Coastal wetland environment and bird inhabitation correlation analysis method

Abstract

The invention provides a coastal wetland environment and bird inhabitation correlation analysis method, and relates to the technical field of environment correlation analysis methods, and the method comprises the steps: obtaining and preprocessing wetland environment data, and obtaining the preprocessed wetland environment data; analyzing the preprocessed wetland environment data through a principal component analysis method, and determining key environment factors influencing bird inhabitation; according to a clustering algorithm, inhabitation modes and seasonal changes of the birds are analyzed; and constructing a correlation model between the wetland environment factors and the bird inhabitation characteristics through a machine learning algorithm, and outputting an analysis result according to the correlation model. By realizing quantitative analysis of environmental factor influence and extracting key environmental factors having significant influence on bird inhabitation, subjectivity of factor screening in a conventional method is avoided, bird inhabitation modes and seasonal change rules are disclosed, bird inhabitation modes and seasonal changes are analyzed, and differences of bird behaviors under different time and space conditions are disclosed. And reference is provided for wetland management and protection.

Description

Technical Field

[0001] The present invention relates to the technical field of environmental association analysis methods, and in particular to a method for analyzing the association between coastal wetland environment and bird habitats. Background Art

[0002] Wetland ecosystem is one of the most important ecosystems on Earth. It not only plays an important role in regulating climate, water cycle and material cycle, but also provides a habitat and breeding environment for many birds. As an important part of wetland types, coastal wetlands occupy an important position in the global ecosystem with their complex habitats and rich biodiversity.

[0003] However, with the intensification of human activities and the impact of climate change, the coastal wetland environment is undergoing significant changes, including water level fluctuations, decreased vegetation coverage and increased salinity. These changes in environmental factors have a profound impact on the quality and sustainability of bird habitats.

[0004] Conventional wetland and bird habitat research methods often rely on long-term manual observations, which are inefficient, have incomplete data, and are highly subjective in analysis results. Although some studies have attempted to introduce big data and statistical analysis tools, there are still the following problems in revealing the specific impacts of wetland environmental changes on bird habitats: First, the specific impact mechanism of environmental factors on bird habitat has not yet been effectively quantified.

[0005] Second, there is a lack of comprehensive analysis methods that can simultaneously combine temporal, spatial and multi-dimensional data characteristics.

[0006] Third, the ability to process and visualize complex data is limited, making it difficult to generate intuitive and scientific protection strategies and optimization recommendations. Summary of the invention

[0007] The present invention provides a method for analyzing the correlation between coastal wetland environment and bird habitats, aiming to solve the technical problem that the specific impact mechanism of environmental factors on bird habitats in related technologies has not been effectively quantified, avoid the subsequent problem of lack of comprehensive analysis that can simultaneously combine time, space and multi-dimensional data characteristics, and solve the problem of limited use of processing and visualization capabilities of complex data.

[0008] To achieve the above object, the present invention provides the following technical solutions: A method for analyzing the association between coastal wetland environment and bird habitats comprises the following steps: Wetland environmental data are acquired and preprocessed to obtain preprocessed wetland environmental data.

[0009] The pre-processed wetland environmental data were analyzed using principal component analysis to determine the key environmental factors affecting bird habitats.

[0010] Based on the clustering algorithm, the habitat patterns and seasonal changes of birds are analyzed.

[0011] A correlation model between wetland environmental factors and bird habitat characteristics is constructed through machine learning algorithms, and the analysis results are output based on the correlation model.

[0012] Provide a visual map of the analysis results and generate optimization suggestions based on the analysis results.

[0013] As a preferred solution of the present invention, wetland environmental data is obtained and preprocessed to obtain preprocessed wetland environmental data, specifically including: Wetland environmental data include: wetland water level, wetland vegetation coverage, wetland salinity and bird habitat data; bird habitat data include: bird population number, habitat area and duration of habitat activities.

[0014] Preprocessing includes data format conversion, data cleaning, data normalization and data completion; Data format conversion: the collected wetland environmental data is converted into a unified format, and the converted data is expressed as follows: ; In the formula, Represents the i-th original data record.

[0015] Data cleaning, detect and remove outliers in wetland environmental data. The outlier determination formula is: ; In the formula, is the data record value, is the mean of similar data, is the standard deviation, k is the preset abnormal threshold coefficient; the cleaned data is expressed as: ; Data normalization: normalize the cleaned wetland environmental data to eliminate the impact of different indicator units and magnitudes. The normalization formula is: ; In the formula, x is the original data value, is the minimum value of the same indicator data, is the maximum value of the same indicator data, is the normalized data value; the normalized data is expressed as: ; Data completion: missing values ​​in the normalized wetland environmental data are completed by interpolation algorithm. The interpolation algorithm adopts Lagrange interpolation method, and its calculation formula is: ; In the formula, , are known data points, is the missing value for the prediction.

[0016] Furthermore, when obtaining wetland environmental data, a variety of data collection methods are used, including: deploying wetland monitoring sensors to obtain real-time wetland water level, salinity and vegetation coverage data; using drones and satellite remote sensing technology to collect wetland panoramic images and vegetation coverage information; combining manual data recording at bird observation stations to collect bird population numbers, habitat area and duration of habitat activities. At the same time, in order to ensure the timeliness and accuracy of the data, the collected data is verified and integrated through multi-source data fusion technology.

[0017] As a preferred solution of the present invention, the pre-processed wetland environmental data is analyzed by principal component analysis to determine the key environmental factors affecting bird habitats, specifically including: Data matrix construction, representing the preprocessed wetland environmental data in matrix form: ; In the formula, m is the number of samples, n is the number of environmental factors, Represents the value of the jth environmental factor in the ith sample.

[0018] Compute the covariance matrix of the data matrix , and its calculation formula is: ; In the formula, represents the i-th row in the data matrix, is the mean vector of all samples, and C represents the covariance relationship between environmental factors.

[0019] Perform eigendecomposition on the covariance matrix C to obtain the eigenvalue set and the corresponding set of eigenvectors , where represents the variance contribution of the ith principal component, Representation and Eigenvalues The corresponding feature vector.

[0020] Principal component extraction, sort the principal components according to the size of the eigenvalues, and select the eigenvectors corresponding to the first k eigenvalues ​​to form the principal component matrix: , the following conditions are met: ; In the formula, is the threshold of the cumulative variance contribution rate, and k is the number of principal components that meet the threshold.

[0021] Principal component mapping, mapping the original data matrix X to the principal component space to obtain the principal component data matrix , and its calculation formula is: ; In the formula, Z represents the principal component data after dimensionality reduction, is the selected principal component matrix.

[0022] Determine the key factors based on the principal component matrix The coefficient value of the filter is used to screen the environmental factors that contribute most to the principal component and determine the key factor set that affects bird habitat. , where t represents the number of key factors.

[0023] Furthermore, in the principal component analysis method, in order to ensure the scientific nature of the principal component extraction, the characteristic decomposition results of the covariance matrix are verified, mainly by checking whether the cumulative variance contribution rate reaches the set threshold And whether the independence between the principal components meets the requirements; in addition, for the selection of eigenvectors, a secondary screening is carried out in combination with the significance of the actual environmental factors to eliminate possible redundant components.

[0024] As a preferred solution of the present invention, the habitat patterns and seasonal changes of birds are analyzed according to a clustering algorithm, specifically including: The k-means clustering algorithm is used, the number of clusters is set to K, and the sample classification is achieved by minimizing the square distance from the sample point in the cluster to the cluster center. The objective function is: ; In the formula, is the kth cluster; is the cluster center.

[0025] Determine the optimal number of clusters and calculate the objective function value according to the time rule The trend of changing with the number of clusters K, select The position where the value decreases significantly and tends to be flat is taken as the optimal number of clusters.

[0026] Extract pattern features and calculate feature mean for each cluster , extract the characteristics of bird habitat patterns, and the extraction formula is: ; In the formula, Indicates the number of samples in the cluster.

[0027] Furthermore, when analyzing bird habitat patterns based on clustering algorithms, the initialization method of cluster center points has a great influence on the final clustering results. The initial center points are preferably selected by the K-Means++ method to reduce the randomness of the algorithm and improve the stability of the clustering results. At the same time, in determining the optimal number of clusters K, the elbow rule and silhouette coefficient method are combined for double verification to ensure the rationality of the division.

[0028] As a preferred solution of the present invention, a correlation model between wetland environmental factors and bird habitat characteristics is constructed through a machine learning algorithm, and the analysis results are output according to the correlation model, specifically including: The random forest regression algorithm is used to build the association model. The model structure consists of multiple decision trees. The prediction formula of random forest is: ; In the formula, is the predicted value, T is the total number of decision trees, is the prediction result of the tth decision tree.

[0029] The dataset is split into a training set and a test set. The training set is used for model training, and the test set is used to verify the model performance. The mean square error is used as the loss function to train the model.

[0030] By calculating the impact of each environmental factor on the model prediction results, the importance of the factor is evaluated. The feature importance is determined by reducing the contribution value of the mean square error. The determination formula is: ; In the formula, Representation Factor The importance of is the inclusion factor The mean square error when To remove the factor The mean square error when .

[0031] The new wetland environmental data is predicted based on the trained model to generate prediction results of bird habitat characteristics. Based on the prediction results of bird habitat characteristics and combined with feature importance analysis, optimization suggestions are generated.

[0032] Furthermore, in the random forest regression algorithm, in order to improve the prediction accuracy of the model, the hyperparameter optimization technology is used to tune the forest depth, the number of decision trees and the minimum number of samples for splitting nodes; preferably, the optimal parameter combination is determined through grid search or Bayesian optimization method to improve the performance consistency of the model on the training set and the test set; in addition, in the feature importance analysis of the model, the stability of the key factors is further verified in combination with the cross-validation results.

[0033] As a preferred solution of the present invention, new wetland environmental data is predicted based on the trained model to generate prediction results of bird habitat characteristics. Based on the prediction results of bird habitat characteristics, combined with feature importance analysis, optimization suggestions are generated, specifically including: Identify environmental factors that significantly affect bird habitats and analyze the deviation between the predicted results and the target values. If the predicted results show that the population of a certain bird species is lower than expected, adjust the key environmental factors based on the characteristic significance and propose optimization strategies for the environmental factors that significantly affect bird habitats. Specifically: Water level adjustment, based on feature significance analysis, if water level is a key factor, it is recommended to maintain the water level within a certain range.

[0034] Vegetation restoration: If the vegetation coverage rate has a significant impact on the bird habitat area, it is recommended to increase wetland vegetation planting.

[0035] Salinity management, where high salinity has a negative impact on bird activity duration, propose operational methods to reduce salinity.

[0036] Furthermore, when generating optimization recommendations, specific optimization strategies are formulated based on the prediction results of the association model and the ranking of feature importance; optimization measures need to be combined with the actual management objectives of the wetlands. In water level adjustment, it is necessary to consider both the habitat needs of birds and the balance of other ecological service functions of the wetlands; for vegetation restoration, it is recommended to select native species suitable for wetland habitats for planting to ensure the stability and sustainability of the ecosystem.

[0037] As a preferred solution of the present invention, the analysis results are provided with a visual mapping, and optimization suggestions are generated based on the analysis results, specifically including: Include the analysis results in the prediction of bird habitat characteristics and environmental factor characteristics Summarized as a mapping dataset, the dataset is represented as: ; In the formula, represents the jth environmental factor, Representation Factor The feature importance of Represents the prediction result of the i-th sample.

[0038] Select the corresponding mapping dimension based on the analysis results, specifically: Spatial mapping, based on wetland geospatial information, generates distribution heat maps of bird habitat patterns.

[0039] Time mapping, based on time series data, shows seasonal change trends.

[0040] Significance mapping, based on the feature significance sorting, generates a factor contribution map to bird habitat.

[0041] The wetland area is divided into grid units, and the habitat characteristic prediction value corresponding to each unit reflects the bird population and habitat area through the depth of color to achieve visual mapping.

[0042] Furthermore, the design of heat maps for visual mapping needs to consider the expression effects of different data dimensions, such as distinguishing the distribution of bird populations by color depth, and displaying key environmental factors in different regions through legends and annotations; for time mapping, combined with dynamic interactive functions, the evolution trend of bird habitat characteristics at multiple time points can be displayed to intuitively reflect seasonal changes. Feature importance bar charts should be marked with specific values ​​and contribution ratios to facilitate wetland managers to quickly identify key factors.

[0043] As a preferred solution of the present invention, the wetland area is divided into grid units, and the habitat characteristic prediction value corresponding to each unit reflects the bird population and habitat area through the color depth to achieve visual mapping, which specifically includes: Spatial mapping, the heat map formula is: ; In the formula, is the color value of the grid cell, is the contribution weight of the i-th sample to the grid unit, is the predicted value.

[0044] Time mapping, draw a time series line graph to show the changing trends of bird habitat characteristics in different time periods.

[0045] Saliency mapping, generating feature importance bar charts, according to Sort from high to low to intuitively display the key factors.

[0046] Generate optimization recommendations, analyze deviations, and compare forecast results With target value Deviation of the environmental factors to be optimized is determined by the deviation formula: ; In the formula, It is indicated that bird habitat features need to be enhanced. It indicates the need to reduce the negative impact of environmental factors on birds.

[0047] Environmental factors for bias, combined with feature importance Based on the historical data, we propose adjustment strategies, specifically: Water level optimization: If the water level has a significant impact on the prediction results, it is recommended to adjust the water level to the optimal range.

[0048] Vegetation restoration: If there is a significant positive correlation between vegetation coverage and habitat area, it is recommended to increase coverage through planting.

[0049] Salinity control: if salinity affects the habitat time, a reasonable salinity adjustment method should be proposed.

[0050] All analysis results and optimization suggestions are organized into interactive visualization reports, which include: map heat maps showing the distribution of bird habitat patterns in different regions; feature importance charts to identify key environmental factors; and optimization plan lists to provide targeted suggestions based on deviation analysis.

[0051] Furthermore, when generating interactive visualization reports, the map heat map contained in the report should be highly consistent with the actual geographic information of the wetland, and the detailed information of a specific area can be zoomed in or out as needed; feature importance charts and optimization plan lists need to provide export functions to facilitate archiving or sharing by wetland protection managers; at the same time, it is recommended to add a deviation analysis module between the predicted results and the actual results in the report to verify the reliability of the model.

[0052] Compared with the prior art, the present invention has the following beneficial effects: 1. Realize the quantitative analysis of the impact of environmental factors. Through the principal component analysis method, extract the key environmental factors that have a significant impact on bird habitats, avoid the subjectivity of factor screening in traditional methods, and improve the scientificity and accuracy of the analysis.

[0053] 2. Reveal the habitat patterns and seasonal changes of birds. Through clustering algorithms, the habitat patterns and seasonal changes of birds are analyzed, revealing the differences in bird behavior under different time and space conditions, providing key references for wetland management and protection.

[0054] 3. Build accurate association models and generate scientific optimization suggestions. Use machine learning algorithms to build association models between wetland environmental factors and bird habitat characteristics. Use algorithms such as random forests to achieve efficient prediction and feature importance analysis, providing a quantitative basis for environmental factor optimization. Based on the prediction results of the association model and combined with feature importance analysis, put forward targeted optimization suggestions, such as adjusting water levels, restoring vegetation coverage, controlling salinity, etc., which significantly enhances the scientific nature and operability of decision-making.

[0055] 4. Provide intuitive visualization results and form interactive protection reports. Based on visualization techniques such as heat maps, time series graphs and feature importance bar charts, it clearly displays the spatial and temporal distribution and changes of bird habitat characteristics, provides intuitive support for the formulation of protection strategies and forms interactive protection reports; integrates analysis results and optimization suggestions to generate interactive visualization reports, providing wetland protection managers with a set of scientific, comprehensive and easy-to-understand decision support tools.

[0056] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments of the present application will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0058] Figure 1 A flow chart of a method for analyzing the association between coastal wetland environment and bird habitat provided in an embodiment of the present application. DETAILED DESCRIPTION

[0059] The technical solutions in the embodiments of the present application will be described below in conjunction with the drawings in the embodiments of the present application.

[0060] See also Figure 1 , Figure 1 A flowchart of a method for analyzing the association between a coastal wetland environment and bird habitat is provided for an embodiment of the present application.

[0061] In this embodiment, a method for analyzing the association between coastal wetland environment and bird habitat includes step S100, step S200, step S300, step S400 and step S500.

[0062] Step S100, obtaining wetland environmental data and preprocessing it to obtain preprocessed wetland environmental data.

[0063] Specifically, they include: Wetland environmental data include: wetland water level, wetland vegetation coverage, wetland salinity and bird habitat data; bird habitat data include: bird population number, habitat area and duration of habitat activities.

[0064] Preprocessing includes data format conversion, data cleaning, data normalization and data completion; Data format conversion: the collected wetland environmental data is converted into a unified format, and the converted data is expressed as follows: ; In the formula, Represents the i-th original data record.

[0065] Data cleaning, detect and remove outliers in wetland environmental data. The outlier determination formula is: ; In the formula, is the data record value, is the mean of similar data, is the standard deviation, k is the preset abnormal threshold coefficient; the cleaned data is expressed as: ; Data normalization: normalize the cleaned wetland environmental data to eliminate the impact of different indicator units and magnitudes. The normalization formula is: ; In the formula, x is the original data value, is the minimum value of the same indicator data, is the maximum value of the same indicator data, is the normalized data value; the normalized data is expressed as: ; Data completion: missing values ​​in the normalized wetland environmental data are completed by interpolation algorithm. The interpolation algorithm adopts Lagrange interpolation method, and its calculation formula is: ; In the formula, , are known data points, is the missing value for the prediction.

[0066] Furthermore, when obtaining wetland environmental data, a variety of data collection methods are used, including: deploying wetland monitoring sensors to obtain real-time wetland water level, salinity and vegetation coverage data; using drones and satellite remote sensing technology to collect wetland panoramic images and vegetation coverage information; combining manual data recording at bird observation stations to collect bird population numbers, habitat area and duration of habitat activities. At the same time, in order to ensure the timeliness and accuracy of the data, the collected data is verified and integrated through multi-source data fusion technology.

[0067] It should be noted that when performing data cleaning, the outlier detection threshold is set in combination with the statistical results of historical data. The specific choice of the threshold can be determined according to the natural fluctuation range of the wetland environment. For example, k=3 is set in combination with the water level variation range of the actual wetland environment to exclude outliers of non-natural fluctuations to the greatest extent; in addition, outliers in bird habitat data are verified through time series change trends to avoid analysis deviations caused by anomalies caused by short-term interference.

[0068] It should be further explained that the interpolation algorithm used for data completion selects the optimal strategy according to the characteristics of the missing data. When the distribution of missing data is random, the Lagrange interpolation method is preferred; when the missing data has obvious time series characteristics, linear interpolation or moving average method is used; at the same time, for key data missing, the spatial characteristics and historical data of the surrounding area are combined for completion.

[0069] Step S200: Analyze the pre-processed wetland environmental data by principal component analysis to determine key environmental factors that affect bird habitat.

[0070] Specifically, they include: Data matrix construction, representing the preprocessed wetland environmental data in matrix form: ; In the formula, m is the number of samples, n is the number of environmental factors, Represents the value of the jth environmental factor in the ith sample.

[0071] Compute the covariance matrix of the data matrix , and its calculation formula is: ; In the formula, represents the i-th row in the data matrix, is the mean vector of all samples, and C represents the covariance relationship between environmental factors.

[0072] Perform eigendecomposition on the covariance matrix C to obtain the eigenvalue set and the corresponding set of eigenvectors , where represents the variance contribution of the ith principal component, Representation and Eigenvalues The corresponding feature vector.

[0073] Principal component extraction, sort the principal components according to the size of the eigenvalues, and select the eigenvectors corresponding to the first k eigenvalues ​​to form the principal component matrix: , the following conditions are met: ; In the formula, is the threshold of the cumulative variance contribution rate, and k is the number of principal components that meet the threshold.

[0074] Principal component mapping, mapping the original data matrix X to the principal component space to obtain the principal component data matrix , and its calculation formula is: ; In the formula, Z represents the principal component data after dimensionality reduction, is the selected principal component matrix.

[0075] Determine the key factors based on the principal component matrix The coefficient value of the filter is used to screen the environmental factors that contribute most to the principal component and determine the key factor set that affects bird habitat. , where t represents the number of key factors.

[0076] It should be noted that in the principal component analysis method, in order to ensure the scientific nature of the principal component extraction, the characteristic decomposition results of the covariance matrix are verified, mainly by checking whether the cumulative variance contribution rate reaches the set threshold And whether the independence between the principal components meets the requirements; in addition, for the selection of eigenvectors, a secondary screening is carried out in combination with the significance of the actual environmental factors to eliminate possible redundant components.

[0077] Step S300: Analyze the habitat patterns and seasonal changes of birds according to the clustering algorithm.

[0078] Specifically, they include: The k-means clustering algorithm is used, the number of clusters is set to K, and the sample classification is achieved by minimizing the square distance from the sample point in the cluster to the cluster center. The objective function is: ; In the formula, is the kth cluster; is the cluster center.

[0079] Determine the optimal number of clusters and calculate the objective function value according to the time rule The trend of changing with the number of clusters K, select The position where the value decreases significantly and tends to be flat is taken as the optimal number of clusters.

[0080] Extract pattern features and calculate feature mean for each cluster , extract the characteristics of bird habitat patterns, and the extraction formula is: ; In the formula, Indicates the number of samples in the cluster.

[0081] It should be noted that when analyzing bird habitat patterns based on clustering algorithms, the initialization method of cluster center points has a great influence on the final clustering results. It is preferred to select the initial center points through the K-Means++ method to reduce the randomness of the algorithm and improve the stability of the clustering results. At the same time, in determining the optimal number of clusters K, the elbow rule and silhouette coefficient method are combined for double verification to ensure the rationality of the division. Step S400: construct a correlation model between wetland environmental factors and bird habitat characteristics through a machine learning algorithm, and output analysis results based on the correlation model.

[0082] Specifically, they include: The random forest regression algorithm is used to build the association model. The model structure consists of multiple decision trees. The prediction formula of random forest is: ; In the formula, is the predicted value, T is the total number of decision trees, is the prediction result of the tth decision tree.

[0083] The dataset is split into a training set and a test set. The training set is used for model training, and the test set is used to verify the model performance. The mean square error is used as the loss function to train the model.

[0084] By calculating the impact of each environmental factor on the model prediction results, the importance of the factor is evaluated. The feature importance is determined by reducing the contribution value of the mean square error. The determination formula is: ; In the formula, Representation Factor The importance of is the inclusion factor The mean square error when To remove the factor The mean square error when .

[0085] It should be noted that in the random forest regression algorithm, in order to improve the prediction accuracy of the model, hyperparameter optimization technology is used to tune the forest depth, the number of decision trees and the minimum number of samples for splitting nodes; preferably, the best parameter combination is determined through grid search or Bayesian optimization method to improve the performance consistency of the model on the training set and the test set; in addition, in the feature importance analysis of the model, the stability of the key factors is further verified in combination with the cross-validation results.

[0086] The trained model is used to predict new wetland environmental data and generate prediction results of bird habitat characteristics. Based on the prediction results of bird habitat characteristics and combined with feature importance analysis, optimization suggestions are generated, including: Identify environmental factors that significantly affect bird habitats and analyze the deviation between the predicted results and the target values. If the predicted results show that the population of a certain bird species is lower than expected, adjust the key environmental factors based on the characteristic significance and propose optimization strategies for the environmental factors that significantly affect bird habitats. Specifically: Water level adjustment, based on feature significance analysis, if water level is a key factor, it is recommended to maintain the water level within a certain range.

[0087] Vegetation restoration: If the vegetation coverage rate has a significant impact on the bird habitat area, it is recommended to increase wetland vegetation planting.

[0088] Salinity management, where high salinity has a negative impact on bird activity duration, propose operational methods to reduce salinity.

[0089] It should be noted that when generating optimization recommendations, specific optimization strategies are formulated based on the prediction results of the association model and the ranking of feature importance; optimization measures need to be combined with the actual management objectives of the wetlands. In water level adjustments, it is necessary to consider both the habitat needs of birds and the balance of other ecological service functions of the wetlands; for vegetation restoration, it is recommended to select native species suitable for wetland habitats for planting to ensure the stability and sustainability of the ecosystem.

[0090] Step S500: Provide a visual mapping of the analysis results and generate optimization suggestions based on the analysis results.

[0091] Specifically, they include: Include the analysis results in the prediction of bird habitat characteristics and environmental factor characteristics Summarized as a mapping dataset, the dataset is represented as: ; In the formula, represents the jth environmental factor, Representation Factor The feature importance of Represents the prediction result of the i-th sample.

[0092] Select the corresponding mapping dimension based on the analysis results, specifically: Spatial mapping, based on wetland geospatial information, generates distribution heat maps of bird habitat patterns.

[0093] Time mapping, based on time series data, shows seasonal change trends.

[0094] Significance mapping, based on the feature significance sorting, generates a factor contribution map to bird habitat.

[0095] It should be noted that the design of heat maps for visual mapping needs to consider the expression effects of different data dimensions, such as distinguishing the distribution of bird populations by color depth, and displaying key environmental factors in different regions through legends and annotations; for time mapping, combined with dynamic interactive functions, the evolution trend of bird habitat characteristics at multiple time points can be displayed to intuitively reflect seasonal changes. Feature importance bar charts should be marked with specific values ​​and contribution ratios to facilitate wetland managers to quickly identify key factors.

[0096] The wetland area is divided into grid cells, and the habitat feature prediction value corresponding to each cell reflects the bird population and habitat area through the depth of color to achieve visual mapping, including: Spatial mapping, the heat map formula is: ; In the formula, is the color value of the grid cell, is the contribution weight of the i-th sample to the grid unit, is the predicted value.

[0097] Time mapping, draw a time series line graph to show the changing trends of bird habitat characteristics in different time periods.

[0098] Saliency mapping, generating feature importance bar charts, according to Sort from high to low to intuitively display the key factors.

[0099] Generate optimization recommendations, analyze deviations, and compare forecast results With target value Deviation of the environmental factors to be optimized is determined by the deviation formula: ; In the formula, It is indicated that bird habitat features need to be enhanced. It indicated that the negative impact of environmental factors on birds needs to be reduced; Environmental factors for bias, combined with feature importance Based on the historical data, we propose adjustment strategies, specifically: Water level optimization: If the water level has a significant impact on the prediction results, it is recommended to adjust the water level to the optimal range.

[0100] Vegetation restoration: If there is a significant positive correlation between vegetation coverage and habitat area, it is recommended to increase coverage through planting.

[0101] Salinity control: if salinity affects the habitat time, a reasonable salinity adjustment method should be proposed.

[0102] All analysis results and optimization suggestions are organized into interactive visualization reports, which include: map heat maps showing the distribution of bird habitat patterns in different regions; feature importance charts to identify key environmental factors; and optimization plan lists to provide targeted suggestions based on deviation analysis.

[0103] It should be noted that when generating interactive visualization reports, the map heat map contained in the report should be highly consistent with the actual geographic information of the wetland, and the detailed information of the specific area can be zoomed in or out as needed; the feature importance chart and optimization plan list must provide export function to facilitate archiving or sharing by wetland protection managers; at the same time, it is recommended to add a deviation analysis module between the predicted results and the actual results in the report to verify the reliability of the model.

[0104] Please refer to Figure 1 , exemplarily, the following will substitute a set of experimental data to verify and demonstrate the specific process of this method.

[0105] Wetland environmental data: collect wetland water level (2.5m~4.0m), vegetation coverage (30%~70%), salinity .

[0106] Bird habitat data: collected bird population number (50,200), habitat area (520 square kilometers), and habitat activity duration (2 to 10 hours).

[0107] Total data sample size: 300 records.

[0108] The collected wetland environment data and bird habitat data are converted into data formats and uniformly stored in matrix form D; those exceeding the abnormal threshold are eliminated. After cleaning, the data volume is 280; each indicator is normalized to Range, the standardized data matrix is ​​recorded as ; For the 10 missing records of vegetation coverage, the Lagrange interpolation method was used to complete the data.

[0109] Constructing the data matrix , where each column is the wetland water level, vegetation coverage and salinity; Calculate the covariance matrix C, perform eigenvalue decomposition on it, and obtain the eigenvalue set and the corresponding eigenvectors; extract the two principal components whose cumulative variance contribution rate reaches 90% to form the principal component matrix , and reduce the data dimension to ; According to the principal component load, the key factors are determined to be wetland water level and vegetation coverage.

[0110] The k-means clustering algorithm was used, and the number of clusters k was set to 3. The 280 data were divided into three clusters by optimizing the objective function J. The clustering results showed that birds with large populations mainly inhabited areas with high vegetation coverage and moderate water levels. The seasonal variation characteristics showed that the impact of salinity was mainly concentrated in autumn and winter.

[0111] The random forest regression algorithm was used to construct an association model between wetland environmental factors and bird population size; the training set and test set accounted for 80% and 20% of the total data, respectively; the model validation results showed that the mean square error was 5.3, indicating a high prediction accuracy; the feature importance analysis results showed that the weight of wetland water level was 0.45, vegetation coverage was 0.35, and salinity was 0.20.

[0112] Based on the analysis results, a heat map of the wetland is generated to show the distribution of bird populations in different areas; optimization suggestions are put forward: Water level optimization: It is recommended to maintain the wetland water level in the range of 3.0m~3.5m.

[0113] Vegetation restoration: It is recommended to increase wetland plant planting in areas where vegetation coverage is less than 40%.

[0114] Salinity management: Reduce salinity in high-salinity areas to below 0.9g / L through water diversion measures.

[0115] Through the above process, a correlation model between wetland environmental factors and bird habitat characteristics was successfully constructed, key factors were identified and optimization suggestions were put forward, providing a scientific basis for wetland management and bird protection.

[0116] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for analyzing the association between coastal wetland environment and bird habitat, characterized in that: The steps include: Acquire and preprocess wetland environmental data to obtain preprocessed wetland environmental data; The pre-processed wetland environmental data were analyzed by principal component analysis to determine the key environmental factors affecting bird habitats. Analyze bird habitat patterns and seasonal changes based on clustering algorithms; The association model between wetland environmental factors and bird habitat characteristics is constructed through machine learning algorithms, and the analysis results are output based on the association model; Provide a visual map of the analysis results and generate optimization suggestions based on the analysis results.

2. A method for analyzing the association between coastal wetland environment and bird habitat as claimed in claim 1, characterized in that: Obtain wetland environmental data and preprocess them to obtain preprocessed wetland environmental data, including: Wetland environmental data include: wetland water level, wetland vegetation coverage, wetland salinity and bird habitat data; bird habitat data include: bird population, habitat area and duration of habitat activities; Preprocessing includes data format conversion, data cleaning, data normalization and data completion; Data format conversion: the collected wetland environmental data is converted into a unified format, and the converted data is expressed as follows: ； In the formula, Represents the i-th original data record; Data cleaning, detect and remove outliers in wetland environmental data. The outlier determination formula is: ; In the formula, is the data record value, is the mean of similar data, is the standard deviation, k is the preset abnormal threshold coefficient; the cleaned data is expressed as: ; Data normalization: normalize the cleaned wetland environmental data to eliminate the impact of different indicator units and magnitudes. The normalization formula is: ; In the formula, x is the original data value, is the minimum value of the same indicator data, is the maximum value of the same indicator data, is the normalized data value; the normalized data is expressed as: ; Data completion: missing values ​​in the normalized wetland environmental data are completed by interpolation algorithm. The interpolation algorithm adopts Lagrange interpolation method, and its calculation formula is: ; In the formula, , are known data points, is the missing value for the prediction.

3. The method for analyzing the association between coastal wetland environment and bird habitat according to claim 1, characterized in that: The pre-processed wetland environmental data were analyzed by principal component analysis to determine the key environmental factors affecting bird habitats, including: Data matrix construction, representing the preprocessed wetland environmental data in matrix form: ; In the formula, m is the number of samples, n is the number of environmental factors, represents the value of the jth environmental factor in the i-th sample; Compute the covariance matrix of the data matrix , and its calculation formula is: ; In the formula, represents the i-th row in the data matrix, is the mean vector of all samples, and C represents the covariance relationship between environmental factors; Perform eigendecomposition on the covariance matrix C to obtain the eigenvalue set and the corresponding set of eigenvectors , where represents the variance contribution of the ith principal component, Representation and Eigenvalues The corresponding eigenvector; Principal component extraction, sort the principal components according to the size of the eigenvalues, and select the eigenvectors corresponding to the first k eigenvalues ​​to form the principal component matrix: , the following conditions are met: ; In the formula, is the threshold of the cumulative variance contribution rate, k is the number of principal components that meet the threshold; Principal component mapping, mapping the original data matrix X to the principal component space to obtain the principal component data matrix , and its calculation formula is: ; In the formula, Z represents the principal component data after dimensionality reduction, is the selected principal component matrix; Determine the key factors based on the principal component matrix The coefficient value of the filter is used to screen the environmental factors that contribute most to the principal component and determine the key factor set that affects bird habitat. , where t represents the number of key factors.

4. The method for analyzing the association between coastal wetland environment and bird habitat according to claim 1, characterized in that: Based on the clustering algorithm, the bird habitat patterns and seasonal changes are analyzed, including: The k-means clustering algorithm is used, the number of clusters is set to K, and the sample classification is achieved by minimizing the square distance from the sample point in the cluster to the cluster center. The objective function is: ; In the formula, is the kth cluster; is the cluster center; Determine the optimal number of clusters and calculate the objective function value according to the time rule The trend of changing with the number of clusters K, select The position where the value decreases significantly and tends to be flat is taken as the optimal number of clusters; Extract pattern features and calculate feature mean for each cluster , extract the characteristics of bird habitat patterns, and the extraction formula is: ; In the formula, Indicates the number of samples in the cluster.

5. The method for analyzing the association between coastal wetland environment and bird habitat according to claim 1, characterized in that: The association model between wetland environmental factors and bird habitat characteristics is constructed through machine learning algorithms, and the analysis results are output based on the association model, including: The random forest regression algorithm is used to build the association model. The model structure consists of multiple decision trees. The prediction formula of random forest is: ; In the formula, is the predicted value, T is the total number of decision trees, is the prediction result of the tth decision tree; The data set is split into a training set and a test set. The training set is used for model training, and the test set is used to verify the model performance. The mean square error is used as the loss function to train the model. By calculating the impact of each environmental factor on the model prediction results, the importance of the factor is evaluated. The feature importance is determined by reducing the contribution value of the mean square error. The determination formula is: ; In the formula, Representation Factor The importance of is the inclusion factor The mean square error when To remove the factor The mean square error when The new wetland environmental data is predicted based on the trained model to generate prediction results of bird habitat characteristics. Based on the prediction results of bird habitat characteristics and combined with feature importance analysis, optimization suggestions are generated.

6. The method for analyzing the association between coastal wetland environment and bird habitat according to claim 1, characterized in that: The trained model is used to predict new wetland environmental data and generate prediction results of bird habitat characteristics. Based on the prediction results of bird habitat characteristics and combined with feature importance analysis, optimization suggestions are generated, including: Identify environmental factors that significantly affect bird habitats and analyze the deviation between the predicted results and the target values. If the predicted results show that the population of a certain bird species is lower than expected, adjust the key environmental factors based on the characteristic significance and propose optimization strategies for the environmental factors that significantly affect bird habitats. Specifically: Water level adjustment: Based on feature significance analysis, if water level is a key factor, it is recommended to maintain the water level within a certain range; Vegetation restoration: if the vegetation coverage rate has a significant impact on the bird habitat area, it is recommended to increase the planting of wetland vegetation; Salinity management, where high salinity has a negative impact on bird activity duration, propose operational methods to reduce salinity.

7. The method for analyzing the association between coastal wetland environment and bird habitat according to claim 1, characterized in that: Provides a visual map of the analysis results and generates optimization suggestions based on the analysis results, including: Include the analysis results in the prediction of bird habitat characteristics and environmental factor characteristics Summarized as a mapping dataset, the dataset is represented as: ; In the formula, represents the jth environmental factor, Representation Factor The feature importance of Represents the prediction result of the i-th sample; Select the corresponding mapping dimension based on the analysis results, specifically: Spatial mapping, based on wetland geospatial information, generates distribution heat maps of bird habitat patterns; Time mapping, based on time series data, shows seasonal trends; Significance mapping, which generates a factor contribution map to bird habitat based on feature significance sorting; The wetland area is divided into grid units, and the habitat characteristic prediction value corresponding to each unit reflects the bird population and habitat area through the depth of color to achieve visual mapping.

8. The method for analyzing the association between coastal wetland environment and bird habitat according to claim 1, characterized in that: The wetland area is divided into grid cells, and the habitat feature prediction value corresponding to each cell reflects the bird population and habitat area through the depth of color to achieve visual mapping, including: Spatial mapping, the heat map formula is: ; In the formula, is the color value of the grid cell, is the contribution weight of the i-th sample to the grid unit, is the predicted value; Time mapping, drawing a time series line graph to show the changing trends of bird habitat characteristics in different time periods; Saliency mapping, generating feature importance bar charts, according to Sort from high to low to intuitively display key factors; Generate optimization recommendations, analyze deviations, and compare forecast results With target value Deviation, determine the environmental factors to be optimized, the deviation formula is: ; In the formula, It is indicated that bird habitat features need to be enhanced. It indicated that the negative impact of environmental factors on birds needs to be reduced; Environmental factors for bias, combined with feature importance Based on the historical data, we propose adjustment strategies, specifically: Water level optimization: if the water level has a significant impact on the prediction results, it is recommended to adjust the water level to the optimal range; Vegetation restoration: if there is a significant positive correlation between vegetation coverage and habitat area, it is recommended to increase coverage through planting; Salinity control: if salinity affects the roosting time, a reasonable salinity adjustment method is proposed; All analysis results and optimization suggestions are organized into interactive visualization reports, which include: map heat maps showing the distribution of bird habitat patterns in different regions; feature importance charts to identify key environmental factors; and optimization plan lists to provide targeted suggestions based on deviation analysis.

Images