Ecological system health dynamic early warning platform and method based on sky-ground integration

Abstract

The invention discloses an ecological system health dynamic early warning platform and method based on sky-ground integration, and the platform comprises a data preprocessing layer which is used for obtaining sky satellite data, air unmanned plane data and ground sensor data of a target natural region in a preset historical time period; the model analysis layer is used for extracting a plurality of ecological parameters of each preset ecological attention object from the acquired data; dynamically determining the final weight of each preset ecological attention object based on the current environment state at the current moment; performing weighted summation on the plurality of ecological parameters and the final weight to obtain a current ecological health comprehensive value; the health early warning layer is used for determining a change trend curve graph of the ecological health comprehensive value in a future preset time period according to the current ecological health comprehensive value and the current environment state; and based on the change trend curve graph, determining whether to carry out early warning on the target natural area. By adopting the method and the device, the timeliness requirement of ecological protection can be met, and meanwhile, the accuracy of an early warning result is improved.

Description

Technical Field

[0001] This application relates to the field of ecosystem protection and management technology, and in particular to an integrated sky-ground ecosystem health dynamic early warning platform and method. Background Technology

[0002] In the field of ecosystem protection and management, especially in the ecological protection of nature reserves, accurately assessing the health status of ecosystems and predicting their future trends is crucial for achieving scientific management and sustainable development. For example, in forest ecosystems, managers need to understand the dynamic changes in indicators such as vegetation cover, species diversity, and water conservation capacity in order to take preventative measures to address potential ecological degradation, pest and disease outbreaks, or human disturbance.

[0003] Currently, researchers regularly conduct field surveys, collecting data on soil, vegetation, and other factors to analyze and issue assessment reports. This method is time-consuming and labor-intensive, typically measured in years, and suffers from poor data real-time performance, making it difficult to meet the timely needs of ecological protection. Secondly, existing systems are based on fixed-weight assessment models, which ignore dynamic environmental changes and thus reduce the accuracy of early warning results. Summary of the Invention

[0004] This application provides an integrated sky-ground ecosystem health dynamic early warning platform. To provide a basic understanding of some aspects of the disclosed embodiments, a brief summary is given below. This summary is not intended as a general commentary, nor is it intended to identify key / important components or describe the scope of protection of these embodiments. Its sole purpose is to present some concepts in a simple form as a prelude to the detailed description that follows.

[0005] In a first aspect, embodiments of this application provide an integrated sky-ground ecosystem health dynamic early warning platform, the platform comprising: Data preprocessing layer, model analysis layer, and health early warning layer; among them, The data preprocessing layer is used to periodically obtain satellite data, UAV data, and ground sensor data of the target natural area from the time series database within a preset historical time period. The model analysis layer is used to extract multiple ecological parameters for each preset ecological concern object from satellite data, aerial drone data, and ground sensor data; dynamically determine the final weight of each preset ecological concern object based on the current environmental state at the current moment; and perform a weighted summation of multiple ecological parameters and final weights to obtain the current comprehensive ecological health value of the target natural area. The health early warning layer is used to determine the trend curve of the ecological health comprehensive value within a preset period of time based on the current ecological health comprehensive value and the current environmental status; based on the trend curve, it is determined whether to issue an early warning for the target natural area.

[0006] Optionally, based on the current comprehensive ecological health value and the current environmental status, determine a trend curve of the comprehensive ecological health value over a future preset period, including: To obtain the historical comprehensive value of ecological health and the historical environmental status of a target natural region at the same historical moment; The historical ecological health comprehensive value and historical environmental status at each historical moment are combined into a historical binary array; Combine the current comprehensive ecological health value and the current environmental status into a current binary array; Arrange each historical binary array and the current binary array in chronological order to obtain a time-series binary array sequence. Based on the temporal binary array sequence, generate multiple target ecological health comprehensive values ​​for a future preset time period; By projecting multiple target comprehensive ecological health values ​​onto a coordinate graph in chronological order, a trend curve of the comprehensive ecological health values ​​over a future preset time period is obtained.

[0007] Optionally, based on the time-series binary array sequence, generate multiple target ecological health comprehensive values ​​for a future preset time period, including: The numerical code of each historical moment in the time-series binary array sequence is used as the independent variable, and the historical ecological health comprehensive value of each historical moment in the time-series binary array sequence is used as the dependent variable. Based on each independent variable and the dependent variable of each independent variable, the rate of change of the comprehensive ecological health value over time and the predicted comprehensive ecological health value when time t=0 are obtained by combining the least squares method. The rate of change is used as the slope of the linear regression equation, and the predicted comprehensive ecological health value at time t=0 is used as the intercept of the linear regression equation. A linear regression equation for the trend curve is constructed using the slope and intercept. By substituting the numerical code of each target time within a future preset time period into the independent variables of the linear regression equation, the comprehensive ecological health value of multiple targets within the future preset time period is calculated.

[0008] Optionally, based on the time-series binary array sequence, generate multiple target ecological health comprehensive values ​​for a future preset time period, including: The time-series binary array sequence is input into the pre-trained ecological health comprehensive value prediction model. The pre-trained ecological health comprehensive value prediction model is obtained by machine learning based on the Transformer time series model. The Transformer time series model can capture time-series dependencies and achieve multivariate time series prediction. Output the comprehensive ecological health values ​​of multiple targets within a preset future time period; among them... Generate a pre-trained comprehensive ecological health value prediction model according to the following steps: Obtain the first comprehensive ecological health value and the first environmental state of different natural regions at the same historical moment; The first comprehensive ecological health value and the first environmental state at each historical moment are combined into a binary array for each sample. A comprehensive ecological health value prediction model was established using the Transformer time series model. Input each sample binary array into the ecological health comprehensive value prediction model and output the model loss value; When the model loss value reaches its minimum, a pre-trained ecological health comprehensive value prediction model is generated; or when the model loss value has not reached its minimum, the model parameters are updated, and the steps of inputting each sample binary array into the ecological health comprehensive value prediction model are continued until the model loss value reaches its minimum.

[0009] Optionally, the current environmental status includes the current season type, rainfall level, soil moisture level, temperature level, human activity level, and natural disaster level; Based on the current environmental state at the current moment, the final weight of each preset ecological concern object is dynamically determined, including: From the pre-built mapping relationship between environmental status and vegetation health score, obtain the target vegetation health score corresponding to the current season type, rainfall level, soil moisture level, temperature level, human activity level, and natural disaster level; From the pre-built mapping relationship between environmental conditions and water conservation scores, obtain the target water conservation scores corresponding to the current season type, rainfall level, soil moisture level, temperature level, human activity level, and natural disaster level; From the pre-built mapping relationship between environmental status and biodiversity score, obtain the target biodiversity score corresponding to the current season type, rainfall level, soil moisture level, temperature level, human activity level, and natural disaster level; The final weight of each preset ecological concern object is determined based on the vegetation health score, water conservation score, and biodiversity score of each target.

[0010] Optionally, based on the vegetation health score, water conservation score, and biodiversity score of each target, the final weight of each preset ecological concern object is determined, including: The vegetation health scores of each target are summed to obtain the total vegetation health score; The scores for water conservation of each target are summed to obtain the total score for water conservation. The biodiversity scores for each objective are summed to obtain the total biodiversity score. The total score is obtained by summing the total scores for vegetation health, water conservation, and biodiversity. The ratios between the total vegetation health score, the total water conservation score, the total biodiversity score, and the comprehensive value are calculated separately and used as the final weights for each preset ecological concern object.

[0011] Optionally, based on the trend curve, determine whether to issue an early warning for the target natural area, including: Calculate the final average of the comprehensive ecological health values ​​for the last n time points in the trend curve graph according to the chronological order of time. Calculate the slope of the trend curve; When the slope is less than a preset slope threshold and the ending mean is less than the lower limit of a preset interval, the target area is determined to be high-risk; or, when the slope is greater than or equal to a preset slope threshold and the ending mean is greater than or equal to the upper limit of a preset interval, the target area is determined to be low-risk. When the target area is considered high-risk, early warning information for the target natural area is generated to provide early warning for the target natural area.

[0012] Optionally, before periodically acquiring satellite data, drone data, and ground sensor data for the target natural area within a preset historical time period, the following may also be included: Receive remote sensing image data from satellites, and filter the remote sensing images that cover the natural area of ​​the target from the remote sensing image data to obtain the raw satellite data in the sky; Receive raw aerial drone data collected by drones performing monitoring missions in the target natural area; Receive raw ground sensor data from fixed sensor networks and mobile patrol terminals deployed in the target natural area; The raw satellite data, drone data, and ground sensor data are aligned, cleaned, and standardized according to time series to obtain the final satellite data, drone data, and ground sensor data. Store the final satellite data, the final drone data, and the final ground sensor data in a time-series database.

[0013] Optionally, preset ecological concerns include vegetation health, water conservation, and biodiversity; Extract multiple ecological parameters for each preset ecological interest object, including: Based on satellite data, leaf area index and temporal slope reflecting growth trend are determined as multiple ecological parameters of vegetation health. Based on data from aerial drones, the frequency of species occurrence detected by infrared cameras and the number of endemic species recorded in patrol logs are determined as multiple ecological parameters of biodiversity. Based on ground sensor data, the average soil moisture content, soil humidity, and temperature are determined as multiple ecological parameters for water conservation.

[0014] Secondly, a dynamic early warning method for ecosystem health based on integrated sky-ground system is proposed, the method comprising: From the time-series database, periodically acquire satellite data, drone data, and ground sensor data for the target natural area within a preset historical time period; Multiple ecological parameters for each preset ecological concern object are extracted from satellite data, aerial drone data, and ground sensor data; the final weight of each preset ecological concern object is dynamically determined based on the current environmental state at the current moment; and the multiple ecological parameters and the final weight are weighted and summed to obtain the current comprehensive ecological health value of the target natural area. Based on the current comprehensive ecological health value and the current environmental status, determine the trend curve of the comprehensive ecological health value within a preset future period; based on the trend curve, determine whether to issue an early warning for the target natural area.

[0015] In this embodiment, on the one hand, periodically acquiring satellite data, UAV data, and ground sensor data of the target natural area enables the rapid and efficient collection of large amounts of real-time data. This avoids the significant investment of manpower and time, and the slow data updates associated with traditional methods. Based on multi-source real-time data, ecological parameters are extracted and trend curves are fitted. These trend curves characterize the changing trend of the comprehensive ecological health value over a future period and provide timely warnings. This process significantly shortens the data collection and analysis cycle, improves data real-time performance, and meets the timeliness requirements of ecological protection. On the other hand, by dynamically determining the final weight of each preset ecological concern object based on the current environmental state, the dynamic weight determination method fully considers the impact of real-time changes in environmental factors on different ecological concerns objects. This allows the weighted summation of the comprehensive ecological health value to more accurately reflect the actual ecological condition of the target natural area, enabling more precise determination of whether a warning is needed and improving the accuracy of the warning results.

[0016] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0018] Figure 1 This is a schematic diagram of the system structure of an integrated sky-ground ecosystem health dynamic early warning platform provided in an embodiment of this application; Figure 2 This is a schematic diagram of real-time data storage in a time-series database provided in an embodiment of this application; Figure 3 This is a schematic diagram of an ecological health monitoring result provided in an embodiment of this application; Figure 4 This is a schematic diagram of a trend curve provided in an embodiment of this application; Figure 5 This is a schematic flowchart illustrating the process of generating a trend curve provided in an embodiment of this application. Figure 6 This is a schematic diagram of the method flow for a dynamic early warning method for ecosystem health based on an integrated sky-ground system, provided in an embodiment of this application. Figure 7 This is a flowchart illustrating a training method for an ecological health comprehensive value prediction model provided in an embodiment of this application; Figure 8 This is a schematic diagram of the structure of an integrated sky-ground ecosystem health dynamic early warning device provided in an embodiment of this application; Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0019] The following description and accompanying drawings fully illustrate specific embodiments of this application to enable those skilled in the art to practice them.

[0020] It should be understood that the described embodiments are merely some, not all, of the embodiments in this application. All other embodiments obtained by those skilled in the art based on the embodiments in this application without inventive effort are within the scope of protection of this application.

[0021] In the following description, when referring to the accompanying drawings, the same numbers in different drawings denote the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of systems and methods consistent with some aspects of this application as detailed in the appended claims.

[0022] In the description of this application, it should be understood that the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances. Furthermore, in the description of this application, unless otherwise stated, "multiple" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship.

[0023] Currently, researchers regularly conduct field surveys, collecting data on soil, vegetation, and other factors to analyze and generate assessment reports. Alternatively, they may use a fixed-weight assessment model to create the reports.

[0024] The inventors realized that manual methods are time-consuming and labor-intensive, often taking years, and the data lacks real-time responsiveness, making it difficult to meet the timely needs of ecological protection. Secondly, existing systems are based on fixed-weight assessment models, which ignore dynamic environmental changes and thus reduce the accuracy of early warning results.

[0025] In this application embodiment, on the one hand, periodically acquiring satellite data, UAV data, and ground sensor data of the target natural area enables the rapid and efficient collection of large amounts of real-time data, avoiding the huge investment of manpower and time and the slow data updates in traditional methods. Based on multi-source real-time data, ecological parameters are extracted and trend change curves are fitted. These trend change curves can characterize the changing trend of the comprehensive ecological health value over a future period and issue timely warnings. This process greatly shortens the data collection and analysis cycle, improves data real-time performance, and meets the timeliness requirements of ecological protection. On the other hand, by dynamically determining the final weight of each preset ecological concern object based on the current environmental state, the dynamic weight determination method can fully consider the impact of real-time changes in environmental factors on different ecological concern objects. This makes the weighted summation of the comprehensive ecological health value more accurately reflect the actual ecological status of the target natural area, enabling more precise determination of whether a warning is needed and improving the accuracy of the warning results. The following is a detailed description using exemplary embodiments.

[0026] Please see Figure 1 , Figure 1 This is a schematic diagram of the system structure of an integrated sky-ground ecosystem health dynamic early warning platform provided in this application embodiment. The platform includes: a data preprocessing layer, a model analysis layer, and a health early warning layer.

[0027] In some embodiments of this application, a data preprocessing layer is used to periodically acquire satellite data, UAV data, and ground sensor data of the target natural area within a preset historical time period from a time-series database; a model analysis layer is used to extract multiple ecological parameters for each preset ecological concern object from the satellite data, UAV data, and ground sensor data; dynamically determine the final weight of each preset ecological concern object based on the current environmental state at the current moment; and perform a weighted summation of multiple ecological parameters and the final weight to obtain the current comprehensive ecological health value of the target natural area; a health early warning layer is used to determine the trend curve of the comprehensive ecological health value within a preset future time period based on the current comprehensive ecological health value and the current environmental state; and determine whether to issue an early warning for the target natural area based on the trend curve.

[0028] The time-series database stores pre-processed (aligned, cleaned, and standardized) multi-source data. Satellite data is standardized data acquired through satellite remote sensing technology, such as standardized vegetation indices and land use classifications. Satellite data is used to analyze large-scale ecological conditions, providing macro-level ecological information. Unmanned aerial vehicle (UAV) data is standardized data collected by sensors on UAVs, such as standardized species distribution frequencies and high-resolution imagery. Unmanned aerial vehicle (UAV) data is used to acquire high-resolution ecological data for local areas, providing micro-level ecological information. Ground sensor data is standardized data collected by ground sensors, such as standardized soil moisture and temperature. Ground sensor data provides real-time monitoring data of ground ecological parameters, supporting ecological health assessments. Preset ecological focus objects are pre-defined ecosystem elements requiring attention, such as vegetation health, water conservation, and biodiversity. Ecological parameters describe specific indicators of ecological focus objects, such as leaf area index and soil moisture content. The comprehensive ecological health value is a comprehensive index obtained by weighted summation of ecological parameters, reflecting the overall ecological health status of the target natural area. The trend curve is a graphical representation of the future changes in the comprehensive ecological health value based on historical and current data. It can intuitively show the future trend of the comprehensive ecological health value and assist in early warning decision-making.

[0029] In some embodiments of this application, the specific process of continuously storing data in the time-series database includes: receiving remote sensing image data from satellites, and filtering remote sensing images covering the target natural area from the remote sensing image data to obtain raw satellite data; receiving raw aerial drone data collected by drones performing monitoring tasks in the target natural area; receiving raw ground sensor data collected by fixed sensor networks and mobile patrol terminals deployed in the target natural area; performing data alignment, data cleaning, and standardization on the raw satellite data, raw aerial drone data, and raw ground sensor data according to time sequence to obtain final satellite data, final aerial drone data, and final ground sensor data; and storing the final satellite data, final aerial drone data, and final ground sensor data in the time-series database.

[0030] The data comprises: remote sensing imagery data, which is image data obtained through satellite capture and contains various information about the Earth's surface, such as vegetation cover and land use type; target natural areas, which are specific natural areas of interest to the platform, such as forest reserves, wetlands, and grasslands; raw satellite data, which has undergone preliminary screening to cover the target natural areas; raw UAV data, which is data collected by UAVs in the target natural areas and includes images, videos, and sensor data; and raw ground sensor data, which is data collected by fixed sensor networks and mobile patrol terminals in the target natural areas, such as soil moisture, temperature, and water quality. A time-series database is a database specifically designed to store time-series data, enabling efficient storage and retrieval of data arranged in chronological order.

[0031] In one possible implementation, when the target natural area is a forest park, the platform needs to monitor and provide early warnings about the ecological health of this area. In this case, satellite imagery data from the past year, covering the entire forest park and its surrounding areas, is acquired from the National Remote Sensing Satellite Data Center. Based on the forest park's geographic coordinates, all satellite imagery containing the forest park is filtered out. For example, high-resolution satellite imagery taken each month is selected. Data collected by drones that regularly fly within the forest park, including high-resolution images and infrared imagery, is also collected. Drones fly twice a month, collecting data including vegetation cover and animal activity traces. Data is also collected from a fixed sensor network within the forest park and from mobile terminals carried by patrol personnel. The sensor network includes soil moisture sensors and temperature sensors, and the data collected regularly by patrol personnel includes water quality samples and plant growth data. The satellite imagery capture time, drone flight time, and sensor data collection time are aligned. For example, satellite imagery from January 1, 2025, drone data from January 2, and sensor data from January 3 are aligned to the same day to ensure temporal consistency. Remove cloud cover from satellite imagery, correct blurred areas in UAV images, and eliminate outliers from sensor data. For example, remove cloud-obscured areas from satellite imagery, correct blur caused by unstable flight altitude in UAV images, and eliminate outliers caused by equipment malfunctions in sensor data. Convert soil moisture from percentage to a standardized value (0-1), and unify vegetation indices from different dimensions to standardized values. For example, convert soil moisture from 50% to 0.5, and unify vegetation indices from different dimensions to 0.8. Store the processed satellite data, UAV data, and ground sensor data in InfluxDB. For example, stored data includes: satellite imagery data from January 1, 2025 (standardized vegetation index 0.8); UAV data from January 2, 2025 (standardized species frequency 0.7); and ground sensor data from January 3, 2025 (standardized soil moisture 0.5). Data stored in the time-series database includes, for example... Figure 2 As shown.

[0032] Among them, the pre-set ecological concerns include vegetation health, water conservation, and biodiversity.

[0033] In some embodiments of this application, the specific process of extracting multiple ecological parameters for each preset ecological interest object includes: determining leaf area index and temporal change slope reflecting growth trend as multiple ecological parameters of vegetation health based on satellite data; determining the frequency of species occurrence monitored by infrared cameras and the number of endemic species in patrol records as multiple ecological parameters of biodiversity based on aerial drone data; and determining the average soil moisture content, soil humidity, and temperature as multiple ecological parameters of water conservation based on ground sensor data.

[0034] Vegetation health reflects plant growth and the vegetation cover of an ecosystem. Water conservation refers to an ecosystem's ability to retain and regulate water resources, including soil water retention capacity. Biodiversity is the richness and diversity of species within an ecosystem. Ecological parameters are specific numerical indicators used to quantify the status of ecologically important objects. Leaf area index (LAI) is the total area of ​​plant leaves per unit land area. The slope of time-series variation is the rate of change of growth trends obtained through time-series analysis. Species occurrence frequency is the number or frequency of occurrences of a particular species within the monitored area.

[0035] In one possible implementation, remote sensing image analysis software (such as ENVI) is used to calculate vegetation indices (such as NDVI) and convert them into leaf area indices. Time-series analysis is performed on multi-year satellite data, and linear regression is used to calculate the slope of vegetation index changes. Infrared camera images are analyzed using image recognition software (such as TensorFlow) to count the frequency of species occurrences and the number of endemic species. The average value of soil moisture sensor data over a period of time (such as one month) is calculated. Real-time data from soil moisture and temperature sensors are read, recorded, and analyzed.

[0036] For example Figure 3 As shown, the target natural area is a mountainous forest, and the platform needs to assess its vegetation health, water conservation, and biodiversity. Using ENVI software to analyze satellite imagery from 2025, the average leaf area index (LAI) of the forest area was calculated to be 3.5. Linear regression analysis showed that the annual slope of the vegetation index was 0.02, indicating a good vegetation growth trend. Using TensorFlow to analyze UAV infrared camera images, it was found that a rare species appeared 3 times per month in 2025. Five endemic species were recorded in 2025, indicating high biodiversity in the area. The average soil moisture sensor data from January to December 2025 was calculated, resulting in an average soil moisture content of 25%. Records show that the average soil moisture in 2025 was 25%, and the average soil temperature was 15℃.

[0037] The current environmental status includes the current season type, rainfall level, soil moisture level, temperature level, human activity level, and natural disaster level.

[0038] In some embodiments of this application, the specific process of dynamically determining the final weight of each preset ecological concern object based on the current environmental state includes: obtaining the target vegetation health score corresponding to the current season type, rainfall level, soil moisture level, temperature level, human activity level, and natural disaster level from the pre-constructed mapping relationship between environmental state and vegetation health score; obtaining the target water conservation score corresponding to the current season type, rainfall level, soil moisture level, temperature level, human activity level, and natural disaster level from the pre-constructed mapping relationship between environmental state and water conservation score; obtaining the target biodiversity score corresponding to the current season type, rainfall level, soil moisture level, temperature level, human activity level, and natural disaster level from the pre-constructed mapping relationship between environmental state and biodiversity score; and determining the final weight of each preset ecological concern object based on each target vegetation health score, each target water conservation score, and each target biodiversity score.

[0039] Specifically, the process of determining the final weight of each preset ecological concern object based on the vegetation health score, water conservation score, and biodiversity score of each target includes: summing the vegetation health scores of each target to obtain a total vegetation health score; summing the water conservation scores of each target to obtain a total water conservation score; summing the biodiversity scores of each target to obtain a total biodiversity score; summing the total vegetation health score, total water conservation score, and total biodiversity score to obtain a comprehensive value; and calculating the ratio between the total vegetation health score, total water conservation score, and total biodiversity score and the comprehensive value, which serves as the final weight of each preset ecological concern object.

[0040] It should be noted that by dynamically determining the final weight of each preset ecological concern object based on the current environmental state at the current moment, the dynamic weight determination method can fully consider the impact of real-time changes in environmental factors on different ecological concern objects, so that the comprehensive ecological health value obtained by weighted summation can more accurately reflect the actual ecological status of the target natural area, and can more accurately determine whether an early warning is needed, thus improving the accuracy of the early warning results.

[0041] In one possible implementation, the current season type is spring, the rainfall level indicates low rainfall, the soil moisture level indicates low soil moisture, the temperature level indicates high temperature, the human activity level indicates no human activity, and the natural disaster level indicates no anomalies. In this case, the scores corresponding to the current season type, rainfall level, soil moisture level, temperature level, human activity level, and natural disaster level can be found by looking up the mapping relationship, as shown in Table 1.

[0042] Table 1

[0043] At this point, the vegetation health score = 0.9 (spring) + 0.7 (drought) + 0.8 (high temperature) + 0.8 (soil dryness) = 3.2; the water conservation score = 0.5 + 1.0 + 0.9 + 1.0 = 3.4; the biodiversity score = 0.6 + 0.5 + 0.6 + 0.4 = 2.1; the overall value = 3.2 + 3.4 + 2.1 = 8.7; the vegetation health weight = 3.2 / 8.7 ≈ 0.37; the water conservation weight = 3.4 / 8.7 ≈ 0.39; and the biodiversity weight = 2.1 / 8.7 ≈ 0.24.

[0044] In some embodiments of this application, the specific process of obtaining the current comprehensive ecological health value of the target natural area by weighted summation of multiple ecological parameters and final weights includes: 0.37 (vegetation health weight), 0.39 (water conservation weight), 0.24 (biodiversity weight), a vegetation health score of 0.85 (leaf area index and growth trend slope calculated from satellite data), a water conservation score of 0.90 (mean soil moisture content, humidity, and temperature calculated from ground sensor data), and a biodiversity score of 0.75 (species occurrence frequency and number of endemic species calculated from UAV data). At this time, the current comprehensive ecological health value of the target natural area = 0.37×0.85+0.39×0.90+0.24×0.75=0.3145+0.351+0.18=0.8455; This value is between 0 and 1, and a value close to 1 indicates good ecological health. 0.8455 indicates that the overall ecological health of the mountain forests is currently relatively good.

[0045] In some embodiments of this application, the specific process of determining the trend curve of the ecological health comprehensive value within a future preset time period based on the current ecological health comprehensive value and the current environmental state includes: obtaining the historical ecological health comprehensive value and historical environmental state of the target natural area at the same historical moment; combining the historical ecological health comprehensive value and historical environmental state at each historical moment into a historical binary array; combining the current ecological health comprehensive value and current environmental state into a current binary array; arranging each historical binary array and current binary array in chronological order to obtain a time-series binary array sequence; generating multiple target ecological health comprehensive values ​​within a future preset time period based on the time-series binary array sequence; and projecting the multiple target ecological health comprehensive values ​​into a coordinate graph in chronological order to obtain the trend curve of the ecological health comprehensive value within the future preset time period. For example, the trend curve of the ecological health comprehensive value within a future preset time period... Figure 4 As shown.

[0046] It should be noted that the trend change curve can characterize the changing trend of the comprehensive ecological health value over a period of time and issue timely warnings. This process greatly shortens the data collection and analysis cycle, improves the real-time nature of the data, and can meet the timeliness requirements of ecological protection.

[0047] Specifically, the process of generating multiple target comprehensive ecological health values ​​for a future preset time period based on a time-series binary array sequence includes: using the numerical code of each historical moment in the time-series binary array sequence as the independent variable and the historical comprehensive ecological health value of each historical moment in the time-series binary array sequence as the dependent variable; based on each independent variable and the dependent variable of each independent variable, using the least squares method to solve for the rate of change of the comprehensive ecological health value over time and the predicted comprehensive ecological health value at time t=0; using the rate of change as the slope of the linear regression equation and the predicted comprehensive ecological health value at time t=0 as the intercept of the linear regression equation; constructing a linear regression equation for a trend curve using the slope and intercept; and substituting the numerical code of each target moment within the future preset time period into the independent variable of the linear regression equation to calculate multiple target comprehensive ecological health values ​​within the future preset time period.

[0048] Least squares is a mathematical optimization method used to fit data points to a straight line, minimizing the sum of the squared perpendicular distances from all data points to the line. By minimizing the error, the parameters (slope and intercept) of the linear regression equation are solved. For example, the slope and intercept can be solved by fitting a linear regression model using the least squares method (using NumPy's polyfit function).

[0049] Specifically, the process of generating multiple target comprehensive ecological health values ​​within a preset future time period based on a time-series binary array sequence includes: inputting the time-series binary array sequence into a pre-trained comprehensive ecological health value prediction model, which is obtained through machine learning based on the Transformer time-series model. The Transformer time-series model can capture time-series dependencies and achieve multivariate time-series prediction; and outputting multiple target comprehensive ecological health values ​​within a preset future time period.

[0050] Specifically, the process of generating a pre-trained ecological health comprehensive value prediction model includes: obtaining the first ecological health comprehensive value and the first environmental state determined for different natural regions at the same historical moment; combining the first ecological health comprehensive value and the first environmental state at each historical moment into a sample binary array; establishing an ecological health comprehensive value prediction model using a Transformer time series model; inputting each sample binary array into the ecological health comprehensive value prediction model and outputting the model loss value; generating the pre-trained ecological health comprehensive value prediction model when the model loss value reaches its minimum; or updating the model parameters and continuing to execute the step of inputting each sample binary array into the ecological health comprehensive value prediction model until the model loss value reaches its minimum when the model loss value has not reached its minimum.

[0051] It's important to note that the Transformer model, through its self-attention mechanism, effectively captures short-term and long-term dependencies in time-series data. This allows the model to better understand the changing trends of the overall ecological health value, especially when dealing with complex patterns such as seasonal variations and periodic fluctuations. In ecological health monitoring, ecosystems may be affected by multiple factors, including seasonal changes and climate change. The Transformer can capture these complex time-series dependencies, thus more accurately predicting future trends.

[0052] For example Figure 5 As shown, Figure 5 This application provides a schematic flowchart illustrating the generation process of a trend curve. First, the comprehensive ecological health value and environmental state of the target natural area at historical moments are obtained. Data from each historical moment are combined into a binary array (time code, comprehensive ecological health value). The comprehensive ecological health value and environmental state at the current moment are obtained. The current data are combined into a binary array. All binary arrays are arranged in chronological order to form a time-series binary array sequence. A linear regression method or a Transformer-based machine learning method can be selected. In the linear regression process: the time code is used as the independent variable, and the comprehensive ecological health value as the dependent variable. The least squares method is used to solve for the rate of change (slope) and the initial value (intercept). A linear regression equation is constructed. Substituting the numerical codes from future moments, the target comprehensive ecological health value is calculated. In the model processing process: the time-series binary array sequence is input into a pre-trained Transformer model. The target comprehensive ecological health value within a preset future time period is output. Finally, the calculated future comprehensive ecological health value is projected onto a coordinate graph to generate a trend curve.

[0053] In some embodiments of this application, the specific process of determining whether to issue an early warning for a target natural area based on a trend curve includes: calculating the final average of the ecological health comprehensive values ​​of the last n time points in the trend curve in chronological order; calculating the slope of the trend curve; determining the target area as high-risk when the slope is less than a preset slope threshold and the final average is less than the lower limit of a preset interval; or determining the target area as low-risk when the slope is greater than or equal to a preset slope threshold and the final average is greater than or equal to the upper limit of a preset interval; and generating early warning information for the target natural area to issue an early warning for the target natural area when the target area is high-risk.

[0054] In one possible implementation, the target natural area is a mountainous forest. The platform needs to assess its ecological health risk based on a trend curve and generate early warning information. The comprehensive ecological health values ​​for the last three time points in the trend curve are 0.78, 0.76, and 0.74, respectively.

[0055] The ending mean = (0.78 + 0.76 + 0.74) / 3 = 0.76. Using linear regression to fit the data points in the trend curve, the slope can be calculated to be -0.02. The preset slope threshold is -0.01, and the preset interval is [0.80, 1.00]. The slope -0.02 is less than the preset slope threshold -0.01. The ending mean 0.76 is less than the lower limit of the preset interval 0.80. Based on these conditions, the target area is determined to be high-risk. The generated warning information is: target area (mountainous forest), current ecological health status (high risk), and recommended measures (immediate measures to improve ecological health status, such as increasing vegetation cover and reducing human activities).

[0056] In this embodiment, the platform can assess the ecological health risks of a target natural area based on a trend curve and generate early warning information when necessary, providing a scientific basis for ecological protection and management.

[0057] In this embodiment, on the one hand, periodically acquiring satellite data, UAV data, and ground sensor data of the target natural area enables the rapid and efficient collection of large amounts of real-time data. This avoids the significant investment of manpower and time, and the slow data updates associated with traditional methods. Based on multi-source real-time data, ecological parameters are extracted and trend curves are fitted. These trend curves characterize the changing trend of the comprehensive ecological health value over a future period and provide timely warnings. This process significantly shortens the data collection and analysis cycle, improves data real-time performance, and meets the timeliness requirements of ecological protection. On the other hand, by dynamically determining the final weight of each preset ecological concern object based on the current environmental state, the dynamic weight determination method fully considers the impact of real-time changes in environmental factors on different ecological concerns objects. This allows the weighted summation of the comprehensive ecological health value to more accurately reflect the actual ecological condition of the target natural area, enabling more precise determination of whether a warning is needed and improving the accuracy of the warning results.

[0058] Please see Figure 6 This document provides a flowchart illustrating a method for dynamic early warning of ecosystem health based on an integrated sky-ground system, as described in this application. Figure 6 As shown, the detection method in this application embodiment may include the following steps: S101 periodically acquires satellite data, drone data, and ground sensor data of the target natural area from the time series database within a preset historical time period; S102: Extract multiple ecological parameters for each preset ecological concern object from satellite data, aerial drone data, and ground sensor data; dynamically determine the final weight of each preset ecological concern object based on the current environmental state at the current moment; and sum the multiple ecological parameters and the final weight to obtain the current comprehensive ecological health value of the target natural area. S103. Based on the current comprehensive ecological health value and the current environmental status, determine the trend curve of the comprehensive ecological health value within a preset future period; based on the trend curve, determine whether to issue an early warning for the target natural area.

[0059] In this embodiment, on the one hand, periodically acquiring satellite data, UAV data, and ground sensor data of the target natural area enables the rapid and efficient collection of large amounts of real-time data. This avoids the significant investment of manpower and time, and the slow data updates associated with traditional methods. Based on multi-source real-time data, ecological parameters are extracted and trend curves are fitted. These trend curves characterize the changing trend of the comprehensive ecological health value over a future period and provide timely warnings. This process significantly shortens the data collection and analysis cycle, improves data real-time performance, and meets the timeliness requirements of ecological protection. On the other hand, by dynamically determining the final weight of each preset ecological concern object based on the current environmental state, the dynamic weight determination method fully considers the impact of real-time changes in environmental factors on different ecological concerns objects. This allows the weighted summation of the comprehensive ecological health value to more accurately reflect the actual ecological condition of the target natural area, enabling more precise determination of whether a warning is needed and improving the accuracy of the warning results.

[0060] Please see Figure 7 This is a flowchart illustrating a training method for an ecological health comprehensive value prediction model, provided in an embodiment of this application. Figure 7 As shown, the process includes the following steps: S201, to obtain the first comprehensive ecological health value and the first environmental state of different natural regions at the same historical moment; S202 combines the first comprehensive ecological health value and the first environmental state at each historical moment into a binary array for each sample. S203, using the Transformer time series model, establishes a comprehensive ecological health value prediction model; S204: Input each sample binary array into the ecological health comprehensive value prediction model and output the model loss value; S205: When the model loss value reaches the minimum, generate a pre-trained ecological health comprehensive value prediction model; or when the model loss value has not reached the minimum, update the model parameters and continue to execute the step of inputting each sample binary array into the ecological health comprehensive value prediction model until the model loss value reaches the minimum.

[0061] In this embodiment, on the one hand, periodically acquiring satellite data, UAV data, and ground sensor data of the target natural area enables the rapid and efficient collection of large amounts of real-time data. This avoids the significant investment of manpower and time, and the slow data updates associated with traditional methods. Based on multi-source real-time data, ecological parameters are extracted and trend curves are fitted. These trend curves characterize the changing trend of the comprehensive ecological health value over a future period and provide timely warnings. This process significantly shortens the data collection and analysis cycle, improves data real-time performance, and meets the timeliness requirements of ecological protection. On the other hand, by dynamically determining the final weight of each preset ecological concern object based on the current environmental state, the dynamic weight determination method fully considers the impact of real-time changes in environmental factors on different ecological concerns objects. This allows the weighted summation of the comprehensive ecological health value to more accurately reflect the actual ecological condition of the target natural area, enabling more precise determination of whether a warning is needed and improving the accuracy of the warning results.

[0062] The following are embodiments of the apparatus described in this application, which can be used to execute the embodiments of the method described in this application. For details not disclosed in the apparatus embodiments of this application, please refer to the embodiments of the method described in this application.

[0063] Please see Figure 8 This illustration shows a schematic diagram of a space-ground integrated ecosystem health dynamic early warning device provided in an exemplary embodiment of this application. This space-ground integrated ecosystem health dynamic early warning system can be implemented as all or part of an electronic device through software, hardware, or a combination of both. The space-ground integrated ecosystem health dynamic early warning device 1 includes a multi-source data sensing module 10, a current ecological health comprehensive value generation module 20, and an early warning module 30.

[0064] The multi-source data sensing module 10 is used to periodically acquire satellite data, drone data, and ground sensor data of the target natural area from a time-series database within a preset historical time period. The current ecological health comprehensive value generation module 20 is used to extract multiple ecological parameters for each preset ecological concern object from satellite data, aerial drone data, and ground sensor data; dynamically determine the final weight of each preset ecological concern object based on the current environmental state at the current moment; and perform a weighted summation of multiple ecological parameters and final weights to obtain the current ecological health comprehensive value of the target natural area. The early warning module 30 is used to determine the trend curve of the ecological health comprehensive value within a preset period of time based on the current ecological health comprehensive value and the current environmental status; and to determine whether to issue an early warning for the target natural area based on the trend curve.

[0065] It should be noted that the above-described embodiments of the integrated sky-ground ecosystem health dynamic early warning device, when executing the integrated sky-ground ecosystem health dynamic early warning method, only illustrate the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. Furthermore, the integrated sky-ground ecosystem health dynamic early warning device and the integrated sky-ground ecosystem health dynamic early warning method embodiments belong to the same concept, and their implementation process is detailed in the method embodiments, which will not be repeated here.

[0066] The sequence numbers of the embodiments in this application are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0067] In this embodiment, on the one hand, periodically acquiring satellite data, UAV data, and ground sensor data of the target natural area enables the rapid and efficient collection of large amounts of real-time data. This avoids the significant investment of manpower and time, and the slow data updates associated with traditional methods. Based on multi-source real-time data, ecological parameters are extracted and trend curves are fitted. These trend curves characterize the changing trend of the comprehensive ecological health value over a future period and provide timely warnings. This process significantly shortens the data collection and analysis cycle, improves data real-time performance, and meets the timeliness requirements of ecological protection. On the other hand, by dynamically determining the final weight of each preset ecological concern object based on the current environmental state, the dynamic weight determination method fully considers the impact of real-time changes in environmental factors on different ecological concerns objects. This allows the weighted summation of the comprehensive ecological health value to more accurately reflect the actual ecological condition of the target natural area, enabling more precise determination of whether a warning is needed and improving the accuracy of the warning results.

[0068] This application also provides a computer-readable medium having program instructions stored thereon, which, when executed by a processor, implement the dynamic early warning method for ecosystem health based on the sky-ground integration provided in the above-described method embodiments.

[0069] This application also provides a computer program product containing instructions that, when run on a computer, causes the computer to execute the above-described method embodiments of the integrated sky-ground ecosystem health dynamic early warning method.

[0070] Please see Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 9 As shown, the electronic device 1000 may include: at least one processor 1001, at least one network interface 1004, a user interface 1003, a memory 1005, and at least one communication bus 1002.

[0071] The communication bus 1002 is used to realize the connection and communication between these components.

[0072] The user interface 1003 may include a display screen and a camera. Optionally, the user interface 1003 may also include a standard wired interface and a wireless interface.

[0073] The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0074] The processor 1001 may include one or more processing cores. The processor 1001 connects to various parts within the electronic device 1000 using various interfaces and lines. It executes various functions and processes data by running or executing instructions, programs, code sets, or instruction sets stored in the memory 1005, and by calling data stored in the memory 1005. Optionally, the processor 1001 may be implemented using at least one hardware form selected from Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), and Programmable Logic Array (PLA). The processor 1001 may integrate one or more of the following: a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), and a modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content to be displayed on the screen; and the modem handles wireless communication. It is understood that the modem may also be implemented as a separate chip, without being integrated into the processor 1001.

[0075] The memory 1005 may include random access memory (RAM) or read-only memory. Optionally, the memory 1005 may include a non-transitory computer-readable storage medium. The memory 1005 can be used to store instructions, programs, code, code sets, or instruction sets. The memory 1005 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory 1005 may also be at least one storage system located remotely from the aforementioned processor 1001. Figure 9 As shown, the memory 1005, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and an integrated sky-ground ecosystem health dynamic early warning application.

[0076] exist Figure 9 In the illustrated electronic device 1000, the user interface 1003 is mainly used to provide an input interface for the user and to acquire user input data; while the processor 1001 can be used to call the space-air-ground integrated ecosystem health dynamic early warning application stored in the memory 1005, and specifically perform the following operations: From the time-series database, periodically acquire satellite data, drone data, and ground sensor data for the target natural area within a preset historical time period; Multiple ecological parameters for each preset ecological concern object are extracted from satellite data, aerial drone data, and ground sensor data; the final weight of each preset ecological concern object is dynamically determined based on the current environmental state at the current moment; and the multiple ecological parameters and the final weight are weighted and summed to obtain the current comprehensive ecological health value of the target natural area. Based on the current comprehensive ecological health value and the current environmental status, determine the trend curve of the comprehensive ecological health value within a preset future period; based on the trend curve, determine whether to issue an early warning for the target natural area.

[0077] In this embodiment, on the one hand, periodically acquiring satellite data, UAV data, and ground sensor data of the target natural area enables the rapid and efficient collection of large amounts of real-time data. This avoids the significant investment of manpower and time, and the slow data updates associated with traditional methods. Based on multi-source real-time data, ecological parameters are extracted and trend curves are fitted. These trend curves characterize the changing trend of the comprehensive ecological health value over a future period and provide timely warnings. This process significantly shortens the data collection and analysis cycle, improves data real-time performance, and meets the timeliness requirements of ecological protection. On the other hand, by dynamically determining the final weight of each preset ecological concern object based on the current environmental state, the dynamic weight determination method fully considers the impact of real-time changes in environmental factors on different ecological concerns objects. This allows the weighted summation of the comprehensive ecological health value to more accurately reflect the actual ecological condition of the target natural area, enabling more precise determination of whether a warning is needed and improving the accuracy of the warning results.

[0078] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program for dynamic early warning of ecosystem health based on integrated sky-ground systems can be stored in a computer-readable storage medium. When executed, the program can include the processes of the embodiments of the above methods. The storage medium for the program for dynamic early warning of ecosystem health based on integrated sky-ground systems can be a magnetic disk, optical disk, read-only memory, or random access memory, etc.

[0079] The above-disclosed embodiments are merely preferred embodiments of this application and should not be construed as limiting the scope of this application. Therefore, any equivalent variations made in accordance with the claims of this application shall still fall within the scope of this application.

Claims

1. A dynamic early warning platform for ecosystem health based on a sky-ground integrated system, characterized in that, The platform includes: Data preprocessing layer, model analysis layer, and health early warning layer; among them, The data preprocessing layer is used to periodically obtain satellite data, UAV data, and ground sensor data of the target natural area from the time series database within a preset historical time period. The model analysis layer is used to extract multiple ecological parameters for each preset ecological concern object from the satellite data, aerial drone data, and ground sensor data; dynamically determine the final weight of each preset ecological concern object based on the current environmental state at the current moment; and perform a weighted summation of the multiple ecological parameters and the final weight to obtain the current comprehensive ecological health value of the target natural area. The health early warning layer is used to determine the trend curve of the ecological health comprehensive value within a preset future period based on the current ecological health comprehensive value and the current environmental state; and to determine whether to issue an early warning for the target natural area based on the trend curve.

2. The system according to claim 1, characterized in that, The step of determining the trend curve of the ecological health comprehensive value within a preset future time period based on the current ecological health comprehensive value and the current environmental state includes: Obtain the historical comprehensive ecological health value and historical environmental status of the target natural region at the same historical moment; The historical ecological health comprehensive value and historical environmental status at each historical moment are combined into a historical binary array; Combine the current comprehensive ecological health value and the current environmental state into a current binary array; Arrange each historical binary array and the current binary array in chronological order to obtain a time-series binary array sequence. Based on the time-series binary array sequence, generate multiple target ecological health comprehensive values ​​for a future preset time period; The multiple target comprehensive ecological health values ​​are projected onto a coordinate graph in chronological order to obtain a trend curve of the comprehensive ecological health values ​​within a preset future time period.

3. The system according to claim 2, characterized in that, The step of generating multiple target ecological health comprehensive values ​​for a future preset time period based on the time-series binary array sequence includes: The numerical code of each historical moment in the time-series binary array sequence is used as the independent variable, and the historical ecological health comprehensive value of each historical moment in the time-series binary array sequence is used as the dependent variable. Based on each independent variable and the dependent variable of each independent variable, the rate of change of the comprehensive ecological health value over time and the predicted comprehensive ecological health value when time t=0 are obtained by combining the least squares method. The rate of change is used as the slope of the linear regression equation, and the predicted comprehensive ecological health value at time t=0 is used as the intercept of the linear regression equation. A linear regression equation for the trend curve is constructed using the slope and the intercept. The numerical codes of each target time within the future preset time period are substituted into the independent variables in the linear regression equation to calculate the comprehensive ecological health values ​​of multiple targets within the future preset time period.

4. The system according to claim 2, characterized in that, The step of generating multiple target ecological health comprehensive values ​​for a future preset time period based on the time-series binary array sequence includes: The time-series binary array sequence is input into a pre-trained ecological health comprehensive value prediction model. The pre-trained ecological health comprehensive value prediction model is obtained by machine learning based on the Transformer time-series model. The Transformer time-series model can capture time-series dependencies and achieve multivariate time-series prediction. Output the comprehensive ecological health values ​​of multiple targets within a preset future time period; among them... Generate a pre-trained comprehensive ecological health value prediction model according to the following steps: Obtain the first comprehensive ecological health value and the first environmental state of different natural regions at the same historical moment; The first comprehensive ecological health value and the first environmental state at each historical moment are combined into a binary array for each sample. A comprehensive ecological health value prediction model was established using the Transformer time series model. Input each sample binary array into the ecological health comprehensive value prediction model and output the model loss value; When the model loss value reaches its minimum, a pre-trained ecological health comprehensive value prediction model is generated; or when the model loss value has not reached its minimum, the model parameters are updated, and the step of inputting each sample binary array into the ecological health comprehensive value prediction model is continued until the model loss value reaches its minimum.

5. The system according to claim 1, characterized in that, The current environmental status includes the current season type, rainfall level, soil moisture level, temperature level, human activity level, and natural disaster level; The dynamic determination of the final weight of each preset ecological concern object based on the current environmental state at the current moment includes: From the pre-built mapping relationship between environmental status and vegetation health score, obtain the target vegetation health score corresponding to the current season type, rainfall level, soil moisture level, temperature level, human activity level, and natural disaster level; From the pre-built mapping relationship between environmental conditions and water conservation scores, obtain the target water conservation scores corresponding to the current season type, rainfall level, soil moisture level, temperature level, human activity level, and natural disaster level; From the pre-built mapping relationship between environmental status and biodiversity score, obtain the target biodiversity score corresponding to the current season type, rainfall level, soil moisture level, temperature level, human activity level, and natural disaster level; The final weight of each preset ecological concern object is determined based on the vegetation health score, water conservation score, and biodiversity score of each target.

6. The system according to claim 5, characterized in that, The final weight of each preset ecological concern object is determined based on the target vegetation health score, target water conservation score, and target biodiversity score, including: The vegetation health scores of each target are summed to obtain the total vegetation health score. The scores for each target water conservation objective are summed to obtain the total water conservation score. The biodiversity scores for each objective are summed to obtain the total biodiversity score. The sum of the total scores for vegetation health, water conservation, and biodiversity is used to obtain a comprehensive value. The ratios between the total vegetation health score, the total water conservation score, the total biodiversity score, and the comprehensive value are calculated respectively, and used as the final weights for each preset ecological concern object.

7. The system according to claim 1, characterized in that, The step of determining whether to issue an early warning for the target natural area based on the trend curve includes: Calculate the final average of the comprehensive ecological health values ​​for the last n time points in the trend curve graph according to the chronological order of time. Calculate the slope of the trend curve; When the slope is less than a preset slope threshold and the ending mean is less than the lower limit of a preset interval, the target area is determined to be high-risk; or, when the slope is greater than or equal to a preset slope threshold and the ending mean is greater than or equal to the upper limit of a preset interval, the target area is determined to be low-risk. If the target area is deemed high-risk, an early warning message for the target natural area is generated to issue an early warning for the target natural area.

8. The system according to any one of claims 1-7, characterized in that, Before periodically acquiring satellite data, UAV data, and ground sensor data for the target natural area within a preset historical time period, the process also includes: Receive remote sensing image data from satellites, and filter remote sensing images covering the target natural area from the remote sensing image data to obtain raw satellite data from the sky; Receive raw aerial drone data collected by drones performing monitoring missions in the target natural area; Receive raw ground sensor data from fixed sensor networks and mobile patrol terminals deployed in the target natural area; The original satellite data, the original UAV data, and the original ground sensor data are aligned, cleaned, and standardized according to time sequence to obtain the final satellite data, the final UAV data, and the final ground sensor data. The final satellite data, final UAV data, and final ground sensor data are stored in a time-series database.

9. The system according to any one of claims 1-7, characterized in that, The preset ecological concerns include vegetation health, water conservation, and biodiversity. The extraction of multiple ecological parameters for each preset ecological interest object includes: Based on the satellite data, the leaf area index and the temporal slope reflecting the growth trend are determined as multiple ecological parameters of the vegetation health. Based on the data from the aerial drone, the frequency of species occurrence detected by the infrared camera and the number of endemic species in the patrol records are determined as multiple ecological parameters of the biodiversity. Based on the ground sensor data, the average soil moisture content, soil humidity, and temperature are determined as multiple ecological parameters for water conservation.

10. A dynamic early warning method for ecosystem health based on integrated sky-ground system, characterized in that, The method includes: From the time-series database, periodically acquire satellite data, drone data, and ground sensor data for the target natural area within a preset historical time period; Multiple ecological parameters of each preset ecological concern object are extracted from the satellite data, aerial drone data, and ground sensor data; the final weight of each preset ecological concern object is dynamically determined based on the current environmental state at the current moment; the multiple ecological parameters and the final weight are weighted and summed to obtain the current comprehensive ecological health value of the target natural area. Based on the current comprehensive ecological health value and the current environmental state, a trend curve of the comprehensive ecological health value over a future preset period is determined; based on the trend curve, it is determined whether to issue an early warning for the target natural area.

Images