Multi-scene ecological restoration effect data monitoring decision regulation method and system

Abstract

The application discloses a kind of multi-scene ecological restoration effect data monitoring decision regulation methods and systems, it is related to ecological restoration technical field;The method comprises: obtaining the growth condition data of target green plant in the first scene and the second scene of target monitoring area based on different ecological evaluation rules division;Calculate the state difference parameters between two scenes;According to the parameter diagnosis ecological restoration abnormal factors and corresponding measure defects;According to defect, targeted compensation recovery measures are executed, and when the data of two scenes all reach dynamic minimum evaluation threshold, it is judged that recovery effect reaches standard.The application realizes the fusion and cross-validation of multi-source ecological data through multi-scene parallel monitoring and difference analysis, overcomes the limitations of traditional single scene monitoring, can more accurately diagnose problems and trace to management measures, forms closed-loop management from intelligent monitoring, accurate diagnosis to adaptive regulation, significantly improves the scientificity and effect of ecological restoration management.

Description

Technical Field

[0001] This invention relates to the field of ecological restoration, specifically to a method and system for monitoring, decision-making, and regulating the effects of ecological restoration in multiple scenarios. Background Technology

[0002] Ecological restoration is a crucial process for repairing degraded ecosystems and enhancing their biodiversity and service functions. In this process, continuous and precise monitoring and evaluation of restoration effects, and timely adjustments to restoration measures based on the evaluation results, are core elements to ensure project success and sustainable ecological benefits. However, current monitoring of ecological restoration effects often relies on single scenarios or static evaluation rules. These methods are relatively limited in data collection dimensions and often fail to consider the complex interactions between different ecological factors during analysis. They are also prone to overlooking the inherent conflicts and interactions between different evaluation standards, leading to inaccurate diagnoses of the true state of the ecosystem.

[0003] Another major limitation of existing technical solutions is that their effect assessments often rely on fixed thresholds and linear analytical models. Ecosystems themselves are highly dynamic and complex, and their restoration process is influenced by a combination of factors, including climate, soil, biological factors, and human management. Static assessment systems cannot effectively adapt to changing environmental factors and different restoration stages, resulting in a lack of sufficient scientific rigor and foresight in the final assessment of restoration effectiveness. These problems collectively lead to relatively extensive management of ecological restoration projects, a lack of targeted adjustments, low overall restoration efficiency, and weak long-term sustainability of restoration effects. Therefore, there is an urgent need for an ecological restoration effect monitoring and decision-making method and system that can integrate multi-source data, adapt to dynamic changes, and provide intelligent diagnosis and control. Summary of the Invention

[0004] This invention provides a multi-scenario ecological restoration effect data monitoring, decision-making and control method, which solves the problems of single data collection dimension, ignored conflicting evaluation standards, and inaccurate state diagnosis caused by the reliance on a single scenario or static evaluation rules in traditional ecological restoration effect monitoring methods.

[0005] This invention is achieved through the following technical solution:

[0006] Firstly, this application provides a data monitoring, decision-making, and control method for ecological restoration effects across multiple scenarios, including:

[0007] Acquire growth status data of at least one target green plant in the target monitoring area in a first scenario; wherein, the first scenario is a target scenario for ecological assessment divided based on a first ecological assessment rule;

[0008] The growth status data of the target green plants in the target monitoring area in the second scene is obtained, and the status difference parameters are determined based on the growth status data of the target green plants in the first scene and the second scene; wherein, the second scene is a scene in which the same area is divided based on a second ecological assessment rule that is different from the first ecological assessment rule;

[0009] Based on the state difference parameters, at least one anomalous factor associated with ecological restoration is identified, as well as the defects in ecological restoration measures corresponding to the anomalous factor.

[0010] Based on the deficiencies of the restoration measures, compensatory restoration measures are implemented for the target monitoring area, and when the growth status data of the target monitoring area in both the first scenario and the second scenario meet the preset minimum evaluation threshold, the ecological restoration effect of the target monitoring area is determined to be satisfactory.

[0011] A further optimization scheme is that the target green plants are the plants in the disturbed area that represent the target monitoring area as being in an ecological restoration state in the ecological restoration assessment, and whose weight in the ecological assessment rules is higher than the average weight of all assessment items.

[0012] A further optimized solution is that the ecological assessment rules are generated through an ecological overlap model, which couples an ecological habitat prediction model, a soil erosion quantification model, and a soil nutrient model.

[0013] The ecological overlap model is configured as follows:

[0014] Based on ecological habitat trends and vegetation growth status data, diversity assessment rules are generated.

[0015] Soil erosion assessment rules are generated based on soil quality parameters and vegetation growth status data.

[0016] Soil quality gain assessment rules are generated based on soil fertility parameters and vegetation growth status data.

[0017] A further optimized scheme is as follows: Based on the state difference parameters, determining at least one anomalous factor associated with ecological restoration, and the corresponding deficiencies in ecological restoration measures, includes:

[0018] Based on the state difference parameters, at least one state abnormality feature of the target green plant is determined;

[0019] Based on the abnormal state characteristics, the restoration conflict phenomenon of the target green plant is determined;

[0020] Identify the ecological restoration measures that cause the restoration conflict, and determine the deficiencies of the restoration measures associated with the non-environmental factors based on the ecological restoration measures and the non-environmental factors experienced by the target vegetation.

[0021] A further optimized scheme includes, based on the state difference parameters, determining at least one anomalous factor associated with ecological restoration, and the defects in ecological restoration measures corresponding to the anomalous factor, further comprising:

[0022] The state difference parameters are input into a pre-trained multimodal anomaly identification model to extract anomalies and output the probability distribution of the anomalies; wherein, the anomalies include pest and disease infestation, nutrient imbalance, water stress or pollution accumulation;

[0023] By incorporating the aforementioned anomalous factors into a causal reasoning framework, and using Bayesian network analysis to analyze the correlation between the implementation records of ecological restoration measures and the aforementioned anomalous factors, the corresponding defects in ecological restoration measures can be identified.

[0024] A further optimization scheme is as follows: determining the state difference parameters based on the growth status data of the target green plant in the first scene and the second scene includes:

[0025] The ecological assessment rules for the first and second scenarios are broken down to determine assessment dimensions including growth indicators, environmental factors, and stress factors.

[0026] Based on the evaluation dimensions, determine the sample matching size of the target green plants and the attribution factors for differences in different evaluation dimensions;

[0027] Based on the sample matching quantity, a quartile matrix of the target green plant is constructed on different evaluation dimensions. The quartile matrix is ​​used to compare the evaluation weight, evaluation location, evaluation type and regional green plant content.

[0028] Based on the quartile matrix, the derived higher-order features of the target green plant are determined, and the growth trend difference between the derived higher-order features of the target green plant in the first scene and the second scene is calculated.

[0029] The state difference parameter is determined based on the growth trend difference.

[0030] A further optimized solution is that the compensation and recovery measures include:

[0031] Based on the type and severity of the deficiencies in the aforementioned recovery measures, an adaptive, personalized recovery plan is generated;

[0032] Implement multimodal interventions, including at least one of bioremediation, engineering adjustment, and management optimization.

[0033] A further optimized solution is that determining whether the ecological restoration effect of the target monitoring area meets the standard includes:

[0034] The minimum assessment threshold is dynamically calculated using a multiple regression model based on scenario-specific factors. These scenario-specific factors include differences in ecological assessment rules, characteristics of plant species, seasonal changes, and regional climate conditions.

[0035] The growth status data is decomposed into multiple sub-indicators, and weights are assigned to each sub-indicator. The overall ecological health score is generated by integrating these sub-indicators using a fuzzy logic algorithm. The sub-indicators include biomass growth rate, chlorophyll fluorescence value, soil organic matter content, and biodiversity index.

[0036] When the overall ecological health score exceeds the minimum assessment threshold calculated dynamically, and all core sub-indicators reach their individual thresholds, the ecological restoration effect is deemed to have met the standards.

[0037] Secondly, this application provides a multi-scenario ecological restoration effect data monitoring, decision-making, and control system, including:

[0038] The first scene module is used to acquire growth status data of at least one target green plant in the target monitoring area in the first scene; wherein, the first scene is a target scene for ecological assessment based on the first ecological assessment rule.

[0039] The second scene module is communicatively connected to the first scene module and is used to acquire growth status data of the target green plants in the second scene of the target monitoring area, and to determine the state difference parameters based on the growth status data of the target green plants in the first scene and the second scene; wherein, the second scene is a scene divided into the same area based on a second ecological assessment rule that is different from the first ecological assessment rule;

[0040] The analysis and diagnosis module is communicatively connected to the second scenario module and is used to determine at least one abnormal factor associated with ecological restoration and the defects of ecological restoration measures corresponding to the abnormal factor based on the state difference parameters.

[0041] The decision control module is communicatively connected to the analysis and diagnosis module. It is used to perform compensatory restoration measures on the target monitoring area based on the defects of the restoration measures, and to determine that the ecological restoration effect of the target monitoring area meets the preset minimum evaluation threshold when the growth status data of the target monitoring area in the first scenario and the second scenario both meet the preset minimum evaluation threshold.

[0042] Thirdly, this application provides a computer-readable storage medium storing a multi-scenario ecological restoration effect data monitoring, decision-making, and control program, wherein when the multi-scenario ecological restoration effect data monitoring, decision-making, and control program is executed by a processor, it implements the steps of the multi-scenario ecological restoration effect data monitoring, decision-making, and control method described above.

[0043] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0044] By employing parallel monitoring of the same area using two scenarios based on different ecological assessment rules and calculating their state difference parameters, this method achieves the fusion analysis and cross-validation of multi-source, multi-dimensional ecological data, overcoming the limitations of traditional single-scenario or static rule monitoring. This approach can more comprehensively and accurately diagnose abnormal factors in the ecological restoration process, trace their origins to specific deficiencies in restoration measures, and then generate and execute targeted compensation and restoration plans. Finally, the effectiveness is validated by requiring the growth data from both scenarios to meet a minimum assessment threshold for dynamic adaptation, forming a closed-loop management process from intelligent monitoring and precise diagnosis to adaptive regulation, thus improving the scientific rigor, relevance, and overall effectiveness of ecological restoration project management. Attached Figure Description

[0045] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort. In the drawings:

[0046] Figure 1 A flowchart of the multi-scenario ecological restoration effect data monitoring, decision-making, and control method provided in the embodiments of this application;

[0047] Figure 2 A schematic diagram illustrating the feature extraction and hybrid encoding generation process provided in the embodiments of this application;

[0048] Figure 3 A functional module block diagram of the multi-scenario ecological restoration effect data monitoring, decision-making and control system provided in the embodiments of this application. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0050] First, some of the technical terms used in this application will be explained to help those skilled in the art understand this application.

[0051] CNN: Convolutional Neural Network;

[0052] LSTM: Long Short-Term Memory;

[0053] AHP: Analytic Hierarchy Process;

[0054] pH: Potential of Hydrogen, hydrogen ion concentration index;

[0055] API: Application Programming Interface;

[0056] DNA: Deoxyribonucleic acid;

[0057] Firstly, such as Figure 1 As shown, this application provides a data monitoring, decision-making, and control method for ecological restoration effects in multiple scenarios, including the following steps:

[0058] Step S1: Obtain growth status data of at least one target green plant in the target monitoring area in the first scene; wherein, the first scene is a target scene for ecological assessment divided based on the first ecological assessment rule;

[0059] Step S2: Obtain the growth status data of the target green plants in the second scene of the target monitoring area, and determine the status difference parameter based on the growth status data of the target green plants in the first scene and the second scene; wherein, the second scene is a scene divided into the same area based on a second ecological assessment rule that is different from the first ecological assessment rule;

[0060] Step S3: Based on the state difference parameters, determine at least one abnormal factor associated with ecological restoration, and the defects of ecological restoration measures corresponding to the abnormal factor;

[0061] Step S4: Based on the deficiencies of the restoration measures, implement compensatory restoration measures for the target monitoring area, and determine that the ecological restoration effect of the target monitoring area meets the preset minimum evaluation threshold when the growth status data of the target monitoring area in both the first scenario and the second scenario meet the preset minimum evaluation threshold.

[0062] This embodiment uses two different sets of ecological assessment rules to divide and collect data in the same monitoring area in parallel, which establishes the basis for multi-perspective comparative analysis and can capture the true state and abnormal fluctuations of the ecosystem more comprehensively and sensitively.

[0063] Based on the quantified state differences, a deep diagnosis was made of the specific ecological restoration measures that led to the anomalies, thus achieving a leap from monitoring to root cause localization.

[0064] Based on the diagnostic results, targeted compensation and restoration were implemented. Innovatively, a threshold dynamically adjusted based on multi-scenario specific factors was used for dual effect verification, ensuring the effectiveness of each intervention measure and the scientific and robust nature of the final restoration effect judgment. This formed a complete decision support system from intelligent monitoring and precise diagnosis to optimized regulation and closed-loop verification, improving the level of refined management and the reliability of the results of ecological restoration projects.

[0065] In one embodiment, step S1: acquiring growth status data of at least one target green plant in the target monitoring area in the first scene, specifically includes the following steps:

[0066] Step S11: From the plant community in the target monitoring area, based on the weight system of the ecological assessment rules, select plants with a weight ratio higher than the average as target green plants, and include their data in a dynamically managed green plant ecological pool; wherein, the green plant ecological pool is a dynamic data storage and management system; as the applied ecological assessment rules change, the system automatically triggers data operations through a rule engine or a specific algorithm: target green plant data under the new rules is entered into the database, while data that is no longer relevant under the old rules is removed from the database (i.e., removed from the database), to ensure that the data in the pool always accurately matches the current assessment scenario;

[0067] Specifically, the target vegetation refers to plants in disturbed areas that represent the ecological restoration status of the target monitoring area in the ecological restoration assessment, and whose weight percentage of each assessment indicator is higher than the average weight of all assessment items in the ecological assessment rules used; the target monitoring area includes degraded areas affected by human or natural disturbances, such as along power transmission lines, mines, wetlands or deserts; the assessment indicators typically include biomass, coverage, etc.

[0068] Specifically, the green plant ecological pool is a dynamic data system for storing and managing data on the growth status of target green plants. As the applied ecological assessment rules change, the system automatically triggers data operations through a rule engine or a specific algorithm: target green plant data under the new rules is entered into the database, while data that is no longer relevant under the old rules is removed from the database, so as to ensure that the data in the pool always accurately matches the current assessment scenario.

[0069] Preferably, in ecological assessment, the weighting system of assessment rules can be established through multi-index weighting calculation methods, such as the analytic hierarchy process (AHP) or the entropy weighting method; selecting plants with weights higher than the mean as target plants can ensure that monitoring data efficiently reflects the core dynamics of ecological restoration; for example, in mine ecological restoration, tolerant plants that are highly indicative of heavy metal enrichment and have a high proportion in the weighting system can be given priority as target plants.

[0070] Step S12: Call the ecological overlap model that couples the ecological habitat prediction model, the soil erosion quantification model, and the soil nutrient model; the model generates diversity assessment rules, soil erosion assessment rules, and soil nutrient assessment rules based on ecological habitat trends, soil quantification parameters, and soil fertility material parameters, combined with vegetation growth status data; then, the target monitoring area is divided according to one of the generated rules to obtain the first scene.

[0071] Specifically, the ecological assessment rules are generated by an ecological overlap model, which is a composite model coupling the following three sub-models:

[0072] Ecological habitat prediction model: used to simulate and predict the suitability, distribution and changing trends of species habitats within a target area;

[0073] Soil erosion quantification model: used to calculate and assess the rate and amount of soil loss due to external forces such as water and wind;

[0074] Soil nutrient model: used to describe and predict the cycling, content and plant availability of nutrients in soil, such as nitrogen, phosphorus and potassium.

[0075] Specifically, the ecological overlap model generates three types of core evaluation rules for scenario segmentation based on different inputs and logic:

[0076] Generate diversity assessment rules: These rules are generated based on ecological habitat trends (e.g., species suitability distribution) and combined with historical or real-time vegetation growth data. Preferably, when the model detects a decline in habitat suitability and a decrease in species evenness, the weight of diversity restoration measures in the overall assessment can be automatically increased, and corresponding rules can be generated accordingly.

[0077] Soil erosion assessment rules are generated based on soil quality parameters (e.g., erosion modulus S) and vegetation growth status data (e.g., vegetation interception rate V). Specifically, the judgment is made by establishing a Y=f(S,V) model. If the calculated Y value exceeds the preset critical value, an assessment rule emphasizing the need to implement physical soil stabilization measures is generated.

[0078] Soil quality gain assessment rules are generated based on soil fertility parameters and vegetation growth status data; specifically, the judgment is made by calculating the nutrient cycling rate ΔN. ,in, To input nutrients, To lose nutrients; if If this occurs, it will trigger the generation of an assessment rule requiring additional fertilizer compensation;

[0079] Step S13: Within the defined first scene area, for the selected target green plants, use the deployed sensor network to collect growth data such as plant height, biomass, leaf area index, or chlorophyll content.

[0080] Specifically, based on a set of evaluation rules selected in step S12, such as soil attenuation rules or rules based on the degree of heavy metal pollution, the target monitoring area is spatially divided to clarify the geographical boundary of the first scenario; within this scenario, key growth indicators of the target green plants selected in step S11 are collected through pre-deployed sensor networks, remote sensing equipment, etc.

[0081] Specifically, the collected growth data includes, but is not limited to, directly measurable physical indicators such as plant height and canopy coverage, biomass estimated based on models, and physiological indicators obtained through techniques such as spectral analysis, such as leaf area index (LAI) and chlorophyll content. These comprehensive and multi-dimensional data together constitute the initial, quantifiable baseline information of the ecological restoration status in this first scenario.

[0082] In one embodiment, step S2: acquiring growth status data of the target green plant in the second scene within the target monitoring area, and determining the state difference parameter based on the growth status data of the target green plant in the first scene and the second scene, specifically includes the following steps:

[0083] Step S21: Apply another set of ecological assessment rules different from the first scenario to divide the same target monitoring area into a second scenario, and simultaneously collect the growth status data of the same target green plants in this scenario;

[0084] Specifically, within the same target monitoring area, another set of ecological assessment rules, different from those used in step S1, is invoked to re-divide the area, resulting in a second scenario. For example, if the first scenario is divided based on soil attenuation rules, the second scenario can be divided based on soil gain rules or diversity assessment rules. Specifically, through the deployed sensor network that is co-calibrated with the first scenario, the growth status data of the same target green plants in the second scenario are collected synchronously to ensure that the data is comparable in time.

[0085] Step S22: The first and second branch channels in parallel are used to process the data of the two scenes respectively. The first channel extracts the global features of the target green plants, and the second channel calculates the difference features relative to the global features. Linear weighted fusion and residual connection are used to verify the consistency of the features. If the verification is inconsistent, a hybrid encoding that integrates the interaction parameters of local features with different rules and the cross-scale feature fusion parameters is constructed to accurately represent ecological differences.

[0086] Specifically, a first branch channel and a second branch channel are set up in parallel; the first branch channel processes the first scene data and calculates the global feature vector of the target green plant through a feature extraction network. For example, convolutional neural networks (CNNs) can extract aggregate indicators such as canopy density and community connectivity that reflect the overall growth and spatial pattern.

[0087] Specifically, the second branch channel outputs the first channel's output. Using this as a benchmark, calculate the difference feature vector of the second scene data relative to this benchmark. The calculation formula is as follows:

[0088] Equation (1)

[0089] Specifically, the features of the two channels are linearly weighted and fused to obtain the fusion difference. Subsequently, the original difference characteristics are calculated through residual connections. fusion difference residuals between ;

[0090] Specifically, a consistency threshold is set. ;like If linear fusion is deemed sufficient to characterize the differences, then it can be used directly. or Further analysis will be conducted; if This indicates a complex nonlinear interaction between the features of the two scenes, triggering the construction of hybrid encoding. ;

[0091] Specifically, the hybrid encoding It is constructed using the following formula:

[0092] Equation (2)

[0093] in, This represents the interaction parameters of local features that are integrated from different ecological assessment rules (such as the soil rules in the first scenario and the hydrological rules in the second scenario). It represents a cross-scale feature fusion parameter that integrates target green plant features from different scales, such as microscopic (e.g., leaf chlorophyll) to macroscopic (e.g., canopy coverage); Represents the activation function. and These represent operators for feature interaction and cross-scale fusion, respectively.

[0094] Preferably, by constructing and applying hybrid coding, it is possible to capture latent ecological degradation or conflict situations that simple linear models cannot detect, making the features ultimately used to characterize ecological differences, i.e., state differences, more accurate and reliable; the specific process is as follows: Figure 2 As shown;

[0095] Step S23: Decompose the evaluation rules of the two scenarios to determine the evaluation dimensions. After completing the sample matching, construct the quartile matrix of the target green plant on different evaluation dimensions. Based on this, derive higher-order features and calculate the growth trend difference between the two scenarios. Finally, determine the quantified state difference parameters.

[0096] Specifically, the ecological assessment rules supporting the first and second scenarios are atomized and decomposed, and uniformly mapped to multiple assessment dimensions, including growth indicators such as plant height and biomass, environmental factors such as soil moisture and pH value, stress factors such as pest and disease index, indicator definition, data type, and collection frequency, to ensure cross-scenario comparability.

[0097] Specifically, cross-scenario sample matching is performed; for example, by calculating the covariate scores of each sample point on multiple evaluation dimensions, the sample with the closest covariate scores in the second scenario is selected for each sample in the first scenario to form a matching pair in order to control confounding factors.

[0098] Specifically, based on the matched sample data, a quartile matrix is ​​constructed for the target green plants; this matrix organizes the data from four core perspectives for systematic comparison: assessment weight (i.e., importance of each dimension), assessment location (i.e., geographical coordinates of the sampling point), assessment type (i.e., plant species), and assessment area target green plant content (i.e., abundance or coverage).

[0099] Specifically, based on the quartile matrix, higher-order features are derived through mathematical transformations, such as calculating the interaction terms and rates of change between different dimensional indicators. These features can more profoundly reflect the ecological state.

[0100] Specifically, for the same high-order feature, its changing trend with time series in the first and second scenarios is calculated respectively, and then the difference in growth trend between the two is obtained;

[0101] Finally, the state difference parameter D is calculated by weighted aggregation of multiple growth trend differences, or by using the following normalized distance formula:

[0102] Equation (3)

[0103] in, For the first The weight of each growth indicator, and These are the normalized data vectors for the first and second scenarios, respectively.

[0104] Furthermore, to assist in subsequent anomaly diagnosis, the specific dimensions of abnormal fluctuations in the state difference parameter D value can be correlated with environmental factors such as soil moisture content and pH value. If the correlation coefficient is greater than a preset threshold, for example, r is greater than 0.8, then the environmental factor can be initially identified as a potential abnormal correlation factor.

[0105] In one embodiment, step S3: Based on the state difference parameters, determining at least one anomalous factor associated with ecological restoration, and the deficiencies in ecological restoration measures corresponding to the anomalous factor, specifically includes the following steps:

[0106] Step S31: Analyze the abnormal state characteristics reflected by the state difference parameters, determine whether there is a restoration conflict, and combine non-environmental factors to initially trace the source to the corresponding ecological restoration measures defects;

[0107] Specifically, the abnormal state characteristics refer to the specific anomalies exhibited by the target green plants in cross-scene comparisons, quantified by the state difference parameters, such as a significantly lower biomass growth rate or increased spatial heterogeneity of chlorophyll fluorescence values.

[0108] Specifically, the aforementioned restoration conflict phenomenon refers to the phenomenon in which, during the ecological restoration process, multiple human-implemented restoration measures (i.e., non-environmental factors) fail to coordinate or even mutually restrict each other, or a single measure, while solving one problem, triggers a new one, resulting in the aforementioned abnormal state of the target green plants.

[0109] The key to identifying the restoration conflict phenomenon lies in determining whether the abnormal state is caused by the incoordination or side effects between human-implemented, controllable ecological restoration measures (i.e., non-environmental factors). Specifically, by associating and matching the identified abnormal state characteristics with a preset knowledge base of restoration measure effects or historical cases, it is possible to preliminarily infer the defects of ecological restoration measures that may cause conflicts. For example, excessive irrigation to conserve water may cause rhizosphere hypoxia, or the introduction of fast-growing species to achieve rapid greening may lead to nutrient competition with the target green plants.

[0110] Step S32: Input the state difference parameters and environmental variable data of the same period into the pre-trained multimodal abnormal factor identification model. The model adopts a structure combining convolutional neural network and long short-term memory network, and outputs abnormal factors such as pest and disease invasion and nutritional imbalance and their probability distribution.

[0111] Specifically, a multimodal anomaly identification model is constructed. The training dataset for this model comes from historical ecological restoration projects and integrates key data from multiple heterogeneous sources: First, there are state difference parameters, which serve as the core analysis object and quantify the degree of deviation of the ecological state under different assessment scenarios; second, there are various environmental variable data from the same period, including temperature, humidity, specific concentrations of nutrients such as nitrogen, phosphorus, and potassium in the soil, as well as precipitation pattern data that finely characterizes the temporal distribution and spatial intensity; finally, deeper plant physiological indicators need to be included, such as chlorophyll fluorescence parameters that reflect the functional status of the photosynthetic system and photosynthetic rate that directly characterizes light energy conversion efficiency; these data together constitute the factual basis for model training and optimization.

[0112] Specifically, the model employs a hybrid architecture combining a Convolutional Neural Network (CNN) and a Long Short-Term Memory (LSTM) network. Further, the CNN branch is responsible for extracting features from the spatial dimension, such as identifying patterns like lesion distribution and nutrient spatial heterogeneity in remote sensing images or sensor network data; the LSTM branch is responsible for extracting features from the time-series dimension to capture periodic or trend changes in growth indicators and environmental factors. The model processes the fused spatiotemporal features and outputs a probability distribution covering multiple potential abnormal factors, including pest and disease infestation, nutrient imbalance, water stress, and pollution accumulation. The output format, for example, shows a probability of 0.75 for water stress and 0.20 for nutrient imbalance, rather than a single determination, thus quantifying the likelihood of each abnormal factor.

[0113] Step S33: Introduce the identified abnormal factors and historical implementation records into the Bayesian network causal reasoning framework. By analyzing the conditional probability associations between variables, infer the specific ecological restoration measures that caused the abnormality, such as excessive fertilization or insufficient irrigation, or unreasonable vegetation configuration.

[0114] Specifically, a Bayesian network causal reasoning framework is constructed. The nodes in the network include: high-probability anomalies output in step S32, historical implementation records of various measures, such as the date, duration, and amount of irrigation, the type, time, and dosage of fertilizer, the species and density of vegetation configuration, and the defects of ecological restoration measures to be inferred, such as insufficient irrigation, excessive fertilization, and unreasonable vegetation configuration.

[0115] Specifically, based on historical data and expert knowledge, directed edges, i.e. causal relationships, and their conditional probability tables are predefined between network nodes. During inference, identified anomalous factors are input into the network as evidence, i.e., observed variables, such as water stress. Combined with known implementation records of measures, the posterior probability of each potential measure defect node is calculated using Bayesian inference methods such as variable elimination or Gibbs sampling. The measure defect node with the highest posterior probability is determined as the most likely root cause of the anomaly. For example, if water stress is observed and irrigation records show that the irrigation frequency in the past week was only 50% of the planned value, the network may output insufficient irrigation as the root measure defect with a high probability.

[0116] In one embodiment, step S4: Based on the diagnosed deficiencies in the restoration measures, compensatory restoration measures are implemented on the target monitoring area, and when the growth status data of the area in both the first and second scenarios reach the preset minimum evaluation threshold, the ecological restoration effect is determined to be satisfactory; this step specifically includes the following operations:

[0117] Step S41: Based on the specific type and quantified severity of the deficiencies in the recovery measures, formulate and implement an adaptive personalized recovery plan. This plan guides the implementation of multimodal intervention measures, including bioremediation, engineering adjustments, and management optimization. Specifically, the adaptive strategy is implemented through an intelligent decision-making system built on a rule base or machine learning model. This system takes the type and severity of the deficiencies in the recovery measures as input, matches a preset intervention measure template library, and integrates real-time environmental data of the target monitoring area to dynamically generate a personalized recovery plan.

[0118] Specifically, the adaptive compensation strategy is an intelligent decision-making system built on a rule base or machine learning model. The system's input is the type of restoration measure defect diagnosed in step S3, such as insufficient irrigation or excessive fertilization, and the severity of the defect. The severity of the defect can be quantified into a level of 1 to 5, or divided based on a specific range of the state difference parameter D value. Based on this input information, the system matches a preset intervention measure template library and integrates real-time environmental data of the target monitoring area, such as soil moisture and future weather forecasts, to dynamically generate a personalized restoration plan. For example, for a severe insufficient irrigation defect, the generated plan might include: immediately increasing the irrigation frequency to once a day and adjusting the irrigation time to the evening to reduce water evaporation, which is a management optimization; and simultaneously, planning to check and repair any potential water pipeline leaks in the area within a week, which is an engineering adjustment.

[0119] The multimodal intervention measures mainly include the following three categories:

[0120] Bioremediation refers to improving the soil microecological environment and promoting nutrient cycling by introducing native plant species that are native to the area, highly adaptable, or have specific remediation functions such as heavy metal enrichment or nitrogen fixation, or by inoculating beneficial microbial communities such as mycorrhizal fungi and nitrogen-fixing bacteria, thereby fundamentally enhancing the long-term resilience of the ecosystem.

[0121] Engineering adjustments: refer to directly changing site conditions through physical or chemical means, such as improving soil by importing soil to adjust soil pH, building infiltration ditches or reservoirs to repair hydrological structures, and installing shading facilities to reduce the strong light stress on seedlings.

[0122] Management optimization refers to adjusting the processes, plans, or agreements of human management activities, such as revising irrigation schedules and irrigation volumes, adjusting the types and timing of fertilizer application, and updating monitoring and integrated pest management standards.

[0123] Preferred adaptive compensation strategies may employ decision trees or case-based reasoning mechanisms; for example, when the diagnosed defect type is unreasonable vegetation configuration and related to nutrient competition, the system may prioritize recommending thinning of overly dense plants or replanting with symbiotic plants of the same family as bioremediation options, rather than simply suggesting supplemental fertilization.

[0124] Step S42: Taking into account the differences in ecological assessment rules, the characteristics of green plant species, seasonal and climatic conditions and other scenario-specific factors, a multiple regression model is used to dynamically calculate the minimum assessment threshold; at the same time, the growth status data obtained after implementing compensation measures is decomposed into multiple sub-indicators, and the overall ecological health score is calculated by integrating them using a fuzzy logic algorithm.

[0125] Specifically, a dynamic minimum evaluation threshold system is first defined; the core of this system is a multiple regression model, the formula of which can be expressed as:

[0126] Equation (4)

[0127] in, The minimum evaluation threshold is defined as follows; the independent variable X represents a series of scenario-specific factors, including the quantified value of the difference between the ecological assessment rules used in the first and second scenarios. Characteristics of green plant species For example, the coding of fast-growing or slow-growing varieties; seasonal changes. For example, the growing season or dormant season; regional climate conditions. For example, drought index;

[0128] The model's input parameters are further expanded to include historical ecological baseline data (i.e., background values), predicted future precipitation trends from climate model outputs, and socioeconomic factors such as project budget constraints that affect the setting of restoration targets. Through this model, the minimum threshold for achieving the target can be calculated to fit the current specific spatiotemporal conditions, thus eliminating the need to use fixed thresholds.

[0129] Specifically, a multi-dimensional comprehensive evaluation framework is used to quantitatively assess the effects of the compensation measures. This framework first decomposes the newly collected growth data into several independently measurable sub-indicators, such as the biomass growth rate reflecting plant productivity and the chlorophyll fluorescence value characterizing plant physiological health. / The indicators include soil organic matter content, which indicates soil fertility, and a biodiversity index, which measures ecosystem structure and function; each sub-indicator is assigned a weight based on its relative importance in overall ecological health. ;

[0130] Specifically, fuzzy logic algorithms are used to integrate the aforementioned sub-indicators. This process first defines fuzzy sets such as healthy, sub-healthy, and unhealthy for each sub-indicator, along with their corresponding membership functions. The measured value of each sub-indicator is converted into its membership degree to each fuzzy set. Subsequently, the system, based on a predefined fuzzy rule base (e.g., "if the biomass growth rate is healthy and the chlorophyll fluorescence value is healthy, then overall health is high"), combines the weights of each sub-indicator. Fuzzy reasoning is performed; finally, by using defuzzification methods such as the centroid method, a comprehensive overall ecological health score ranging from 0 to 100 is output.

[0131] This dynamic threshold model demonstrates stronger adaptability to different scenarios. For example, in arid and semi-arid regions, the model can automatically assign higher weights to sub-indicators such as water use efficiency, while automatically lowering the absolute threshold requirement for biomass growth rate during the dry season, in order to scientifically reflect the objective constraints of environmental stress on recovery potential.

[0132] Step S43: Compare the calculated overall ecological health score with the dynamically calculated minimum assessment threshold, and verify whether the measured value of each core sub-indicator reaches its individual threshold; only when the overall score exceeds the dynamic threshold and all core sub-indicators meet the standard can the ecological restoration effect of the target monitoring area be finally determined to be up to standard.

[0133] Specifically, a dual-condition judgment logic is implemented. First, the overall ecological health score calculated in step S42 is compared with the minimum assessment threshold dynamically calculated based on the scenario conditions during the same period. The overall score must be greater than the dynamic threshold. At the same time, it is necessary to verify whether the measured values ​​of each core sub-indicator, such as biomass growth rate and chlorophyll fluorescence value, have reached their respective preset individual thresholds. These individual thresholds can be determined based on the physiological limits of plant species, regional historical ecological baseline data, or the minimum recovery requirements set by the project. Only when both of the above conditions, namely, the overall health score meets the standard and all core sub-indicators meet the individual standards, are met simultaneously, will the system finally determine that the ecological restoration effect of the target monitoring area meets the standard. This dual judgment logic of overall compliance and key indicator qualification effectively prevents a serious shortcoming in a certain key ecological dimension from being masked by a high overall comprehensive score, thereby ensuring the comprehensiveness and robustness of the restoration effect judgment.

[0134] A preferred judgment strategy can also be set to a continuous monitoring mode; for example, the system can require that after the target monitoring area meets the dual judgment criteria for the first time, it must maintain the compliant state for several consecutive monitoring cycles, such as two months, before the recovery project can be finally confirmed as closed, thereby avoiding misjudgments caused by short-term fluctuations in the ecosystem or accidental factors.

[0135] Secondly, such as Figure 3 As shown, this application provides a multi-scenario ecological restoration effect data monitoring, decision-making, and control system, including:

[0136] The first scene module 100 is used to acquire growth status data of at least one target green plant in the target monitoring area in the first scene; wherein, the first scene is a target scene for ecological assessment based on the first ecological assessment rule.

[0137] The second scene module 200 is communicatively connected to the first scene module 100 and is used to acquire growth status data of the target green plants in the second scene of the target monitoring area, and determine the state difference parameters based on the growth status data of the target green plants in the first scene and the second scene; wherein, the second scene is a scene divided into the same area based on a second ecological assessment rule that is different from the first ecological assessment rule;

[0138] The analysis and diagnosis module 300 is communicatively connected to the second scenario module 200 and is used to determine at least one abnormal factor associated with ecological restoration and the defects of ecological restoration measures corresponding to the abnormal factor based on the state difference parameters.

[0139] The decision control module 400 is communicatively connected to the analysis and diagnosis module 300. It is used to perform compensatory restoration measures on the target monitoring area based on the defects of the restoration measures, and to determine that the ecological restoration effect of the target monitoring area meets the preset minimum evaluation threshold when the growth status data of the target monitoring area in the first scenario and the second scenario both meet the preset minimum evaluation threshold.

[0140] The functions of each module in the above-mentioned multi-scenario ecological restoration effect data monitoring, decision-making and control system correspond to the steps in the above-mentioned multi-scenario ecological restoration effect data monitoring, decision-making and control method embodiment. Their functions and implementation processes will not be described in detail here.

[0141] Thirdly, this application provides a multi-scenario ecological restoration effect data monitoring, decision-making and control device, which can be a personal computer (PC), laptop, server or other device with data processing capabilities.

[0142] In this embodiment, the multi-scenario ecological restoration effect data monitoring, decision-making, and control device may include a processor, a memory, a communication interface, and a communication bus.

[0143] The communication bus can be of any type and is used to interconnect the processor, memory, and communication interface.

[0144] The communication interface includes input / output (I / O) interfaces, physical interfaces, and logical interfaces. These interfaces are used for interconnecting internal components of the multi-scenario ecological restoration effect data monitoring, decision-making, and control equipment, as well as for interconnecting the equipment with other devices (such as other computing devices or user equipment). Physical interfaces can be Ethernet interfaces, fiber optic interfaces, ATM interfaces, etc.; user equipment can be displays, keyboards, etc.

[0145] Memory can be various types of storage media, such as random access memory (RAM), read-only memory (ROM), non-volatile RAM (NVRAM), flash memory, optical storage, hard disk, programmable ROM (PROM), erasable PROM (EPROM), electrically erasable PROM (EEPROM), etc.

[0146] The processor can be a general-purpose processor, which can call the multi-scenario ecological restoration effect data monitoring, decision-making, and control program stored in the memory, and execute the multi-scenario ecological restoration effect data monitoring, decision-making, and control method provided in the embodiments of this application. For example, the general-purpose processor can be a central processing unit (CPU). The method executed when the multi-scenario ecological restoration effect data monitoring, decision-making, and control program is called can be referred to the various embodiments of the multi-scenario ecological restoration effect data monitoring, decision-making, and control method of this application, and will not be repeated here.

[0147] Fourthly, embodiments of this application also provide a readable storage medium.

[0148] This application stores a multi-scenario ecological restoration effect data monitoring, decision-making, and control program on a readable storage medium. When the multi-scenario ecological restoration effect data monitoring, decision-making, and control program is executed by a processor, it implements the steps of the multi-scenario ecological restoration effect data monitoring, decision-making, and control method described above.

[0149] The method implemented when the multi-scenario ecological restoration effect data monitoring, decision-making and control procedure is executed can be referred to in various embodiments of the multi-scenario ecological restoration effect data monitoring, decision-making and control method of this application, and will not be repeated here.

[0150] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for monitoring, deciding, and controlling ecological restoration effects across multiple scenarios, characterized in that: include: Acquire growth status data of at least one target green plant in the target monitoring area in a first scenario; wherein, the first scenario is a target scenario for ecological assessment divided based on a first ecological assessment rule; The growth status data of the target green plants in the target monitoring area in the second scene is obtained, and the status difference parameters are determined based on the growth status data of the target green plants in the first scene and the second scene; wherein, the second scene is a scene in which the same area is divided based on a second ecological assessment rule that is different from the first ecological assessment rule; Based on the state difference parameters, at least one anomalous factor associated with ecological restoration is identified, as well as the defects in ecological restoration measures corresponding to the anomalous factor. Based on the deficiencies of the restoration measures, compensatory restoration measures are implemented for the target monitoring area, and when the growth status data of the target monitoring area in both the first scenario and the second scenario meet the preset minimum evaluation threshold, the ecological restoration effect of the target monitoring area is determined to be satisfactory.

2. The multi-scenario ecological restoration effect data monitoring, decision-making, and control method according to claim 1, characterized in that, The target green plants are those that represent the ecological restoration status of the target monitoring area in the ecological restoration assessment, and whose weight in the ecological assessment rules is higher than the average weight of all assessment items in the disturbed area.

3. The multi-scenario ecological restoration effect data monitoring, decision-making, and control method according to claim 1, characterized in that, The ecological assessment rules are generated through an ecological overlap model, which couples an ecological habitat prediction model, a soil erosion quantification model, and a soil nutrient model. The ecological overlap model is configured as follows: Based on ecological habitat trends and vegetation growth status data, diversity assessment rules are generated. Soil erosion assessment rules are generated based on soil quality parameters and vegetation growth status data. Soil quality gain assessment rules are generated based on soil fertility parameters and vegetation growth status data.

4. The multi-scenario ecological restoration effect data monitoring, decision-making, and control method according to claim 1, characterized in that, The step of determining at least one anomalous factor associated with ecological restoration, and the corresponding deficiencies in ecological restoration measures, based on the state difference parameters, includes: Based on the state difference parameters, at least one state abnormality feature of the target green plant is determined; Based on the abnormal state characteristics, the restoration conflict phenomenon of the target green plant is determined; Identify the ecological restoration measures that cause the restoration conflict, and determine the deficiencies of the restoration measures associated with the non-environmental factors based on the ecological restoration measures and the non-environmental factors experienced by the target vegetation.

5. The multi-scenario ecological restoration effect data monitoring, decision-making, and control method according to claim 4, characterized in that, The step of determining at least one anomalous factor associated with ecological restoration, and the corresponding deficiencies in ecological restoration measures, based on the state difference parameters, further includes: The state difference parameters are input into a pre-trained multimodal anomaly identification model to extract anomalies and output the probability distribution of the anomalies; wherein, the anomalies include pest and disease infestation, nutrient imbalance, water stress or pollution accumulation; By incorporating the aforementioned anomalous factors into a causal reasoning framework, and using Bayesian network analysis to analyze the correlation between the implementation records of ecological restoration measures and the aforementioned anomalous factors, the corresponding defects in ecological restoration measures can be identified.

6. The multi-scenario ecological restoration effect data monitoring, decision-making, and control method according to claim 1, characterized in that, The step of determining the state difference parameters based on the growth status data of the target green plant in the first scene and the second scene includes: The ecological assessment rules for the first and second scenarios are broken down to determine assessment dimensions including growth indicators, environmental factors, and stress factors. Based on the evaluation dimensions, determine the sample matching size of the target green plants and the attribution factors for differences in different evaluation dimensions; Based on the sample matching quantity, a quartile matrix of the target green plant is constructed on different evaluation dimensions. The quartile matrix is ​​used to compare the evaluation weight, evaluation location, evaluation type and regional green plant content. Based on the quartile matrix, the derived higher-order features of the target green plant are determined, and the growth trend difference between the derived higher-order features of the target green plant in the first scene and the second scene is calculated. The state difference parameter is determined based on the growth trend difference.

7. The multi-scenario ecological restoration effect data monitoring, decision-making, and control method according to claim 1, characterized in that, The compensation and recovery measures include: Based on the type and severity of the deficiencies in the aforementioned recovery measures, an adaptive, personalized recovery plan is generated; Implement multimodal interventions, including at least one of bioremediation, engineering adjustment, and management optimization.

8. The multi-scenario ecological restoration effect data monitoring, decision-making, and control method according to claim 1, characterized in that, The determination that the ecological restoration effect of the target monitoring area meets the standard includes: The minimum assessment threshold is dynamically calculated using a multiple regression model based on scenario-specific factors. These scenario-specific factors include differences in ecological assessment rules, characteristics of plant species, seasonal changes, and regional climate conditions. The growth status data is decomposed into multiple sub-indicators, and weights are assigned to each sub-indicator. The overall ecological health score is generated by integrating these sub-indicators using a fuzzy logic algorithm. The sub-indicators include biomass growth rate, chlorophyll fluorescence value, soil organic matter content, and biodiversity index. When the overall ecological health score exceeds the minimum assessment threshold calculated dynamically, and all core sub-indicators reach their individual thresholds, the ecological restoration effect is deemed to have met the standards.

9. A multi-scenario ecological restoration effect data monitoring, decision-making, and control system, characterized in that, include: The first scene module is used to acquire growth status data of at least one target green plant in the target monitoring area in the first scene; wherein, the first scene is a target scene for ecological assessment based on the first ecological assessment rule. The second scene module is communicatively connected to the first scene module and is used to acquire growth status data of the target green plants in the second scene of the target monitoring area, and to determine the state difference parameters based on the growth status data of the target green plants in the first scene and the second scene; wherein, the second scene is a scene divided into the same area based on a second ecological assessment rule that is different from the first ecological assessment rule; The analysis and diagnosis module is communicatively connected to the second scenario module and is used to determine at least one abnormal factor associated with ecological restoration and the defects of ecological restoration measures corresponding to the abnormal factor based on the state difference parameters. The decision control module is communicatively connected to the analysis and diagnosis module. It is used to perform compensatory restoration measures on the target monitoring area based on the defects of the restoration measures, and to determine that the ecological restoration effect of the target monitoring area meets the preset minimum evaluation threshold when the growth status data of the target monitoring area in the first scenario and the second scenario both meet the preset minimum evaluation threshold.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a multi-scenario ecological restoration effect data monitoring, decision-making, and control program, wherein when the multi-scenario ecological restoration effect data monitoring, decision-making, and control program is executed by a processor, it implements the steps of the multi-scenario ecological restoration effect data monitoring, decision-making, and control method as described in any one of claims 1 to 8.

Images