An automated monitoring method for dynamic changes of plant diversity in mountain forest regions

Abstract

This invention relates to an automated monitoring method for the dynamic changes of plant diversity in mountainous forest areas, comprising: constructing a three-dimensional model of the target mountainous area and deploying a sensor network; collecting microclimate data and soil characteristic data to determine a spatial distribution matrix; acquiring historical plant distribution data of the target mountainous area, mapping it to the spatial distribution matrix to construct a comprehensive distribution map of environmental factors, and determining a correlation model between plant distribution and environmental factors; acquiring real-time satellite remote sensing image data of the target mountainous area, using the correlation model to analyze the characteristics of changes in plant cover and plant diversity distribution, and locating several key plant distribution areas; identifying plant species and counting their numbers in the key plant distribution areas, updating the comprehensive distribution map of environmental factors, and completing the automated monitoring of the dynamic changes of plant diversity in the target mountainous area. This invention integrates multi-source environmental data to achieve automated analysis of plant diversity distribution.

Description

Technical Field

[0001] This invention relates to the field of plant diversity dynamic change monitoring technology, and in particular to an automated monitoring method for plant diversity dynamic changes in mountainous forest areas. Background Technology

[0002] Plant diversity research in mountain forest regions is a crucial component of ecological conservation and resource management, directly impacting ecosystem stability and the sustainable use of biological resources. Due to their complex topography and diverse climatic conditions, mountain environments are key areas for biodiversity distribution and play an irreplaceable role in enhancing global ecological balance and ecosystem stability. Traditional methods for monitoring plant diversity in mountain forests largely rely on manual field surveys. These methods have significant limitations in coverage, data accuracy, and real-time performance, especially in the complex topography of mountain environments, making it difficult to comprehensively capture the dynamic relationship between plant distribution changes and environmental changes.

[0003] In recent years, the rapid development of technologies such as Geographic Information Systems (GIS), remote sensing, sensor networks, and machine learning has provided new means and methods for monitoring plant diversity. These technologies enable large-scale, high-frequency, and automated data collection and analysis, significantly improving monitoring efficiency and accuracy. However, they still face several challenges, such as the difficulty of data integration (data from different sources differs in format, resolution, and time scale, requiring complex processing and analysis methods); the accuracy and reliability of relevant models (constructing relevant models requires substantial data support, and the accuracy and reliability of the models directly affect the accuracy of monitoring results); and the real-time and automated nature of monitoring (real-time monitoring and automated updates of plant diversity require efficient data processing and analysis workflows, as well as reliable automated monitoring systems). Therefore, addressing the limitations and challenges of existing technologies, this invention proposes an automated monitoring method for plant diversity in mountainous forest areas, improving monitoring efficiency and accuracy, and providing strong technical support for the protection and management of mountainous forest ecosystems. Summary of the Invention

[0004] The purpose of this invention is to provide an automated monitoring method for plant diversity in mountainous forest areas, enabling precise digital representation of the mountainous environment and integrating multi-source environmental data to support automated analysis of plant diversity distribution.

[0005] To achieve the above objectives, the present invention provides the following solution:

[0006] An automated monitoring method for dynamic changes in plant diversity in mountainous forest areas includes:

[0007] Construct a 3D model of the target mountainous region and deploy a sensor network based on the 3D model;

[0008] The sensor network is used to collect microclimate data and soil property data to determine the spatial distribution matrix of microclimate and soil properties.

[0009] Historical plant distribution data of the target mountainous area is obtained, mapped with the spatial distribution matrix, and a comprehensive distribution map of environmental factors is constructed. Based on the comprehensive distribution map of environmental factors, the vertical distribution pattern of plant diversity is analyzed, and the correlation model between plant distribution and environmental factors is determined.

[0010] Real-time satellite remote sensing image data of the target mountainous area is acquired, and the correlation model is used to analyze the characteristics of plant cover change and plant diversity distribution change, and to locate several key plant distribution areas.

[0011] Plant species and quantities are identified and counted in key areas of plant distribution, and updated to the comprehensive distribution map of environmental factors, thus completing the automated monitoring of dynamic changes in plant diversity in the target mountainous area.

[0012] Optionally, a three-dimensional model of the target mountainous region is constructed, including:

[0013] Acquire high-resolution satellite imagery of the target mountainous region and extract mountain imagery data from the high-resolution satellite imagery;

[0014] The mountain image data is fused with a digital elevation model to determine the spatial structure information of the mountain terrain, and a three-dimensional modeling technique is used for digital reconstruction to construct a three-dimensional landform model.

[0015] The slope of the three-dimensional landform model is analyzed and optimized, and the optimized three-dimensional landform model is stored in multiple layers to obtain a three-dimensional dataset.

[0016] Optionally, by collecting microclimate data and soil property data through the sensor network, the spatial distribution matrix of microclimate and soil properties is determined, including:

[0017] A sensor network was deployed in the target mountainous area using an optimized 3D terrain model to collect microclimate and soil property data and construct an environmental dataset.

[0018] Based on the three-dimensional dataset, the environmental dataset is subjected to stratified sampling according to differences in altitude and slope aspect to obtain grouping results;

[0019] Based on the grouping results, the spatial distribution continuity of microclimate data and soil characteristic data is estimated using spatial interpolation method, a spatial distribution feature map is constructed, and the missing areas of the spatial distribution feature map are filled in using the neighborhood averaging method to obtain the final spatial distribution feature map.

[0020] Based on the final spatial distribution feature map, a spatial distribution matrix of microclimate and soil properties is constructed by combining altitude and slope aspect differences. The spatial distribution matrix is ​​then subjected to data standardization processing to obtain a dimension-reduced spatial distribution matrix.

[0021] Optionally, historical plant distribution data of the target mountainous area is obtained and mapped with the spatial distribution matrix to construct a comprehensive distribution map of environmental factors, including:

[0022] Collect historical plant distribution data for the target mountainous area;

[0023] By using fusion technology, the spatial distribution matrix is ​​correlated and mapped with historical plant distribution data, and the mapping results are corrected to obtain an environmental distribution dataset.

[0024] Outlier detection and removal are performed on the environmental distribution dataset, and the data is smoothed using spatial neighborhood analysis to obtain an environmental distribution view;

[0025] Kriging interpolation was used to fill in missing areas and perform spatial calibration on the environmental distribution view to obtain a comprehensive distribution map of environmental factors.

[0026] Optionally, based on the comprehensive distribution map of environmental factors, the vertical distribution pattern of plant diversity is analyzed to determine the correlation model between plant distribution and environmental factors, including:

[0027] By combining historical plant distribution data and a comprehensive distribution map of environmental factors, the random forest algorithm is used to predict the vertical distribution pattern of plant diversity and analyze the variable weights of each environmental factor.

[0028] Based on the variable weights, a preliminary correlation model between plant distribution and environmental factors is constructed.

[0029] Obtain the parameters and correlation indicators of the preliminary association model, optimize the preliminary association model based on the parameters and correlation indicators, obtain the final association model, and determine the association relationship and influencing factors.

[0030] Optionally, real-time satellite remote sensing image data of the target mountainous area is acquired, and the correlation model is used to analyze the characteristics of vegetation cover change and vegetation diversity distribution change, including:

[0031] Based on real-time satellite remote sensing image data, time series analysis methods are used to quantify the changing trend of plant cover and determine the dynamic change characteristics of cover.

[0032] Based on the dynamic change characteristics of the coverage, the correlation model is used to conduct correlation analysis on the influencing factors of plant diversity distribution to identify key environmental driving factors.

[0033] Based on the analysis of the key environmental driving factors, the interaction between plant diversity and environmental factors is analyzed, a dynamically changing distribution feature mapping is generated, and a comparative analysis is conducted with historical plant distribution data to determine the dynamic change characteristics of plant diversity distribution.

[0034] Optionally, after determining the dynamic change characteristics of the plant diversity distribution, the method further includes: if the dynamic change characteristics of the plant diversity distribution exceed a preset change threshold, then acquiring the latest microclimate data and soil characteristic data, recalculating the spatial distribution matrix, and updating the comprehensive distribution map of environmental factors.

[0035] Optionally, several key areas of plant distribution can be located, including:

[0036] By combining the plant cover variation characteristics and spatial distribution matrix, distribution characteristic information is obtained;

[0037] Based on the aforementioned distribution characteristics, spatial statistical analysis methods are used to quantify the relationship between plant cover and plant diversity distribution, and to determine the range of key plant distribution areas.

[0038] Spatial distribution data of key plant distribution areas are obtained and layered to obtain boundary delineation information of key plant distribution areas.

[0039] Optionally, the key areas of plant distribution are used for plant species identification and quantity statistics, including:

[0040] The image set is generated by using drones to collect high-frequency images of each key area of ​​plant distribution, followed by noise reduction and enhancement processing.

[0041] The image set is divided into several regional units, and the plant species in each regional unit are identified by a pre-trained convolutional neural network model to obtain the distribution results of plant species in each regional unit.

[0042] Based on the plant species distribution results, the distribution results of plant species in the key areas of plant distribution are statistically analyzed.

[0043] Optionally, updating the comprehensive distribution map of environmental factors includes:

[0044] The division of key plant distribution areas is embedded in the comprehensive distribution map of environmental factors;

[0045] The plant species identification and quantity statistics results corresponding to the key plant distribution areas are fused with microclimate data and soil characteristic data, and then superimposed as a new layer onto the comprehensive distribution map of environmental factors.

[0046] The plant diversity information in the comprehensive distribution map of environmental factors is updated based on real-time changes in the types and quantities of plants, and the updated comprehensive distribution map of environmental factors is then visualized.

[0047] The beneficial effects of this invention are as follows:

[0048] (1) Improved monitoring precision and accuracy: This invention, through the construction of a three-dimensional model and the deployment of a sensor network, can accurately collect microclimate and soil characteristic data, thereby determining their spatial distribution matrix. The comprehensive distribution map of environmental factors constructed by combining historical plant distribution data with the correlation model between plant distribution and environmental factors makes the analysis of dynamic changes in plant diversity based on satellite remote sensing image data more accurate. This method of multi-dimensional data fusion and precise analysis effectively improves the precision and accuracy of monitoring dynamic changes in plant diversity, avoiding monitoring errors caused by incomplete or inaccurate data.

[0049] (2) Enhanced monitoring efficiency and real-time performance: The automated monitoring process of this invention eliminates the need for significant manpower in the entire process from data acquisition to analysis and processing, greatly improving monitoring efficiency. Simultaneously, real-time acquisition of satellite remote sensing imagery data combined with correlation models enables timely detection of changes in plant cover and plant diversity distribution, rapid location of key plant distribution areas, and timely updates to plant species and quantity information. This highly efficient real-time monitoring capability helps to promptly grasp the dynamic changes in plant diversity in mountainous forest areas, providing strong support for the formulation and implementation of ecological protection measures.

[0050] (3) Providing a scientific basis for biodiversity conservation and management: This invention can clearly analyze the vertical distribution patterns of plant diversity and determine the correlation model between plant distribution and environmental factors, providing a scientific basis for ecological protection and management in mountainous forest areas. Through precise location and dynamic monitoring of key areas of plant distribution, more targeted protection measures can be taken, such as rationally planning forest resource development and formulating ecological restoration plans, which helps maintain the stability and biodiversity of mountainous forest ecosystems and promotes ecological balance and sustainable development.

[0051] (4) Reduced monitoring costs and manpower: Compared with traditional manual field surveys and monitoring methods, this invention reduces reliance on manpower and lowers the errors and risks caused by manual operation. At the same time, by acquiring data through sensor networks and satellite remote sensing technologies, there is no need to frequently enter mountainous and forested areas for field surveys, saving a lot of manpower, material resources and time costs, improving the economy and feasibility of monitoring work, and making long-term, large-scale monitoring of dynamic changes in plant diversity possible. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 This is a flowchart of an automated monitoring method for dynamic changes in plant diversity in mountainous forest areas, according to an embodiment of the present invention.

[0054] Figure 2 This is a flowchart illustrating the method for constructing a three-dimensional model according to an embodiment of the present invention.

[0055] Figure 3 This is a flowchart of a method for constructing a spatial distribution matrix according to an embodiment of the present invention;

[0056] Figure 4 This is a flowchart illustrating the method for constructing a comprehensive distribution map of environmental factors according to an embodiment of the present invention.

[0057] Figure 5 This is a flowchart illustrating the method for constructing an association model according to an embodiment of the present invention;

[0058] Figure 6 This is a flowchart of a method for analyzing the characteristics of changes in plant cover and the characteristics of changes in plant diversity distribution according to an embodiment of the present invention;

[0059] Figure 7 This is a flowchart of a method for locating key areas of plant distribution according to an embodiment of the present invention;

[0060] Figure 8 This is a flowchart illustrating the method for updating the comprehensive distribution map of environmental factors according to an embodiment of the present invention. Detailed Implementation

[0061] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0062] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0063] This embodiment provides an automated monitoring method for the dynamic changes of plant diversity in mountainous forest areas, such as... Figure 1 As shown, it includes:

[0064] Construct a 3D model of the target mountainous region and deploy a sensor network based on the 3D model;

[0065] The sensor network is used to collect microclimate data and soil property data to determine the spatial distribution matrix of microclimate and soil properties.

[0066] Historical plant distribution data of the target mountainous area is obtained, mapped with the spatial distribution matrix, and a comprehensive distribution map of environmental factors is constructed. Based on the comprehensive distribution map of environmental factors, the vertical distribution pattern of plant diversity is analyzed, and the correlation model between plant distribution and environmental factors is determined.

[0067] Real-time satellite remote sensing image data of the target mountainous area is acquired, and the correlation model is used to analyze the characteristics of plant cover change and plant diversity distribution change, and to locate several key plant distribution areas.

[0068] Plant species and quantities are identified and counted in key areas of plant distribution, and updated to the comprehensive distribution map of environmental factors, thus completing the automated monitoring of dynamic changes in plant diversity in the target mountainous area.

[0069] Specifically, this embodiment achieves efficient, accurate, and automated monitoring of plant diversity by constructing a 3D model, deploying a sensor network, fusing multi-source data, building correlation models, and performing real-time remote sensing image analysis. Through multi-dimensional data fusion and automated processing, monitoring efficiency and accuracy are significantly improved, providing strong technical support for the protection and management of mountain forest ecosystems. This embodiment can achieve real-time monitoring and dynamic updating of key areas of plant distribution within mountainous regions, promptly identifying trends in plant diversity changes and providing a scientific basis for the formulation and implementation of ecological protection measures.

[0070] Furthermore, such as Figure 2As shown, constructing a three-dimensional model of the target mountainous region includes:

[0071] Acquire high-resolution satellite imagery of the target mountainous region and extract mountain imagery data from the high-resolution satellite imagery;

[0072] The mountain image data is fused with a digital elevation model to determine the spatial structure information of the mountain terrain, and a three-dimensional modeling technique is used for digital reconstruction to construct a three-dimensional landform model.

[0073] The slope of the three-dimensional landform model is analyzed and optimized, and the optimized three-dimensional landform model is stored in multiple layers to obtain a three-dimensional dataset.

[0074] Specifically, this embodiment focuses on data collection and monitoring in forested mountainous areas. By combining high-resolution satellite imagery with a digital elevation model (DEM), and considering the complex terrain of mountainous areas, detailed geomorphological feature data of the region is first acquired. Satellite imagery data with a resolution of 0.5 meters is combined with a DEM with a spatial resolution of 1 meter. Image processing software is used to perform geometric and radiometric corrections on the satellite imagery, ensuring that the spatial registration accuracy between the imagery and elevation data is within 0.2 meters. Subsequently, image fusion algorithms such as Principal Component Analysis (PCA) are used to overlay multispectral imagery with elevation data, extracting surface texture and height information to form a preliminary geomorphological feature dataset. Next, to accurately digitally represent topographic relief, slope, and aspect, 3D modeling technology was employed. Using the slope analysis tool in ArcGIS software, the slope value of each grid cell was calculated based on DEM data. For example, in a certain mountainous study area, the slope ranged from 0 to 45 degrees, with an average slope of 18.5 degrees. Simultaneously, the orientation of each grid cell was determined using an aspect analysis algorithm, obtaining the aspect distribution, with north-facing slopes accounting for approximately 30% and south-facing slopes accounting for approximately 25%. Combining the topographic relief characteristics, the TIN (Triangular Irregular Network) modeling method was used to convert elevation points into a 3D surface model, i.e., an optimized 3D geomorphological model, accurately depicting the mountain's undulating morphology with an error controlled within 0.3 meters. Finally, the above analysis results were integrated to generate a 3D dataset of mountain geomorphological features.

[0075] Furthermore, such as Figure 3 As shown, microclimate data and soil property data are collected through the sensor network to determine the spatial distribution matrix of microclimate and soil properties, including:

[0076] A sensor network was deployed in the target mountainous area using an optimized 3D terrain model to collect microclimate and soil property data and construct an environmental dataset.

[0077] Based on the three-dimensional dataset, the environmental dataset is subjected to stratified sampling according to differences in altitude and slope aspect to obtain grouping results;

[0078] Based on the grouping results, the spatial distribution continuity of microclimate data and soil characteristic data is estimated using spatial interpolation method, a spatial distribution feature map is constructed, and the missing areas of the spatial distribution feature map are filled in using the neighborhood averaging method to obtain the final spatial distribution feature map.

[0079] Based on the final spatial distribution feature map, a spatial distribution matrix of microclimate and soil properties is constructed by combining altitude and slope aspect differences. The spatial distribution matrix is ​​then subjected to data standardization processing to obtain a dimension-reduced spatial distribution matrix.

[0080] Specifically, this embodiment first extracts terrain parameters such as elevation, slope, and aspect from the optimized 3D terrain model. Assuming the dataset covers an area of ​​10 square kilometers with a resolution of 5 meters, the terrain analysis tools in ArcGIS software are used to calculate slope values ​​(range 0-45 degrees) and aspect values ​​(0-360 degrees), generating a 3D terrain grid matrix. Next, a ground sensor network is deployed, with 100 sensor nodes arranged in the target mountainous area, spaced approximately 300 meters apart, covering an elevation range of 500-1500 meters. Each node is equipped with a temperature and humidity sensor (accuracy ±0.5℃, ±2%RH) and a soil moisture sensor (accuracy ±3%). Data is transmitted in real time to a cloud server via LoRa wireless communication technology, forming a microclimate and soil characteristic dataset. Subsequently, stratified sampling was performed to address the differences in altitude and slope aspect. Altitude was divided into 10 levels per 100 meters, and slope aspect was divided into 8 intervals at 45 degrees. A Python script was used to perform stratified clustering analysis on the sensor data, calculating the mean temperature (e.g., 18.2℃ at 800 meters altitude, north-facing slope) and mean soil moisture (e.g., 35.6%) within each level and slope aspect interval. Finally, the spatial distribution matrix of microclimate and soil properties was determined. Kriging interpolation was used to spatially interpolate the stratified sampling data, generating a 10-meter resolution grid map of temperature and soil moisture distribution. Correlation analysis (Pearson coefficient calculation, e.g., a correlation coefficient of -0.85 between temperature and altitude) revealed the influence of altitude and slope aspect on microclimate. Combined with the topographic grid matrix, a comprehensive spatial distribution matrix was formed.

[0081] Furthermore, such as Figure 4 As shown, historical plant distribution data for the target mountainous area is obtained and mapped to the spatial distribution matrix to construct a comprehensive distribution map of environmental factors, including:

[0082] Collect historical plant distribution data for the target mountainous area;

[0083] By using fusion technology, the spatial distribution matrix is ​​correlated and mapped with historical plant distribution data, and the mapping results are corrected to obtain an environmental distribution dataset.

[0084] Outlier detection and removal are performed on the environmental distribution dataset, and the data is smoothed using spatial neighborhood analysis to obtain an environmental distribution view;

[0085] Kriging interpolation was used to fill in missing areas and perform spatial calibration on the environmental distribution view to obtain a comprehensive distribution map of environmental factors.

[0086] Specifically, this embodiment collects historical plant distribution data from a mountainous area with an altitude range of 500 meters to 3000 meters, including 1000 sampling points. Each sampling point records environmental variables such as the number of plant species, soil pH (range 4.5 to 7.5), average annual rainfall (500 mm to 2000 mm), average annual temperature (5 to 20 degrees Celsius), and slope (0 to 45 degrees Celsius) as historical plant distribution data. The spatial distribution matrix is ​​correlated and mapped with the historical plant distribution data from the preliminary survey. Assuming that the survey data covers 60% of the mountain forest area and records the distribution location and density of 100 plant species, such as a density of 200 conifers per hectare in a certain area, the correlation coefficient between environmental factors and plant distribution is calculated through spatial overlay analysis. It is found that the correlation coefficient between temperature and conifer distribution is 0.75, indicating that temperature has a significant impact on its distribution. Subsequently, for areas with missing data, Kriging spatial interpolation was used to fill in the gaps. Assuming the missing areas accounted for 30% of the total area, and based on known point data, a search radius of 5 kilometers was set. After interpolation, the predicted temperature for the missing areas was 15.0 degrees Celsius, and the predicted soil pH was 6.4, with the error controlled within 5% to ensure data continuity. Finally, by combining the above data, a comprehensive distribution map of environmental factors in the mountainous forest area was generated using geographic information system software, overlaying information such as temperature and soil pH to form a multi-layered visualization result.

[0087] Furthermore, such as Figure 5 As shown, based on the comprehensive distribution map of environmental factors, the vertical distribution pattern of plant diversity is analyzed, and the correlation model between plant distribution and environmental factors is determined, including:

[0088] By combining historical plant distribution data and a comprehensive distribution map of environmental factors, the random forest algorithm is used to predict the vertical distribution pattern of plant diversity and analyze the variable weights of each environmental factor.

[0089] Based on the variable weights, a preliminary correlation model between plant distribution and environmental factors is constructed.

[0090] Obtain the parameters and correlation indicators of the preliminary association model, optimize the preliminary association model based on the parameters and correlation indicators, obtain the final association model, and determine the association relationship and influencing factors.

[0091] Specifically, this embodiment integrates historical plant distribution data into a comprehensive environmental factor distribution map using a Geographic Information System (GIS), then imports it into a Python environment. A random forest model is constructed using the RandomForestRegressor module from the sklearn library, with 100 trees and a maximum depth of 10. Model performance is evaluated through cross-validation (5-fold), yielding a mean squared error (MSE) of 0.25, indicating high predictive accuracy. Subsequently, feature importance scores are extracted, revealing that annual average rainfall is 0.35, soil pH is 0.25, annual average temperature is 0.20, slope is 0.15, and altitude is 0.05, indicating that rainfall has the greatest impact on the vertical distribution of plant diversity. Based on this, when constructing the correlation model, the regression relationship between rainfall and the number of plant species is prioritized. Linear regression analysis shows that for every 100 mm increase in rainfall, the number of plant species increases by an average of 2.3, with a correlation coefficient R0. 2 The correlation coefficient was 0.68, confirming a strong correlation. To further optimize the model, an interaction term analysis was performed on rainfall with other variables (such as soil pH). The analysis revealed a significant interaction effect between rainfall and pH (p-value less than 0.05), thus constructing the final multivariate correlation model and improving the prediction accuracy to an MSE of 0.20.

[0092] Furthermore, such as Figure 6 As shown, real-time satellite remote sensing image data of the target mountainous area is acquired, and the correlation model is used to analyze the characteristics of changes in vegetation cover and vegetation diversity distribution, including:

[0093] Based on real-time satellite remote sensing image data, time series analysis methods are used to quantify the changing trend of plant cover and determine the dynamic change characteristics of cover.

[0094] Based on the dynamic change characteristics of the coverage, the correlation model is used to conduct correlation analysis on the influencing factors of plant diversity distribution to identify key environmental driving factors.

[0095] Based on the analysis of the key environmental driving factors, the interaction between plant diversity and environmental factors is analyzed, a dynamically changing distribution feature mapping is generated, and a comparative analysis is conducted with historical plant distribution data to determine the dynamic change characteristics of plant diversity distribution.

[0096] Specifically, this embodiment uses time series analysis and the statsmodels library in Python to perform trend decomposition, breaking down NDVI data into three parts: trend, seasonality, and residuals. Then, by combining environmental factors such as rainfall (annual average of 500 mm) and temperature (annual average of 15 degrees Celsius) data, the analysis using a correlation model shows that rainfall contributes 60% to NDVI and temperature contributes 30%, thus identifying the key environmental driving factors.

[0097] This study analyzes the interaction between plant diversity and environmental factors based on key environmental drivers, and compares this with historical plant distribution data to determine the dynamic changes in plant diversity distribution. In this embodiment, changes in the Shannon diversity index, Simpson diversity index, and Fisher's α index are used to determine the dynamic changes in plant diversity distribution. Specifically, the Shannon diversity index is:

[0098]

[0099] Simpson diversity index:

[0100]

[0101] Fisher's alpha index:

[0102]

[0103] Among them, P i Let be the relative abundance of the i-th species. n i Let N be the number of individuals of the i-th species, N be the total number of individuals, and S be the number of species.

[0104] Furthermore, after determining the dynamic change characteristics of the plant diversity distribution, the method further includes: if the dynamic change characteristics of the plant diversity distribution exceed a preset change threshold, then acquiring the latest microclimate data and soil characteristic data, recalculating the spatial distribution matrix, and updating the comprehensive distribution map of environmental factors.

[0105] Specifically, in this embodiment, when a significant or rapid change in the dynamic characteristics of plant diversity distribution is detected within a short period of time, such as when the species richness index drops from 3.5 to 2.8, which is lower than the preset threshold of 3.0, historical data and real-time data are called, and the plant distribution change area is spatially divided using the K-means clustering algorithm. If the area with significant change is found to account for 15.2%, the update of the comprehensive distribution map of environmental factors is automatically triggered, and the latest microclimate data is obtained by calling the meteorological station API interface.

[0106] Furthermore, such as Figure 7 As shown, several key areas of plant distribution are located, including:

[0107] By combining the plant cover variation characteristics and spatial distribution matrix, distribution characteristic information is obtained;

[0108] Based on the aforementioned distribution characteristics, spatial statistical analysis methods are used to quantify the relationship between plant cover and plant diversity distribution, and to determine the range of key plant distribution areas.

[0109] Spatial distribution data of key plant distribution areas are obtained and layered to obtain boundary delineation information of key plant distribution areas.

[0110] Specifically, this embodiment first integrates plant cover change characteristic data with environmental factor data in the spatial distribution matrix to construct a comprehensive feature matrix. For example, it combines the plant cover change rate with the values ​​of environmental factors such as temperature and humidity to form a new feature vector. Statistical analysis methods (such as Pearson correlation coefficient and Spearman rank correlation coefficient) are used to calculate the correlation between plant cover change characteristics and various environmental factors to identify environmental factors that significantly affect plant cover change. Based on the correlation analysis results, the distribution characteristic information of environmental factors significantly related to plant cover change is extracted, such as temperature gradient, humidity range, and light intensity changes. Then, appropriate spatial statistical analysis methods are selected, such as geographically weighted regression (GWR), spatial autocorrelation analysis (such as Moran's I index), and hotspot analysis, to quantify the relationship between plant cover and plant diversity distribution. Based on the quantitative analysis results, the scope of key plant distribution areas is determined. In this embodiment, areas with significant changes in plant diversity index and close relationships between plant cover change and environmental factors are designated as key areas. For example, hotspot analysis identifies areas with high plant diversity clusters; these areas are the key plant distribution areas. Based on distribution characteristics, the data is stratified. By analyzing the stratified data, boundary delineation information of key plant distribution areas is extracted. For example, cluster analysis methods (such as K-means clustering) or threshold-based segmentation methods can be used to divide the key areas into different sub-regions and determine the boundaries of each sub-region. For instance, based on a threshold of the plant diversity index, the key areas can be divided into high-diversity, medium-diversity, and low-diversity areas, and the boundary information of these areas can be extracted.

[0111] Furthermore, such as Figure 8 As shown, plant species distribution and quantity statistics were performed in key areas of plant distribution, and the results were updated in the comprehensive environmental factor distribution map. Firstly, plant species distribution and quantity statistics were performed in key areas of plant distribution, including:

[0112] The image set is generated by using drones to collect high-frequency images of each key area of ​​plant distribution, followed by noise reduction and enhancement processing.

[0113] The image set is divided into several regional units, and the plant species in each regional unit are identified by a pre-trained convolutional neural network model to obtain the distribution results of plant species in each regional unit.

[0114] Based on the plant species distribution results, the distribution results of plant species in the key areas of plant distribution are statistically analyzed.

[0115] Then update the comprehensive distribution map of environmental factors, including:

[0116] The division of key plant distribution areas is embedded in the comprehensive distribution map of environmental factors;

[0117] The plant species identification and quantity statistics results corresponding to the key plant distribution areas are fused with microclimate data and soil characteristic data, and then superimposed as a new layer onto the comprehensive distribution map of environmental factors.

[0118] The plant diversity information in the comprehensive distribution map of environmental factors is updated based on real-time changes in the types and quantities of plants, and the updated comprehensive distribution map of environmental factors is then visualized.

[0119] Specifically, this embodiment targets key areas of plant distribution. A drone equipped with a high-resolution camera is used to collect high-frequency image data at a grid density of 100 meters per square kilometer. The flight altitude is set at 150 meters to ensure an image resolution of 0.05 meters per pixel and a coverage rate of over 95%. Using the acquired image data, the U-Net algorithm from deep learning is employed for image segmentation, separating vegetated and non-vegetated areas within the key regions. The training dataset consists of 5000 labeled mountain forest images, with 10,000 training iterations. The segmentation accuracy reaches over 90%, ensuring clear boundaries of vegetated areas.

[0120] For the segmented vegetation areas, a convolutional neural network (CNN) model was used to classify plant species. The pre-trained model was ResNet-50, which was fine-tuned by combining images of 100 common mountain plants collected on-site. The classification accuracy target was 85%. By classifying and statistically analyzing the plant pixels in each image, a distribution map of plant species in a specific area was generated, including the coverage area percentage of each plant. For example, in a certain area, coniferous forests accounted for 40%, broad-leaved forests accounted for 35%, and shrubs accounted for 25%.

[0121] This embodiment first embeds the range and boundary information of each key plant distribution area into the comprehensive environmental factor distribution map. Then, it integrates the distribution results of each key plant distribution area. For example, if 50 plant species are recorded in a key plant distribution area, with conifers accounting for 40% and broad-leaved trees accounting for 60%, a Geographic Information System (GIS) is used to associate these data with latitude and longitude coordinates and fuse them with microclimate data and soil characteristic data from the comprehensive environmental factor distribution map. This fusion is then overlaid as a new layer onto the comprehensive environmental factor distribution map. A set time period (e.g., one quarter) is established to obtain real-time changes in plant species and quantities, as well as microclimate and soil characteristic data, and update them on the comprehensive environmental factor distribution map.

[0122] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. An automated monitoring method for the dynamic changes of plant diversity in mountainous forest areas, characterized in that, include: Construct a 3D model of the target mountainous region and deploy a sensor network based on the 3D model; The sensor network is used to collect microclimate data and soil property data to determine the spatial distribution matrix of microclimate and soil properties. By collecting microclimate and soil property data through the sensor network, the spatial distribution matrix of microclimate and soil properties is determined, including: A sensor network was deployed in the target mountainous area using an optimized 3D terrain model to collect microclimate and soil property data and construct an environmental dataset. Based on the 3D dataset, the environmental dataset is stratified and sampled according to differences in altitude and slope aspect to obtain grouping results; Based on the grouping results, the spatial distribution continuity of microclimate data and soil characteristic data is estimated using spatial interpolation method, a spatial distribution feature map is constructed, and the missing areas of the spatial distribution feature map are filled in using the neighborhood averaging method to obtain the final spatial distribution feature map. Based on the final spatial distribution feature map, a spatial distribution matrix of microclimate and soil properties is constructed by combining altitude and slope aspect differences. The spatial distribution matrix is ​​then subjected to data standardization processing to obtain a dimension-reduced spatial distribution matrix. Historical plant distribution data of the target mountainous area is obtained, mapped with the spatial distribution matrix, and a comprehensive distribution map of environmental factors is constructed. Based on the comprehensive distribution map of environmental factors, the vertical distribution pattern of plant diversity is analyzed, and the correlation model between plant distribution and environmental factors is determined. Based on the comprehensive distribution map of environmental factors, the vertical distribution pattern of plant diversity is analyzed, and the correlation model between plant distribution and environmental factors is determined, including: By combining historical plant distribution data and comprehensive distribution maps of environmental factors, the random forest algorithm is used to predict the vertical distribution pattern of plant diversity and analyze the variable weights of each environmental factor. Based on the variable weights, a preliminary correlation model between plant distribution and environmental factors is constructed. Obtain the parameters and correlation indicators of the preliminary association model, optimize the preliminary association model based on the parameters and correlation indicators, obtain the final association model, and determine the association relationship and influencing factors; Real-time satellite remote sensing image data of the target mountainous area is acquired, and the correlation model is used to analyze the characteristics of plant cover change and plant diversity distribution change, and to locate several key plant distribution areas. Plant species and quantities are identified and counted in key areas of plant distribution, and updated to the comprehensive distribution map of environmental factors, thus completing the automated monitoring of dynamic changes in plant diversity in the target mountainous area.

2. The automated monitoring method for dynamic changes in plant diversity in mountainous forest areas according to claim 1, characterized in that, Constructing a three-dimensional model of the target mountainous region includes: Acquire high-resolution satellite imagery of the target mountainous region and extract mountain imagery data from the high-resolution satellite imagery; The mountain image data is fused with a digital elevation model to determine the spatial structure information of the mountain terrain, and a three-dimensional modeling technique is used for digital reconstruction to construct a three-dimensional landform model. The slope of the three-dimensional landform model is analyzed and optimized, and the optimized three-dimensional landform model is stored in multiple layers to obtain a three-dimensional dataset.

3. The automated monitoring method for dynamic changes in plant diversity in mountainous forest areas according to claim 1, characterized in that, Historical plant distribution data of the target mountainous area is obtained and mapped to the spatial distribution matrix to construct a comprehensive distribution map of environmental factors, including: Collect historical plant distribution data for the target mountainous area; By using fusion technology, the spatial distribution matrix is ​​correlated and mapped with historical plant distribution data, and the mapping results are corrected to obtain an environmental distribution dataset. Outlier detection and removal are performed on the environmental distribution dataset, and the data is smoothed using spatial neighborhood analysis to obtain an environmental distribution view; Kriging interpolation was used to fill in missing areas and perform spatial calibration on the environmental distribution view to obtain a comprehensive distribution map of environmental factors.

4. The automated monitoring method for dynamic changes in plant diversity in mountainous forest areas according to claim 1, characterized in that, Real-time acquisition of satellite remote sensing image data of the target mountainous area, and analysis of vegetation cover change characteristics and vegetation diversity distribution change characteristics using the correlation model, including: Based on real-time satellite remote sensing image data, time series analysis methods are used to quantify the changing trend of plant cover and determine the dynamic change characteristics of cover. Based on the dynamic change characteristics of the coverage, the correlation model is used to conduct correlation analysis on the influencing factors of plant diversity distribution to identify key environmental driving factors. Based on the analysis of the key environmental driving factors, the interaction between plant diversity and environmental factors is analyzed, a dynamically changing distribution feature mapping is generated, and a comparative analysis is conducted with historical plant distribution data to determine the dynamic change characteristics of plant diversity distribution.

5. The automated monitoring method for dynamic changes in plant diversity in mountainous forest areas according to claim 4, characterized in that, After determining the dynamic change characteristics of the plant diversity distribution, the method further includes: if the dynamic change characteristics of the plant diversity distribution exceed a preset change threshold, then acquiring the latest microclimate data and soil characteristic data, recalculating the spatial distribution matrix, and updating the comprehensive distribution map of environmental factors.

6. The automated monitoring method for dynamic changes in plant diversity in mountainous forest areas according to claim 1, characterized in that, Locate key areas of plant distribution, including: By combining the plant cover variation characteristics and spatial distribution matrix, distribution characteristic information is obtained; Based on the aforementioned distribution characteristics, spatial statistical analysis methods are used to quantify the relationship between plant cover and plant diversity distribution, and to determine the range of key plant distribution areas. Spatial distribution data of key plant distribution areas are obtained and layered to obtain boundary delineation information of key plant distribution areas.

7. The automated monitoring method for dynamic changes in plant diversity in mountainous forest areas according to claim 1, characterized in that, The key areas of plant distribution were used for plant species identification and quantity statistics, including: The image set is generated by using drones to collect high-frequency images of each key area of ​​plant distribution, followed by noise reduction and enhancement processing. The image set is divided into several regional units, and the plant species in each regional unit are identified by a pre-trained convolutional neural network model to obtain the distribution results of plant species in each regional unit. Based on the plant species distribution results, the distribution results of plant species in the key areas of plant distribution are statistically analyzed.

8. The automated monitoring method for dynamic changes in plant diversity in mountainous forest areas according to claim 1, characterized in that, Update the comprehensive distribution map of environmental factors, including: The division of key plant distribution areas is embedded in the comprehensive distribution map of environmental factors; The plant species identification and quantity statistics results corresponding to the key plant distribution areas are integrated with microclimate data and soil characteristic data, and then used as a new layer to be overlaid on the comprehensive distribution map of environmental factors. The plant diversity information in the comprehensive distribution map of environmental factors is updated based on real-time changes in the types and quantities of plants, and the updated comprehensive distribution map of environmental factors is then visualized.

Images