Method for monitoring ecological environment of vegetation in small watershed in arid region

Abstract

The invention discloses an arid region small watershed vegetation ecological environment monitoring method comprising the following steps: obtaining an arid region small watershed earth surface image, processing the arid region small watershed earth surface image, and obtaining a first vegetation distribution image; acquiring a second vegetation distribution image according to matching of the first vegetation distribution image and a spectral feature library; extracting spatial distribution characteristics of a vegetation community from the second vegetation distribution image, calculating a dynamic change trend of the spatial distribution characteristics, and obtaining a community structure change data set; based on the community structure change data set, predicting a short-term evolution trend of the vegetation community structure by adopting a long-short-term memory network, and obtaining a dynamic change prediction result of the vegetation community structure; according to the dynamic change prediction result, combining the community structure change data set to quantify the biodiversity level to obtain a biodiversity evaluation result; key indexes are extracted from the biodiversity evaluation result, and vegetation ecological environment state data are monitored in real time according to the key indexes.

Description

Technical Field

[0001] The invention belongs to the technical field of environmental monitoring, and in particular relates to a method for monitoring the ecological environment of vegetation in a small watershed in arid areas. Background Art

[0002] Vegetation ecological environment monitoring is a core area of ecological protection and sustainable development. It is crucial for maintaining ecological balance and assessing environmental changes. This is especially true in small watersheds in arid areas with scarce water resources and fragile ecosystems. The state of vegetation directly reflects the health of the ecosystem. However, existing monitoring methods mostly rely on manual surveys or traditional remote sensing technology. Manual surveys are time-consuming and labor-intensive, with limited coverage. Traditional remote sensing technology has difficulty accurately identifying drought-resistant vegetation types unique to arid areas due to insufficient resolution or confusion of vegetation spectral characteristics. These limitations result in low monitoring accuracy and efficiency, making it impossible to dynamically track changes in vegetation community structure. Summary of the Invention

[0003] In order to solve the above technical problems, the present invention proposes a method for monitoring the ecological environment of vegetation in small watersheds in arid areas, which can realize the accurate identification, dynamic monitoring and prediction of the ecological environment of vegetation in small watersheds in arid areas, and provide a scientific basis for ecological protection and resource management.

[0004] The present invention provides a method for monitoring the ecological environment of vegetation in a small watershed in an arid area, comprising:

[0005] Acquiring a surface image of a small watershed in an arid area, and processing the surface image of the small watershed in the arid area to obtain a first vegetation distribution image;

[0006] Acquiring a second vegetation distribution image based on matching the first vegetation distribution image with the spectral feature library;

[0007] Extracting the spatial distribution characteristics of the vegetation community from the second vegetation distribution image, calculating the dynamic change trend of the spatial distribution characteristics, and obtaining a community structure change data set;

[0008] Based on the community structure change dataset, a long short-term memory network is used to predict the short-term evolution trend of the vegetation community structure and obtain the dynamic change prediction results of the vegetation community structure;

[0009] Based on the dynamic change prediction results, combined with the community structure change dataset, the biodiversity level is quantified to obtain the biodiversity assessment results;

[0010] Key indicators are extracted from the biodiversity assessment results, and vegetation ecological environment status data is monitored in real time based on the key indicators.

[0011] Optionally, processing the surface image of the small watershed in the arid area to obtain the first vegetation distribution image includes:

[0012] Using preprocessing technology to denoise and enhance the surface image of the small watershed in the arid area to obtain the first surface image;

[0013] Based on the first surface image, a deep convolutional neural network is used to extract features of complex terrain and sparse vegetation, segment the vegetation area, and obtain a first vegetation distribution image.

[0014] Optionally, obtaining the second vegetation distribution image according to the matching of the first vegetation distribution image with the spectral feature library includes:

[0015] Using the spectral feature library to extract features from the first vegetation image to obtain an initial matching feature set;

[0016] constructing a drought-tolerant vegetation classification model based on the initial matching feature set to obtain a classification model;

[0017] If the matching rate of the classification model to the first vegetation image is lower than a preset threshold, the classification model is optimized using a gradient descent method to obtain an optimized classification model;

[0018] The first vegetation image is classified using the optimized classification model to obtain a second vegetation image.

[0019] Optionally, extracting the spatial distribution characteristics of the vegetation community from the second vegetation distribution image, calculating the dynamic change trend of the spatial distribution characteristics, and obtaining the community structure change dataset includes:

[0020] extracting spatial distribution features of vegetation communities from the second vegetation distribution image using an image segmentation algorithm to generate a spatial feature data set;

[0021] If the pixel value distribution of the spatial feature data set meets the preset threshold, the spatial features are classified by a clustering algorithm to obtain a community structure classification result;

[0022] Based on the community structure classification results, a spatiotemporal analysis algorithm is used to calculate the dynamic changes of the community structure and generate a dynamic change data set;

[0023] Analyze the time trend of the dynamic change data set by a linear regression algorithm to obtain a change trend data set;

[0024] Extracting significantly changed community structure features from the change trend data set to generate a structure change feature set;

[0025] According to the structural change feature set, data fusion technology is used to integrate spatial characteristics and change trends to generate the community structure change dataset.

[0026] Optionally, based on the community structure change dataset, a long short-term memory network is used to predict the short-term evolution trend of the vegetation community structure, and the dynamic change prediction results of the vegetation community structure are obtained, including:

[0027] Obtain a dataset of vegetation community structure changes, determine the data characteristics through time series analysis, and obtain a standardized dataset;

[0028] Using data processing technology to preprocess the standardized data set to obtain a preprocessed data set;

[0029] The preprocessed data set is trained using a long short-term memory network to obtain a prediction model;

[0030] Based on the prediction result, the prediction result is obtained.

[0031] Optionally, based on the dynamic change prediction results, combined with the community structure change dataset, the biodiversity level is quantified to obtain a biodiversity assessment result including:

[0032] The species diversity index calculation method was used, combined with the community dynamics trend, to calculate the species diversity level and obtain preliminary quantitative results;

[0033] If the preliminary quantitative results are inconsistent with the preset ecosystem dynamic threshold, the species distribution feature weights are adjusted using the random forest algorithm to obtain an optimized quantitative result;

[0034] Based on the optimized quantitative results, analyze the impact of environmental factors, extract key environmental driving factors, and determine the characteristics of biodiversity changes;

[0035] By using the characteristics of biodiversity changes and adopting time series analysis methods, the dynamic trends of the community are predicted to obtain the biodiversity assessment results.

[0036] Optionally, key indicators are extracted from the biodiversity assessment results, and real-time monitoring of vegetation ecological environment status data based on the key indicators includes:

[0037] Extract key indicators from biodiversity assessment results through data cleaning to obtain a standardized indicator set;

[0038] Perform real-time analysis on standardized indicator sets to obtain real-time data streams;

[0039] By comparing the real-time data stream with historical data, calculating the magnitude of the change, and determining whether it exceeds a preset threshold;

[0040] If the change exceeds the preset threshold, the dynamic update mechanism is triggered to generate updated vegetation ecological environment status data;

[0041] Based on the updated vegetation ecological environment status data, time series analysis is used to obtain the trend of environmental status changes;

[0042] The changing trends were classified by clustering algorithms to obtain the vegetation ecological characteristics of different regions;

[0043] Visualization technology is used to dynamically present the ecological characteristics of vegetation and obtain real-time monitoring data.

[0044] Compared with the existing technology, the present invention has the following advantages and technical effects: the present invention obtains vegetation spectral characteristics through high-resolution multispectral remote sensing data, constructs a drought-resistant vegetation spectral feature library, and uses support vector machines and random forest algorithms to establish a classification model. Combined with high-resolution drone images, deep convolutional neural networks and transfer learning techniques are used to realize vegetation type identification and distribution image generation. Spatiotemporal analysis algorithms and long short-term memory networks are further used to predict the changing trends of vegetation community structure and calculate the biodiversity index. Finally, through real-time data stream processing and visualization technology, the vegetation ecological environment monitoring results are dynamically updated and output. The present invention realizes the accurate identification, dynamic monitoring and prediction of the vegetation ecological environment in small watersheds in arid areas, providing a scientific basis for ecological protection and resource management. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The accompanying drawings, which constitute part of this application, are intended to provide a further understanding of this application. The exemplary embodiments and descriptions of this application are intended to explain this application and do not constitute an improper limitation on this application. In the accompanying drawings:

[0046] Figure 1 This is a flow chart of a method for monitoring the ecological environment of vegetation in a small watershed in an arid area according to an embodiment of the present invention. DETAILED DESCRIPTION

[0047] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0048] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0049] This embodiment proposes a method for monitoring the ecological environment of vegetation in a small watershed in arid areas. Figure 1 As shown, the specific steps include:

[0050] Acquiring a surface image of a small watershed in the arid area, processing the surface image of the small watershed in the arid area, and acquiring a first vegetation distribution image;

[0051] Acquiring a second vegetation distribution image based on matching the first vegetation distribution image with the spectral feature library;

[0052] Extracting the spatial distribution characteristics of the vegetation community from the second vegetation distribution image, calculating the dynamic change trend of the spatial distribution characteristics, and obtaining a community structure change data set;

[0053] Based on the community structure change dataset, a long short-term memory network is used to predict the short-term evolution trend of vegetation community structure and obtain the prediction results of the dynamic change of vegetation community structure;

[0054] Based on the dynamic change prediction results, combined with the community structure change dataset, the biodiversity level is quantified to obtain the biodiversity assessment results;

[0055] Extract key indicators from the biodiversity assessment results and monitor vegetation ecological environment status data in real time based on the key indicators.

[0056] Specifically, this embodiment obtains vegetation spectral characteristics through high-resolution multispectral remote sensing data, constructs a drought-resistant vegetation spectral feature library, and uses support vector machines and random forest algorithms to establish a classification model. Combined with high-resolution drone images, deep convolutional neural networks and transfer learning techniques are used to realize vegetation type recognition and distribution image generation. Spatiotemporal analysis algorithms and long short-term memory networks are further used to predict the changing trend of vegetation community structure and calculate the biodiversity index. Finally, through real-time data stream processing and visualization technology, the vegetation ecological environment monitoring results are dynamically updated and output. The present invention realizes the accurate identification, dynamic monitoring and prediction of the vegetation ecological environment in small watersheds in arid areas, providing a scientific basis for ecological protection and resource management.

[0057] Furthermore, processing the surface image of the small watershed in the arid area to obtain a first vegetation distribution image includes:

[0058] Using preprocessing technology to denoise and enhance the surface image of the small watershed in the arid area to obtain the first surface image;

[0059] Based on the first surface image, a deep convolutional neural network is used to extract features of complex terrain and sparse vegetation, segment the vegetation area, and obtain a first vegetation distribution image.

[0060] Specifically, when acquiring surface images of small watersheds in arid regions using drone-mounted high-resolution cameras, it is preferable to use a multispectral camera to capture the surface's reflective properties in different wavelengths. For example, visible and near-infrared wavelengths can be selected, with a resolution of 0.1 meter per pixel, ensuring that details of sparse vegetation and complex terrain are captured. The sun's altitude should be taken into consideration during acquisition to avoid shadow interference, and image capture should preferably be performed between 10:00 AM and 2:00 PM. Preprocessing techniques for image denoising and enhancement include, specifically, median filtering to remove salt-and-pepper noise while preserving vegetation edge information; histogram equalization is then used to enhance image contrast. For example, for images with uneven grayscale distribution, adjustments can be made to make the pixel value difference between vegetation and bare soil more distinct, facilitating subsequent segmentation. This approach improves image quality and provides reliable data for feature extraction. When using deep convolutional neural networks for feature extraction and vegetation segmentation, the U-Net model is understandably often chosen due to its adaptability to small sample sizes. For example, during training, a dataset containing 1,000 annotated images is input. The network extracts the texture and spectral features of vegetation through multi-layer convolution and outputs a first vegetation distribution image. During segmentation, the model can distinguish between sparse shrubs and bare soil, generating a pixel-level vegetation distribution map. For example, areas with a Normalized Difference Vegetation Index (NDVI) greater than 0.3 can be designated as vegetation areas. For example, an area with an NDVI value of 0.4 is marked as vegetation; an area with an NDVI value below 0.3 is marked as non-vegetation.

[0061] Furthermore, obtaining a second vegetation distribution image according to the matching of the first vegetation distribution image with the spectral feature library includes:

[0062] Using the spectral feature library to extract features from the first vegetation image to obtain an initial matching feature set;

[0063] Building a drought-tolerant vegetation classification model based on the initial matching feature set to obtain a classification model;

[0064] If the matching rate of the classification model to the first vegetation image is lower than a preset threshold, the classification model is optimized using a gradient descent method to obtain an optimized classification model;

[0065] The first vegetation image is classified by optimizing the classification model to obtain a second vegetation image.

[0066] Specifically, the spectral feature library is used to extract features of the first vegetation image, aiming to match them using known vegetation spectral feature templates. For example, in a small watershed in an arid area, the spectral feature library may contain spectral data for a variety of drought-resistant vegetation, such as camel thorn and sea buckthorn, with wavelengths ranging from visible light to near-infrared. In one embodiment, the system extracts features with a matching degree of more than 80% by comparing the spectral curves of the image pixels with the templates in the library to form an initial matching feature set. This ensures that the subsequent classification model can focus on the spectral characteristics of typical vegetation in the area. Transfer learning technology is used to construct a drought-resistant vegetation classification model and train it based on the initial matching feature set. It can be understood that transfer learning reduces the need for large amounts of labeled data by reusing the deep feature extraction capabilities of pre-trained models such as ResNet. In one possible implementation, assuming that the initial matching feature set contains 1,000 samples, transfer learning can transfer the convolutional layer parameters of the pre-trained model to the new model, and only fine-tune the fully connected layer to generate a classification model suitable for the specific classification task of drought-resistant vegetation. If the matching rate of the classification model is lower than a preset threshold, such as 85%, the gradient descent method is used for fine-tuning to obtain an optimized classification model. For example, the fine-tuning process may adjust the learning rate of the model to 0.001, and optimize the loss function through multiple iterations to improve the classification accuracy of the model for sparse vegetation. Specifically, the fine-tuning may adjust the weight of the near-infrared band in the spectral characteristics to better distinguish healthy vegetation from non-vegetated areas. The first vegetation image is classified by optimizing the classification model to obtain a second vegetation image. For example, the classification model may divide the image pixels into three categories: drought-resistant vegetation, non-vegetation, and transition areas, and generate a second vegetation image containing category labels.

[0067] Furthermore, extracting the spatial distribution characteristics of the vegetation community from the second vegetation distribution image, calculating the dynamic change trend of the spatial distribution characteristics, and obtaining the community structure change dataset includes:

[0068] extracting spatial distribution features of vegetation communities from the second vegetation distribution image using an image segmentation algorithm to generate a spatial feature data set;

[0069] If the pixel value distribution of the spatial feature dataset meets the preset threshold, the spatial features are classified using a clustering algorithm to obtain the community structure classification results;

[0070] Based on the results of community structure classification, spatiotemporal analysis algorithms are used to calculate the dynamic changes of community structure and generate a dynamic change data set;

[0071] Analyze the time trend of dynamic changes in a dynamically changing data set through a linear regression algorithm to obtain a change trend data set;

[0072] Extract significantly changed community structure features from the change trend dataset to generate a structure change feature set;

[0073] According to the structural change feature set, data fusion technology is used to integrate spatial characteristics and change trends to generate a community structure change dataset.

[0074] Specifically, for example, an image segmentation algorithm is used to extract the spatial distribution characteristics of vegetation communities from the second vegetation distribution image. A segmentation algorithm based on region growing can be used. In principle, the algorithm expands from a seed point based on the grayscale value or color characteristics of the pixels, merging similar pixels to form community areas. For example, in a vegetation image of an arid area, the seed point is set to a green pixel, and the threshold is a grayscale difference of 10. Community areas dominated by drought-resistant shrubs are segmented, generating a spatial feature dataset containing community boundaries and areas, with the area unit being square kilometers. In one possible implementation, if the pixel value distribution of the spatial feature dataset meets a preset threshold, such as a pixel value variance of less than 50, the spatial features can be classified using a K-means clustering algorithm. K-means clustering divides the community into high-density shrub areas and sparse grass areas based on the spatial coordinates and color characteristics of the pixels. For example, if the K value is set to 3, the clustering results show that the shrub area accounts for 60%, the grass area accounts for 30%, and the bare land accounts for 10%, clearly reflecting the community structure. Specifically, based on the community structure classification results, a spatiotemporal analysis algorithm is used to calculate the dynamic changes in community structure. Spatiotemporal analysis quantifies area changes by comparing community boundaries in multi-temporal images. For example, by analyzing images from 2023 and 2024, it was found that the area of the shrub area increased from 100 square kilometers to 120 square kilometers. A dynamic change dataset containing the change amplitude and timestamp is generated, revealing the trend of community expansion. Preferably, if the time series of the dynamic change dataset meets the continuity condition, such as collecting data for 6 consecutive months every month, the time trend can be analyzed by a linear regression algorithm. Linear regression is used to fit the area change curve to obtain a change trend dataset with a positive slope. For example, the area of the shrub area increases by 2 square kilometers per month, indicating that the community is expanding steadily, providing a basis for ecological monitoring. In one embodiment, community structure features that have changed significantly are extracted from the change trend dataset. The change rate threshold can be set to 5%, and the shrub area expansion features are extracted to generate a structural change feature set. For example, the shrub area boundary expands 10 kilometers to the north, reflecting the impact of improved water sources or increased rainfall, providing data support for vegetation restoration research. For example, data fusion technology is used to integrate spatial characteristics and change trends to generate a community structure change dataset.

[0075] Furthermore, based on the community structure change dataset, a long short-term memory network is used to predict the short-term evolution trend of vegetation community structure. The prediction results of the dynamic change of vegetation community structure include:

[0076] Obtain a dataset of vegetation community structure changes, determine the data characteristics through time series analysis, and obtain a standardized dataset;

[0077] Using data processing technology to preprocess the standardized data set to obtain a preprocessed data set;

[0078] The preprocessed data set is trained using a long short-term memory network to obtain a prediction model;

[0079] Based on the prediction result, the prediction result is obtained.

[0080] Specifically, after obtaining a dataset on vegetation community structure changes, time series analysis is a key step for extracting data features. For example, time series analysis can use decomposition methods to extract trends, seasonality, and random fluctuations. For example, analyzing monthly vegetation coverage data for a region over a 10-year period can identify seasonal fluctuations in coverage, such as an increase of approximately 15% in spring and a decrease of approximately 10% in autumn. This analysis helps understand the dynamics of vegetation communities and provides a foundation for subsequent modeling. In one possible implementation, generating a standardized dataset involves converting the raw data to a format with a mean of 0 and a standard deviation of 1. Specifically, assuming that the mean of vegetation community coverage data is 50% and the standard deviation is 5%, normalization converts the data to a standard normal distribution, facilitating model training. Standardization ensures comparability of data across different dimensions and improves model convergence. It should be noted that data cleaning is a crucial step in removing outliers. For example, the coverage recorded at a certain point in time may be as high as 90%, far exceeding the normal range of 40%-60%. Outliers are identified and removed through boxplot analysis. The cleaned preprocessed dataset better reflects the true changing trends of vegetation communities and avoids model bias due to noisy data. Long short-term memory networks are preferably used to train preprocessed datasets because they excel at capturing long- and short-term dependencies in time series. In one embodiment, a three-layer network is set up for data on a specific forest community. Coverage data from the past five years is input, and the model is trained to predict trends over the next 12 months. The initial prediction model may predict a 2% decrease in coverage, but the actual data only shows a 0.5% decrease, exceeding the threshold. By adjusting the network learning rate to 0.001 and adding hidden layer nodes, the optimized model's prediction error is reduced to within 0.3%. For example, dynamic change prediction results require consistency verification through time series analysis. Specifically, the prediction results indicate that the coverage of a grassland community will decrease by 5% over the next three years. Analyzing historical data using an autoregressive model confirms that the downward trend is consistent with the fluctuation pattern of the past 10 years, verifying the credibility of the prediction. This verification ensures that the prediction results are consistent with actual ecological processes. It is understandable that optimizing the prediction model's ability to predict short-term evolutionary trends is of great significance to ecological management. For example, if a wetland community's coverage is predicted to drop by 3% in the short term, protective measures can be implemented in advance, such as restricting development or increasing irrigation. This predictive capability supports precise decision-making and reduces ecological risks. In one embodiment, when adjusting the parameters of a long-short-term memory network, a grid search can be used to determine the optimal parameter combination. For example, a learning rate of 0.001-0.01 and a hidden layer number of 50-200 can be tried to find the configuration with the lowest error. This approach ensures that the model adapts to the data characteristics of different vegetation communities and improves prediction accuracy.

[0081] Furthermore, based on the dynamic change prediction results and combined with the community structure change dataset, the biodiversity level was quantified, and the biodiversity assessment results included:

[0082] The species diversity index calculation method was used, combined with the community dynamics trend, to calculate the species diversity level and obtain preliminary quantitative results;

[0083] If the preliminary quantitative results are inconsistent with the preset ecosystem dynamic threshold, the species distribution feature weights are adjusted using the random forest algorithm to obtain an optimized quantitative result;

[0084] Based on the optimized quantitative results, analyze the impact of environmental factors, extract key environmental driving factors, and determine the characteristics of biodiversity changes;

[0085] By analyzing the changing characteristics of biodiversity and using time series analysis methods, we can predict community dynamic trends and obtain biodiversity assessment results.

[0086] Specifically, in the field of analyzing changes in vegetation community structure, data cleaning technology is a key step in ensuring data quality. For example, for a dataset of changes in community structure, outliers can be detected using statistical methods. For example, suppose a dataset records the species density of 100 sample plots in a certain area. If the species density of a certain sample plot is found to be far greater than 3 times the standard deviation of the mean, it can be treated as noise data and removed. After cleaning, the data can better reflect the true community characteristics, which helps to improve the accuracy of subsequent analysis. Specifically, the acquisition of a standardized dataset requires unifying data of different dimensions. For example, indicators such as species density and cover are standardized using Z-scores and converted into datasets with a mean of 0 and a standard deviation of 1. This eliminates dimensional differences and facilitates time series analysis. In one embodiment, data from a forest plot contains species density and soil moisture. After standardization, their changing trends can be directly compared to reveal the dynamic characteristics of the community. Preferably, time series analysis is used to extract species distribution characteristics. For example, by analyzing five years of species abundance data of a grassland community using an autoregressive model, it can be found that the abundance of a dominant species fluctuates periodically, suggesting that the community dynamic trend is related to seasonal rainfall. In one possible implementation, the Shannon index can be used to calculate the species diversity index. For example, suppose a wetland community contains 10 plant species. After calculating the relative abundance of each species, a diversity index of 2.3 indicates high community diversity. If this value falls below the preset threshold of 2.5, further analysis of community stability is required. This method intuitively quantifies community structure and facilitates ecological assessment.

[0087] Furthermore, key indicators are extracted from the biodiversity assessment results. Based on these key indicators, the real-time monitoring of vegetation ecological environment status data includes:

[0088] Extract key indicators from biodiversity assessment results through data cleaning to obtain a standardized indicator set;

[0089] Perform real-time analysis on standardized indicator sets to obtain real-time data streams;

[0090] By comparing real-time data streams with historical data, the magnitude of change is calculated to determine whether it exceeds the preset threshold;

[0091] If the change exceeds the preset threshold, the dynamic update mechanism is triggered to generate updated vegetation ecological environment status data;

[0092] Based on the updated vegetation ecological environment status data, time series analysis is used to obtain the trend of environmental status changes;

[0093] The changing trends were classified by clustering algorithms to obtain the vegetation ecological characteristics of different regions;

[0094] Visualization technology is used to dynamically present the ecological characteristics of vegetation and obtain real-time monitoring data.

[0095] Specifically, missing data is filled using the average of neighboring sampling points, while abnormally high-density records due to human interference are removed. The cleaned data forms a standardized set of indicators, such as species richness, evenness, and community density, providing a reliable foundation for subsequent analysis. In one possible implementation, stream processing technology performs real-time analysis of this standardized set of indicators. Stream processing can process continuously input data streams, making it suitable for dynamically monitoring ecosystem changes. Compared to traditional batch processing, stream processing can respond to ecological anomalies more quickly, improving monitoring efficiency. Specifically, the real-time data stream is compared with historical data to calculate the magnitude of change. Historical data can come from monthly records over the past five years, while the real-time data stream represents the current hourly monitoring value. For example, if the species richness of a forest area drops from a historical average of 50 species to a real-time value of 40 species, the change is 20%. If the preset threshold is 15%, a dynamic update mechanism is triggered. This comparison can promptly identify ecosystem anomalies and provide a basis for protective measures. Preferably, the dynamic update mechanism generates updated data on the status of the vegetation ecosystem. For example, when the vegetation coverage of a grassland area decreases due to drought, the system will update the status based on real-time data and generate comprehensive status data including coverage, species diversity and soil moisture. These data provide precise guidance for ecological restoration and avoid waste of resources. In one embodiment, time series analysis is used to predict the trend of changes in environmental status. For example, based on the vegetation coverage data of the past three years, seasonal fluctuations and long-term downward trends are analyzed, and it is predicted that the coverage rate may decrease by 5% in the next year. This prediction helps to formulate protection plans in advance and optimize resource allocation. For example, clustering algorithms classify changing trends to identify regional characteristics.

[0096] The above are merely preferred embodiments of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A method for monitoring the ecological environment of vegetation in a small watershed in an arid area, characterized in that: include: Acquiring a surface image of a small watershed in an arid area, and processing the surface image of the small watershed in the arid area to obtain a first vegetation distribution image; Acquiring a second vegetation distribution image based on matching the first vegetation distribution image with the spectral feature library; Extracting the spatial distribution characteristics of the vegetation community from the second vegetation distribution image, calculating the dynamic change trend of the spatial distribution characteristics, and obtaining a community structure change data set; Based on the community structure change dataset, a long short-term memory network is used to predict the short-term evolution trend of the vegetation community structure and obtain the dynamic change prediction results of the vegetation community structure; Based on the dynamic change prediction results, combined with the community structure change dataset, the biodiversity level is quantified to obtain the biodiversity assessment results; Key indicators are extracted from the biodiversity assessment results, and vegetation ecological environment status data is monitored in real time based on the key indicators.

2. The method for monitoring the ecological environment of vegetation in a small watershed in arid areas according to claim 1, characterized in that: Processing the surface image of the small watershed in the arid area to obtain the first vegetation distribution image includes: Using preprocessing technology to denoise and enhance the surface image of the small watershed in the arid area to obtain the first surface image; Based on the first surface image, a deep convolutional neural network is used to extract features of complex terrain and sparse vegetation, segment the vegetation area, and obtain a first vegetation distribution image.

3. The method for monitoring the ecological environment of vegetation in a small watershed in arid areas according to claim 1, characterized in that: Acquiring a second vegetation distribution image according to matching the first vegetation distribution image with the spectral feature library includes: Using the spectral feature library to extract features from the first vegetation image to obtain an initial matching feature set; constructing a drought-tolerant vegetation classification model based on the initial matching feature set to obtain a classification model; If the matching rate of the classification model to the first vegetation image is lower than a preset threshold, the classification model is optimized using a gradient descent method to obtain an optimized classification model; The first vegetation image is classified using the optimized classification model to obtain a second vegetation image.

4. The method for monitoring the ecological environment of vegetation in a small watershed in an arid area according to claim 1, characterized in that: Extracting the spatial distribution characteristics of vegetation communities from the second vegetation distribution image, calculating the dynamic change trend of the spatial distribution characteristics, and obtaining a community structure change dataset includes: extracting spatial distribution features of vegetation communities from the second vegetation distribution image using an image segmentation algorithm to generate a spatial feature data set; If the pixel value distribution of the spatial feature data set meets the preset threshold, the spatial features are classified by a clustering algorithm to obtain a community structure classification result; Based on the community structure classification results, a spatiotemporal analysis algorithm is used to calculate the dynamic changes of the community structure and generate a dynamic change data set; Analyze the time trend of the dynamic change data set by a linear regression algorithm to obtain a change trend data set; Extracting significantly changed community structure features from the change trend data set to generate a structure change feature set; According to the structural change feature set, data fusion technology is used to integrate spatial characteristics and change trends to generate the community structure change dataset.

5. The method for monitoring the ecological environment of vegetation in a small watershed in arid areas according to claim 1, characterized in that: Based on the community structure change dataset, a long short-term memory network is used to predict the short-term evolution trend of vegetation community structure. The dynamic change prediction results of vegetation community structure are obtained, including: Obtain a dataset of vegetation community structure changes, determine the data characteristics through time series analysis, and obtain a standardized dataset; Using data processing technology to preprocess the standardized data set to obtain a preprocessed data set; The preprocessed data set is trained using a long short-term memory network to obtain a prediction model; Based on the prediction result, the prediction result is obtained.

6. The method for monitoring the ecological environment of vegetation in a small watershed in arid areas according to claim 1, characterized in that: Based on the dynamic change prediction results and combined with the community structure change dataset, the biodiversity level was quantified, and the biodiversity assessment results were obtained, including: The species diversity index calculation method was used, combined with the community dynamics trend, to calculate the species diversity level and obtain preliminary quantitative results; If the preliminary quantitative results are inconsistent with the preset ecosystem dynamic threshold, the species distribution feature weights are adjusted using the random forest algorithm to obtain an optimized quantitative result; Based on the optimized quantitative results, the impact of environmental factors is analyzed, key environmental driving factors are extracted, and the characteristics of biodiversity changes are determined; By using the characteristics of biodiversity changes and adopting time series analysis methods, the dynamic trends of the community are predicted to obtain the biodiversity assessment results.

7. The method for monitoring the ecological environment of vegetation in a small watershed in arid areas according to claim 1, characterized in that: Key indicators are extracted from the biodiversity assessment results. Based on these key indicators, real-time monitoring of vegetation ecological environment status data includes: Extract key indicators from biodiversity assessment results through data cleaning to obtain a standardized indicator set; Perform real-time analysis on standardized indicator sets to obtain real-time data streams; By comparing the real-time data stream with historical data, calculating the magnitude of the change, and determining whether it exceeds a preset threshold; If the change exceeds the preset threshold, the dynamic update mechanism is triggered to generate updated vegetation ecological environment status data; Based on the updated vegetation ecological environment status data, time series analysis is used to obtain the trend of environmental status changes; The changing trends were classified by clustering algorithms to obtain the vegetation ecological characteristics of different regions; Visualization technology is used to dynamically present the ecological characteristics of vegetation and obtain real-time monitoring data.

Images