrically corrected multi-temporal imaging data, and identify the water body range of each time period based on the difference in reflectivity between the water body and other land features using the threshold segmentation method to obtain water body range maps for multiple time periods.

[0024] First, radiometric correction was performed on a subset of multi-temporal imaging data. This correction process included atmospheric correction, sensor correction, and topographic correction. Atmospheric correction employed a dark target method, identifying dark target pixels in the image, calculating atmospheric influence parameters, and removing the effects of atmospheric scattering and absorption. Sensor correction corrected the sensor's systematic errors based on the satellite sensor's response function. Topographic correction incorporated digital elevation model data of a specific wetland to eliminate the impact of terrain undulations on imaging radiometrics. After correction, water reflectance characteristics were extracted from the imaging data of each time period. Water bodies exhibit significantly lower reflectance in the near-infrared band than vegetation and soil. Based on this difference, a threshold segmentation method was used to set a reflectance threshold, classifying pixels below the threshold as water pixels. This generated a water body extent map for each time period, presented as a binary image where water areas are represented by specific pixel values ​​and non-water areas by different pixel values.

[0025] Step S1132: Compare the water body extent maps of adjacent time periods, calculate the expansion or contraction area of ​​the water body extent, record the direction and amount of change of the expansion or contraction of the water body extent, and obtain the time series change data of the water body extent.

[0026] The generated water body extent maps for multiple time periods are arranged chronologically, and the water body extent maps of adjacent time periods are compared sequentially. Image overlay and region calculation algorithms are used to determine the overlapping and difference areas of the water body extent between two time periods. For the difference area, if the water body extent of the later time period is larger than that of the earlier time period, the expansion area is calculated; if it is smaller than that of the earlier time period, the contraction area is calculated. At the same time, the specific direction of expansion or contraction is determined through spatial coordinate analysis, such as expansion towards the north or south of the wetland, and the change in each direction is recorded. The expansion area, contraction area, direction of change, and change amount of all adjacent time periods are integrated to form a time-series data of water body extent variation arranged in chronological order.

[0027] Step S1133: Extract the vegetation cover area reflection characteristics of each time period from the radiometrically corrected multi-temporal imaging data, identify the vegetation cover area of ​​each time period based on the near-infrared high reflectivity of vegetation, and obtain vegetation cover range maps for multiple time periods.

[0028] Reflectance characteristics of vegetation cover areas for each time period were extracted from radiometrically corrected multi-temporal imaging data. Vegetation exhibits high reflectance in the near-infrared band, while its reflectance is relatively low in the visible light band. Based on this characteristic, a vegetation index calculation method was employed to enhance the vegetation signal through reflectance combinations in specific bands. Then, a vegetation index threshold was set, and pixels exceeding the threshold were identified as vegetation-covered pixels, thereby generating a vegetation cover map for each time period. The vegetation cover map is also presented in a binarized image format, with vegetation-covered areas represented by specific pixel values ​​and non-vegetation-covered areas by different pixel values, and each vegetation cover map is temporally consistent with the corresponding water body map.

[0029] Step S1134: Analyze the spatial overlap between the vegetation cover map and the water body map of the same period, determine the range of water accumulation area within the vegetation cover area, the range of water accumulation area within the vegetation cover area reflects the hydrological connectivity within the wetland, and obtain the temporal variation data of the water accumulation area.

[0030] Spatial overlay analysis was performed on vegetation cover maps and water body maps for each time period. Image spatial analysis algorithms were used to identify overlapping pixel regions in the two images; these overlapping regions represent waterlogged areas within the vegetation cover area. The area and spatial distribution coordinates of the waterlogged areas were calculated for each time period, and the location changes of these waterlogged areas within the vegetation cover area were recorded. The information on the area and distribution coordinates of the waterlogged areas from different time periods was integrated chronologically to form temporal variation data of the waterlogged areas. This temporal variation data reflects the hydrological connectivity between vegetation areas and water bodies within a wetland; for example, the expansion or contraction of waterlogged areas indicates an increase or decrease in hydrological connectivity.

[0031] Step S1135: Calculate the connectivity ratio between the water body area and the water accumulation area area for each time period. The connectivity ratio is the ratio of the water accumulation area area to the total water body area. The connectivity ratio is used to reflect the degree of hydrological connectivity within the wetland.

[0032] For each time period, the total area of ​​the water body and the area of ​​the water accumulation zone are extracted from the water body extent map. A proportional calculation method is used, dividing the area of ​​the water accumulation zone by the total area of ​​the water body to obtain the connectivity ratio for that time period. The connectivity ratio ranges from 0 to 1; the closer the value is to 1, the stronger the connectivity between the water accumulation zone within the vegetation cover area of ​​the wetland and the overall water body; the closer the value is to 0, the weaker the connectivity. The connectivity ratio values ​​for each time period are arranged in chronological order to form a connectivity ratio sequence data.

[0033] Step S1136: Arrange the time-series change data of water body range, time-series change data of water accumulation area and connectivity ratio data according to the monitoring time period, and convert them into a continuous signal with the monitoring time period as the index and the change amount or ratio as the signal value to obtain the original hydrological connectivity signal.

[0034] The temporal variation data of water body extent, temporal variation data of water accumulation area, and connectivity ratio data are integrated and sorted according to the monitoring time period. Using the monitoring time period as an index, the corresponding changes in water body expansion or contraction, changes in water accumulation area, and connectivity ratio value for each time period are used as signal values. The above data are converted into a continuous analog or digital signal sequence through a signal encoding algorithm, which can continuously reflect the changes in the hydrological connectivity status of a wetland at different time periods, thus obtaining the original hydrological connectivity signal.

[0035] Step S114: Perform band feature analysis on the subset of spectral radiation data, extract spectral parameters related to vegetation growth stage and vegetation type, and convert them into original signals that can reflect the vegetation succession state to obtain the original vegetation succession signal.

[0036] Band feature analysis is performed on a subset of spectral radiation data. First, the spectral radiation data is decomposed by band, and the radiation value of each band is extracted. Then, a spectral matching algorithm is used to compare the extracted spectral radiation values ​​with a standard spectral library of known vegetation species to identify the vegetation species in a specific wetland. Simultaneously, by analyzing changes in spectral parameters in specific bands, such as the red edge position of vegetation and the depth of chlorophyll absorption bands, the growth stage of the vegetation, such as seedling stage, maturity stage, and decline stage, is determined. The vegetation species information and growth stage change information are then converted into a continuous signal using a signal conversion algorithm, thus obtaining the original vegetation succession signal.

[0037] Step S115: Perform scattering coefficient inversion on the microwave scattering data subset to obtain information on soil moisture distribution and water depth changes in the target wetland area, and convert it into a raw signal that can reflect the interaction state between soil and water, thus obtaining the raw signal of soil-water interaction.

[0038] A scattering coefficient inversion operation was performed on a subset of microwave scattering data. Using a radar scattering model, soil moisture distribution data and water depth variation data were obtained by inverting the backscattering coefficients in the microwave scattering data and combining them with information on the topography and land cover types of a specific wetland. The soil moisture distribution data was presented in spatial grid units, with each grid corresponding to a soil moisture value; the water depth variation data recorded the water depth values ​​of specific areas at different time periods. The spatial variation of soil moisture distribution and the temporal variation of water depth were converted into continuous signals using a signal conversion algorithm to obtain the original soil-water interaction signal.

[0039] Step S116: Perform signal purification processing on the original hydrological connectivity signal, original vegetation succession signal, and original soil-water interaction signal, and integrate the purified hydrological connectivity signal, purified vegetation succession signal, and purified soil-water interaction signal according to the spatial partitioning and time sequence of the target wetland area to form an ecological process signal set of the target wetland area. The ecological process signal set includes hydrological connectivity process signal, vegetation succession process signal, and soil-water interaction process signal.

[0040] Signal filtering algorithms were employed to purify the original hydrological connectivity signals, vegetation succession signals, and soil-water interaction signals. These algorithms remove noise components, such as signal fluctuations caused by remote sensing equipment errors or atmospheric interference. For the hydrological connectivity signals, moving average filtering was used to remove short-term fluctuation noise; for the vegetation succession and soil-water interaction signals, Kalman filtering was used for noise suppression. After purification, a wetland was divided into multiple spatial zones, each based on topography and ecological function. The purified signals were then integrated according to spatial zones and time periods, with each spatial zone and time period corresponding to a set of data from all three types of signals. This resulted in a comprehensive ecological process signal set containing signals from hydrological connectivity, vegetation succession, and soil-water interaction processes.

[0041] Step S120: Establish a mapping relationship model between ecological process signals and wetland ecological security status. Use the mapping relationship model to perform state matching on the set of ecological process signals to obtain a description of the current ecological security status of the target wetland area. The mapping relationship model is obtained by associating ecological process signals interpreted from historical remote sensing data with the results of field surveys of wetland ecological security during the same period.

[0042] In this embodiment, historical remote sensing data and corresponding field survey results from multiple historical wetland areas are collected. A mapping relationship model is constructed through data association and model training. This model employs a neural network architecture, using historical ecological process signals as input and historical wetland ecological security status as output. The model parameters are optimized through iterative training. After training, the set of ecological process signals from a specific wetland is input into the mapping relationship model. The model, through internal feature association and matching calculations, outputs a description of the current ecological security status of the wetland, such as the hydrological system integrity level, vegetation community stability level, and soil-water interaction health level.

[0043] Step S121: Collect historical remote sensing data of multiple historical wetland areas, including historical multi-temporal imaging data, historical spectral radiation data and historical microwave scattering data of each historical wetland area.

[0044] Historical remote sensing data were collected from multiple historical wetland areas of different types, encompassing wetlands across different climatic zones and ecological categories. The historical remote sensing data was acquired through remote sensing data archives and scientific research databases, including multi-temporal historical imaging data, historical spectral radiometric data, and historical microwave scattering data for each historical wetland area over several consecutive years. Specifically, the historical multi-temporal imaging data records the seasonal changes of each historical wetland in different years; the historical spectral radiometric data includes the spectral characteristics of vegetation and soil at different periods; and the historical microwave scattering data reflects the historical changes in soil moisture and water distribution in each historical wetland. During the collection process, the data was screened to ensure temporal continuity and spatial integrity.

[0045] Step S122: Interpret the historical remote sensing data for ecological process signals to obtain a set of historical ecological process signals for each historical wetland area. The set of historical ecological process signals includes historical hydrological connectivity process signals, historical vegetation succession process signals, and historical soil-water interaction process signals.

[0046] The same remote sensing data interpretation method as in step S110 is used to process the collected historical remote sensing data. For each historical wetland area, historical hydrological connectivity process signals are extracted from historical multi-temporal imaging data; band feature analysis is performed on historical spectral radiometric data to obtain historical vegetation succession process signals; and scattering coefficient inversion is performed on historical microwave scattering data to obtain historical soil-water interaction process signals. These three types of historical signals are integrated to form a historical ecological process signal set for each historical wetland area, and each historical ecological process signal set is associated with the corresponding historical wetland area and time period.

[0047] Step S123: Collect the results of the field survey on wetland ecological security for each historical wetland area during the corresponding period of the historical remote sensing data. The results of the field survey on wetland ecological security include the assessment of the integrity of the hydrological system, the assessment of the stability of the vegetation community, and the assessment of the health of the soil-water interaction.

[0048] By reviewing historical survey literature, research reports, and ecological monitoring records, we collected field survey results on wetland ecological security for each historical wetland area during the corresponding periods of historical remote sensing data. The field survey results were obtained by a professional survey team through on-site sampling, instrumental measurements, and visual assessments. Specifically, the hydrological system integrity assessment included indicators such as water connectivity and water supply stability; the vegetation community stability assessment covered vegetation species diversity and community structure integrity; and the soil-water interaction health assessment involved parameters such as soil pollution levels and water-soil nutrient exchange efficiency. The field survey results for each period are presented in the form of an assessment report, including specific assessment indicators and classifications.

[0049] Step S124: Bind the set of historical ecological process signals for each historical wetland area with the results of the field survey of wetland ecological security during the same period to form multiple sets of historical correlation data. Each set of historical correlation data includes a set of historical ecological process signals and the corresponding wetland ecological security status assessment results.

[0050] For each historical wetland area, its historical ecological process signal set is matched and bound with the field survey results of wetland ecological security during the same period, according to time periods. First, the corresponding time period for each signal in the historical ecological process signal set is determined, and then the corresponding field survey results for that time period are found. Through timestamp matching and spatial location verification, it is ensured that the historical ecological process signals and the field survey results belong to the same time period of the same historical wetland area. The matched set of historical ecological process signals (including historical hydrological connectivity process signals, historical vegetation succession process signals, and historical soil-water interaction process signals) is combined with the corresponding wetland ecological security status assessment results to form a set of historical correlation data. This operation is performed for all time periods of all historical wetland areas to generate multiple sets of historical correlation data.

[0051] Step S1241: Divide the historical ecological process signal set of each historical wetland area into time periods. Divide the continuous historical ecological process signal set into multiple historical ecological process signal subsets at fixed time intervals. Each historical ecological process signal subset corresponds to a specific monitoring time period.

[0052] For each historical wetland area, a continuous set of historical ecological process signals is divided into segments at fixed time intervals using a time-segmentation algorithm. The fixed time interval is set based on the acquisition frequency of historical remote sensing data, such as monthly or quarterly intervals. During the segmentation process, based on the signal's time index, the continuous signal sequence is divided into multiple independent subsets of historical ecological process signals. Each subset corresponds to a specific monitoring period and includes signal segments of historical hydrological connectivity processes, historical vegetation succession processes, and historical soil-water interaction processes within that period.

[0053] Step S1242: Collect the field survey results of wetland ecological security for each historical wetland area during each monitoring period, and perform feature extraction on the subset of historical ecological process signals for each monitoring period. Extract the average intensity of historical hydrological connectivity process signals, the rate of change of historical vegetation succession process signals, and the distribution uniformity of historical soil-water interaction process signals during the monitoring period to obtain the historical signal feature group for the monitoring period.

[0054] Field survey results on wetland ecological security for each historical wetland area were collected for each monitoring period to ensure accurate correspondence between the survey results and the monitoring period. Simultaneously, feature extraction was performed on a subset of historical ecological process signals for each monitoring period. For the historical hydrological connectivity process signal subset, the average signal intensity within that period was calculated to reflect the overall level of hydrological connectivity. For the historical vegetation succession process signal subset, the rate of change was calculated using the signal slope to reflect the speed of vegetation succession. For the historical soil-water interaction process signal subset, spatial distribution analysis algorithms were used to calculate the distribution evenness to reflect the spatial consistency of soil-water interaction. These three feature parameters were combined to form the historical signal feature set for that monitoring period.

[0055] Step S1243: Quantify the results of the field survey on wetland ecological security during the same period, and convert the hydrological system integrity assessment, vegetation community stability assessment and soil-water interaction health assessment into quantitative scores to obtain the safety status quantitative score group for the monitoring period. The quantitative score of each assessment dimension corresponds to a set numerical range.

[0056] Specifically, for the hydrological system integrity assessment, based on the level descriptions of indicators such as water connectivity and water supply stability in the assessment report, the five levels of "intact," "relatively intact," "average," "poor," and "extremely poor" are each mapped to different quantitative scores. For the vegetation community stability assessment, based on the assessment results of vegetation species diversity and community structure integrity, "stable," "relatively stable," "average," "unstable," and "extremely unstable" are similarly converted into corresponding quantitative scores. For the soil-water interaction health assessment, based on the assessment conclusions of parameters such as soil pollution level and water-soil nutrient exchange efficiency, "healthy," "relatively healthy," "average," "unhealthy," and "extremely unhealthy" are mapped to corresponding quantitative scores. The quantitative scores for each assessment dimension are within the same numerical range, ensuring the comparability of scores across different dimensions. After completing the quantitative conversion of the three assessment dimensions, the three quantitative scores are combined to form a quantitative score group for the safety status during the monitoring period.

[0057] Step S1244: Bind the historical signal feature group and the safety status quantitative scoring group for the same monitoring period to form a set of historical associated data.

[0058] Historical signal feature groups and safety status quantitative scoring groups corresponding to the same monitoring period are identified. Matching and verification are performed using the timestamps of the monitoring period and the spatial identifiers of the historical wetland area to ensure that the two sets of data belong to the same historical wetland area and the same time period. After successful verification, the historical signal feature groups (including the average intensity of historical hydrological connectivity process signals, the rate of change of historical vegetation succession process signals, and the distribution uniformity of historical soil-water interaction process signals) are associated and bound with the safety status quantitative scoring groups (including hydrological system integrity quantitative scores, vegetation community stability quantitative scores, and soil-water interaction health quantitative scores) to form a set of historical associated data containing input features and output labels.

[0059] Step S1245: Perform time period segmentation, wetland ecological security field survey results collection, historical signal feature group extraction, safety status quantitative scoring group quantification and data binding operations for all monitoring periods of all historical wetland areas to obtain multiple sets of historical associated data.

[0060] The process iterates through all collected historical wetland areas. For each monitoring period in each historical wetland area, the following steps are executed sequentially: time period segmentation (S1241), field survey result collection and historical signal feature extraction (S1242), safety status quantification and scoring (S1243), and data binding (S1244). During execution, data quality is checked at each stage, removing invalid data with missing information, time mismatches, or logical contradictions. All valid data that passes quality checks are combined to ultimately obtain multiple sets of historical correlation data.

[0061] Step S125: Divide the multiple sets of historical correlation data into a training dataset and a validation dataset. The training dataset is used for training the parameters of the mapping relationship model, and the validation dataset is used for validating the effect of the mapping relationship model.

[0062] A data partitioning algorithm was employed to divide multiple sets of historical correlation data. Random sampling was followed during partitioning, while ensuring consistency between the training and validation datasets in terms of historical wetland area type, monitoring period distribution, and safety status level coverage. This prevented imbalanced data distribution from negatively impacting model training performance. The partitioning ratio was determined based on the total amount of data; typically, the majority of historical correlation data was allocated to the training dataset, and a smaller portion to the validation dataset. After partitioning, the training and validation datasets were stored independently, labeled with their data sources and partition identifiers for retrieval during subsequent model training and validation.

[0063] Step S126: Construct the basic architecture of the mapping relationship model. The basic architecture includes a signal input module, a feature association module, and a state output module. The historical ecological process signals in the training dataset are input into the signal input module. The feature association module establishes the correlation between the features of the historical ecological process signals and the wetland ecological security status assessment results. The state output module outputs the ecological security status prediction results.

[0064] The underlying architecture of the mapping relationship model adopts a deep learning neural network architecture, specifically including a signal input module, a feature association module, and a state output module. The signal input module receives historical signal feature groups from the training dataset, converts them into vector forms that the neural network can process, and standardizes the input vectors to ensure the feature data falls within a uniform data distribution range. The feature association module consists of multiple hidden layers, including convolutional layers, pooling layers, and fully connected layers. The convolutional layers extract local features from the input vector, the pooling layers reduce the dimensionality of the extracted local features to reduce data redundancy, and the fully connected layers integrate the dimensionality-reduced features to establish a complex nonlinear correlation between the features and the wetland ecological security status assessment results. The state output module uses the softmax activation function to convert the feature vectors output by the feature association module into probability distributions for different ecological security status levels, outputting the ecological security status prediction results.

[0065] Step S127: Iteratively optimize the correlation parameters of the mapping relationship model infrastructure using the training dataset, so that the deviation between the ecological security status prediction results output by the state output module and the actual wetland ecological security field survey results in the training dataset is gradually reduced.

[0066] The training dataset is input into the mapping model in batches. For each batch, the model calculates the ecological security status prediction result through forward propagation. The cross-entropy loss function is used to calculate the deviation between the prediction result and the actual security status quantification score group in the training dataset. The deviation signal is then propagated from the output layer to the input layer via backpropagation, adjusting the weights and biases of each hidden layer layer by layer. During iterative optimization, a learning rate parameter is set to control the magnitude of each parameter adjustment, while a momentum optimization algorithm is used to accelerate model convergence and avoid the model getting trapped in local optima. After a certain number of iterations, the overall loss value of the current model on the training dataset is calculated. The parameter optimization process is paused when the loss value no longer decreases significantly after multiple iterations.

[0067] Step S128: Validate the optimized mapping relationship model using the validation dataset. Input the historical ecological process signals from the validation dataset into the optimized mapping relationship model, and compare the prediction results output by the optimized mapping relationship model with the actual wetland ecological security field survey results in the validation dataset. If the deviation is within the preset acceptable range, the training of the mapping relationship model is completed, and the mapping relationship model between ecological process signals and wetland ecological security status is obtained.

[0068] The validation dataset is input into the iteratively optimized mapping model, and the model outputs the corresponding ecological security status prediction results. Evaluation metrics such as accuracy, precision, recall, and F1 score are used to compare the prediction results with the actual security status quantitative score groups in the validation dataset. The value of each evaluation metric is calculated. If the values ​​of all evaluation metrics reach the preset acceptable threshold (i.e., the deviation is within the preset acceptable range), the model training effect is deemed satisfactory, and the training of the mapping model is completed. If the evaluation metrics do not reach the preset threshold, the process returns to step S127, and training parameters such as the learning rate are adjusted. Iterative parameter optimization continues until the model's performance on the validation dataset meets the requirements, ultimately obtaining the mapping model between ecological process signals and wetland ecological security status.

[0069] Step S129: Input the set of ecological process signals of the target wetland area into the mapping relationship model, calculate the degree of matching between the hydrological connectivity process signals, vegetation succession process signals and soil-water interaction process signals in the ecological process signal set and different ecological security states through the feature association module, and output the ecological security state with the highest degree of matching through the state output module to obtain the current ecological security state description of the target wetland area.

[0070] The ecological process signal set of a certain wetland is preprocessed to extract feature parameters of hydrological connectivity, vegetation succession, and soil-water interaction signals for each spatial partition and time period, forming an input feature vector consistent with the training data format. This input feature vector is then fed into the trained mapping model, where the signal input module converts it into a model-processable form and standardizes it. The feature association module extracts key information from the input features through calculations in each hidden layer, calculating the degree of matching between the feature and different ecological security status levels. The status output module outputs a probability distribution based on the degree of matching, selecting the ecological security status level with the highest probability value as the prediction result. Combining the predicted level and the feature contribution of each assessment dimension, a current ecological security status description is generated, including specific descriptions of hydrological system integrity, vegetation community stability, and soil-water interaction health.

[0071] Step S130: Perform cross-process correlation analysis on each signal in the ecological process signal set, identify the abnormal transmission relationship between different ecological process signals, determine the initiating process signal, intermediate transmission process signal and final influencing process signal of abnormal transmission, and form an abnormal transmission chain of ecological processes.

[0072] In this embodiment, for a set of ecological process signals of a wetland, hydrological connectivity process signals, vegetation succession process signals, and soil-water interaction process signals are extracted. The correlation degree among these three signals is calculated through time-series correlation analysis and compared with the standard correlation degree range to identify anomalous correlation combinations. The anomalous start time of the signals in the anomalous correlation combinations is traced to determine the transmission order between the signals, thereby identifying the initial process signal, intermediate transmission process signal, and final influencing process signal, which are then connected in the transmission order to form an anomalous ecological process transmission chain.

[0073] Step S131: Extract the hydrological connectivity process signal, vegetation succession process signal, and soil-water interaction process signal from the ecological process signal set, and generate the time-series variation curves of the hydrological connectivity process signal, vegetation succession process signal, and soil-water interaction process signal, wherein the time-series variation curves are plotted with the monitoring period as the horizontal axis and the signal intensity as the vertical axis.

[0074] From the ecological process signal set of a wetland, signals of hydrological connectivity, vegetation succession, and soil-water interaction were separated. For each signal, signal intensity values ​​were extracted sequentially according to the monitoring time period. With the monitoring time period as the horizontal axis and the signal intensity value as the vertical axis, a curve fitting algorithm was used to generate time-series variation curves for each signal. These curves visually reflect the intensity change trends of each ecological process signal over different monitoring time periods. For example, the time-series variation curve of the hydrological connectivity signal can reflect the fluctuations in the wetland's hydrological connectivity status, while the time-series variation curve of the vegetation succession signal can demonstrate the dynamic process of vegetation succession.

[0075] Step S132: Calculate the temporal correlation between the hydrological connectivity process signal and the vegetation succession process signal, calculate the temporal correlation between the hydrological connectivity process signal and the soil-water interaction process signal, and calculate the temporal correlation between the vegetation succession process signal and the soil-water interaction process signal. The temporal correlation is used to reflect the consistency of the changing trends of the two ecological process signals within the same time period.

[0076] Using correlation analysis algorithms in time series analysis, the temporal correlation degree between three pairs of ecological process signals was calculated. For the hydrological connectivity process signal and the vegetation succession process signal, the signal intensity values ​​of the two at the same time period were extracted, and the correlation coefficient between the two was calculated to measure the temporal correlation degree. The larger the absolute value of the correlation coefficient, the stronger the consistency of the change trend of the two at the same time period. Using the same method, the correlation coefficients between the hydrological connectivity process signal and the soil-water interaction process signal, and between the vegetation succession process signal and the soil-water interaction process signal were calculated to obtain the corresponding temporal correlation degree, thereby reflecting the dynamic correlation characteristics between each pair of signals.

[0077] Step S133: Retrieve the standard correlation range between hydrological connectivity process signals and vegetation succession process signals, hydrological connectivity process signals and soil-water interaction process signals, and vegetation succession process signals and soil-water interaction process signals in a normal wetland ecosystem. The standard correlation range is obtained through statistical analysis of a large amount of historical monitoring data from normal wetlands.

[0078] The standard correlation ranges are retrieved from the data storage system. These ranges are determined through historical monitoring of a large number of wetlands in normal ecological security states, collecting signals related to hydrological connectivity, vegetation succession, and soil-water interaction. The temporal correlation between each pair of signals is calculated, and then statistical analysis is employed to determine the correlation ranges. Specifically, the standard correlation ranges include a first standard correlation range between hydrological connectivity and vegetation succession signals, a second standard correlation range between hydrological connectivity and soil-water interaction signals, and a third standard correlation range between vegetation succession and soil-water interaction signals. Each range includes an upper and lower limit value.

[0079] Step S134: Compare the calculated actual temporal correlation degree with the corresponding standard correlation degree range, and mark the abnormal correlation combinations that exceed the standard correlation degree range. Each abnormal correlation combination contains two ecological process signals with abnormal correlation.

[0080] The actual temporal correlation degree of the three pairs of signals calculated in step S132 is compared one by one with the corresponding standard correlation degree range retrieved in step S133. If the actual temporal correlation degree of a pair of signals is lower than the lower limit of the standard correlation degree range or higher than the upper limit, the correlation relationship of the pair of signals is determined to be abnormal, and it is marked as an abnormal correlation combination. Each abnormal correlation combination clearly records the types of two ecological process signals included and the actual temporal correlation degree value. For example, if the actual correlation degree between the hydrological connectivity process signal and the vegetation succession process signal exceeds the first standard correlation degree range, the combination is marked as an abnormal correlation combination.

[0081] Step S1341: Retrieve from the data storage system the first standard correlation range between hydrological connectivity process signals and vegetation succession process signals in a normal wetland ecosystem, the second standard correlation range between hydrological connectivity process signals and soil-water interaction process signals, and the third standard correlation range between vegetation succession process signals and soil-water interaction process signals.

[0082] Through the query interface of the data storage system, input a data retrieval command, specifying the type of standard correlation range to be obtained. The system then extracts relevant standard data for normal wetland ecosystems from the database according to the command. The first standard correlation range is the statistical range of the time-series correlation between hydrological connectivity process signals and vegetation succession process signals under normal conditions; the second standard correlation range is the statistical range of the time-series correlation between hydrological connectivity process signals and soil-water body interaction process signals under normal conditions; and the third standard correlation range is the statistical range of the time-series correlation between vegetation succession process signals and soil-water body interaction process signals under normal conditions. After retrieval, the data is formatted to match the format of the actual calculated time-series correlation results, facilitating subsequent comparison operations.

[0083] Step S1342: Compare the actual temporal correlation between the calculated hydrological connectivity process signal and the vegetation succession process signal with the first standard correlation range. If the actual temporal correlation between the hydrological connectivity process signal and the vegetation succession process signal is less than the lower limit of the first standard correlation range or greater than the upper limit of the first standard correlation range, then the correlation combination is determined to be an abnormal correlation combination and marked as the first abnormal correlation combination. The first abnormal correlation combination includes the hydrological connectivity process signal and the vegetation succession process signal.

[0084] The actual temporal correlation between the hydrological connectivity process signal and the vegetation succession process signal is compared with the lower and upper limits of the first standard correlation range. If the actual value is lower than the lower limit, it indicates that the consistency of their changing trends is significantly lower than normal; if the actual value is higher than the upper limit, it indicates that the consistency of their changing trends is significantly higher than normal, both of which indicate an abnormal correlation. When either of the above two conditions is met, the correlation combination is determined to be an abnormal correlation combination and marked as the first abnormal correlation combination, while the deviation between the actual value and the standard range is recorded.

[0085] Step S1343: Compare the actual temporal correlation between the calculated hydrological connectivity process signal and the soil-water interaction process signal with the second standard correlation range. If the actual temporal correlation between the hydrological connectivity process signal and the soil-water interaction process signal exceeds the upper and lower limits of the second standard correlation range, then the correlation combination is determined to be an abnormal correlation combination and marked as the second abnormal correlation combination. The second abnormal correlation combination includes the hydrological connectivity process signal and the soil-water interaction process signal.

[0086] Using the same comparison method as in step S1342, the actual temporal correlation between the hydrological connectivity process signal and the soil-water interaction process signal is compared with the upper and lower limits of the second standard correlation range. If the actual temporal correlation is lower than the lower limit or higher than the upper limit of the second standard correlation range, i.e., it exceeds the standard range, the correlation combination is determined to be an abnormal correlation combination, marked as the second abnormal correlation combination, and the specific magnitude and direction of the actual value exceeding the standard range are recorded in detail.

[0087] Step S1344: Compare the actual temporal correlation between the calculated vegetation succession process signal and the soil-water interaction process signal with the third standard correlation range. If the actual temporal correlation between the vegetation succession process signal and the soil-water interaction process signal exceeds the upper and lower limits of the third standard correlation range, then the correlation combination is determined to be an abnormal correlation combination and marked as the third abnormal correlation combination. The third abnormal correlation combination includes the vegetation succession process signal and the soil-water interaction process signal.

[0088] Compare the actual temporal correlation between vegetation succession signals and soil-water interaction signals with the range of the third standard correlation. If the actual values ​​are not between the upper and lower limits of the third standard correlation range, i.e., exceed the standard range, it indicates that the correlation between the signals does not conform to the characteristics of a normal wetland ecosystem, and is judged as an abnormal correlation combination, marked as the third abnormal correlation combination, and the relevant comparison results data are recorded.

[0089] Step S1345: Record the actual temporal correlation degree value, the corresponding standard correlation degree range, and the magnitude of exceeding the standard correlation degree range for each abnormal correlation combination.

[0090] For the marked first, second, and third anomalous association combinations, anomaly record entries are established. Each entry includes the identifier of the anomalous association combination, the types of two ecological process signals within the combination, the actual time-series correlation value, the corresponding standard correlation range (lower and upper limits), and the magnitude of exceeding the standard range. The magnitude of exceeding the standard range is determined by calculating the difference between the actual value and the nearest boundary value of the standard range. If the actual value is lower than the lower limit, the magnitude of exceeding the standard range is the difference between the lower limit and the actual value; if the actual value is higher than the upper limit, the magnitude of exceeding the standard range is the difference between the actual value and the upper limit.

[0091] Step S1346: Analyze the relationship between the excess magnitude of each anomalous association combination and the trend of ecological process signal changes, and mark potential anomalous association combinations and confirmed anomalous association combinations based on the analysis results.

[0092] For each anomalous association combination, the relationship between the magnitude of the exceedance and the time-series variation curves of the corresponding two ecological process signals is analyzed. If the exceedance is small and the signal variation trend deviates from normal only in individual periods, it is identified as a potential anomalous association combination, indicating that it may be a temporary anomaly caused by short-term disturbance factors. If the exceedance is large and the signal variation trend shows obvious anomalous deviations in multiple consecutive periods, it is identified as a confirmed anomalous association combination, indicating that it may be a persistent association anomaly caused by abnormal processes within the wetland ecosystem.

[0093] Step S1347: Integrate all marked first, second, and third anomalous association combinations, distinguish between potential and confirmed anomalous association combinations, and form a list of anomalous association combinations. Each anomalous association combination contains two ecological process signals with anomalous associations and a description of the degree of anomalousness.

[0094] The first, second, and third marked abnormal association combinations are summarized and categorized into potential abnormal association combinations and confirmed abnormal association combinations. An abnormality severity description is added to each abnormal association combination. The severity is determined based on the duration of the abnormality exceeding the amplitude and signal change trend, and is divided into three levels: "minor abnormality," "moderate abnormality," and "severe abnormality." The categorized abnormal association combinations and their corresponding severity descriptions are integrated to form an abnormal association combination list. The list is arranged in descending order of severity to facilitate prioritizing the processing of severe abnormal association combinations.

[0095] Step S135: For each abnormal association combination, trace the start time of the abnormal change of the two ecological process signals with abnormal association, and determine the ecological process signal that changes abnormally first as the potential starting process signal and the ecological process signal that changes abnormally later as the potential affected process signal.

[0096] For each anomalous association combination in the list, the time-series variation curves and corresponding signal intensity data of the two ecological process signals within the combination are extracted. Anomaly detection algorithms are used to identify anomalous data points deviating from the normal fluctuation range in each signal intensity data, determining the moment when each signal first shows an anomalous data point, i.e., the start time of the anomalous change. By comparing the start times of the anomalous changes of the two signals, the signal with the earlier start time is identified as the potential initiating process signal, and the signal with the later start time is identified as the potential affected process signal. A preliminary judgment is made that the anomaly may have propagated from the potential initiating process signal to the potential affected process signal.

[0097] Step S136: Analyze the correlation between the potential initiation process signal and other ecological process signals. If the potential initiation process signal affects not only the marked potential affected process signal, but also a third ecological process signal through the marked potential affected process signal, then the marked potential affected process signal is determined as the intermediate transmission process signal, and the third ecological process signal is determined as the final influencing process signal.

[0098] For a identified potential initiation process signal, time-series correlation data and anomaly records between it and two other ecological process signals in a specific wetland ecological process signal set are retrieved. Using the potential initiation process signal as the core, a signal correlation network is constructed to analyze its correlation path with a third ecological process signal besides the marked potentially affected process signal. The transmission sequence is determined by the sequential changes in the time-series curves. If the potential initiation process signal first exhibits anomalies, triggering anomalies in the marked potentially affected process signal, and the anomaly in the potentially affected process signal further leads to anomalies in the third ecological process signal, then a transmission chain of "potential initiation process signal—potentially affected process signal—third ecological process signal" exists. In this case, the marked potentially affected process signal is defined as the intermediate transmission process signal, and the third ecological process signal is defined as the final influencing process signal, clarifying the roles of the three in the anomaly transmission.

[0099] Step S137: If the potential initiation process signal directly affects two other ecological process signals, then the two other ecological process signals are determined as intermediate transmission process signals and final influencing process signals according to the order in which they are affected.

[0100] When an anomalous change in a potential initiating process signal directly causes anomalies in two other ecological process signals within a wetland's ecological process signal set without being relayed by other signals, the onset times of these anomalous changes in these two ecological process signals are extracted. By comparing the onset times of the two signals, the ecological process signal with the earlier onset of the anomaly is identified as the intermediate transmission process signal, and the other ecological process signal with the later onset of the anomaly is identified as the final influencing process signal. This indicates that the anomalous impact of the potential initiating process signal is direct and progressive, with the two affected signals sequentially carrying over the anomalous transmission in chronological order.

[0101] Step S138: Connect the determined initial process signal, intermediate transmission process signal and final impact process signal according to the order of abnormal transmission to form an ecological process abnormal transmission chain. The ecological process abnormal transmission chain presents a complete transmission path of the abnormality from the initial link to the final impact link.

[0102] The initial process signal, intermediate transmission process signal, and final influencing process signal, determined through the above steps, are connected in series according to the chronological order of anomaly propagation. The initial process signal serves as the starting point of the propagation chain, the intermediate transmission process signal as the transmission intermediary, and the final influencing process signal as the transmission endpoint. The direction of propagation is marked with arrows or sequence numbers, forming a visualized ecological process anomaly propagation chain. This ecological process anomaly propagation chain clearly presents the complete path of the anomaly's gradual transmission from its initial occurrence stage to its final influencing stage, while also marking the time of anomaly occurrence and signal change characteristics at each stage.

[0103] Step S140: Based on the ecological process anomaly transmission chain and the current ecological security status description, construct an early warning trigger node system, combine the early warning trigger node system with the ecological process dynamic characteristics of the target wetland area, and generate wetland ecological security early warning rules. The early warning trigger node system includes the early warning activation conditions corresponding to each anomaly transmission link.

[0104] In this embodiment, based on the description of the abnormal transmission chain of an ecological process in a certain wetland and its current ecological security status, the abnormal transmission links in the chain are broken down, and early warning activation conditions are set for each link, integrating them to form an early warning trigger node system. Combining the dynamic changes in the wetland's hydrology, vegetation, and soil-water interaction, the conditions and timeliness of the early warning trigger nodes are adjusted, ultimately generating wetland ecological security early warning rules covering spatial zoning, trigger conditions, and response timing.

[0105] Step S141: Disassemble the abnormal transmission chain of the ecological process, separate each abnormal transmission link, and each abnormal transmission link contains the transmission relationship between the signal of the previous process and the signal of the next process, thus obtaining multiple abnormal transmission links.

[0106] The established ecological process anomaly transmission chains are structurally decomposed and divided into multiple independent anomaly transmission links according to the transmission sequence. Each anomaly transmission link corresponds to the interaction relationship between two adjacent process signals in the transmission chain; that is, the specific link where an abnormal change in the preceding process signal triggers an anomaly in the following process signal. For example, if the transmission chain is "hydrological connectivity process signal—vegetation succession process signal—soil-water interaction process signal," it is decomposed into two anomaly transmission links: "hydrological connectivity process signal → vegetation succession process signal" and "vegetation succession process signal → soil-water interaction process signal." Each link records the specific type, transmission direction, and anomaly correlation characteristics of the preceding and following process signals, forming a list of anomaly transmission links.

[0107] Step S142: For each abnormal transmission link, determine the critical value of abnormal change of the previous process signal and the critical value of response change of the next process signal in the abnormal transmission link. The critical value of abnormal change is the limit value of the previous process signal deviating from the normal range, and the critical value of response change is the limit value of the next process signal deviating from the normal range after being affected.

[0108] For each anomalous transmission stage, the signal intensity range of the preceding process is retrieved under normal wetland conditions. Statistical analysis determines the upper and lower limits of normal signal fluctuation, and the threshold values ​​exceeding these limits are set as anomalous change thresholds. Similarly, for the subsequent process signal, based on its signal characteristics under normal conditions and the degree of influence of the preceding process signal anomaly, the threshold value at which it begins to deviate from the normal range after being affected is determined; this is the response change threshold. The setting of the anomalous change threshold and the response change threshold should refer to the signal change thresholds of that transmission stage in historical anomalous data, and be adjusted in conjunction with the level assessment results in the current ecological security status description to ensure that the threshold values ​​accurately reflect the initiation point of the anomalous transmission.

[0109] Step S143: Set the condition that the previous process signal reaches the critical value of abnormal change and the subsequent process signal begins to approach the critical value of response change as the early warning activation condition corresponding to the abnormal transmission link. Each early warning activation condition includes the type of ecological process signal, the critical value of the ecological process signal and the triggering time.

[0110] By combining the critical values ​​for abnormal changes and response changes at each abnormal transmission stage, an early warning activation condition is constructed for that stage. An early warning activation condition is met when the intensity of the preceding process signal reaches or exceeds the critical value for abnormal change, and the intensity of the following process signal shows a trend towards the critical value for response change (e.g., fluctuating towards the critical value over several consecutive time periods). Each early warning activation condition clearly indicates the types of preceding and following ecological process signals involved, the corresponding critical values ​​for abnormal changes and response changes, and the triggering timing requirement of "the preceding signal meeting the standard and the following signal approaching it," ensuring the targeted and accurate activation of the early warning.

[0111] Step S144: Integrate the early warning activation conditions corresponding to all abnormal transmission links, arrange them in the order of the abnormal transmission chain of the ecological process, and form initial early warning trigger nodes. Each initial early warning trigger node corresponds to an abnormal transmission link and the corresponding early warning activation conditions.

[0112] The warning activation conditions for all abnormal transmission links are summarized and sorted according to the transmission sequence of the ecological process abnormal transmission chain. Each warning activation condition serves as an independent initial warning trigger node, with the node number consistent with the order of the abnormal transmission link in the transmission chain. Each initial warning trigger node is associated with corresponding abnormal transmission link information, including transmission direction, involved signals, and critical value parameters, forming a sequence of initial warning trigger nodes arranged in the transmission order, which intuitively reflects the correspondence between warning nodes and abnormal transmission links.

[0113] Step S145: Based on the current ecological security status description of the target wetland area, adjust the early warning activation conditions of the initial early warning trigger node to obtain the adjusted early warning trigger node.

[0114] Referring to the current ecological security status description of a certain wetland, if the description indicates that a certain assessment dimension (such as the integrity of the hydrological system) is at a "poor" level, then the critical value for abnormal changes or response changes of the initial warning trigger node involving the ecological process signal of that dimension should be appropriately lowered to make the warning activation more sensitive. If a certain assessment dimension is at a "relatively complete" or "stable" level, the critical value can be appropriately increased to avoid false triggering. For warning trigger nodes that transmit across dimensions, the matching degree of the critical value should be adjusted according to the synergistic relationship of each dimension in the current ecological security status to ensure that the warning activation conditions of the adjusted warning trigger node are compatible with the current basic ecological security status of the wetland.

[0115] Step S146: Extract the dynamic characteristics of ecological processes in the target wetland area. The dynamic characteristics of ecological processes include the seasonal fluctuation patterns of hydrological connectivity processes, the growth cycle patterns of vegetation succession processes, and the annual variation patterns of soil-water interaction processes.

[0116] By analyzing the long-term variation data of a wetland's historical ecological process signal set, the dynamic characteristics of each ecological process are extracted. For hydrological connectivity processes, the variation patterns of water body range and connectivity ratio in different seasons are statistically analyzed to determine the fluctuation characteristics of the rainy and dry seasons. For vegetation succession processes, the signal variation patterns of vegetation at different growth stages such as germination, growth, and withering are identified to clarify the characteristic nodes within the growth cycle. For soil-water interaction processes, the interaction variation patterns of soil moisture and water depth in different months of the year are analyzed to summarize the influence characteristics of seasonal precipitation and evaporation on the interaction process. These patterns are organized into seasonal fluctuation curves of hydrological connectivity processes, vegetation succession growth cycle maps, and annual variation models of soil-water interaction, forming a dynamic characteristic database of ecological processes.

[0117] Step S147: Analyze the compatibility between the early warning activation conditions of each adjusted early warning trigger node and the dynamic characteristics of the ecological processes in the target wetland area. If the critical value of the ecological process signal in the early warning activation conditions of the adjusted early warning trigger node conflicts with the seasonal fluctuation pattern, then adjust the critical value of the ecological process signal according to the season and supplement the effective early warning period for each adjusted early warning trigger node.

[0118] The critical values ​​in the early warning activation conditions of each adjusted early warning trigger node are compared with the corresponding dynamic characteristics of the ecological process. If an early warning trigger node involves hydrological connectivity process signals, its critical value may be within the normal fluctuation range during the dry season, but may frequently exceed it during the rainy season, resulting in a conflict with the seasonal fluctuation pattern. To address this, the critical values ​​are reset according to seasonal segments, setting appropriate abnormal change critical values ​​and response change critical values ​​for the rainy and dry seasons respectively. Simultaneously, based on the active periods of each process signal in the dynamic characteristics of the ecological process, an effective early warning period is added to each adjusted early warning trigger node. For example, for early warning trigger nodes related to vegetation succession, the effective period is set to the vigorous vegetation growth period to avoid false early warnings caused by natural signal fluctuations during the withering period.

[0119] Step S148: Integrate the adjusted early warning activation conditions, early warning effective time periods, and corresponding abnormal transmission links to form an early warning trigger node system. The early warning trigger node system includes the early warning activation conditions and timeliness requirements corresponding to each abnormal transmission link.

[0120] The seasonally adjusted early warning activation conditions and supplemented effective early warning periods are linked and integrated with the corresponding information on abnormal transmission links. Each early warning trigger node includes the adjusted abnormal change threshold, response change threshold, effective early warning period, corresponding abnormal transmission link number, and transmission direction. Arranged in the order of the ecological process abnormal transmission chain, a complete early warning trigger node system is formed, which is presented in the form of tables or graphs to show the early warning standards and time constraints for each abnormal transmission link.

[0121] Step S149: Associate each early warning trigger node in the early warning trigger node system with the spatial partition of the target wetland area to define the early warning trigger nodes corresponding to different spatial partitions, integrate the early warning trigger node information of all spatial partitions, and generate wetland ecological security early warning rules. The wetland ecological security early warning rules include spatial partitions, early warning trigger nodes, early warning activation conditions, and early warning effective time periods.

[0122] Based on the spatial zoning of a wetland, each node in the early warning triggering node system is associated with a suitable spatial zone. The applicability of each early warning triggering node to each zone is determined according to the differences in ecological process characteristics across different spatial zones. For example, for spatial zones in the core water area of ​​the wetland, early warning triggering nodes related to hydrological connectivity processes are prioritized; for spatial zones in densely vegetated areas, early warning triggering nodes related to vegetation succession and soil-water interaction are prioritized. By integrating each spatial zone and its associated early warning triggering nodes, activation conditions, and effective time periods, a wetland ecological security early warning rule encompassing spatial positioning and early warning standards is formed.

[0123] Step S1491: Based on the topographic features and ecological function differences of the target wetland area, the target wetland area is divided into multiple spatial zones, each of which has relatively uniform ecological process characteristics.

[0124] Using digital elevation model data of a wetland, we analyzed its topographic features, such as differences in depressions, mudflats, and water bodies. Simultaneously, based on ecological function indicators such as vegetation distribution and hydrological connectivity, the wetland was divided into multiple spatial zones. For example, it was divided into a core water area, a marsh vegetation area, a mudflat transition zone, and a surrounding buffer zone. The topographic features and ecological processes (such as hydrology, vegetation, and soil-water interaction) within each spatial zone remained relatively uniform. A unique zone identifier was assigned to each spatial zone, and its geographic boundary coordinates and descriptions of major ecological processes were recorded.

[0125] Step S1492: For each spatial partition, collect historical ecological process signal data within that spatial partition, analyze the normal range of hydrological connectivity process signals, vegetation succession process signals, and soil-water interaction process signals within that spatial partition, and obtain the normal range of ecological process signals for each spatial partition.

[0126] For each spatial zone, hydrological connectivity process signals, vegetation succession process signals, and soil-water interaction process signals are extracted from the historical ecological process signal set of a specific wetland. Statistical analysis methods are used to calculate the mean, standard deviation, and fluctuation range of each signal within the zone, determining the numerical range of each signal under normal conditions within that zone, i.e., the normal range of ecological process signals. The normal ranges differ between different spatial zones; for example, the normal range of hydrological connectivity process signals in the core water area differs from that in the surrounding buffer zone, and therefore needs to be calculated and stored separately.

[0127] Step S1493: Analyze each early warning trigger node in the early warning trigger node system, and extract the early warning activation conditions corresponding to the early warning trigger node. The early warning activation conditions include the types of ecological process signals involved, the critical values ​​of ecological process signals, and the triggering time.

[0128] Each early warning trigger node in the early warning trigger node system is subjected to structured analysis, and the early warning activation condition elements contained therein are extracted using a node analysis algorithm. Specifically, the types of ecological process signals involved (such as hydrological connectivity process signals), the corresponding critical values ​​for abnormal changes and response changes, and the triggering timing requirement of "the previous signal meets the standard and the subsequent signal approaches it" are extracted to form an early warning activation condition list for each early warning trigger node. Each element in the list is associated with a node identifier.

[0129] Step S1494: Compare the critical value of the ecological process signal in the early warning activation condition of each early warning trigger node with the normal range of the ecological process signal of each spatial partition. If the normal range of the ecological process signal of any spatial partition matches the critical value of the ecological process signal in the early warning activation condition of the early warning trigger node, then associate the early warning trigger node with that spatial partition.

[0130] The abnormal change threshold and response change threshold of each early warning trigger node are compared with the normal range of ecological process signals corresponding to each spatial partition. If the normal range of a certain type of ecological process signal in a spatial partition matches the threshold value of the signal involved in the early warning trigger node in numerical logic (i.e., the threshold value is outside the normal range of the partition, which meets the anomaly judgment logic), then the early warning trigger node is determined to be compatible with the spatial partition, and the node is associated with the spatial partition. One early warning trigger node can be associated with multiple spatial partitions, and one spatial partition can also be associated with multiple early warning trigger nodes. The specific association relationship is determined based on the compatibility between the threshold value and the normal range of the partition.

[0131] Step S1495: For each spatial partition, count all early warning triggering nodes associated with that spatial partition, sort all early warning triggering nodes associated with that spatial partition according to the order of the ecological process abnormal transmission chain, and determine the monitoring priority of each early warning triggering node in that spatial partition.

[0132] For each spatial partition, all associated early warning triggering nodes are aggregated and sorted according to their order in the ecological process anomaly transmission chain. Early warning triggering nodes at the front end of the transmission chain correspond to the early stages of anomaly transmission and are assigned a "high" monitoring priority; nodes at the back end of the transmission chain correspond to the later stages of anomaly transmission and are assigned a "medium" or "low" monitoring priority. Simultaneously, the priorities are fine-tuned based on the current ecological security status description of the spatial partition. If an ecological process involving a node is in a vulnerable state within that partition, its monitoring priority is appropriately increased.

[0133] Step S1496: Supplement the geographic boundary coordinates of each spatial partition and its associated early warning triggering node. For each spatial partition's early warning triggering node, supplement the corresponding early warning adjustment parameters based on the dynamic characteristics of the spatial partition's ecological processes.

[0134] Add the geographic boundary coordinates of each spatial partition and its associated early warning triggering nodes to enable the early warning triggering nodes to be located within a specific spatial range. Based on the dynamic characteristics of the ecological processes of each spatial partition (such as the hydrological seasonal fluctuation characteristics of the core water area), supplement early warning adjustment parameters for each early warning triggering node within that partition. For example, for the core water area during the rainy season, supplement the critical value fine-tuning coefficient for hydrologically related early warning triggering nodes so that the critical value can adapt to the seasonal dynamic changes of the partition.

[0135] Step S1497: Integrate the geographic boundary coordinates of each spatial partition, the associated early warning triggering nodes, the monitoring priority of the early warning triggering nodes, and the early warning adjustment parameters to form an early warning sub-rule for each spatial partition.

[0136] The geographic boundary coordinates of each spatial partition, the list of associated early warning trigger nodes, the monitoring priority of each node, and the early warning adjustment parameters are integrated and packaged to form an independent early warning sub-rule for that spatial partition. The early warning sub-rule clearly defines the early warning nodes that need to be monitored within the partition, the order of monitoring, the adjustment standards for node parameters, and the spatial range.

[0137] Step S1498: Collect all warning sub-rules for all spatial partitions, and sort all warning sub-rules for all spatial partitions according to the ecological function priority of the spatial partitions.

[0138] Based on the ecological function importance of each spatial zone of a wetland, ecological function priorities are assigned; for example, the ecological function priority of the core water area is higher than that of the surrounding buffer zone. All early warning sub-rules for all spatial zones are then sorted in descending order of ecological function priority to ensure that, given limited resources, early warning monitoring is prioritized for zones with important ecological functions.

[0139] Step S1499: Integrate the early warning sub-rules of all spatial partitions after sorting, supplement the overall early warning response process framework, determine the response order when early warning trigger nodes of multiple spatial partitions are triggered simultaneously, and generate wetland ecological security early warning rules. The wetland ecological security early warning rules include spatial partition information, early warning trigger node information, early warning adjustment parameters, and response order.

[0140] All sorted spatial partition early warning sub-rules are integrated to form a unified rule system. Based on this, an overall early warning response process framework is added. This framework clarifies that when multiple spatial partitions' early warning trigger nodes simultaneously meet the activation conditions, the response order is determined according to the ecological function priority of the spatial partitions and the position of the early warning trigger nodes in the transmission chain; that is, partitions with higher ecological function priority and nodes at the front end of the transmission chain are responded to first. The final generated wetland ecological security early warning rules cover spatial partition information, details of early warning trigger nodes for each partition, early warning adjustment parameters, and the response order for multiple nodes triggering simultaneously.

[0141] Step S150: Obtain real-time multi-dimensional remote sensing data of the target wetland area, interpret the real-time multi-dimensional remote sensing data for ecological process signals to obtain a set of real-time ecological process signals, match the set of real-time ecological process signals with wetland ecological security early warning rules, and generate wetland ecological security early warning information for the target wetland area. The wetland ecological security early warning information includes abnormal ecological process links, abnormal transmission paths, and early warning response nodes.

[0142] In this embodiment, multi-dimensional remote sensing data of a wetland is acquired in real time using a remote sensing data receiving device, and the data is interpreted to obtain a set of real-time ecological process signals. After the real-time signals are split into spatial partitions, they are matched and checked against the warning trigger nodes of each partition in the wetland ecological security early warning rules. Nodes that meet the activation conditions are identified, and their corresponding abnormal ecological process links and abnormal transmission paths are traced. Combined with the early warning rules, the early warning response nodes are determined, and finally, wetland ecological security early warning information containing the above information is generated.

[0143] For example, step S151: acquire real-time multi-dimensional remote sensing data of the target wetland area, wherein the real-time multi-dimensional remote sensing data includes real-time multi-temporal imaging data of the target wetland area, real-time spectral radiation data of the target wetland area, and real-time microwave scattering data of the target wetland area.

[0144] Real-time multi-dimensional remote sensing data of a wetland was acquired by establishing a real-time data transmission link with a remote sensing satellite ground receiving station. Real-time multi-temporal imaging data was transmitted back in real-time by an optical remote sensing satellite at a set shooting frequency, covering current and recent optical imaging results of the wetland. Each frame of the image is accompanied by a precise shooting timestamp and geographic coordinates. Real-time spectral radiometric data was acquired in real-time by a hyperspectral remote sensing sensor, containing real-time radiometric values ​​of different land features within the wetland in multiple spectral bands, and the data was transmitted in real-time in band order. Real-time microwave scattering data was acquired in real-time by a synthetic aperture radar, generating scattering data reflecting the current soil moisture and water distribution of the wetland based on the scattering characteristics of radar waves. The data includes metadata information such as polarization mode and imaging resolution. During the acquisition process, a data verification mechanism was used to check the integrity of the data and transmission errors in real time, ensuring that the real-time data could be directly used for subsequent processing.

[0145] Step S152: Perform preprocessing on the real-time multi-dimensional remote sensing data, interpret the preprocessed real-time remote sensing data to obtain a set of real-time ecological process signals, which includes real-time hydrological connectivity process signals, real-time vegetation succession process signals and real-time soil-water interaction process signals.

[0146] First, preprocessing operations are performed on the real-time multi-dimensional remote sensing data. For real-time multi-temporal imaging data, the same radiometric correction method as in step S1131 is used, including atmospheric correction, sensor correction, and topographic correction, to eliminate interference factors in the real-time imaging process. For real-time spectral radiometric data, band calibration and radiometric normalization are performed to ensure that the radiometric values ​​of different bands are in a uniform dimension. For real-time microwave scattering data, noise removal and geometric correction are performed to correct geometric distortions in the radar imaging process. After preprocessing, the preprocessed real-time remote sensing data is interpreted using the same interpretation method as in step S110: real-time hydrological connectivity process signals are obtained from the real-time multi-temporal imaging data through temporal feature extraction; real-time vegetation succession process signals are obtained from the real-time spectral radiometric data through band feature analysis; and real-time soil-water interaction process signals are obtained from the real-time microwave scattering data through scattering coefficient inversion. These three types of real-time signals are integrated to form a set of real-time ecological process signals, each signal carrying a time and spatial identifier corresponding to the real-time data.

[0147] Step S153: Read the wetland ecological security early warning rules, extract multiple spatial partitions, the early warning trigger nodes associated with each spatial partition, and the early warning activation conditions corresponding to each early warning trigger node from the wetland ecological security early warning rules.

[0148] The generated wetland ecological security early warning rules are read through the rule parsing module. These rules are stored in a structured data format, including the rule version number, generation time, and complete rule content. During parsing, the system first extracts all spatial partition information for a given wetland as defined in the rules, including the geographic boundary coordinates, partition identifier, and ecological function description for each partition. Then, for each spatial partition, all associated early warning triggering nodes are extracted, including node number, corresponding anomaly transmission link, and monitoring priority. Finally, for each early warning triggering node, the early warning activation conditions are extracted, including the types of ecological process signals involved, anomaly change thresholds, response change thresholds, and effective warning period. The extracted information is then categorized and organized by spatial partition to form structured rule extraction results, facilitating matching with real-time ecological process signal sets.

[0149] Step S154: Split the real-time ecological process signal set according to spatial partitions to obtain a subset of real-time ecological process signals for each spatial partition. Each subset of real-time ecological process signals includes real-time hydrological connectivity process signals, real-time vegetation succession process signals, and real-time soil-water interaction process signals within the spatial partition.

[0150] Based on the spatial partition geographic boundary coordinates extracted from the results according to the rules, the real-time ecological process signal set is spatially segmented. A spatial matching algorithm is used to compare the spatial coordinates of each signal in the real-time ecological process signal set with the geographic boundaries of each spatial partition to determine the spatial partition to which the signal belongs. Real-time hydrological connectivity process signals, real-time vegetation succession process signals, and real-time soil-water interaction process signals belonging to the same spatial partition are categorized and integrated to form a subset of real-time ecological process signals for that spatial partition. Each subset is labeled with a corresponding spatial partition identifier and contains complete data of the three types of real-time signals within that partition, ensuring that the real-time signals of each spatial partition can be independently used for early warning matching within that partition.

[0151] Step S155: For each spatial partition, the real-time ecological process signal subset is matched with the early warning activation conditions of the early warning triggering node associated with that spatial partition. Check whether the real-time ecological process signal in the real-time ecological process signal subset reaches the ecological process signal threshold value in the early warning activation conditions and whether it is within the effective period of the early warning activation conditions.

[0152] For each spatial partition's real-time ecological process signal subset, the warning activation conditions of each warning trigger node associated with that partition are sequentially retrieved for matching checks. First, it is checked whether the current time falls within the valid warning period specified in the warning activation conditions. If not, the warning trigger node is directly determined not to meet the activation conditions. If it does fall within the valid period, the real-time ecological process signals corresponding to the warning activation conditions are further extracted from the real-time ecological process signal subset. The extracted real-time signal strength is compared with the abnormal change threshold and response change threshold in the warning activation conditions to check whether the real-time signal has reached the abnormal change threshold and whether the corresponding associated signals show a trend towards the response change threshold. For example, if the warning trigger node involves the transmission relationship between hydrological connectivity process signals and vegetation succession process signals, it is checked whether the real-time hydrological connectivity process signal has reached the abnormal change threshold, and simultaneously, whether the real-time vegetation succession process signal shows a trend towards the response change threshold.

[0153] Step S156: If the real-time ecological process signal in the real-time ecological process signal subset of any spatial partition reaches the warning activation condition of any warning trigger node associated with that spatial partition, then mark the warning trigger node as a triggered state and record the abnormal transmission link corresponding to the warning trigger node in the triggered state.

[0154] During the matching and checking process, if a subset of real-time ecological process signals in a spatial partition is found to meet all the early warning activation conditions of an early warning trigger node associated with that partition, the early warning trigger node is marked as triggered using the status marking module, and a trigger timestamp and its spatial partition identifier are added to it. Simultaneously, based on the association between the early warning trigger node and the anomaly propagation link, the anomaly propagation link information corresponding to the trigger node is extracted from the early warning rules, including the preceding process signal, the following process signal, and the propagation direction. This information, along with the trigger node's identifier and trigger time, is recorded in the trigger event log to ensure that each triggered node has a corresponding traceable anomaly propagation link record.

[0155] Step S157: Track the position of the early warning trigger node in the ecological process anomaly transmission chain, and determine the anomaly transmission path by combining the changes of other real-time ecological process signals in the real-time ecological process signal set.

[0156] Based on the abnormal transmission link corresponding to the warning trigger node in the trigger state, the specific position of this link in the abnormal transmission chain of the ecological process is located, clarifying whether it is the starting link, intermediate link, or terminal link of the transmission chain. If the trigger node corresponds to the starting link, other real-time signals related to the subsequent process signal in the real-time ecological process signal set are retrieved to check for any abnormal change trends in subsequent signals. If the trigger node corresponds to the intermediate link, the real-time signal changes of its preceding transmission links and subsequent transmission links are traced simultaneously. Combining these real-time signal changes, a complete transmission sequence from the starting process signal to the final influencing process signal is identified, for example, "abnormal real-time hydrological connectivity process signal → abnormal real-time vegetation succession process signal → abnormal real-time soil-water interaction process signal." This transmission sequence is the current abnormal transmission path. During the determination process, it is necessary to verify whether the order of occurrence of abnormal real-time signals in each link is consistent with the logical order of the transmission chain to ensure the accuracy of the abnormal transmission path.

[0157] Step S158: Based on the early warning triggering node and abnormal transmission path of the triggering state, determine the corresponding early warning response node, wherein the early warning response node is the ecological link that needs to implement monitoring enhancement or intervention measures.

[0158] Based on the location of the triggering node in the abnormal transmission chain and the complete sequence of the abnormal transmission path, the early warning response node is determined. If the triggering node is located at the beginning of the transmission chain, the early warning response node is set as the ecological process corresponding to the beginning of the chain and the immediately following subsequent links. High-frequency real-time monitoring of these links is required to track the progress of abnormal transmission in a timely manner. If the triggering node is located in the middle of the transmission chain, the early warning response node is set as the middle link and the subsequent links that have not experienced abnormalities. In addition to enhanced monitoring, intervention plans for this link need to be prepared. If the triggering node is located at the end of the transmission chain, the early warning response node is set as the end link and other related ecological processes that may be indirectly affected by it. Enhanced monitoring needs to be initiated immediately and the necessity of intervention measures needs to be assessed. Each early warning response node has a clearly defined corresponding monitoring frequency adjustment plan and potential intervention measure type to ensure that the response measures are targeted.

[0159] Step S159: Integrate the marked abnormal transmission links, the determined abnormal transmission paths and the early warning response nodes to generate wetland ecological security early warning information for the target wetland area. The wetland ecological security early warning information includes abnormal ecological process links, abnormal transmission paths and early warning response nodes.

[0160] The system integrates the recorded abnormal transmission links corresponding to the triggering nodes of the early warning system, the determined complete abnormal transmission paths, and the set early warning response nodes to generate wetland ecological security early warning information for a specific wetland in a standardized format. The early warning information includes: a detailed list of the abnormal ecological process links, the abnormal characteristics, and the occurrence time of each abnormal link; a visual sequence diagram of the complete transmission logic from start to finish, with the real-time signal anomaly level of each link marked; and a section on early warning response nodes specifying the exact location of each response node, the required monitoring enhancement plan, and intervention recommendations. Furthermore, an early warning level is added to the early warning information, determined comprehensively based on the number of triggering nodes, the length of the abnormal transmission path, and the degree of anomaly, categorized into "Level 1 Early Warning," "Level 2 Early Warning," and "Level 3 Early Warning." The generated early warning information is pushed to the wetland ecological monitoring and management platform via a data interface.

[0161] Figure 2 The following is a schematic diagram of the hardware structure of a remote sensing-based wetland ecological security early warning system 100 for implementing the above-described remote sensing-based wetland ecological security early warning method, provided by an embodiment of the present invention. Figure 2 As shown, the wetland ecological security early warning system 100 based on remote sensing analysis may include a processor 110, a machine-readable storage medium 120, a bus 130, and a communication unit 140.

[0162] In the specific implementation process, one or more processors 110 execute computer-executable instructions stored in machine-readable storage medium 120, so that processor 110 can execute the wetland ecological security early warning method based on remote sensing analysis as described in the above method embodiment. Processor 110, machine-readable storage medium 120 and communication unit 140 are connected through bus 130. Processor 110 can be used to control the sending and receiving actions of communication unit 140.

[0163] The specific implementation process of processor 110 can be found in the various method embodiments executed by the wetland ecological security early warning system 100 based on remote sensing analysis, which are similar in principle and technical effect. This embodiment will not be repeated here.

[0164] It should be noted that, in order to simplify the description of this invention and thus aid in the understanding of one or more embodiments, the foregoing description of the embodiments of this invention sometimes combines multiple features into a single embodiment, drawing, or description thereof. Similarly, it should be noted that, in order to simplify the description of this invention and thus aid in the understanding of one or more embodiments, the foregoing description of the embodiments of this invention sometimes combines multiple features into a single embodiment, drawing, or description thereof.

Claims

1. A wetland ecological security early warning method based on remote sensing analysis, characterized in that, The method includes: A multi-dimensional remote sensing data set of the target wetland area is acquired, and ecological process signals are interpreted on the multi-dimensional remote sensing data set to obtain an ecological process signal set of the target wetland area. The multi-dimensional remote sensing data set includes multi-temporal imaging data, spectral radiation data and microwave scattering data of the target wetland area. The ecological process signal set includes hydrological connectivity process signals, vegetation succession process signals and soil-water interaction process signals. A mapping relationship model between ecological process signals and wetland ecological security status is established. The mapping relationship model is used to perform state matching on the set of ecological process signals to obtain a description of the current ecological security status of the target wetland area. The mapping relationship model is obtained by associating ecological process signals interpreted from historical remote sensing data with the results of field surveys of wetland ecological security during the same period. Cross-process correlation analysis is performed on each signal in the ecological process signal set to identify abnormal transmission relationships between different ecological process signals, determine the initiating process signal, intermediate transmission process signal and final influencing process signal of abnormal transmission, and form an abnormal transmission chain of ecological processes; Based on the described ecological process anomaly transmission chain and the current ecological security status description, an early warning trigger node system is constructed. This system is then combined with the dynamic characteristics of the ecological processes in the target wetland area to generate wetland ecological security early warning rules. The early warning trigger node system includes early warning activation conditions corresponding to each anomaly transmission link. Specifically, the ecological process anomaly transmission chain is disassembled, and each anomaly transmission link is separated. Each anomaly transmission link includes the transmission relationship between the previous process signal and the next process signal, resulting in multiple anomaly transmission links. Corresponding early warning activation conditions are set for each anomaly transmission link. Each early warning activation condition includes the type of ecological process signal, the ecological process signal threshold, and the triggering time. All early warning activation conditions corresponding to the anomaly transmission links are integrated and arranged in the order of the ecological process anomaly transmission chain to form initial early warning trigger nodes. Each initial early warning trigger node corresponds to an anomaly transmission link and a corresponding early warning activation condition. Based on the current ecological security status description of the target wetland area, the early warning activation conditions of the initial early warning trigger nodes are adjusted to obtain adjusted early warning trigger nodes. Real-time multi-dimensional remote sensing data of the target wetland area is acquired, and ecological process signals are interpreted from the real-time multi-dimensional remote sensing data to obtain a set of real-time ecological process signals. The set of real-time ecological process signals is matched with wetland ecological security early warning rules to generate wetland ecological security early warning information for the target wetland area. The wetland ecological security early warning information includes abnormal ecological process links, abnormal transmission paths, and early warning response nodes.

2. The wetland ecological security early warning method based on remote sensing analysis according to claim 1, characterized in that, The process of acquiring a multi-dimensional remote sensing data set of the target wetland area, interpreting the ecological process signals of the multi-dimensional remote sensing data set, and obtaining a set of ecological process signals of the target wetland area includes: The multi-dimensional remote sensing data set of the target wetland area is retrieved from the data storage system. The multi-dimensional remote sensing data set includes multi-temporal imaging data, spectral radiation data and microwave scattering data of the target wetland area. The multi-dimensional remote sensing data set is split according to data type, separating independent multi-temporal imaging data subsets, spectral radiometric data subsets, and microwave scattering data subsets; Temporal feature extraction is performed on the subset of multi-temporal imaging data to track the temporal change trajectory of water body boundaries and vegetation coverage areas within the target wetland area. The temporal change trajectory of water body boundaries and vegetation coverage areas within the target wetland area is converted into a raw signal that can reflect the hydrological connectivity status, thus obtaining the hydrological connectivity raw signal. Band feature analysis is performed on the subset of spectral radiation data to extract spectral parameters related to vegetation growth stage and vegetation type, and these parameters are converted into original signals that can reflect the vegetation succession state to obtain the original vegetation succession signal. Scattering coefficient inversion is performed on the microwave scattering data subset to obtain information on soil moisture distribution and water depth changes in the target wetland area, and then converted into a raw signal that can reflect the interaction state between soil and water, thus obtaining the raw signal of soil-water interaction. The original hydrological connectivity signals, vegetation succession signals, and soil-water interaction signals are purified and processed. Then, according to the spatial division and time sequence of the target wetland area, the purified hydrological connectivity signals, purified vegetation succession signals, and purified soil-water interaction signals are integrated to form an ecological process signal set for the target wetland area. The ecological process signal set includes hydrological connectivity process signals, vegetation succession process signals, and soil-water interaction process signals.

3. The wetland ecological security early warning method based on remote sensing analysis according to claim 2, characterized in that, The process involves extracting temporal features from the subset of multi-temporal imaging data, tracking the temporal changes in water body boundaries and vegetation cover areas within the target wetland region, and converting these temporal changes into raw signals that reflect the hydrological connectivity status. This yields the raw hydrological connectivity signal, which includes: Radiometric correction is performed on the subset of multi-temporal imaging data. Water body reflection features for each time period are extracted from the radiometrically corrected multi-temporal imaging data. Based on the difference in reflectivity between water bodies and other land features, a threshold segmentation method is used to identify the water body range for each time period, resulting in water body range maps for multiple time periods. By comparing the water body extent maps of adjacent time periods, the expansion or contraction area of ​​the water body extent is calculated, and the direction and amount of the expansion or contraction of the water body extent are recorded to obtain the time series change data of the water body extent. The vegetation cover reflectance features of each time period are extracted from the radiometrically corrected multi-temporal imaging data. Based on the near-infrared high reflectance characteristics of vegetation, the vegetation cover area of ​​each time period is identified, and vegetation cover range maps of multiple time periods are obtained. By analyzing the spatial overlap between the vegetation cover map and the water body map of the same period, the range of water accumulation area within the vegetation cover area is determined. The range of water accumulation area within the vegetation cover area reflects the hydrological connectivity within the wetland, and the temporal variation data of the water accumulation area is obtained. Calculate the connectivity ratio between the water body area and the water accumulation area area for each time period. The connectivity ratio is the ratio of the water accumulation area area to the total water body area. The connectivity ratio is used to reflect the degree of hydrological connectivity within the wetland. Arrange the time-series change data of water body range, water accumulation area, and connectivity ratio data according to the monitoring time period, and convert them into a continuous signal with the monitoring time period as the index and the change or ratio as the signal value to obtain the original hydrological connectivity signal.

4. The wetland ecological security early warning method based on remote sensing analysis according to claim 1, characterized in that, The process involves establishing a mapping model between ecological process signals and wetland ecological security status. This model is then used to perform state matching on the set of ecological process signals to obtain a description of the current ecological security status of the target wetland area, including: Historical remote sensing data of multiple historical wetland areas were collected. The historical remote sensing data included historical multi-temporal imaging data, historical spectral radiometric data and historical microwave scattering data of each historical wetland area. The historical remote sensing data is interpreted to obtain a set of historical ecological process signals for each historical wetland area. The set of historical ecological process signals includes historical hydrological connectivity process signals, historical vegetation succession process signals, and historical soil-water interaction process signals. Collect the results of field surveys on wetland ecological security in various historical wetland areas during the corresponding periods of historical remote sensing data. The results of the field surveys on wetland ecological security include assessments of hydrological system integrity, vegetation community stability, and soil-water interaction health. By binding the set of historical ecological process signals of each historical wetland area with the results of the field survey of wetland ecological security during the same period, multiple sets of historical correlation data are formed. Each set of historical correlation data includes a set of historical ecological process signals and the corresponding wetland ecological security status assessment results. The multiple sets of historical correlation data are divided into training datasets and validation datasets. The training datasets are used for training the parameters of the mapping relationship model, and the validation datasets are used for validating the performance of the mapping relationship model. The basic architecture for constructing the mapping relationship model includes a signal input module, a feature association module, and a state output module. Historical ecological process signals from the training dataset are input into the signal input module. The feature association module establishes the correlation between the features of the historical ecological process signals and the wetland ecological security status assessment results. The state output module outputs the ecological security status prediction results. The correlation parameters of the mapping relationship model infrastructure are iteratively optimized using the training dataset, so that the deviation between the ecological security status prediction results output by the state output module and the actual wetland ecological security field survey results in the training dataset is gradually reduced. The optimized mapping model is validated using a validation dataset. Historical ecological process signals from the validation dataset are input into the optimized mapping model. The prediction results output by the optimized mapping model are compared with the actual wetland ecological security field survey results in the validation dataset. If the deviation is within the preset acceptable range, the training of the mapping model is completed, and the mapping model between ecological process signals and wetland ecological security status is obtained. The set of ecological process signals of the target wetland area is input into the mapping relationship model. The feature association module calculates the degree of matching between the hydrological connectivity process signal, vegetation succession process signal, and soil-water interaction process signal in the ecological process signal set and different ecological security states. The state output module outputs the ecological security state with the highest degree of matching, thus obtaining a description of the current ecological security state of the target wetland area.

5. The wetland ecological security early warning method based on remote sensing analysis according to claim 4, characterized in that, The process involves binding the historical ecological process signal set of each historical wetland area with the results of field surveys on wetland ecological security during the same period to form multiple sets of historically correlated data, including: The historical ecological process signal set of each historical wetland area is divided into time periods. The continuous historical ecological process signal set is divided into multiple time period subsets of historical ecological process signal at fixed time intervals, and each historical ecological process signal subset corresponds to a specific monitoring time period. We collected the results of field surveys on wetland ecological security in each historical wetland area during each monitoring period, and performed feature extraction on the subset of historical ecological process signals for each monitoring period. We extracted the average intensity of historical hydrological connectivity process signals, the rate of change of historical vegetation succession process signals, and the distribution uniformity of historical soil-water interaction process signals during the monitoring period to obtain the historical signal feature group for the monitoring period. The results of the field survey on wetland ecological security during the same period were quantitatively processed. The assessment of the integrity of the hydrological system, the stability of the vegetation community, and the health of the soil-water interaction were converted into quantitative scores to obtain a quantitative score group of the safety status during the monitoring period. The quantitative score of each assessment dimension corresponds to a set numerical range. The historical signal feature groups of the same monitoring period are bound with the safety status quantitative scoring groups to form a set of historical associated data; For all monitoring periods in all historical wetland areas, time period segmentation, collection of wetland ecological security field survey results, extraction of historical signal feature groups, quantification of security status scoring groups, and data binding operations were performed to obtain multiple sets of historical associated data.

6. The wetland ecological security early warning method based on remote sensing analysis according to claim 1, characterized in that, The step of performing cross-process correlation analysis on each signal in the ecological process signal set to identify abnormal transmission relationships between different ecological process signals, determine the initiating process signal, intermediate transmission process signal, and final influencing process signal of abnormal transmission, and form an abnormal transmission chain of ecological processes includes: Extract the hydrological connectivity process signal, vegetation succession process signal, and soil-water interaction process signal from the ecological process signal set, and generate time-series variation curves of the hydrological connectivity process signal, vegetation succession process signal, and soil-water interaction process signal. The time-series variation curves are plotted with the monitoring period as the horizontal axis and the signal intensity as the vertical axis. The temporal correlation between hydrological connectivity process signals and vegetation succession process signals is calculated, the temporal correlation between hydrological connectivity process signals and soil-water interaction process signals is calculated, and the temporal correlation between vegetation succession process signals and soil-water interaction process signals is calculated. The temporal correlation is used to reflect the consistency of the changing trends of the two ecological process signals within the same time period. The standard correlation range between hydrological connectivity process signals and vegetation succession process signals, hydrological connectivity process signals and soil-water interaction process signals, and vegetation succession process signals and soil-water interaction process signals in normal wetland ecosystems was retrieved. The standard correlation range was obtained through statistical analysis of a large amount of historical monitoring data from normal wetlands. The calculated actual temporal correlation degree is compared with the corresponding standard correlation degree range, and abnormal correlation combinations that exceed the standard correlation degree range are marked. Each abnormal correlation combination contains two ecological process signals with abnormal correlation. For each abnormal association combination, trace the start time of the abnormal changes of the two ecological process signals with abnormal associations, and determine the ecological process signal that changes abnormally first as the potential initiating process signal and the ecological process signal that changes abnormally later as the potential affected process signal. Analyze the correlation between potential initiation process signals and other ecological process signals. If a potential initiation process signal affects not only the labeled potential affected process signal but also a third type of ecological process signal through the labeled potential affected process signal, then the labeled potential affected process signal is identified as an intermediate transmission process signal, and the third type of ecological process signal is identified as the final influencing process signal. If the potential initiation process signal directly affects two other ecological process signals, then based on the order in which the two other ecological process signals are affected, they are respectively identified as intermediate transmission process signals and final influencing process signals; According to the order of abnormal transmission, the determined initial process signal, intermediate transmission process signal and final impact process signal are connected to form an ecological process abnormal transmission chain. The ecological process abnormal transmission chain presents a complete transmission path of the abnormality from the initial link to the final impact link.

7. The wetland ecological security early warning method based on remote sensing analysis according to claim 6, characterized in that, The step of comparing the calculated actual temporal correlation degree with the corresponding standard correlation degree range, and marking abnormal correlation combinations that exceed the standard correlation degree range, includes: The data storage system retrieves the first standard correlation range between hydrological connectivity process signals and vegetation succession process signals in normal wetland ecosystems, the second standard correlation range between hydrological connectivity process signals and soil-water body interaction process signals, and the third standard correlation range between vegetation succession process signals and soil-water body interaction process signals. The actual temporal correlation between the calculated hydrological connectivity process signal and the vegetation succession process signal is compared with the first standard correlation range. If the actual temporal correlation between the hydrological connectivity process signal and the vegetation succession process signal is less than the lower limit of the first standard correlation range or greater than the upper limit of the first standard correlation range, then the correlation combination is determined to be an abnormal correlation combination and marked as the first abnormal correlation combination. The first abnormal correlation combination includes the hydrological connectivity process signal and the vegetation succession process signal. The actual temporal correlation between the calculated hydrological connectivity process signal and the soil-water interaction process signal is compared with the second standard correlation range. If the actual temporal correlation between the hydrological connectivity process signal and the soil-water interaction process signal exceeds the upper and lower limits of the second standard correlation range, the correlation combination is determined to be an abnormal correlation combination and marked as the second abnormal correlation combination. The second abnormal correlation combination includes the hydrological connectivity process signal and the soil-water interaction process signal. The actual temporal correlation between the calculated vegetation succession process signal and the soil-water interaction process signal is compared with the third standard correlation range. If the actual temporal correlation between the vegetation succession process signal and the soil-water interaction process signal exceeds the upper and lower limits of the third standard correlation range, the correlation combination is determined to be an abnormal correlation combination and marked as the third abnormal correlation combination. The third abnormal correlation combination includes the vegetation succession process signal and the soil-water interaction process signal. Record the actual temporal correlation degree value, the corresponding standard correlation degree range, and the extent to which the correlation degree exceeds the standard correlation degree range for each abnormal correlation combination; Analyze the relationship between the excess magnitude of each anomalous association combination and the changing trend of ecological process signals, and mark potential anomalous association combinations and confirmed anomalous association combinations based on the analysis results; Integrate all marked first, second, and third anomalous association combinations, distinguish between potential and confirmed anomalous association combinations, and form a list of anomalous association combinations. Each anomalous association combination contains two ecological process signals with anomalous associations and a description of the degree of anomalousness.

8. The wetland ecological security early warning method based on remote sensing analysis according to claim 1, characterized in that, Based on the description of the abnormal transmission chain of the ecological process and the current ecological security status, an early warning triggering node system is constructed. This system is then combined with the dynamic characteristics of the ecological processes in the target wetland area to generate wetland ecological security early warning rules, including: The abnormal transmission chain of the ecological process is disassembled, and each abnormal transmission link is separated. Each abnormal transmission link contains the transmission relationship between the signal of the previous process and the signal of the next process, resulting in multiple abnormal transmission links. For each abnormal transmission link, the critical value of abnormal change of the previous process signal and the critical value of response change of the next process signal in the abnormal transmission link are determined. The critical value of abnormal change is the limit value of the previous process signal deviating from the normal range, and the critical value of response change is the limit value of the next process signal deviating from the normal range after being affected. The condition that the previous process signal reaches the critical value of abnormal change and the subsequent process signal begins to approach the critical value of response change is set as the early warning activation condition corresponding to the abnormal transmission link. Each early warning activation condition includes the type of ecological process signal, the critical value of the ecological process signal and the triggering time. Integrate the early warning activation conditions corresponding to all abnormal transmission links, arrange them in the order of the abnormal transmission chain of ecological process, and form initial early warning trigger nodes. Each initial early warning trigger node corresponds to an abnormal transmission link and the corresponding early warning activation conditions. Based on the description of the current ecological security status of the target wetland area, the warning activation conditions of the initial warning trigger node are adjusted to obtain the adjusted warning trigger node; Extract the dynamic characteristics of ecological processes in the target wetland area. These dynamic characteristics include the seasonal fluctuation patterns of hydrological connectivity processes, the growth cycle patterns of vegetation succession processes, and the annual variation patterns of soil-water interaction processes. The adaptation of the early warning activation conditions of each adjusted early warning trigger node to the dynamic characteristics of the ecological processes in the target wetland area is analyzed. If the critical value of the ecological process signal in the early warning activation conditions of the adjusted early warning trigger node conflicts with the seasonal fluctuation pattern, the critical value of the ecological process signal is adjusted in segments according to the season, and an effective early warning period is added to each adjusted early warning trigger node. The integrated and adjusted early warning activation conditions, effective early warning periods, and corresponding anomaly transmission links form an early warning trigger node system. The early warning trigger node system includes the early warning activation conditions and timeliness requirements corresponding to each anomaly transmission link. Each early warning trigger node in the early warning trigger node system is associated with a spatial partition of the target wetland area to define early warning trigger nodes corresponding to different spatial partitions. The early warning trigger node information of all spatial partitions is integrated to generate wetland ecological security early warning rules. The wetland ecological security early warning rules include spatial partitions, early warning trigger nodes, early warning activation conditions, and early warning effective time periods.

9. The wetland ecological security early warning method based on remote sensing analysis according to claim 8, characterized in that, The process involves associating each early warning trigger node in the early warning trigger node system with a spatial partition of the target wetland area to define early warning trigger nodes corresponding to different spatial partitions, integrating the early warning trigger node information of all spatial partitions, and generating wetland ecological security early warning rules, including: Based on the topographic features and ecological function differences of the target wetland area, the target wetland area is divided into multiple spatial zones, each with relatively uniform ecological process characteristics. For each spatial partition, historical ecological process signal data within that spatial partition are collected, and the normal ranges of hydrological connectivity process signals, vegetation succession process signals, and soil-water interaction process signals within that spatial partition are analyzed to obtain the normal ranges of ecological process signals for each spatial partition. Each early warning trigger node in the early warning trigger node system is analyzed, and the early warning activation conditions corresponding to the early warning trigger node are extracted. The early warning activation conditions include the types of ecological process signals involved, the critical values ​​of ecological process signals, and the triggering time. The critical value of the ecological process signal in the early warning activation condition of each early warning trigger node is compared with the normal range of the ecological process signal of each spatial partition. If the normal range of the ecological process signal of any spatial partition matches the critical value of the ecological process signal in the early warning activation condition of the early warning trigger node, then the early warning trigger node is associated with that spatial partition. For each spatial partition, all early warning triggering nodes associated with that spatial partition are counted, and all early warning triggering nodes associated with that spatial partition are sorted according to the order of the abnormal transmission chain of ecological processes to determine the monitoring priority of each early warning triggering node in that spatial partition. For each spatial partition and its associated early warning triggering node, the geographic boundary coordinates of the spatial partition are supplemented. For each spatial partition's early warning triggering node, the corresponding early warning adjustment parameters are supplemented based on the dynamic characteristics of the spatial partition's ecological processes. The geographic boundary coordinates of each spatial partition, the associated early warning triggering nodes, the monitoring priority of the early warning triggering nodes, and the early warning adjustment parameters are integrated to form the early warning sub-rules for each spatial partition; Collect all warning sub-rules for all spatial partitions, and sort all warning sub-rules for all spatial partitions as a whole according to the ecological function priority of the spatial partitions; The early warning sub-rules of all spatial partitions are integrated and sorted to supplement the overall early warning response process framework, so as to determine the response order when the early warning trigger nodes of multiple spatial partitions are triggered simultaneously, and generate wetland ecological security early warning rules. The wetland ecological security early warning rules include spatial partition information, early warning trigger node information, early warning adjustment parameters and response order.

10. A wetland ecological security early warning system based on remote sensing analysis, characterized in that, The wetland ecological security early warning system based on remote sensing analysis includes a processor and a memory, the memory and the processor being connected. The memory is used to store programs, instructions or code, and the processor is used to run the programs, instructions or code in the memory to implement the wetland ecological security early warning method based on remote sensing analysis as described in any one of claims 1-9.

Images