Space capability prediction method based on space-time sequence remote sensing data
Through the spatial capability prediction method of spatiotemporal sequence remote sensing data, combined with time series decomposition, geographic information tools and deep learning models, the problem of insufficient analysis of trend components, seasonal components and random fluctuations in the existing technology is solved, and efficient spatiotemporal sequence data is achieved Forecasting and recognition of surface phenomena trends improve prediction accuracy and data utilization efficiency.
Patent Information
- Application Number
- CN202410933326.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-12
- Publication Date
- 2024-11-08
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing spatiotemporal data prediction methods lack trend components, seasonal components, random fluctuations, geographic information tools and image processing, and cannot effectively model and analyze spatiotemporal sequence remote sensing data, resulting in low prediction accuracy and data utilization efficiency.
Adopting spatial capability prediction method based on spatiotemporal sequence remote sensing data, through data acquisition and preprocessing, spatiotemporal sequence analysis, spatial analysis, prediction model construction and results display and application, including time series decomposition, geographic information tools, image processing, statistical models, and machines Learn models and deep learning models to build a comprehensive prediction system.
Multi-dimensional analysis and prediction of spatiotemporal sequence data is realized, prediction accuracy and data utilization efficiency are improved, surface processes and phenomena are recognized, map and chart results are provided, and decision-making support is facilitated.
Smart Images

Figure CN118916631A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of geographic information systems, and in particular to a spatial capability prediction method based on spatiotemporal sequence remote sensing data. Background Art
[0002] The spatial capability prediction system based on spatiotemporal sequence remote sensing data has broad application potential in many fields such as environmental protection, disaster monitoring, urban planning, agricultural development, etc. For example, by analyzing historical remote sensing image data, the probability of natural disasters such as floods, fires, and land degradation can be predicted; in agriculture, crop yields and the distribution of pests and diseases can be predicted; in urban planning, the changing trends of population growth and traffic flow can be predicted, etc.
[0003] In the prior art, there is a method for predicting spatiotemporal data and a data acquisition monitoring system, the method comprising: establishing a topological space graph based on the collected spatiotemporal data, recording the graph signal sequence of each node of the topological space graph at different times; constructing a spatiotemporal data prediction model based on a spatiotemporal generative adversarial network, using the historical graph signal sequence as a training set to train the spatiotemporal data prediction model; the spatiotemporal generative adversarial network includes a generator and a discriminator, the generator is used to model the spatiotemporal dependency relationship in the spatiotemporal data, and the discriminator is used to regularize the spatiotemporal generative adversarial network; the trained model is used to predict spatiotemporal data to obtain the graph signal sequence of each node at a preset time step in the future. The present invention introduces adversarial loss into the objective function of the prediction model to model the uncertainty in the data, and learns the real spatiotemporal data distribution through the adversarial process, strengthens the ability of the prediction model to learn data representation, and thus alleviates the problem of too fast growth of multi-step prediction errors.
[0004] However, the existing spatiotemporal data prediction can only be modeled and analyzed through spatiotemporal data prediction methods, lacking trend components, seasonal components, random fluctuations, geographic information tools and image processing. Therefore, a spatial capability prediction method based on spatiotemporal series remote sensing data is needed.
[0005] Spatial analysis Summary of the invention
[0006] The present invention provides a spatial capability prediction method based on spatiotemporal sequence remote sensing data to solve the problems in the background technology.
[0007] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a method for predicting spatial capability of remote sensing data based on spatiotemporal sequence, including a spatial capability prediction system based on spatiotemporal sequence remote sensing data, the spatial capability prediction system based on spatiotemporal sequence remote sensing data including data acquisition and preprocessing, spatiotemporal sequence analysis, spatial analysis, prediction model construction, result display and application, and system maintenance and update, the data acquisition and preprocessing internally include remote sensing data acquisition, data feature extraction, data cleaning, and data assimilation, the spatiotemporal sequence analysis includes time series decomposition, time series feature extraction, and trend analysis, the spatial analysis includes geographic information tools and image processing, the prediction model construction internally includes statistical models, machine learning models, and deep learning models, the result display and application internally include map boxes, charts, and decision support systems, and the system maintenance and update internally include database management and system optimization.
[0008] Furthermore, the remote sensing data collection, data cleaning, data assimilation and data feature extraction, remote sensing data collection: acquiring multispectral images, high-resolution images, radar data, etc. of the earth's surface through remote sensing platforms such as satellites, drones, and airplanes; data cleaning: removing noise points, missing values, outliers, etc.; data assimilation: calibrating and fusing data from different times and different sensors to ensure data consistency and comparability; data feature extraction: extracting useful features from preprocessed data, such as texture, shape, size, position, etc., for subsequent analysis and prediction.
[0009] Furthermore, the time series decomposition, time series feature extraction and trend analysis, time series decomposition: decomposing the time series data into trend components, seasonal components and random fluctuation components; time series feature extraction: calculating various statistical indicators, such as mean, standard deviation, peak value, valley value, etc.; trend analysis: identifying the upward, downward or periodic trends of surface processes and phenomena.
[0010] Furthermore, the geographic information tools and image processing, geographic information system tools: use geographic information system software to perform spatial data analysis, including buffer analysis, overlay analysis, network analysis, etc.; image processing: apply image processing techniques, such as filtering, classification, object segmentation, etc., to extract surface coverage information and features.
[0011] Furthermore, the statistical models, machine learning models and deep learning models include: statistical models: such as autoregression, moving average, autoregressive moving average, autoregressive integrated moving average, etc. in time series analysis; machine learning models: including decision trees, random forests, support vector machines, neural networks, etc.; deep learning models: such as convolutional neural networks, recurrent neural networks, long short-term memory networks, etc.
[0012] Furthermore, the maps and charts and decision support system, maps and charts: visualize the prediction results in the form of maps, charts, etc., which are easy for users to understand and interpret. Decision support system: integrate the prediction results into the decision support system to provide information support and prediction services for governments, enterprises and the public.
[0013] Furthermore, the database management and system optimization include: database management: regularly updating remote sensing data and prediction models to reflect the latest surface processes and phenomena; system optimization: continuously optimizing system performance and prediction accuracy based on user feedback and technological advances.
[0014] Compared with the prior art, the present invention provides a spatial capacity prediction method based on spatiotemporal sequence remote sensing data, which has the following beneficial effects:
[0015] 1. The spatial capability prediction system based on spatiotemporal sequence remote sensing data sets up spatiotemporal sequence analysis. The time series decomposition in the spatiotemporal sequence analysis: decomposing the time series data into trend components, seasonal components and random fluctuation components; time series feature extraction: calculating various statistical indicators, such as mean, standard deviation, peak value, valley value, etc.; trend analysis: identifying the rising, falling or periodic trends of surface processes and phenomena.
[0016] 2. The spatial capability prediction system based on spatiotemporal sequence remote sensing data sets up spatial analysis, geographic information tools in spatial analysis, and image processing. Geographic Information System Tools: Use geographic information system software to perform spatial data analysis, including buffer analysis, overlay analysis, network analysis, etc.; Image processing: Apply image processing techniques, such as filtering, classification, object segmentation, etc., to extract surface coverage information and features.
[0017] 3. The spatial capability prediction system based on spatiotemporal series remote sensing data is constructed by setting up prediction models, including statistical models, machine learning models and deep learning models in prediction model construction. Statistical models: such as autoregression, moving average, autoregressive moving average, autoregressive integrated moving average, etc. in time series analysis; machine learning models: including decision trees, random forests, support vector machines, neural networks, etc.; deep learning models: such as convolutional neural networks, recurrent neural networks, long short-term memory networks, etc. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] Figure 1 It is a schematic diagram of the overall structure of the spatial capability prediction system based on spatiotemporal sequence remote sensing data of the present invention.
[0019] In the figure: 1. Spatial capability prediction system based on spatiotemporal sequence remote sensing data; 2. Data collection and preprocessing; 3. Spatiotemporal sequence analysis; 4. Spatial analysis; 5. Prediction model construction; 6. Result display and application; 7. System maintenance and update; 201. Remote sensing data collection; 202. Data feature extraction; 203. Data cleaning; 204. Data assimilation; 301. Time series decomposition; 302. Time series feature extraction; 303. Trend analysis; 401. Geographic information tools; 402. Image processing; 501. Statistical model; 502. Machine learning model; 503. Deep learning model; 601. Maps and charts; 602. Decision support system; 701. Database management; 702. System optimization. DETAILED DESCRIPTION
[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0021] See also Figure 1 The present invention discloses a spatial capability prediction method based on spatiotemporal sequence remote sensing data, including a spatial capability prediction system 1 based on spatiotemporal sequence remote sensing data, the spatial capability prediction system 1 based on spatiotemporal sequence remote sensing data includes data acquisition and preprocessing 2, spatiotemporal sequence analysis 3, spatial analysis 4, prediction model construction 5, result display and application 6 and system maintenance and update 7, the data acquisition and preprocessing 2 internally includes remote sensing data acquisition 201, data feature extraction 202, data cleaning 203 and data assimilation 204, the spatiotemporal sequence analysis 3 includes time series decomposition 301, time series feature extraction 302 and trend analysis 303, the spatial analysis 4 includes geographic information tools 401 and image processing 402, the prediction model construction 5 internally includes a statistical model 501, a machine learning model 502 and a deep learning model 503, the result display and application 6 internally includes a map box chart and a decision support system 602, and the system maintenance and update 7 internally includes database management 701 and system optimization 702.
[0022] Specifically, remote sensing data collection 201, data cleaning 203, data assimilation 204 and data feature extraction 202, remote sensing data collection 201: obtaining multi-spectral images, high-resolution images, radar data, etc. of the earth's surface through remote sensing platforms such as satellites, UAVs, and airplanes; data cleaning 203: removing noise points, missing values, outliers, etc.; data assimilation 204: calibrating and fusing data from different times and different sensors to ensure data consistency and comparability; data feature extraction 202: extracting useful features from the pre-processed data, such as texture, shape, size, position, etc., for subsequent analysis and prediction.
[0023] In this embodiment, remote sensing data collection 201 in data collection and preprocessing 2: This stage involves collecting remote sensing data from different sensors and platforms (such as satellites, drones, aircraft, etc.), which can be multispectral images, radar data, altimeter data or other types of spatial data. Selecting appropriate sensors and acquisition parameters is crucial for subsequent analysis; data feature extraction 202: Once the raw data is collected, the next step is to characterize the data in order to extract the most valuable information for the prediction task. This may include extracting spectral features, texture features, shape features, etc. from images, or identifying patterns and trends from time series data; data cleaning 203: The goal of this stage is to improve the quality of the data, including processing missing values, correcting errors, eliminating noise, and removing outliers. Data cleaning is a key step to ensure the accuracy and consistency of analysis results; data assimilation 204: In some cases, especially for applications that require real-time or near-real-time monitoring, data assimilation technology can be used to fuse data from different sources and time points. This helps to provide more comprehensive and accurate prediction results.
[0024] Specifically, time series decomposition 301, time series feature extraction 302 and trend analysis 303, time series decomposition 301: decomposing time series data into trend components, seasonal components and random fluctuation components; time series feature extraction 302: calculating various statistical indicators, such as mean, standard deviation, peak value, valley value, etc.; trend analysis 303: identifying the upward, downward or periodic trend of surface processes and phenomena.
[0025] In this embodiment, the spatiotemporal series analysis 3 includes time series decomposition 301: time series decomposition 301 is to decompose a complex time series data into its components, such as trends, seasonality and random fluctuations; this process can help understand the main dynamics of the data and provide a basis for further analysis. For example, the time series is decomposed using an additive model or a multiplicative model; time series feature extraction 302: feature extraction is to identify and extract useful information from time series data, which helps to describe the key attributes and behaviors of the data; this includes cycle length, peaks, valleys, slopes, curvatures, entropy and other statistical indicators; the selection of features directly affects the performance of the prediction model; trend analysis 303: trend analysis 303 aims to identify the trend of time series data over time; this is an upward, downward or stable trend, or it can be a more complex change pattern, such as an S-shaped curve or a periodic rise / fall; understanding trends is crucial to predicting future events or assessing the impact of past events.
[0026] Specifically, geographic information tools 401 and image processing 402, geographic information system tools: use geographic information system software to perform spatial data analysis, including buffer analysis, overlay analysis, network analysis, etc.; image processing 402: apply image processing 402 techniques, such as filtering, classification, object segmentation, etc., to extract surface cover information and features.
[0027] In this implementation, spatial analysis 4: geographic information tools 401: use GIS technology to perform spatial analysis on data 4; image processing 402: apply image processing 402 technology to enhance and analyze remote sensing images; predictive model construction 5.
[0028] Specifically, statistical model 501, machine learning model 502 and deep learning model 503, statistical model 501: such as autoregression, moving average, autoregressive moving average, autoregressive integrated moving average, etc. in time series analysis; machine learning model: including decision tree, random forest, support vector machine, neural network, etc.; deep learning model 503: such as convolutional neural network, recurrent neural network, long short-term memory network, etc.
[0029] In this embodiment, the statistical model 501: relies on mathematical statistical methods to establish the relationship between input data and output predictions. Some common statistical models include linear regression, multivariate regression, time series analysis (such as ARIMA model), etc. These models are usually easy to understand and implement, but may be limited when dealing with complex nonlinear problems; machine learning model 502: compared with statistical models, machine learning models focus more on automatically learning and extracting features from data rather than relying on manually defined features. This includes algorithms such as decision trees, random forests, support vector machines (SVM), k-nearest neighbors (k-NN), etc. Machine learning models are able to handle complex data relationships to a certain extent, but may not be suitable for tasks that require deep feature extraction; deep learning model 503: deep learning is a special type of machine learning method that relies on deep neural network structures to simulate the way the human brain processes information. Through deep structures (such as convolutional neural networks CNN, recursive neural networks RNN, long short-term memory networks LSTM, etc.), deep learning models can automatically identify and learn complex features and patterns in remote sensing data. Deep learning has achieved remarkable results in image and video recognition, natural language processing, etc. In many cases, deep learning models are able to surpass traditional machine learning methods.
[0030] Specifically, the map box chart and decision support system 602, map and chart 601: visualize the prediction results in the form of maps, charts, etc., which are convenient for users to understand and interpret. Decision support system 602: integrate the prediction results into the decision support system 602 to provide information support and prediction services for governments, enterprises and the public.
[0031] In this embodiment, maps and charts 601: visualize the results as maps and charts 601 so that users can intuitively understand the data and prediction results; decision support system 602: provide real-time and historical data analysis to help decision makers make data-based decisions.
[0032] Specifically, database management 701 and system optimization 702, database management 701: regularly update remote sensing data and prediction models to reflect the latest surface processes and phenomena; system optimization 702: continuously optimize system performance and prediction accuracy based on user feedback and technological progress.
[0033] In this implementation plan, database management: Since remote sensing data is usually large and constantly updated, an effective database management system is required to store, organize and retrieve this data. Database management also includes data cleaning, calibration and fusion to ensure data quality and consistency; system optimization: over time, the system may need to be optimized due to hardware performance degradation, software bugs or the development of new algorithms and technologies. System optimization may include adjustments to algorithms, hardware upgrades or replacements, and reconfiguration of the entire system to improve efficiency; update and expand functions: in order to respond to new needs or improve existing functions, the system may need to be updated and expanded regularly. This may involve the integration of new algorithms, improvements to the user interface, or the addition of new data analysis and visualization tools; security and privacy protection: as the amount of data grows and security threats evolve, it becomes increasingly important to ensure data security and user privacy protection. System maintenance and updates should include measures such as fixing security vulnerabilities, strengthening access control and data encryption; user support and training: in order to ensure that users can use the system effectively, it is necessary to provide timely technical support and user training. This can help users solve problems encountered during operation and allow them to understand new features and best practices.
[0034] When in use, the spatial capability prediction system based on spatiotemporal remote sensing data 1 is a comprehensive data analysis and prediction platform that uses time series information from remote sensing data to monitor and predict surface processes and phenomena; the main components of the system include: Data acquisition and preprocessing 2: Remote sensing data acquisition 201: Collect data on the earth's surface through remote sensing platforms such as satellites or drones; Data cleaning 203: Remove noise and outliers to ensure data quality; Data assimilation 204: Combine data from different sources to create a more comprehensive data set, Data feature extraction 202: Extract useful features from raw data for subsequent analysis; Spatiotemporal series analysis 3: Time series decomposition 301: Decompose complex time series data into components such as trends, seasonality, and residuals; Time series feature extraction 302: Identify and extract key features in time series; Trend analysis 303: Analyze the trend of data changes over time; Spatial analysis 4: Geographic information tools 401: Use GIS technology to perform spatial analysis on data 4; Image processing 402: Apply image processing 402 technology to enhance and analyze remote sensing images; Prediction model construction 5: Statistical model 501: Such as ARI MA, used for predictions based on statistical data; machine learning models 502: such as random forests or support vector machines, applied to data-driven predictions; deep learning models 503: such as convolutional neural networks or recurrent neural networks, used to learn and predict from large amounts of complex data; result presentation and application 6: maps and charts 601: visualize the results as maps and charts 601 so that users can intuitively understand the data and prediction results; decision support system 602: provide real-time and historical data analysis to help decision makers make data-based decisions; system maintenance and update 7: database management 701: ensure effective storage and retrieval of data; system optimization 702: regularly update the system to improve performance and accuracy; the advantages of the system include: multi-scale monitoring: capable of monitoring surface processes from local to global scales; high-frequency updates: providing high-frequency monitoring services through continuous data collection and updates; multi-parameter analysis: multiple parameters and variables can be processed and analyzed simultaneously.
[0035] In summary, the spatial capability prediction system based on spatiotemporal sequence remote sensing data, by setting up spatiotemporal sequence analysis 3, time series decomposition 301 in spatiotemporal sequence analysis 3: decomposing time series data into trend components, seasonal components and random fluctuation components; time series feature extraction 302: calculating various statistical indicators, such as mean, standard deviation, peak, valley, etc.; trend analysis 303: identifying the rising, falling or periodic trend of surface processes and phenomena; by setting up spatial analysis 4, geographic information tools 401 and image processing 402 in spatial analysis 4, geographic information system tools: using geographic information system software to perform spatial data analysis, including buffer zone analysis analysis, overlay analysis, network analysis, etc.; image processing 402: applying image processing 402 technology, such as filtering, classification, object segmentation, etc., to extract surface cover information and features; by setting up prediction model construction 5, the statistical model 501, machine learning model 502 and deep learning model 503 in the prediction model construction 5, statistical model 501: such as autoregression, moving average, autoregressive moving average, autoregressive integrated moving average, etc. in time series analysis; machine learning model: including decision tree, random forest, support vector machine, neural network, etc.; deep learning model 503: such as convolutional neural network, recurrent neural network, long short-term memory network, etc.
[0036] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. A method for predicting spatial capability based on spatiotemporal sequence remote sensing data, comprising a system for predicting spatial capability based on spatiotemporal sequence remote sensing data (1), characterized in that: The spatial capability prediction system (1) based on spatiotemporal sequence remote sensing data includes data collection and preprocessing (2), spatiotemporal sequence analysis (3), spatial analysis (4), prediction model construction (5), result display and application (6) and system maintenance and update (7). The data collection and preprocessing (2) includes remote sensing data collection (201), data feature extraction (202), data cleaning (203) and data assimilation (204). The spatiotemporal sequence analysis (3) includes time series decomposition (301), time series feature extraction (302) and trend analysis (303). The spatial analysis (4) includes geographic information tools (401) and image processing (402). The prediction model construction (5) includes statistical models (501), machine learning models (502) and deep learning models (503). The result display and application (6) includes map box charts and decision support systems (602). The system maintenance and update (7) includes database management (701) and system optimization (702).
2. The method for predicting spatial capability based on spatiotemporal sequence remote sensing data according to claim 1, characterized in that: The remote sensing data acquisition (201), data cleaning (203), data assimilation (204) and data feature extraction (202), remote sensing data acquisition (201): acquiring multi-spectral images, high-resolution images, radar data, etc. of the earth's surface through remote sensing platforms such as satellites, unmanned aerial vehicles, and airplanes; data cleaning (203): removing noise points, missing values, outliers, etc.; data assimilation (204): calibrating and fusing data from different times and different sensors to ensure data consistency and comparability; data feature extraction (202): extracting useful features such as texture, shape, size, position, etc. from the pre-processed data for subsequent analysis and prediction.
3. The method for predicting spatial capability based on spatiotemporal sequence remote sensing data according to claim 1, characterized in that: The time series decomposition (301), time series feature extraction (302) and trend analysis (303), time series decomposition (301): decomposing time series data into trend components, seasonal components and random fluctuation components; time series feature extraction (302): calculating various statistical indicators, such as mean value, standard deviation, peak value, valley value, etc.; trend analysis (303): identifying the upward, downward or periodic trend of surface processes and phenomena.
4. The method for predicting spatial capability based on spatiotemporal sequence remote sensing data according to claim 1, characterized in that: The geographic information tool (401) and image processing (402), geographic information system tool: using geographic information system software to perform spatial data analysis, including buffer analysis, overlay analysis, network analysis, etc.; image processing (402): applying image processing (402) technology, such as filtering, classification, object segmentation, etc., to extract surface coverage information and features.
5. The method for predicting spatial capability based on spatiotemporal sequence remote sensing data according to claim 1, characterized in that: The statistical model (501), machine learning model (502) and deep learning model (503), statistical model (501): such as autoregression, moving average, autoregressive moving average, autoregressive integrated moving average, etc. in time series analysis; machine learning model: including decision tree, random forest, support vector machine, neural network, etc.; deep learning model (503): such as convolutional neural network, recurrent neural network, long short-term memory network, etc.
6. The method for predicting spatial capability based on spatiotemporal sequence remote sensing data according to claim 1, characterized in that: The map box chart and decision support system (602) are as follows: Maps and charts (601): The prediction results are visualized in the form of maps, charts, etc., which are convenient for users to understand and interpret. Decision support system (602): The prediction results are integrated into the decision support system (602) to provide information support and prediction services for governments, enterprises and the public.
7. The method for predicting spatial capability based on spatiotemporal sequence remote sensing data according to claim 1, characterized in that: The database management (701) and system optimization (702) include: database management (701): regularly updating remote sensing data and prediction models to reflect the latest surface processes and phenomena; system optimization (702): continuously optimizing system performance and prediction accuracy based on user feedback and technological progress.
Citation Information
Cited By
Data analysis system based on integrated remote sensing data acquisition
CN119516404A
