Land space improvement area ecological restoration effect evaluation system based on big data

Abstract

The invention relates to the technical field of territorial space ecological restoration, and discloses a territorial space improvement area ecological restoration effect evaluation system based on big data. The system continuously captures original monitoring data of ecological restoration through a multi-source ecological data acquisition engine to form a multi-source ecological data stream. And the time-space serialization processor carries out serialization recombination and standardization conversion on the data flow according to the time dimension and the space dimension to generate an ecological parameter sequence with unified time-space reference. And the ecological association network constructor analyzes the ecological elements in the sequence and constructs an ecological association network of an interaction relationship between the elements. And the repair path digger extracts a plurality of potential repair paths from the network, carries out path fusion and optimization, and generates a repair path set. And the dynamic evaluation generator calculates and outputs an ecological restoration effect evaluation index based on the path set and the real-time ecological state data. According to the system, dynamic and accurate evaluation of the ecological restoration effect is realized.

Description

Technical Field

[0001] This invention relates to the field of land space ecological restoration technology, specifically a big data-based system for evaluating the effectiveness of land space remediation regional ecological restoration. Background Technology

[0002] Current assessments of the effectiveness of territorial spatial ecological restoration primarily employ a combination of periodic on-site monitoring and remote sensing technology. These methods rely on limited data sources, such as ground sampling data or satellite imagery, and the assessment process is often based on static snapshots, lacking continuous dynamic monitoring capabilities. Existing technologies often use simple statistical models or geographic information system tools, processing data at single points in time or local areas, making it difficult to integrate multi-source heterogeneous data, such as sensor data, meteorological data, and soil data. Inconsistent data formats and spatiotemporal benchmarks lead to fragmented assessment results, failing to reflect the overall changing trends of the ecosystem.

[0003] Existing technologies have significant shortcomings in the data processing stage. Multi-source ecological data fails to undergo effective spatiotemporal alignment and standardization, fusion between different data sources is difficult, assessment models are simplistically linear, and the complex interactions between ecological elements are ignored. Restoration path planning typically relies on historical experience or qualitative analysis, lacking data-driven scientific mining and exhibiting weak dynamic adjustment capabilities. These deficiencies result in delayed assessment results, limited accuracy, and difficulty in guiding real-time restoration decisions. This invention addresses the difficulties in spatiotemporal integration of multi-source data and insufficient characterization of ecological element correlations in ecological restoration assessments. It aims to achieve unified spatiotemporal serialization processing of multi-source data and construct an ecological element interaction network to improve the accuracy and dynamism of the assessment. Summary of the Invention

[0004] The purpose of this invention is to provide a big data-based evaluation system for the ecological restoration effect of land space remediation areas, in order to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, this invention provides a big data-based system for evaluating the effectiveness of regional ecological restoration in land and space remediation, the system comprising: The system includes a multi-source ecological data acquisition engine, a spatiotemporal serialization processor, an ecological association network builder, a restoration path mining machine, and a dynamic evaluation generator; The multi-source ecological data acquisition engine continuously captures original monitoring data related to ecological restoration within the national land space, forming a multi-source ecological data stream; The spatiotemporal serialization processor receives the multi-source ecological data stream, performs serialization, recombination, and standardization transformation according to the time and spatial dimensions, and generates an ecological parameter sequence with a unified spatiotemporal reference. The ecological association network builder parses the ecological elements in the ecological parameter sequence and constructs an ecological association network that expresses the interaction relationships between the elements; The restoration path mining tool extracts multiple potential restoration paths from the ecological association network, and performs path fusion and optimization to generate a set of restoration paths; The dynamic evaluation generator calculates and outputs ecological restoration effect evaluation indicators based on the set of restoration paths and real-time ecological status data.

[0006] Preferably, the multi-source ecological data acquisition engine is configured to access satellite remote sensing data interfaces, ground sensor networks, and manual survey databases, and to perform data integrity verification and conflict resolution on the received raw monitoring data. The raw monitoring data includes at least vegetation coverage, soil physicochemical properties, water quality, and biodiversity indicators, which are then fused to form the multi-source ecological data stream.

[0007] Preferably, the spatiotemporal serialization processor is configured to perform time alignment processing on the multi-source ecological data stream, unifying data from different sources into the same timestamp sequence, and simultaneously perform spatial rasterization processing to map the data to a unified geographic grid cell, and perform missing value imputation and noise filtering on the ecological parameters within each grid cell to generate the ecological parameter sequence.

[0008] Preferably, the ecological association network builder is configured to identify nodes representing different ecological factors from the ecological parameter sequence, the ecological factors including vegetation type, soil moisture, and species abundance, analyze the statistical correlation or causal relationship between nodes, connect nodes with significant correlations with directed edges, and assign weights to each edge to represent the correlation strength, thereby constructing the ecological association network.

[0009] Preferably, the repair path miner is configured to execute a path search algorithm on the ecological association network, starting from a specified starting ecological node, traversing all possible paths to the target ecological node, calculating the overall weight and complexity of each path, sorting the paths according to a preset path quality criterion, filtering out key repair paths and merging redundant branches to form the repair path set.

[0010] Preferably, the step of traversing all possible paths to the target ecological node and calculating the overall weight and complexity of each path includes: starting from the initial ecological node, traversing the ecological association network using a depth-first search algorithm, and recording all complete paths that can reach the target ecological node; for each traversed path, summing the weights of all edges contained therein to obtain the overall weight of the path; simultaneously, counting the total number of nodes traversed by the path, and combining the total number of nodes with the overall weight to obtain the complexity of the path through a preset complexity calculation model.

[0011] Preferably, the dynamic evaluation generator is configured to obtain a snapshot of the ecological parameter sequence at the current moment, calculate the matching degree between the snapshot and the expected ecological state of each path in the set of restoration paths, and use a multi-index fusion algorithm to aggregate the matching degree results of each path into a comprehensive evaluation score, which is the ecological restoration effect evaluation index.

[0012] Preferably, the step of aggregating the matching degree results of each path into a comprehensive evaluation score using a multi-index fusion algorithm includes: obtaining the matching degree calculation result corresponding to each key restoration path in the restoration path set; assigning a preset weight coefficient to each ecological state indicator; multiplying the matching degree value of each ecological state indicator of each path by its corresponding weight coefficient to obtain a weighted matching degree; summing the weighted matching degrees of all ecological state indicators to calculate the individual evaluation score of the path; and arithmetically averaging the individual evaluation scores of all key restoration paths to obtain the comprehensive evaluation score used to characterize the overall restoration effect.

[0013] Preferably, the system further includes a restoration effect predictor, which is configured to receive the set of restoration paths and historical ecological parameter sequences, use a time series prediction model to simulate the changing trends of ecological parameters as they evolve along each restoration path at different future time points, and output restoration effect prediction data.

[0014] Preferably, the system further includes a visualization engine configured to overlay the ecological restoration effect assessment indicators and the restoration effect prediction data onto an electronic map to generate an interactive visualization report containing heat maps, path flow maps, and indicator change curves.

[0015] Compared with the prior art, the beneficial effects of the present invention are: The spatiotemporal serialization processor serializes and standardizes multi-source ecological data streams along both temporal and spatial dimensions, generating a sequence of ecological parameters with a unified spatiotemporal reference. This process, through temporal alignment and spatial interpolation, transforms heterogeneous data into a continuous and consistent sequence, eliminating format and benchmark differences between data sources. In effect, data consistency and comparability are enhanced, enabling long-term time-series and cross-regional analysis, reducing data noise and bias during the assessment process, and improving data utilization efficiency.

[0016] An ecological association network builder analyzes ecological elements in ecological parameter sequences and constructs an ecological association network expressing the interactions between these elements. This network employs graph models or similarity metrics to quantify the nonlinear relationships between elements such as vegetation cover and water quality parameters, dynamically capturing the feedback mechanisms of the ecosystem. In terms of effectiveness, the complex relationships of ecological processes are visualized, the assessment model more closely reflects actual ecosystem behavior, the scientific rigor and relevance of restoration strategies are enhanced, and more precise path optimization is supported. Attached Figure Description

[0017] Figure 1 This is a schematic diagram illustrating the working principle of the big data-based regional ecological restoration effect evaluation system for land space remediation as described in this invention. Figure 2 This is a schematic diagram illustrating the working principle of a spatiotemporal serialization processor. Figure 3 A diagram illustrating the working principle of an ecological network builder. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Please see Figure 1 This invention provides a big data-based system for evaluating the effectiveness of regional ecological restoration in land and space remediation. The system includes a multi-source ecological data acquisition engine, a spatiotemporal serialization processor, an ecological association network builder, a restoration path mining machine, and a dynamic evaluation generator.

[0020] A multi-source ecological data acquisition engine continuously captures raw monitoring data related to ecological restoration within the national land space, forming a multi-source ecological data stream. A spatiotemporal serialization processor receives the multi-source ecological data stream, performs serialization, recombination, and standardization transformation according to time and spatial dimensions, generating an ecological parameter sequence with a unified spatiotemporal reference. An ecological association network builder analyzes the ecological elements in the ecological parameter sequence and constructs an ecological association network expressing the interactions between these elements. A restoration path mining tool extracts multiple potential restoration paths from the ecological association network, performs path fusion and optimization, and generates a set of restoration paths. A dynamic evaluation generator, based on the set of restoration paths and combined with real-time ecological status data, calculates and outputs ecological restoration effectiveness evaluation indicators.

[0021] Example 1: See Figure 2The multi-source ecological data acquisition engine is configured to access satellite remote sensing data interfaces, ground sensor networks, and manual survey databases. It performs data integrity checks and conflict resolution on the received raw monitoring data, which includes at least vegetation cover, soil physicochemical properties, water quality, and biodiversity indicators. After fusion, a multi-source ecological data stream is formed. The spatiotemporal serialization processor is configured to perform time alignment processing on the multi-source ecological data stream, unifying data from different sources to the same timestamp sequence. Simultaneously, it performs spatial rasterization processing, mapping the data to unified geographic grid cells. For ecological parameters within each grid cell, it performs missing value imputation and noise filtering to generate an ecological parameter sequence.

[0022] In practice, the multi-source ecological data acquisition engine connects to various data sources through a pre-defined application programming interface, the satellite remote sensing data interface regularly receives optical and radar image data from various Earth observation satellites, the ground sensor network aggregates readings from various environmental monitoring sensors deployed in the field through the Internet of Things protocol, and the manual survey database stores the results of field verification and sampling analysis entered through mobile terminals. The multi-source ecological data acquisition engine performs data integrity verification on the received raw monitoring data, checking whether each data packet contains all the necessary fields and marking data packets missing key timestamps or geographic coordinates. At the same time, it performs conflict resolution. When there are significant differences in the values ​​of the same ecological indicator from different sources, the data is discarded or weighted averaged according to the preset reliability level of the data source. The raw monitoring data includes at least vegetation cover derived from the normalized vegetation index, soil physicochemical properties such as soil pH and organic matter content obtained from laboratory analysis, water quality indicators such as chemical oxygen demand and ammonia nitrogen content obtained from water quality monitoring instruments, and biodiversity indicators such as species quantity and distribution recorded through quadrat surveys. After verification and parsing, the data is integrated into a continuous multi-source ecological data stream.

[0023] In some embodiments, the spatiotemporal serialization processor performs time alignment processing on multi-source ecological data streams. Using a unified sampling frequency as a standard, it resamples data points with different original sampling frequencies, unifying all data onto the same timestamp sequence, for example, aligning all data to the time point of midnight each day. Simultaneously, the spatiotemporal serialization processor performs spatial rasterization processing, selecting a standard geographic coordinate system and a fixed spatial resolution to divide the entire study area into regular geographic grid cells, mapping each data point to its corresponding grid cell based on its geographic coordinates. It can be understood that the spatiotemporal serialization processor imputes missing values ​​for ecological parameters within each grid cell using the Kriging spatial interpolation method, estimating missing values ​​using known values ​​from surrounding grid cells; simultaneously, it performs noise filtering, applying a moving average filter to smooth out outliers caused by short-term fluctuations or measurement errors, ultimately generating a spatiotemporally consistent sequence of ecological parameters.

[0024] It is understandable that the processing quality of the spatiotemporal serialization processor directly affects the accuracy of subsequent analysis. Therefore, a data quality scoring function is defined to quantify the data reliability of each grid cell at each time point. The expression for the data quality scoring function is: in, Indicates the first Each grid cell in time Data quality score, The score representing the data integrity of a given unit at a given moment is calculated as the ratio of valid data points to theoretical data points. The data smoothness score is obtained by normalizing the inverse of the variance of the filtered data. and These are pre-defined weighting coefficients for completeness and smoothness. The spatiotemporal serialization processor will append the data quality score as metadata to the ecological parameter sequence for reference by subsequent modules.

[0025] In some embodiments, the data transfer between the multi-source ecological data acquisition engine and the spatiotemporal serialization processor adopts an asynchronous message queue mechanism. The multi-source ecological data acquisition engine publishes the processed multi-source ecological data stream to a message topic, and the spatiotemporal serialization processor, as a subscriber, consumes data from the topic for processing. This decoupled design allows the two modules to be expanded and maintained independently. During the data integrity verification phase, the multi-source ecological data acquisition engine generates a globally unique identifier for each valid data record and stores this identifier, along with the data itself, data source information, and data quality identifier, in a distributed file system, forming a complete data traceability chain. When performing spatial rasterization, the spatiotemporal serialization processor uses an area weighting method to allocate the contribution of data points falling on the grid boundary to multiple adjacent grid cells to improve the accuracy of spatial allocation.

[0026] Example 2: See Figure 3 The ecological association network builder is configured to identify nodes representing different ecological factors from the ecological parameter sequence. These ecological factors include vegetation type, soil moisture, and species abundance. It analyzes the statistical correlation or causal relationship between nodes, connects nodes with significant correlations with directed edges, and assigns weights to each edge to represent the correlation strength, thereby constructing an ecological association network.

[0027] In its implementation, the ecological association network builder extracts data from the ecological parameter sequence generated by the spatiotemporal serialization processor. Each data point in the ecological parameter sequence corresponds to the ecological parameter measurement value of a specific geographic grid unit at a specific timestamp. The ecological association network builder identifies nodes representing different ecological factors, including vegetation type classification codes interpreted from remote sensing imagery, soil moisture volume percentages obtained from soil sensors, and species abundance counts obtained from field surveys. Each independent ecological factor is defined as a node, and each node has type and numerical attributes.

[0028] In some embodiments, the ecological correlation network builder analyzes the statistical correlation between nodes, using the Pearson correlation coefficient method to perform linear correlation analysis on the values ​​of ecological factor nodes at different time series within the same geographic grid cell. The ecological correlation network builder also analyzes the causal correlation between nodes, using the Granger causality test to examine whether the time series of one ecological factor node can predict the time series changes of another ecological factor node. For node pairs with significant statistical correlation or causal relationship, the ecological correlation network builder connects them with directed edges, the direction of which is determined by the direction of the causal relationship or the setting of independent and dependent variables in the correlation analysis.

[0029] It can be understood that the ecological association network builder assigns a weight value to each directed edge to quantify the association strength. The weight calculation is based on the statistical significance level and effect size of the association analysis. The formula for calculating the association strength weight is: in, Indicates from node Pointing to node The weight of the edge. Represents a node With nodes Pearson correlation coefficient of time series data Represents a node For nodes The F-statistic of the Granger causality test is normalized. The ecological network builder normalizes all weight values ​​to the range of 0 to 1, ultimately constructing an ecological network containing a set of nodes, edges, and weights.

[0030] In some embodiments, the ecological association network builder sets the analysis time window to a configurable parameter, such as calculating the associations between nodes on a quarterly or annual basis, enabling the ecological association network to dynamically reflect changes in the interactions of ecological elements across different seasons or years. The ecological association network builder constructs an independent ecological association network for each geographic grid cell, and also supports spatial aggregation of ecological parameter sequences from adjacent grid cells to construct regional-scale ecological association networks. The network structure data output by the ecological association network builder is stored in an adjacency list format, with each record containing a source node identifier, a target node identifier, and the weight value of the edge connecting them.

[0031] Optionally, the ecological association network builder performs stationarity tests on the ecological parameter sequences before constructing the network and performs differencing on non-stationary time series data to meet the underlying assumptions of statistical methods such as Granger causality tests. The ecological association network builder sets a significance threshold for associations; only when the p-value of the calculated Pearson correlation coefficient or the p-value of the Granger causality test is lower than this threshold is a significant association considered to exist between nodes and a connection established, thereby controlling the sparsity and reliability of the ecological association network.

[0032] Example 3: The repair path mining tool is configured to execute a path search algorithm on the ecological network. Starting from a specified initial ecological node, it traverses all possible paths to the target ecological node, calculates the overall weight and complexity of each path, sorts the paths according to a preset path quality criterion, selects key repair paths, and merges redundant branches to form a repair path set. Starting from the initial ecological node, a depth-first search algorithm is used to traverse the ecological network, recording all complete paths that can reach the target ecological node. For each traversed path, the weights of all its edges are summed to obtain the overall weight of the path. Simultaneously, the total number of nodes traversed by the path is counted, and the total number of nodes is combined with the overall weight to obtain the complexity of the path using a preset complexity calculation model.

[0033] In practice, the restoration path miner receives ecological network graph structure data generated by the ecological network builder. The restoration path miner is configured to execute a path search algorithm on this ecological network. The user or system needs to specify a starting ecological node and a target ecological node. The starting ecological node typically represents an ecological factor requiring intervention or in an unfavorable state, while the target ecological node represents the ecological factor corresponding to the desired state achieved through ecological restoration. The core function of the restoration path miner is to find all feasible sequences connecting these two nodes.

[0034] The repair path miner starts from the initial ecological node and traverses the ecological network using a depth-first search algorithm. The depth-first search algorithm uses a stack data structure to record the current access path, exploring all unvisited neighboring nodes of the current node until it is impossible to continue deeper or reach the target ecological node. During the traversal, the repair path miner records all complete paths that can start from the initial ecological node and ultimately reach the target ecological node. Each such path represents a potential ecological effect transmission or repair action sequence.

[0035] In some embodiments, for each complete traversed path, the path repair miner calculates its overall weight by summing the weights of all directed edges it contains. The path repair miner also counts the total number of nodes traversed by the path. The overall weight and the total number of nodes are input into a preset complexity calculation model, which quantifies the complexity of the path. The formula for calculating path complexity is: in, Representing a path The complexity, Representing a path The total number of nodes traversed. Representing a path Weights of all included edges The sum of these factors. This formula shows that the more nodes a path has and the smaller its total path weight, the higher its complexity value.

[0036] Understandably, the repair path miner sorts all discovered paths according to a preset path quality criterion. This criterion typically prioritizes paths with high overall weight and low complexity, as high weight implies stronger relationships, and low complexity suggests more direct transmission paths. The repair path miner then selects the top-ranked key repair paths and merges redundant branch paths with many overlapping nodes, ultimately producing a concise and efficient set of repair paths as output.

[0037] In some embodiments, the repair path miner sets a maximum search depth threshold to prevent the depth-first search algorithm from getting too deep or creating loops in large-scale ecological networks. During traversal, the repair path miner maintains a set of access markers to avoid repeatedly visiting the same node, ensuring path simplicity. The repair path miner stores the final set of repair paths in a structured data format, with each path recording its node sequence, overall weight, and calculated complexity value.

[0038] Optionally, the path quality criteria supported by the repair path miner can be configurable, allowing users to adjust the overall weight and complexity priority according to specific evaluation objectives. For example, in some scenarios, it may be more inclined to select paths that, although slightly more complex, cover key intermediate nodes. When merging redundant branches, the repair path miner adopts a tree-structure merging method based on the common prefix of the paths, preserving the diversity of path branch points while simplifying the output results.

[0039] Example 4: The dynamic evaluation generator is configured to obtain a snapshot of the ecological parameter sequence at the current moment, calculate the matching degree between this snapshot and the expected ecological state of each path in the restoration path set, and use a multi-index fusion algorithm to aggregate the matching degree results of each path into a comprehensive evaluation score. The comprehensive evaluation score is the ecological restoration effect evaluation index. The matching degree calculation result for each key restoration path in the restoration path set is obtained; a preset weight coefficient is assigned to each ecological state index; the matching degree value of each ecological state index for each path is multiplied by its corresponding weight coefficient to obtain a weighted matching degree; the weighted matching degrees of all ecological state indices are summed to calculate the individual evaluation score for that path; the individual evaluation scores of all key restoration paths are arithmetically averaged to obtain the comprehensive evaluation score used to characterize the overall restoration effect.

[0040] In practice, the dynamic assessment generator periodically retrieves snapshots of the current ecological parameter sequence from the spatiotemporal serialization processor. These snapshots represent the latest ecological state of each grid cell within the national territory at the assessment time. Simultaneously, the dynamic assessment generator loads a set of restoration paths generated by the restoration path miner. Each path in the restoration path set describes a series of expected sequences of ecological state evolution. The core task of the dynamic assessment generator is to compare the current real-time ecological state with the expected state of each restoration path.

[0041] The dynamic evaluation generator performs matching degree calculations. For each path in the restoration path set, it extracts all ecological factor nodes involved in that path and compares the measured values ​​of the corresponding ecological factors in the current ecological parameter sequence snapshot with the expected theoretical values ​​of the path at that stage. The matching degree calculation uses a similarity metric method based on Euclidean distance to calculate the closeness between the measured and theoretical values.

[0042] In some embodiments, the dynamic assessment generator employs a multi-index fusion algorithm to aggregate the matching results of various ecological state indicators along a single path. The multi-index fusion algorithm assigns a preset weight coefficient to each ecological state indicator, reflecting the importance of that indicator in the overall ecological restoration effect assessment. The weight coefficients are typically preset by domain experts based on ecological principles and management objectives, and stored in a configuration file for the dynamic assessment generator to read. The correspondence between ecological state indicators and weights is shown in Table 1 below: Table 1: Correspondence between ecological status indicators and weights It is understandable that the multi-index fusion algorithm performs the following calculation steps for each key restoration path in the restoration path set: obtain the matching degree values ​​of all ecological status indicators of the path; multiply the matching degree value of each indicator by its corresponding weight coefficient to obtain the weighted matching degree of the indicator; sum the weighted matching degrees of all ecological status indicators to obtain a single evaluation score reflecting the achievement of the path.

[0043] After calculating the individual evaluation score for each key restoration path in the restoration path set, the dynamic evaluation generator performs an arithmetic average of these individual evaluation scores to generate a comprehensive evaluation score. This comprehensive evaluation score serves as a holistic measure to characterize the overall effectiveness of regional ecological restoration at the current moment, i.e., it is an ecological restoration effectiveness evaluation index. The calculation formula is: in, This represents the overall evaluation score. This represents the total number of critical repair paths in the repair path set. This represents the total number of ecological status indicators. Indicates the first The first critical repair path The matching degree values ​​of the ecological status indicators, Indicates the first The preset weighting coefficients for each ecological status indicator.

[0044] In some embodiments, the dynamic assessment generator can be configured to automatically execute the assessment process at fixed time intervals, such as monthly or quarterly, to generate a comprehensive assessment score over time, thereby tracking the dynamic trends of ecological restoration effects. The dynamic assessment generator stores the comprehensive assessment score for each assessment, the individual assessment scores for each path, and the original matching degree calculation results into an assessment results database to form a historical assessment record.

[0045] Optionally, the matching degree calculation method supported by the dynamic evaluation generator can be pluggable. In addition to Euclidean distance, it can also support other similarity measurement algorithms such as cosine similarity or Manhattan distance to adapt to the data characteristics of different ecological indicators. When calculating the comprehensive evaluation score, the dynamic evaluation generator can, in addition to a simple arithmetic mean, assign different weights to each restoration path according to its importance and perform a weighted average. The importance of a path can be derived from the overall weight or complexity of the path calculated by the restoration path miner.

[0046] Example 5: The system also includes a restoration effect predictor, configured to receive a set of restoration paths and historical ecological parameter sequences, use a time series prediction model to simulate the changing trends of ecological parameters as they evolve along each restoration path at different future time points, and output restoration effect prediction data. The system also includes a visualization engine, configured to overlay ecological restoration effect assessment indicators and restoration effect prediction data onto an electronic map, generating an interactive visualization report including heatmaps, path flow maps, and indicator change curves.

[0047] In its implementation, the system includes a restoration effect predictor. This predictor receives a set of restoration paths from a restoration path miner and a large number of historical ecological parameter sequences from a spatiotemporal serialization processor via a data interface. The restoration path set defines the potential sequence of ecological state evolution, while the historical ecological parameter sequences provide background patterns of ecological factors changing over time. The restoration effect predictor uses a time series prediction model to simulate the changing trends of ecological parameters along each restoration path at different future time points. This time series prediction model can employ an autoregressive integral moving average model or a long short-term memory neural network model. The model is trained using the historical ecological parameter sequences to learn the dynamic patterns of ecological factors. For each path in the restoration path set, the restoration effect predictor uses the sequence of ecological factor nodes involved in the path as input conditions to drive the time series prediction model to predict forward, generating estimates of each ecological factor at multiple future time points (e.g., predicting the next 1, 3, and 5 years). Finally, it outputs a structured restoration effect prediction dataset containing path identifiers, time points, ecological factor types, and predicted values.

[0048] In some embodiments, before applying the time series forecasting model, the restoration effect predictor performs stationarity tests and seasonal decomposition on historical ecological parameter sequences to ensure that the time series data meets the model's basic assumptions. For non-stationary sequences, differencing is performed, and for sequences with significant seasonality, a seasonal adjustment component is introduced. The restoration effect predictor treats the node sequence of each path in the restoration path set as a set of constraints. During the forecasting process, the time series forecasting model considers the correlations between the ecological factors represented by these nodes, making the predicted trend more consistent with the interaction logic revealed by the ecological network.

[0049] It is understandable that the restoration effect prediction data output by the restoration effect predictor is stored in a standard time series data format. Each record contains a timestamp, geographic grid unit identifier, ecological factor code, and prediction value. It also records the restoration path identifier and prediction model version information used to generate the prediction, so as to ensure the traceability of the prediction results.

[0050] The system also includes a visualization engine, configured to acquire ecological restoration effectiveness assessment indicators from a dynamic assessment generator and restoration effect prediction data from a restoration effect predictor. The core function of the visualization engine is to overlay this data onto a base electronic map, typically supported by a third-party map provider offering geographic maps at different scales. The visualization engine generates interactive visualization reports, accessible to users via a web browser or a dedicated client.

[0051] In some embodiments, the visualization engine generates heatmaps to visually display the spatial distribution of ecological restoration effectiveness evaluation indicators. The heatmap maps the comprehensive evaluation score of each geographic grid unit to different color shades, with colors gradually changing from red (low score) to green (high score), intuitively presenting the pattern of regional restoration effectiveness. The visualization engine also generates path flow maps to visualize the set of restoration paths. The path flow maps locate nodes in the ecological network at their corresponding geographic coordinates, using arrowed lines to represent the direction of the restoration paths, and the thickness of the lines can reflect the weight or priority of the paths.

[0052] Understandably, the visualization engine generates indicator change curves to show the predicted trends of ecological restoration effect assessment indicators and key ecological factors over time. The indicator change curves plot historical actual values ​​and future predicted values ​​on the horizontal axis and vertical axis, distinguished by different line types or colors. The interactive visualization report allows users to dynamically adjust the visualization content by clicking on map elements, selecting time ranges, and filtering ecological factor types, enabling multi-dimensional exploration of the data.

[0053] Optionally, when rendering heatmaps, the visualization engine can use kernel density estimation to smooth the evaluation scores of discrete grid cells, generating a continuous surface representation that makes spatial patterns easier to identify. The visualization engine also supports exporting visualization reports as static images or interactive web pages for convenient reporting or archiving.

[0054] Optionally, the remediation effect predictor can be configured to run multiple prediction scenarios, such as generating multiple sets of remediation effect prediction data based on different sets of remediation paths or different model parameters for comparative analysis. The visualization engine can display the visualization results of different prediction scenarios side by side in the same report to assist decision-makers in comparing and selecting solutions.

[0055] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A big data-based evaluation system for the ecological restoration effect of land space remediation areas, characterized in that: The system includes a multi-source ecological data acquisition engine, a spatiotemporal serialization processor, an ecological association network builder, a restoration path mining machine, and a dynamic evaluation generator; The multi-source ecological data acquisition engine continuously captures original monitoring data related to ecological restoration within the national land space, forming a multi-source ecological data stream; The spatiotemporal serialization processor receives the multi-source ecological data stream, performs serialization, recombination, and standardization transformation according to the time and spatial dimensions, and generates an ecological parameter sequence with a unified spatiotemporal reference. The ecological association network builder parses the ecological elements in the ecological parameter sequence and constructs an ecological association network that expresses the interaction relationships between the elements; The restoration path mining tool extracts multiple potential restoration paths from the ecological association network, and performs path fusion and optimization to generate a set of restoration paths; The dynamic evaluation generator calculates and outputs ecological restoration effect evaluation indicators based on the set of restoration paths and real-time ecological status data.

2. The big data-based regional ecological restoration effect evaluation system for land space remediation as described in claim 1, characterized in that, The multi-source ecological data acquisition engine is configured to access satellite remote sensing data interfaces, ground sensor networks, and manual survey databases. It performs data integrity verification and conflict resolution on the received raw monitoring data. The raw monitoring data includes at least vegetation coverage, soil physicochemical properties, water quality, and biodiversity indicators. After fusion, the multi-source ecological data stream is formed.

3. The big data-based regional ecological restoration effect evaluation system for land space remediation as described in claim 2, characterized in that, The spatiotemporal serialization processor is configured to perform time alignment processing on the multi-source ecological data stream, unifying data from different sources into the same timestamp sequence, and simultaneously perform spatial rasterization processing to map the data to a unified geographic grid cell. It also performs missing value imputation and noise filtering on the ecological parameters within each grid cell to generate the ecological parameter sequence.

4. The big data-based regional ecological restoration effect evaluation system for land space remediation as described in claim 3, characterized in that, The ecological association network builder is configured to identify nodes representing different ecological factors from the ecological parameter sequence, including vegetation type, soil moisture, and species abundance, analyze the statistical correlation or causal relationship between nodes, connect nodes with significant correlations with directed edges, and assign weights to each edge to represent the correlation strength, thereby constructing the ecological association network.

5. The big data-based regional ecological restoration effect evaluation system for land space remediation as described in claim 4, characterized in that, The repair path miner is configured to execute a path search algorithm on the ecological network. Starting from a specified initial ecological node, it traverses all possible paths to the target ecological node, calculates the overall weight and complexity of each path, sorts the paths according to a preset path quality criterion, selects key repair paths, and merges redundant branches to form the repair path set.

6. The big data-based regional ecological restoration effect evaluation system for land space remediation as described in claim 5, characterized in that, The steps of traversing all possible paths to the target ecological node and calculating the overall weight and complexity of each path include: starting from the initial ecological node, traversing the ecological association network using a depth-first search algorithm, and recording all complete paths that can reach the target ecological node; for each traversed path, summing the weights of all edges contained therein to obtain the overall weight of the path; simultaneously, counting the total number of nodes traversed by the path, and combining the total number of nodes with the overall weight to obtain the complexity of the path using a preset complexity calculation model.

7. The big data-based regional ecological restoration effect evaluation system for land space remediation as described in claim 6, characterized in that, The dynamic evaluation generator is configured to obtain a snapshot of the ecological parameter sequence at the current moment, calculate the matching degree between it and the expected ecological state of each path in the set of restoration paths, and use a multi-index fusion algorithm to aggregate the matching degree results of each path into a comprehensive evaluation score, which is the ecological restoration effect evaluation index.

8. The big data-based regional ecological restoration effect evaluation system for land space remediation as described in claim 7, characterized in that, The step of aggregating the matching results of each path into a comprehensive evaluation score using a multi-index fusion algorithm includes: obtaining the matching degree calculation result corresponding to each key restoration path in the restoration path set; assigning a preset weight coefficient to each ecological state indicator; multiplying the matching degree value of each ecological state indicator of each path by its corresponding weight coefficient to obtain a weighted matching degree; summing the weighted matching degrees of all ecological state indicators to calculate the individual evaluation score of that path; and arithmetically averaging the individual evaluation scores of all key restoration paths to obtain the comprehensive evaluation score used to characterize the overall restoration effect.

9. The big data-based regional ecological restoration effect evaluation system for land space remediation as described in claim 1, characterized in that, The system also includes a restoration effect predictor, which is configured to receive the set of restoration paths and historical ecological parameter sequences, use a time series prediction model to simulate the changing trends of ecological parameters as they evolve along each restoration path at different future time points, and output restoration effect prediction data.

10. The big data-based regional ecological restoration effect evaluation system for land space remediation as described in claim 9, characterized in that, The system also includes a visualization engine configured to overlay the ecological restoration effect assessment indicators and the restoration effect prediction data onto an electronic map to generate an interactive visualization report containing heat maps, path flow maps, and indicator change curves.

Images