Lake ecological trend prediction method and system based on big data analysis

Abstract

The application discloses a lake ecological trend prediction method and system based on big data analysis, and the method comprises the following steps: according to a standardized data set, a multi-source data integration framework is constructed, the spatial resolution of remote sensing images and the time series data of sensors are subjected to feature extraction and complementary fusion, and a fused comprehensive data set is obtained; the abnormal data points in the comprehensive data set are subjected to weighted correction, and a corrected dynamic data set is obtained; for the environmental change trend in the prediction result, the short-term fluctuation and long-term trend of the lake ecological system are subjected to hierarchical analysis by adopting a sliding window technology, and the time distribution and spatial distribution of key risk points are determined; according to the latest monitoring key area, real-time data of the corresponding area in the multi-source data integration framework are combined to perform local prediction, and a lake ecological trend prediction result is obtained. The application improves the spatio-temporal coverage and prediction accuracy of lake ecological monitoring.

Description

Technical Field

[0001] This invention belongs to the field of lake ecological trend prediction technology, and in particular relates to a method and system for predicting lake ecological trends based on big data analysis. Background Technology

[0002] Lake ecological protection, as a crucial field of environmental science and water resource management, holds irreplaceable value in maintaining ecological balance and safeguarding the human living environment. Lakes are not only an important component of natural ecosystems but also vital water sources for human production and daily life; their health directly impacts regional sustainable development. However, current methods for lake ecological monitoring and management face numerous shortcomings, failing to address increasingly complex environmental challenges. Traditional methods rely heavily on manual sampling and fixed equipment, resulting in narrow data coverage, slow update speeds, and an inability to meet the demands of dynamic management. Furthermore, the lack of information-sharing mechanisms between different departments leads to low resource utilization efficiency.

[0003] The field of lake ecological conservation faces significant technical challenges due to the limitations of existing methods. The most pressing issue is the integration of multi-source information. Because monitoring data comes from a wide range of sources and takes many forms, such as remote sensing images from the sky, sensor records from the water, and meteorological observations from the ground, there is a lack of unified standards and methods to fuse this information, making it difficult to form a comprehensive ecological profile. This inadequate integration further leads to a deficiency in predictive capabilities. Without a complete data foundation, the predictive models built often fail to accurately capture the complex changes in lake ecosystems, especially when faced with sudden pollution or climate anomalies, where the models' adaptability and accuracy become severely limited.

[0004] Therefore, effectively integrating multi-source data and constructing dynamically adaptive predictive models based on this has become a key issue in lake ecological protection and management. This invention focuses on overcoming the technical barriers to data integration and improving the model's responsiveness to environmental changes through innovative methods, providing reliable support for real-time monitoring and scientific decision-making in lake ecosystems. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention proposes a lake ecological trend prediction method and system based on big data analysis, which improves the spatiotemporal coverage and prediction accuracy of lake ecological monitoring.

[0006] To achieve the above objectives, this invention provides a method for predicting lake ecological trends based on big data analysis, comprising:

[0007] Multi-source monitoring data of the lake is acquired, and the multi-source monitoring data is standardized to obtain a standardized dataset. Based on the standardized dataset, a multi-source data integration framework is constructed, and feature extraction and complementary fusion are performed on the spatial resolution of remote sensing images and the time series data of sensors to obtain a fused comprehensive dataset.

[0008] Based on the integrated dataset, a time-series analysis of the historical change trend of the lake ecosystem is performed to obtain the fluctuation characteristics of key variables; the abnormal data points in the fluctuation characteristics of key variables are weighted and corrected to obtain the corrected dynamic dataset.

[0009] The modified dynamic dataset is combined with environmental variables and input into the multi-source data integration framework for training to obtain prediction results; based on the environmental change trend data in the prediction results, the short-term fluctuations and long-term trends of the lake ecosystem are analyzed in a hierarchical manner to obtain the temporal and spatial distribution of key risk points.

[0010] Based on the temporal and spatial distribution of key risk points, a dynamic monitoring priority map is generated. The priority map is periodically adjusted to obtain the latest key monitoring areas. Based on the latest key monitoring areas, real-time data of the corresponding areas is extracted using the multi-source data integration framework to perform local predictions and obtain lake ecological trend prediction results.

[0011] On the other hand, to achieve the above objectives, the present invention provides a lake ecological trend prediction system based on big data analysis, including: a data standardization module, a multi-source data integration module, a time series analysis module, an anomaly correction module, a prediction result acquisition module, a hierarchical analysis module, a dynamic monitoring priority generation module, and a local prediction module;

[0012] The data standardization module is used to acquire multi-source monitoring data of the lake, and to perform standardization processing on the multi-source monitoring data to obtain a standardized dataset.

[0013] The multi-source data integration module is used to construct a multi-source data integration framework based on the standardized dataset, and to perform feature extraction and complementary fusion of the spatial resolution of remote sensing images and the time series data of sensors to obtain a fused comprehensive dataset.

[0014] The time series analysis module is used to perform time series analysis on the historical change trend of lake ecology based on the fused comprehensive dataset, and to obtain the fluctuation characteristics of key variables;

[0015] The anomaly correction module is used to perform weighted correction on the abnormal data points in the fluctuation characteristics of the key variables to obtain the corrected dynamic dataset.

[0016] The prediction result acquisition module is used to input the corrected dynamic dataset and environmental variables into the multi-source data integration framework for training to obtain the prediction result.

[0017] The hierarchical analysis module is used to perform hierarchical analysis on the short-term fluctuations and long-term trends of the lake ecosystem based on the environmental change trend data in the prediction results, and to obtain the temporal and spatial distribution of key risk points.

[0018] The dynamic monitoring priority generation module is used to generate a dynamic monitoring priority map based on the time and spatial distribution of key risk points, and to periodically adjust the priority map to obtain the latest key monitoring areas.

[0019] The local prediction module is used to perform local predictions based on the latest key monitoring areas and the real-time data of the corresponding areas extracted from the multi-source data integration framework, so as to obtain the lake ecological trend prediction results.

[0020] Technical Effects of this Invention: This invention discloses a method and system for predicting lake ecological trends based on big data analysis. It addresses the problem of insufficient data coverage by integrating and standardizing multi-source data, utilizes a weighted average algorithm for data fusion, employs a long short-term memory network model to analyze ecosystem change trends and identify abnormal patterns, constructs an adaptive prediction tool to dynamically respond to environmental changes, combines sliding window technology to determine the distribution of key risk points, generates a dynamic monitoring priority map and updates it in real time, ultimately achieving accurate prediction and monitoring of potential environmental challenges to lake ecosystems. This invention effectively integrates multi-source heterogeneous data, improves the spatiotemporal coverage and prediction accuracy of lake ecological monitoring, and provides scientific decision support for lake ecological protection. Attached Figure Description

[0021] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0022] Figure 1 This is a flowchart illustrating the lake ecological trend prediction method based on big data analysis, as described in an embodiment of the present invention.

[0023] Figure 2 This is a schematic diagram of the structure of the lake ecological trend prediction system based on big data analysis, according to an embodiment of the present invention. Detailed Implementation

[0024] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0025] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0026] like Figure 1 As shown, this embodiment provides a lake ecological trend prediction method based on big data analysis, including: acquiring multi-source monitoring data of the lake; standardizing the multi-source monitoring data to obtain a standardized dataset; constructing a multi-source data integration framework based on the standardized dataset; and extracting and complementaryly fusing features from the spatial resolution of remote sensing images and the time series data of sensors to obtain a fused comprehensive dataset.

[0027] Based on the integrated dataset, a time-series analysis of the historical change trend of the lake ecosystem is performed to obtain the fluctuation characteristics of key variables; the abnormal data points in the fluctuation characteristics of key variables are weighted and corrected to obtain the corrected dynamic dataset.

[0028] The modified dynamic dataset is combined with environmental variables and input into the multi-source data integration framework for training to obtain prediction results; based on the environmental change trend data in the prediction results, the short-term fluctuations and long-term trends of the lake ecosystem are analyzed in a hierarchical manner to obtain the temporal and spatial distribution of key risk points.

[0029] Based on the temporal and spatial distribution of key risk points, a dynamic monitoring priority map is generated. The priority map is periodically adjusted to obtain the latest key monitoring areas. Based on the latest key monitoring areas, real-time data of the corresponding areas is extracted using the multi-source data integration framework to perform local predictions and obtain lake ecological trend prediction results.

[0030] Furthermore, obtaining a standardized dataset includes:

[0031] To address the differences in the sources of the multi-source monitoring data for the lake, a pre-established data standardization protocol was used to convert the format of the raw data, resulting in an intermediate dataset with a unified format.

[0032] The timestamps of data from different sources in the intermediate dataset are calibrated, and the calibrated dataset is time-aligned using a linear interpolation method to obtain a standardized dataset.

[0033] Specifically, in the scenario of lake ecological protection monitoring, the acquisition and processing of multi-source monitoring data is a core component. Given the differences in data sources—remote sensing imagery, underwater sensors, and ground-based weather stations—the implementation of data standardization protocols is crucial. Assuming remote sensing imagery data is stored in raster format, underwater sensor data is time-series text, and weather station data is in tabular form, standardization protocols convert these data into a unified structured format, such as an intermediate dataset indexed by time and spatial coordinates. This conversion ensures consistency in subsequent processing and reduces errors caused by format differences. In time synchronization processing, timestamp calibration is a critical step. Assuming remote sensing imagery data is updated hourly, underwater sensors record data every 30 minutes, and weather station data is updated every 10 minutes, the timestamp discrepancy could reach 20 minutes. If a preset threshold of 5 minutes is set, time alignment must be achieved using linear interpolation. If underwater sensor data is missing at 10:30, the value at 10:30 can be linearly estimated based on data from 10:00 and 11:00, generating a time-consistent calibration dataset. This method effectively reduces the impact of time discrepancies on analysis and improves data reliability.

[0034] Furthermore, the resulting integrated dataset includes:

[0035] The remote sensing images and sensor data in the standardized dataset are processed in a hierarchical manner. Spatial resolution and time series features are extracted from data from different sources to obtain a hierarchical feature set. The spatial resolution features of the remote sensing images and the time series features of the sensor data in the hierarchical feature set are weighted to obtain a weighted feature combination.

[0036] Based on the weighted feature combination, features from different sources are complementaryly fused, and data barriers that appear during the fusion are corrected using pre-established rules to determine an intermediate fused dataset. Based on the intermediate fused dataset, the fusion parameters are adjusted for optimization to obtain the fused comprehensive dataset.

[0037] Specifically, in the monitoring scenario of lake ecological protection, the application of data classification tools is a crucial step in processing standardized datasets. For hierarchical processing of remote sensing imagery and sensor data, features can be extracted from both spatial resolution and temporal series dimensions. Remote sensing imagery typically has high spatial coverage, reflecting large-scale vegetation distribution or water changes on the lake surface, while sensor data excels at capturing continuous temporal changes, such as fluctuations in water temperature or dissolved oxygen levels. Classification tools can extract the spatial features of remote sensing imagery and the temporal features of sensors separately, forming a hierarchical feature set, laying the foundation for subsequent analysis. Specifically, the spatial resolution features of remote sensing imagery might be divided into grids per square kilometer, while the temporal series features of sensor data record changing trends hourly. In the application of weighted averaging tools, when weighting the hierarchical feature set, weights can be assigned based on the reliability and importance of the data source. For example, if the spatial resolution features of remote sensing imagery are incomplete due to cloud cover, its weight might be set to 0.3, while the sensor data, due to long-term stable operation, might have a weight of 0.7. If the weight of a certain feature is lower than a preset threshold, such as 0.2, the weight ratio needs to be adjusted and the weighted feature combination recalculated. This method can balance the influence of different data sources and ensure that the final feature combination is more representative.

[0038] Furthermore, obtaining the fluctuation characteristics of key variables includes:

[0039] Based on the comprehensive dataset, time-series information on water temperature, rainfall, and evaporation rate is obtained. The time-series information is then segmented to obtain a set of segmented environmental variables. Using the set of segmented environmental variables, cross-validation is performed on water level changes, watershed flow, and climate fluctuations. Variables are marked as anomalies to determine combinations of anomalous variables.

[0040] For the aforementioned combination of abnormal variables, the relationship between lake water quality, pollutant concentration, and dissolved oxygen is analyzed to obtain potential abnormal pattern characteristics from the relationship and identify key influencing factors. Based on the key influencing factors, ecological data on vegetation cover, algae density, and fish population are jointly processed to obtain the fluctuation characteristics of key variables.

[0041] Specifically, in the context of lake ecological monitoring, the application of data filtering tools is a crucial step in processing comprehensive datasets. Data filtering tools can extract time-series information such as water temperature, rainfall, and evaporation rates from historical records, forming a set of environmental variables through segmentation. Assuming the data is segmented monthly, water temperature might show an average of 28.5 degrees Celsius in summer and 5.2 degrees Celsius in winter, while rainfall might be recorded as an average of 200 mm per month in summer and 50 mm per month in winter. This segmentation helps capture the periodic changes in environmental variables, providing a clear time frame for subsequent analysis. When cross-validating the segmented set of environmental variables using data comparison tools, the relationship between water level changes, watershed discharge, and climate fluctuations can be examined. Assuming a preset threshold of monthly water level changes not exceeding 0.5 meters, if the water level rises by 0.8 meters in a given month, while rainfall abnormally exceeds the average by 150 mm, this combination of variables can be identified as an anomaly. This comparison method can quickly identify potential problem areas, laying the foundation for further analysis.

[0042] Furthermore, obtaining the corrected dynamic dataset includes:

[0043] The abnormal data points in the comprehensive dataset are initially screened, and data records exceeding the preset threshold are obtained from environmental variables and water parameters. The fluctuation characteristics and climate impact are cross-checked to obtain the distribution set of abnormal data points.

[0044] Based on the distribution set of the abnormal data points, the values ​​of the abnormal data points are adjusted, and the correlation between ecological indicators and climate impact is combined through data integration to obtain a corrected dynamic dataset.

[0045] Specifically, in the scenario of lake ecosystem monitoring, the screening of outlier data points in a comprehensive dataset can begin with environmental variables and water parameters to initially identify data records exceeding preset thresholds. For example, if the normal range for water temperature is set at 5 to 30 degrees Celsius, and a record shows a temperature of 32.5 degrees Celsius in a certain month, then that data point is marked as an anomaly. Similarly, if the monthly water level change threshold is set at 0.5 meters, and a record shows a change of 0.7 meters, it will also be included in the anomaly range. This screening method can quickly identify potentially problematic data, providing a foundation for subsequent analysis. Regarding the application of data comparison tools, when cross-checking fluctuation characteristics and climate impacts, the relationship between water level changes and rainfall can be considered. For example, if the water level rises abnormally by 0.6 meters over a certain period, while the recorded monthly rainfall during the same period is 300 millimeters, far exceeding the average of 150 millimeters, comparison can suggest that rainfall is the main influencing factor. This verification method helps to clarify the causal distribution of outlier data points, forming a clear distribution set. When using weighted correction tools to adjust outlier data points, the data can be reasonably corrected based on the correlation between historical trends and climate influences. For example, assuming an outlier water level of 1.2 meters, while the historical average for the same period is 0.8 meters, and considering the higher rainfall, the correction weights can be allocated as follows: 60% for rainfall influence and 40% for other factors, thus adjusting the data to a more reasonable range of 0.9 meters. This correction method can reduce data bias and improve the reliability of the analysis.

[0046] Furthermore, the predicted results include:

[0047] Based on the corrected dynamic dataset, the relevant records of environmental variables and ecological changes are classified and organized to obtain a distribution set related to fluctuation characteristics. The outliers in the distribution set are preliminarily processed to obtain the adjusted data combination.

[0048] A second correction is performed based on the adjusted data combination, and weights are assigned to the corrected data points to determine the data subset.

[0049] By combining the data subset with the correlation between environmental variables and ecological changes, key indicators in the data subset are screened to obtain feature combinations; time-series features of the feature combinations are extracted and analyzed to obtain model information related to dynamic response; the model information is compared with climate impact to obtain prediction results.

[0050] Specifically, in the scenario of lake ecosystem monitoring, for the corrected dynamic dataset, data integration tools can be used to classify and organize relevant records of environmental variables and ecological changes. Assuming environmental variables include water temperature and rainfall, and aquatic ecological changes involve algae density and fish activity frequency, the integration tool can classify these data by month, forming a clear distribution set. For outliers in the distribution set, such as a water temperature reaching 33 degrees Celsius in a certain month, exceeding the normal range of 5 to 30 degrees Celsius, a numerical adjustment tool can be used for initial processing, adjusting it to a value close to the historical average of 29 degrees Celsius, forming an adjusted data set. If fluctuations still exist in the adjusted data set, such as water temperatures slightly exceeding the average in some months, a data smoothing tool can be used for secondary correction. For example, assuming a water temperature of 31 degrees Celsius in a certain month, the smoothing tool, combined with the mean of data from the preceding and following months, can adjust it to 30.5 degrees Celsius. If the correlation between the corrected data points and climate impacts exceeds a preset threshold, such as the correlation between rainfall and water temperature exceeding 80%, then the data points are weighted using a correlation analysis tool to determine that rainfall accounts for 70% and other factors account for 30%, thereby selecting a subset of data that meets the requirements of dynamic response.

[0051] Furthermore, determining the temporal and spatial distribution of key risk points includes:

[0052] The environmental change trend data in the prediction results are segmented to obtain short-term fluctuation data segments, and the correlation between the data segments and the long-term trend is compared to obtain the distribution characteristics of short-term fluctuations.

[0053] The distribution characteristics of the short-term fluctuations are overlaid with the geographical location data of the lake ecosystem to determine the spatial distribution of key risks;

[0054] Based on the spatial distribution of the key risks, the records of environmental changes are broken down to obtain the temporal distribution information related to the risk points. The risk points are then classified using a fluctuation feature comparison tool to determine the temporal distribution pattern of the risk points.

[0055] Based on the temporal distribution patterns of the aforementioned risk points, data on short-term fluctuations and long-term trends are stored in layers. Each layer of data is processed to determine the temporal and spatial distribution of key risk points.

[0056] Specifically, in the scenario of lake ecosystem monitoring, historical data on environmental changes can be analyzed and processed using various tools to obtain valuable information. Firstly, the sliding window tool can be understood as a method of segmenting time-series data to capture short-term fluctuations. When processing lake water temperature data, a window size of 3 months can be set, and the sliding window moves monthly to extract segments of water temperature changes within each 3-month period. For example, if the water temperature rises from 25 degrees Celsius to 28 degrees Celsius within a certain window, showing a short-term warming trend, this segment can be used for subsequent analysis. Secondly, time-series comparison tools can be seen as a means of analyzing the correlation between short-term fluctuations and long-term trends. Specifically, assuming the long-term trend of lake water temperature is an average annual increase of 0.5 degrees Celsius, and a 3-month segment extracted through the sliding window shows a temperature increase of 3 degrees Celsius, significantly deviating from the long-term trend, the comparison tool can identify this as an abnormal fluctuation, thus revealing the distribution characteristics of the short-term fluctuations, such as the concentration of fluctuations in summer months.

[0057] Furthermore, the latest key monitoring areas include:

[0058] Based on the temporal and spatial distribution of the key risk points, the distribution characteristics are stratified to obtain a preliminary priority distribution map;

[0059] For the preliminary priority distribution map, the latest environmental change records in the real-time data stream are overlaid. If the latest environmental change records exceed a preset threshold in some areas, the priority of certain areas is dynamically adjusted to determine the updated key monitoring range.

[0060] Based on the updated key monitoring scope, the frequency of periodic adjustments is optimized, the latest time distribution data is obtained from the real-time data stream, and the appropriate update cycle is determined.

[0061] Based on the adapted update cycle, the priority distribution map of dynamic monitoring is continuously updated to obtain the latest regional division information and the latest key monitoring areas.

[0062] Specifically, in the scenario of lake ecosystem monitoring, the application of geographic information mapping (GIS) tools can help to stratify complex distribution characteristics by analyzing the temporal and spatial distribution data of key risk points. The core of GIS tools lies in combining environmental data with geographic location information to form an intuitive and visual distribution map. Assuming a lake is divided into three regions—north, east, and south—this tool can mark risk points such as water quality deterioration or abnormal water temperature on the map. It reveals that the northern region has a high density of risk points, involving water temperature increases of more than 3 degrees Celsius, while the southern region has fewer risk points, with only slight fluctuations. This stratification helps to initially prioritize regions, forming a priority distribution map, with the northern region marked as high priority. For dynamic adjustments to the initial priority distribution map, data integration tools can overlay the latest environmental change records from real-time data streams onto the existing distribution map. For example, if the real-time data stream shows that the water temperature in the northern region has recently increased by 2 degrees Celsius, exceeding the preset threshold of 1.5 degrees Celsius, the tool will automatically increase the priority weight of this region, moving it from high priority to emergency monitoring, while the eastern region, due to stable data, maintains its original priority. This dynamic adjustment ensures the rational allocation of monitoring resources and timely response to sudden changes.

[0063] Furthermore, the predicted results of lake ecological trends include:

[0064] Based on the latest key monitoring areas, real-time environmental data of the areas are extracted, and the real-time environmental data is denoised to obtain a sorted set of environmental data.

[0065] The organized environmental data set is classified according to time and space dimensions. If the data fluctuation exceeds a preset threshold within a certain time period, the time period is marked to determine the key time period range.

[0066] Based on the key time period, corresponding spatial distribution information is extracted from the environmental data set, and the spatial distribution information is visualized to obtain the preliminary impact range of potential environmental changes. Local feature comparison is performed on the historical and real-time data of the preliminary impact range to obtain the lake ecological trend prediction results.

[0067] Specifically, in the scenario of lake ecosystem monitoring, real-time environmental data can be extracted from a multi-source data integration platform for the latest key monitoring areas. Assuming the northern region is identified as a key monitoring area, the platform will collect data on water quality, temperature, dissolved oxygen, etc., for this area. This data may come from different sensors or monitoring stations, and may contain noise interference. Data cleaning tools can remove outliers or duplicate records, such as removing values ​​in water temperature data that significantly deviate from the normal range, ensuring the accuracy of subsequent analysis. For example, for the processed environmental data set, data stratification tools can classify the data according to time and spatial dimensions. Assuming a 24-hour period as a time unit, the tool will analyze the daily water temperature changes in the northern region, while also spatially dividing the lake into different sub-regions. If the water temperature fluctuation exceeds a preset threshold of 2 degrees Celsius on a certain day, that time period will be marked as a key period of concern. This stratified processing facilitates quickly identifying the time window of anomalies, providing a clear direction for subsequent analysis. For example, after determining the range of key periods of concern, the corresponding spatial distribution information can be extracted and visualized using geographic information mapping tools. If, during the marked time period, the water temperature in the northern sub-region near the shore remains consistently higher than normal, the tool will present this distribution characteristic as a heat map, initially delineating the scope of potential environmental changes. This intuitive presentation helps to quickly identify problem areas and improves analysis efficiency.

[0068] like Figure 2 As shown, this embodiment also provides a lake ecological trend prediction system based on big data analysis, including: a data standardization module, a multi-source data integration module, a time series analysis module, an anomaly correction module, a prediction result acquisition module, a hierarchical analysis module, a dynamic monitoring priority generation module, and a local prediction module;

[0069] The data standardization module is used to acquire multi-source monitoring data of the lake, and to perform standardization processing on the multi-source monitoring data to obtain a standardized dataset.

[0070] The multi-source data integration module is used to construct a multi-source data integration framework based on the standardized dataset, and to perform feature extraction and complementary fusion of the spatial resolution of remote sensing images and the time series data of sensors to obtain a fused comprehensive dataset.

[0071] The time series analysis module is used to perform time series analysis on the historical change trend of lake ecology based on the fused comprehensive dataset, and to obtain the fluctuation characteristics of key variables;

[0072] The anomaly correction module is used to perform weighted correction on the abnormal data points in the fluctuation characteristics of the key variables to obtain the corrected dynamic dataset.

[0073] The prediction result acquisition module is used to input the corrected dynamic dataset and environmental variables into the multi-source data integration framework for training to obtain the prediction result.

[0074] The hierarchical analysis module is used to perform hierarchical analysis on the short-term fluctuations and long-term trends of the lake ecosystem based on the environmental change trend data in the prediction results, and to obtain the temporal and spatial distribution of key risk points.

[0075] The dynamic monitoring priority generation module is used to generate a dynamic monitoring priority map based on the time and spatial distribution of key risk points, and to periodically adjust the priority map to obtain the latest key monitoring areas.

[0076] The local prediction module is used to perform local predictions based on the latest key monitoring areas and the real-time data of the corresponding areas extracted from the multi-source data integration framework, so as to obtain the lake ecological trend prediction results.

[0077] This invention discloses a method and system for predicting lake ecological trends based on big data analysis. It addresses the problem of insufficient data coverage by integrating and standardizing multi-source data, utilizes a weighted average algorithm for data fusion, employs a long short-term memory network model to analyze ecosystem change trends and identify abnormal patterns, constructs an adaptive prediction tool to dynamically respond to environmental changes, and combines sliding window technology to determine the distribution of key risk points. A dynamic monitoring priority map is generated and updated in real time, ultimately achieving accurate prediction and monitoring of potential environmental challenges to lake ecosystems. This invention effectively integrates multi-source heterogeneous data, improves the spatiotemporal coverage and prediction accuracy of lake ecological monitoring, and provides scientific decision support for lake ecological protection.

[0078] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for predicting lake ecological trend based on big data analysis, characterized in that, The method comprises the following steps: obtaining multi-source monitoring data of a lake, and performing standardization processing on the multi-source monitoring data to obtain a standardized data set; constructing a multi-source data integration framework according to the standardized data set, performing feature extraction and complementary fusion on spatial resolution of remote sensing images and time series data of sensors to obtain a fused comprehensive data set; performing time series analysis on historical change trends of lake ecology according to the fused comprehensive data set to obtain key variable fluctuation characteristics; performing weighted correction on abnormal data points in the key variable fluctuation characteristics to obtain a corrected dynamic data set; inputting the corrected dynamic data set combined with environmental variables into the multi-source data integration framework for training to obtain a prediction result; performing hierarchical analysis on short-term fluctuations and long-term trends of lake ecology according to environmental change trend data in the prediction result to obtain time distribution and spatial distribution of key risk points; generating a dynamic monitoring priority map according to the time distribution and spatial distribution of the key risk points, periodically adjusting the priority map to obtain the latest monitoring key area; and performing local prediction on real-time data of the corresponding area combined with the multi-source data integration framework to obtain a lake ecology trend prediction result; obtaining the latest monitoring key area comprises: performing hierarchical processing on distribution characteristics according to the time distribution and spatial distribution of the key risk points to obtain a preliminary priority distribution map; superimposing the latest environmental change record in the real-time data stream on the preliminary priority distribution map, and if the latest environmental change record exceeds a preset threshold in some areas, dynamically adjusting the priority of the areas to determine an updated key monitoring range; optimizing the frequency of periodic adjustment according to the updated key monitoring range, obtaining the latest time distribution data from the real-time data stream, and determining an adaptive update period; performing continuous updating on the priority distribution map of dynamic monitoring according to the adaptive update period, obtaining the latest regional division information, and obtaining the latest monitoring key area; obtaining the lake ecology trend prediction result comprises: extracting real-time environmental data of the area according to the latest monitoring key area, performing denoising processing on the real-time environmental data to obtain a sorted environmental data set; classifying the sorted environmental data set according to time dimension and space dimension, marking a time period if the fluctuation of data in the time period exceeds a preset threshold to determine a key attention time period range; extracting corresponding spatial distribution information from the environmental data set according to the key attention time period range, performing visual processing on the spatial distribution information to obtain a preliminary impact range of potential environmental changes; and performing local feature comparison on historical data and real-time data of the preliminary impact range to obtain a lake ecology trend prediction result. 2.The lake ecological trend prediction method based on big data analysis of claim 1, wherein, obtaining the standardized data set comprises: performing format conversion on the original data according to a pre-established data standardization protocol to obtain an intermediate data set in a unified format according to the source differences of the multi-source monitoring data of the lake; Calibrate the timestamps of the different source data in the intermediate data set, and time-align the calibrated data set by a linear interpolation method to obtain a standardized data set. 3.The lake ecological trend prediction method based on big data analysis of claim 1, wherein, Obtaining the integrated data set after fusion includes: Layered processing of remote sensing images and sensor data in the standardized data set, extracting spatial resolution and time series features for different sources of data to obtain a layered feature set; and weighted processing of the spatial resolution features of remote sensing images and the time series features of sensor data in the layered feature set to obtain a weighted feature combination; Based on the weighted feature combination, complementary fusion of features from different sources is performed, and data barriers that occur during fusion are corrected using pre-established rules to determine an intermediate fusion data set; and the integrated data set is optimized by adjusting fusion parameters to obtain an integrated data set after fusion. 4.The lake ecological trend prediction method based on big data analysis of claim 1, wherein, Obtaining key variable fluctuation features includes: Obtaining time series information of water temperature, rainfall data and evaporation rate from the integrated data set, segmenting the time series information to obtain a segmented environmental variable set; cross- verifying water level changes, basin flow and climate fluctuations through the segmented environmental variable set, marking variables as abnormal, and determining an abnormal variable combination; For the abnormal variable combination, analyze the change relationship between lake water quality, pollutant concentration and dissolved oxygen content, obtain potential abnormal pattern features from the change relationship, and determine key influencing factors; and jointly process ecological data of vegetation coverage, algal density and fish population according to the key influencing factors to obtain key variable fluctuation features. 5.The lake ecological trend prediction method based on big data analysis of claim 1, wherein, Obtaining the modified dynamic data set includes: Preliminary screening of abnormal data points in the integrated data set, obtaining data records exceeding a preset threshold from environmental variables and water body parameters, cross-checking fluctuation features and climate influences, and obtaining a distribution set of abnormal data points; According to the distribution set of abnormal data points, the numerical value of the abnormal data points is adjusted, and the correlation between ecological indicators and climate influences is combined through data integration to obtain a modified dynamic data set. 6.The lake ecological trend prediction method based on big data analysis of claim 1, wherein, Obtaining the prediction result includes: According to the modified dynamic data set, classify and arrange the related records of environmental variables and ecological changes, obtain a distribution set related to fluctuation features, preliminarily process abnormal points in the distribution set, and obtain an adjusted data combination; According to the adjusted data combination, perform secondary correction on the corrected data points, and determine a data subset by weight distribution; Combining the correlation between environmental variables and ecological changes, the key indicators in the data subset are screened to obtain a feature combination; the feature combination is extracted and analyzed for time series features to obtain pattern information related to dynamic response, and the pattern information is compared with climate influences to obtain a prediction result. 7.The lake ecological trend prediction method based on big data analysis of claim 1, wherein, Determining the time distribution and spatial distribution of the key risk points includes: According to the environmental change trend data in the prediction result, segment processing is performed, data segments of short-term fluctuations are obtained, and the association of the data segments and long-term trends is compared to obtain distribution characteristics of short-term fluctuations; The distribution characteristics of short-term fluctuations are superimposed with geographical position data of the lake ecology to determine the spatial distribution of key risks; According to the spatial distribution of key risks, the records of environmental changes are disassembled to obtain time distribution information related to risk points, and classification is performed through a fluctuation feature comparison tool to determine the time distribution law of the risk points; For the time distribution law of the risk points, the data of short-term fluctuations and long-term trends are stored in layers, and each layer of data is processed to determine the time distribution and spatial distribution of key risk points.

8. A system for predicting lake ecological trends based on big data analysis according to any one of claims 1-7, characterized in that, It comprises: a data standardization module, a multi-source data integration module, a time series analysis module, an anomaly correction module, a prediction result acquisition module, a hierarchical analysis module, a dynamic monitoring priority generation module, and a local prediction module; The data standardization module is used to obtain multi-source monitoring data of a lake and perform standardization processing on the multi-source monitoring data to obtain a standardized data set; The multi-source data integration module is used to construct a multi-source data integration framework according to the standardized data set, perform feature extraction and complementary fusion on spatial resolution of remote sensing images and time series data of sensors to obtain a fused comprehensive data set; The time series analysis module is used to perform time series analysis on historical change trends of lake ecology according to the fused comprehensive data set to obtain key variable fluctuation characteristics; The anomaly correction module is used to perform weighted correction on abnormal data points in the key variable fluctuation characteristics to obtain a corrected dynamic data set; The prediction result acquisition module is used to input the corrected dynamic data set combined with environmental variables into the multi-source data integration framework for training to obtain a prediction result; The hierarchical analysis module is used to perform hierarchical analysis on short-term fluctuations and long-term trends of lake ecology according to environmental change trend data in the prediction result to obtain time distribution and spatial distribution of key risk points; The dynamic monitoring priority generation module is used to generate a dynamic monitoring priority map according to the time distribution and spatial distribution of key risk points, periodically adjust the priority map, and obtain the latest monitoring key area; The local prediction module is used to perform local prediction according to the latest monitoring key area combined with real-time data of the corresponding area extracted in the multi-source data integration framework to obtain a lake ecology trend prediction result.

Images