Data analysis system based on integrated remote sensing data acquisition

Abstract

The invention discloses a data analysis system based on integrated remote sensing data acquisition. Comprising a data analysis system, and the data analysis system comprises a remote sensing data acquisition module, a data preprocessing module, a data integration and fusion module, a data feature extraction module, a data time sequence analysis module and a data visualization module. The data analysis system is used for quickly processing the integrated mass remote sensing data by adopting a distributed calculation and parallel and streaming processing mode; according to the method, distributed computing and parallel and streaming processing modes are adopted, a computing task is decomposed into a plurality of parallel tasks, and the computing efficiency is improved; when the remote sensing data is preprocessed, the accuracy of the remote sensing data can be kept, and the calculation efficiency is improved; the multi-source remote sensing image data can form clear image information through fusion processing; and the data time sequence analysis module carries out prediction according to the collected feature information and the previous historical feature information, so that the pasturing duration can be accurately controlled.

Description

Technical Field

[0001] The present invention belongs to the technical field of remote sensing data analysis, and more specifically, particularly relates to a data analysis system based on integrated remote sensing data acquisition. Background Art

[0002] The grazing system is an organizational system for grazing management. It follows the laws of energy flow and material circulation in the grassland ecosystem. On the basis of grassland fence construction, livestock drinking water facilities, vegetation composition regulation and grassland improvement, it adjusts the grazing intensity and frequency in time and space to achieve a balance between the growth of forage and the nutritional needs of livestock in terms of quantity and quality. The original purpose of the grazing system was to consider the restoration of degraded grasslands through the rational use of seasonal grazing and leisure. A large number of scientific studies have also shown that a reasonable grazing system can restore grassland vitality, improve grassland production efficiency, maintain grassland ecological balance, and enable the sustainable use of grassland.

[0003] Zoning rotational grazing is a grassland utilization method that usually divides grazing land by building permanent or temporary fences, and strictly controls the feeding time and feeding range of livestock in the grazing land according to a certain grazing plan. Traditional zoning rotational grazing mainly includes general block rotational grazing (a single herd of livestock grazes in a certain order on the demarcated plots of land), alternating grazing of different livestock (different types of herds use the same grassland at the same time or at different times), warm season camp grazing (zoning rotational grazing with strict order in the grazing area during the growing season of forage), daily grazing (using easily movable point fences or other movable fences to control livestock in a smaller range and concentrate on using forage), etc.

[0004] In order to monitor the growth of pasture grass and predict the growth of pasture grass, remote sensing technology is used to obtain pasture data. By monitoring and predicting the growth of pasture grass through the pasture data information collected by remote sensing, it is possible to effectively and accurately control the grazing time, give the pasture a certain amount of recovery time, improve the production efficiency of the pasture, maintain the ecological balance of the grassland, and enable the sustainable use of the grassland.

[0005] However, there are some problems with the existing technology: there are many sources of existing remote sensing data, and the processing efficiency is slow when processing a large amount of remote sensing image data, which puts a lot of computing pressure on the computer and cannot improve the computing efficiency. In addition, the processing of remote sensing image data from multiple sources is relatively simple, which makes it impossible to accurately integrate remote sensing image data, easily causing data loss, and making it inconvenient to accurately predict pastures. In addition, the prediction of pastures cannot be combined with previous data, and it is not convenient to view previous data. Therefore, we propose a data analysis system based on integrated remote sensing data acquisition. Summary of the invention

[0006] In view of the problems existing in the prior art, the purpose of the present invention is to provide a data analysis system based on integrated remote sensing data acquisition, which improves the computing and processing efficiency of integrated remote sensing image data, improves the fusion of multi-source remote sensing image data, and enables clear observation and prediction of pastures, thereby facilitating dynamic display of pasture conditions, by performing distributed computing, parallel and streaming processing on remote sensing image data, and finely processing and fusing multi-source remote sensing image data, as well as monitoring and predicting pastures through time series.

[0007] To achieve the above object, the present invention provides the following technical solutions: a data analysis system based on integrated remote sensing data acquisition, comprising a data analysis system, wherein the data analysis system comprises a remote sensing data acquisition module, a data preprocessing module, a data integration fusion module, a data feature extraction module, a data time series analysis module and a data visualization module; The data analysis system adopts distributed computing and parallel and stream processing to improve the rapid processing of integrated massive remote sensing data; The distributed computing distributes data and computing tasks to multiple computers through Apache Spark for collaborative processing; The parallel and streaming processing simultaneously processes multiple data streams or tasks through Apache Kafka to improve processing efficiency; decomposes the computing task into multiple parallel tasks to improve computing efficiency; The remote sensing data acquisition module is used to receive and collect remote sensing data monitored by a remote sensing platform, wherein the remote sensing platform includes satellites, aircraft and drones, and the remote sensing platform monitors and collects data through sensors carried thereon, and the sensors include optical sensors, radar sensors and laser radars; Optical sensors capture information in the visible and near-infrared spectrum; Radar sensors obtain surface information by actively emitting electromagnetic waves and receiving reflected signals; LiDAR obtains high-precision terrain data and three-dimensional information by scanning the surface with lasers; The data preprocessing module preprocesses the collected remote sensing data, and the preprocessing includes geometric correction, atmospheric correction, radiation correction, noise removal, image enhancement and image fusion; Geometric correction is the process of converting pixel coordinates in remote sensing images into actual geographic coordinates on the surface; Atmospheric correction is to adjust the image radiation value caused by atmospheric scattering and absorption effects in remote sensing images in order to extract real surface information from the images; Radiometric correctionRadiometric correction adjusts the raw radiometric values ​​of remote sensing data to eliminate errors introduced by the sensor and the data collection process; Noise removal can improve the data quality by removing noise or pseudo-images in remote sensing data; Image enhancement helps to better extract information by improving the visualization of images; Image fusion combines image data from different sensors or at different times to obtain richer information; The data integration and fusion module is used to integrate data from different sensors and remote sensing platforms to ensure the consistency of data from different sources in space and time; The data feature extraction module is used to extract features from the fused remote sensing data, so as to facilitate subsequent analysis and processing based on the extracted features; The data time series analysis module applies statistical models to model and predict time series data, combines remote sensing data acquired at different times, and performs time series analysis and change monitoring; The data visualization module is used to display the prediction and monitoring data of the data time series analysis module in a visualized form to facilitate understanding and analysis; it includes integrated maps and dynamic displays. The integrated map overlays the analysis results of different data sets on the map for spatial visualization; the dynamic display displays surface changes and dynamic processes through time series animation.

[0008] Optionally, the noise removal is performed by calculating and processing the remote sensing data through Gaussian filtering, and the calculation formula of Gaussian filtering is as follows: , (1) In formula (1), It's on point The Gaussian function value of is the standard deviation of the Gaussian function, controlling the strength of the smoothing; X and Y are the coordinates relative to the center of the Gaussian function; Standard Deviation The calculation formula is as follows: , (2) In formula (2), is the value of each remote sensing data point, and the value is The two-dimensional value of is the mean of the remote sensing data points, V is the mean error of the remote sensing data points, and N is the number of remote sensing data points; The calculation formula of the mean error V is as follows: , (3) In formula (3), is the actual calculated mean, is the predicted mean.

[0009] Optionally, the separation of the Gaussian filter separates the two-dimensional Gaussian function into a one-dimensional Gaussian function to improve the efficiency of calculation. The separation formula of the Gaussian filter is as follows: , (4) In formula (4), is a one-dimensional Gaussian function on the X-axis, is a one-dimensional Gaussian function on the Y axis; Then the one-dimensional Gaussian function on the X-axis is as follows: ; (5) The one-dimensional Gaussian function on the Y axis is as follows: ; (6) In formulas (5) and (6), x and y are the Gaussian function values, respectively. The values ​​on the X-axis and Y-axis; Then convolve the one-dimensional Gaussian function on the X-axis and Y-axis. The processing process is as follows: ; (7) ; (8) In formulas (7) and (8), and They are the convolution results of remote sensing data points on the X-axis and Y-axis respectively. and are the values ​​of the remote sensing data points on the X-axis and Y-axis respectively, and are the convolution kernels on the X-axis and Y-axis respectively, and L is the length of the convolution kernel.

[0010] Optionally, the data integration fusion module performs fusion processing on remote sensing images collected by different remote sensing platforms or different sensors, and the calculation process of the fusion processing is as follows: , (9) In formula (9), is the pixel value of the fused image, , are the pixel values ​​of the original images to be fused, and w is the weight coefficient, which takes a value between 0 and 1.

[0011] Optionally, the data feature extraction module accelerates the construction of the integral graph by a fast algorithm, and the construction process is calculated as follows: ; (10) In formula (10), is the image at position The original pixel value of is in position The integral image value of the image is the sum of all pixel values ​​in the area from the upper left corner (0,0) to (x,y); and the formula quickly calculates the integral image value of the current position by using the calculated integral image value; and the formula , , Represent the integral image values ​​of the left, top, and upper-left corners respectively.

[0012] Optionally, the integral graph detects key points by using the determinant of the Hessian matrix; and the determinant calculation formula of the Hessian matrix is ​​as follows: ; (11) In formula (11), It is the second-order derivative of the image in the X-axis direction, which represents the curvature of the image; It is the second-order derivative of the image in the Y-axis direction, which represents the curvature of the image; It is the mixed second-order derivative of the image in the X-axis and Y-axis directions, describing the interactive influence of the image in the two directions; The Haar wavelet response is used to generate feature descriptors for the local area of ​​each key point; and the calculation formula of the Haar wavelet response is as follows: ; (12) In formula (12), d is the size of the Haar feature, which is used to calculate the response in the local area; is the integral map value, which is used to speed up the calculation of the response of the local area.

[0013] Optionally, the data time series analysis module performs time series analysis on the collected and processed remote sensing images through a SARIMA model, calculates the features of the remote sensing images through the integral graph established by the data feature extraction module, and performs time series analysis on the remote sensing images to improve the prediction of remote sensing impacts; The calculation formula for the SARIMA model to process seasonal time series data is as follows: ; (13) In formula (13), p is the order of the autoregressive term, which indicates how many lagged values ​​are used in the SARIMA model to predict the current value, q is the order of the moving average term, which indicates how many lagged error terms are used in the SARIMA model to predict the current value, P is the order of the seasonal autoregressive term, which indicates how many seasonal lagged values ​​are used in the SARIMA model, Q is the order of the seasonal moving average term, which indicates how many seasonal lagged error terms are used in the SARIMA model, s is the seasonal cycle, that is, the length of a complete season, where s is selected as 12, indicating that there are 12 months in a year, Φ1, Φ2, …, Φp is the coefficient of the seasonal autoregressive term, which represents the relationship between the current value and the previous P seasonal lagged values, Θ1, Θ2, …, Θ q is the coefficient of the seasonal moving average term, which represents the relationship between the current value and the previous Q seasonal lagged error terms.

[0014] Optionally, when calculating, the SARIMA model calculates the time series data by introducing exogenous variables, and the calculation formula is as follows: ; (14) In formula (14), is an exogenous variable, and is the coefficient of the exogenous variable, and the exogenous variable is used as an explanatory variable to eliminate the sudden changes of external factors on remote sensing images.

[0015] Optionally, the specific steps of dynamically updating parameters of the SARIMA model by using the Bayesian method are as follows: Assuming model parameters The prior distribution of ; Given observation data and model parameters , calculate the likelihood function , the likelihood function of the SARIMA model usually involves the calculation of the residual of the time series: , (15) In formula (15), Based on the current parameters The model prediction value of is the variance of the error; Combining the prior distribution and the likelihood function, the posterior distribution is calculated using Bayes' theorem ; , (16) In formula (16), is the marginal likelihood; With new data Arrived, update the posterior distribution of the parameters to the new prior distribution, and repeat the above steps: ,(17) In formula (17), is the posterior distribution based on the first t observations.

[0016] Optionally, the data visualization module displays the changes of the remote sensing image through dynamic display, that is, displays the changes and dynamic processes of the grass in the grassland through time series animation. The dynamic display steps of the data visualization module are as follows: Data loading: Import or connect processed remote sensing image data to ensure the accuracy and completeness of remote sensing image data; Configure display: set chart type, style and dynamic features, and configure real-time updates and interactivity; Realize dynamic functions: configure real-time update and interactive operation of remote sensing image data, and realize real-time update, interactivity and dynamic display through time series data; Run and monitor: Start the display module to monitor the update of remote sensing image data and user interaction in real time; Adjustment and optimization: Adjustments are made based on actual usage to optimize performance and user experience.

[0017] Technical effects and advantages of the present invention: When the present invention is used, in order to improve the rapid and efficient processing of remote sensing data, distributed computing and parallel and stream processing methods are adopted to distribute remote sensing data and computing tasks to multiple computers for collaborative processing, thereby simultaneously processing multiple remote sensing data streams or tasks to improve processing efficiency; the computing task is decomposed into multiple parallel tasks to improve computing efficiency; When preprocessing remote sensing data, denoising is performed through Gaussian filtering. When performing Gaussian filtering, the standard deviation of the Gaussian function is reduced to control the intensity of smoothing. The standard deviation of the Gaussian function is reduced through the mean and error, which can maintain the accuracy of remote sensing data and reduce the smoothing intensity of the denoising process. In order to reduce the pressure of calculation, the two-dimensional function of the remote sensing image is decomposed into a one-dimensional function to improve the efficiency of calculation. For multi-source remote sensing image data, remote sensing image fusion is performed through fusion processing, so that the multi-source remote sensing image data can form a clear image information. During the fusion processing, the maximum and minimum values ​​in the remote sensing image are weighted and processed to achieve fusion processing of multi-source remote sensing image data; After fusing remote sensing image data from multiple sources, feature extraction is performed to obtain pasture grass changes, so that the data time series analysis module can make predictions based on the collected feature information and previous historical feature information. The pasture grass needs to be harvested and stored, and the time series data is calculated through exogenous variables to accurately control the grazing time, give the pasture a certain recovery time, improve the production efficiency of the pasture, maintain the ecological balance of the grassland, and enable the sustainable use of the grassland. The Bayesian method is introduced to dynamically update parameters, so as to realize real-time updating of remote sensing image data in the process of displaying grassland turf changes and dynamics, and facilitate timely observation and prediction.

[0018] Further features and advantages of the present invention will become apparent from the following detailed description of exemplary embodiments of the present invention with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a schematic diagram of a data analysis system for integrated remote sensing data acquisition provided by the present invention. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0021] like Figure 1 As shown, the data analysis system based on integrated remote sensing data acquisition provided by the embodiment of the present invention is characterized in that it includes a data analysis system, and the data analysis system includes a remote sensing data acquisition module, a data preprocessing module, a data integration and fusion module, a data feature extraction module, a data time series analysis module and a data visualization module; The data analysis system uses distributed computing, parallelism and streaming processing to improve the rapid processing of integrated massive remote sensing data; Distributed computing distributes data and computing tasks to multiple computers through Apache Spark for collaborative processing; parallel and streaming processing processes multiple data streams or tasks simultaneously through Apache Kafka to improve processing efficiency; computing tasks are decomposed into multiple parallel tasks to improve computing efficiency; The remote sensing data acquisition module is used to receive and collect remote sensing data monitored by the remote sensing platform. The remote sensing platform includes satellites, aircraft and drones. The remote sensing platform monitors and collects data through the sensors carried by it, and the sensors include optical sensors, radar sensors and laser radars; Optical sensors capture information in the visible and near-infrared spectrum; Radar sensors obtain surface information by actively emitting electromagnetic waves and receiving reflected signals; LiDAR obtains high-precision terrain data and three-dimensional information by scanning the surface with lasers; The data preprocessing module preprocesses the collected remote sensing data, including geometric correction, atmospheric correction, radiation correction, noise removal, image enhancement and image fusion; Geometric correction is the process of converting pixel coordinates in remote sensing images into actual geographic coordinates on the surface; Atmospheric correction is to adjust the image radiation value caused by atmospheric scattering and absorption effects in remote sensing images in order to extract real surface information from the images; Radiometric correctionRadiometric correction adjusts the raw radiometric values ​​of remote sensing data to eliminate errors introduced by the sensor and the data collection process; Noise removal can improve the data quality by removing noise or pseudo-images in remote sensing data; Image enhancement helps to better extract information by improving the visualization of images; The data integration and fusion module is used to integrate data from different sensors and remote sensing platforms to ensure the consistency of data from different sources in space and time; The data feature extraction module is used to extract features from the fused remote sensing data, so as to facilitate subsequent analysis and processing based on the extracted features; Image fusion combines image data from different sensors or at different times to obtain richer information; The data time series analysis module applies statistical models to model and predict time series data, combines remote sensing data acquired at different times, and conducts time series analysis and change monitoring; The data visualization module is used to visualize the prediction and monitoring data of the data time series analysis module for easy understanding and analysis. It includes integrated maps and dynamic displays. The integrated map overlays the analysis results of different data sets on the map for spatial visualization. The dynamic display shows surface changes and dynamic processes through time series animation.

[0022] In this embodiment, preferably, the noise removal is performed by calculating and processing the remote sensing data through Gaussian filtering, and the calculation formula of Gaussian filtering is as follows: , (1) In formula (1), It's on point The Gaussian function value of is the standard deviation of the Gaussian function, controlling the strength of the smoothing; X and Y are the coordinates relative to the center of the Gaussian function; Standard Deviation The calculation formula is as follows: , (2) In formula (2), is the value of each remote sensing data point, and the value is The two-dimensional value of is the mean of the remote sensing data points, V is the mean error of the remote sensing data points, and N is the number of remote sensing data points; The calculation formula of mean error v is as follows: , (3) In formula (3), is the actual calculated mean, is the predicted mean; The collected remote sensing image data is filtered through Gaussian filtering to improve the accuracy of the remote sensing image data. In order to maintain the smoothness of the denoising process, the standard deviation of the Gaussian function is calculated at a low value. The mean and error of the remote sensing image data are calculated to maintain the stability of the standard deviation of the Gaussian function, which facilitates smooth denoising.

[0023] Furthermore, the separation of Gaussian filtering separates the two-dimensional Gaussian function into a one-dimensional Gaussian function to improve the efficiency of calculation. The separation formula of Gaussian filtering is as follows: , (4) In formula (4), is a one-dimensional Gaussian function on the X-axis, is a one-dimensional Gaussian function on the Y axis; Then the one-dimensional Gaussian function on the X-axis is as follows: ; (5) The one-dimensional Gaussian function on the Y axis is as follows: ; (6) In formulas (5) and (6), x and y are the Gaussian function values. The values ​​on the X-axis and Y-axis; Then convolve the one-dimensional Gaussian function on the X-axis and Y-axis. The processing process is as follows: ; (7) ; (8) In formulas (7) and (8), and They are the convolution results of remote sensing data points on the X-axis and Y-axis respectively. and are the values ​​of the remote sensing data points on the X-axis and Y-axis respectively, and are the convolution kernels on the X-axis and Y-axis respectively, and L is the length of the convolution kernel; In order to reduce the complexity of Gaussian filtering and improve the efficiency and pressure of calculation, the two-dimensional Gaussian function is converted into a one-dimensional Gaussian function, which can improve the efficiency of calculation.

[0024] In this embodiment, preferably, the data integration fusion module performs fusion processing on remote sensing images collected by different remote sensing platforms or different sensors, and the calculation process of the fusion processing is as follows: , (9) In formula (9), is the pixel value of the fused image, , are the pixel values ​​of the original images to be fused, w is the weight coefficient, and its value is between 0 and 1; By fusing remote sensing image data from different sources, using the maximum and minimum values ​​of the remote sensing image data for calculation and processing, and performing weight distribution processing on the maximum and minimum values, the remote sensing image data from different sources can be fused to form a clear remote sensing image information.

[0025] In this embodiment, preferably, the data feature extraction module accelerates the construction of the integral graph through a fast algorithm, and the construction process is calculated as follows: ; (10) In formula (10), is the image at position The original pixel value of is in position The integral image value of the image is the sum of all pixel values ​​in the area from the upper left corner (0,0) to (x,y); and the formula quickly calculates the integral image value of the current position by using the calculated integral image value; and the formula , , Represent the integral image values ​​of the left, top, and upper-left corners respectively; The accelerated integral map for feature extraction is constructed through a fast algorithm, the integral map value is calculated for the original pixel value, and the calculated integral map value is used to quickly calculate the integral map value of the current position, so as to realize fast calculation and processing of the integral map value and improve the feature extraction rate.

[0026] Furthermore, it is characterized in that: the integral image detects key points through the determinant of the Hessian matrix; and the determinant calculation formula of the Hessian matrix is ​​as follows: ; (11) In formula (11), It is the second-order derivative of the image in the X-axis direction, which represents the curvature of the image; It is the second-order derivative of the image in the Y-axis direction, which represents the curvature of the image; It is the mixed second-order derivative of the image in the X-axis and Y-axis directions, describing the interactive influence of the image in the two directions; The Haar wavelet response is used to generate feature descriptors for the local area of ​​each key point; and the calculation formula of the Haar wavelet response is as follows: ; (12) In formula (12), d is the size of the Haar feature, which is used to calculate the response in the local area; is the integral map value, which is used to accelerate the calculation of the response of the local area; For the integral image, the key points are detected by the determinant of the Hessian matrix, which is convenient for generating feature descriptors for the local area of ​​each key point, realizing the calculation of the features in the local area, and realizing the accelerated calculation of the response of the local area.

[0027] In this embodiment, preferably, the data time series analysis module performs time series analysis on the collected and processed remote sensing images through the SARIMA model, calculates the features of the remote sensing images through the integral graph established by the data feature extraction module, and performs time series analysis on the remote sensing images to improve the prediction of remote sensing impacts; The calculation formula for the SARIMA model to process seasonal time series data is as follows: ; (13) In formula (13), p is the order of the autoregressive term, which indicates how many lagged values ​​are used in the SARIMA model to predict the current value, q is the order of the moving average term, which indicates how many lagged error terms are used in the SARIMA model to predict the current value, P is the order of the seasonal autoregressive term, which indicates how many seasonal lagged values ​​are used in the SARIMA model, Q is the order of the seasonal moving average term, which indicates how many seasonal lagged error terms are used in the SARIMA model, s is the seasonal cycle, that is, the length of a complete season, where s is selected as 12, indicating that there are 12 months in a year, Φ1, Φ2, …, Φ p is the coefficient of the seasonal autoregressive term, which represents the relationship between the current value and the previous P seasonal lagged values, Θ1, Θ2, …, Θ q is the coefficient of the seasonal moving average term, which represents the relationship between the current value and the previous Q seasonal lagged error terms; Through time series analysis and processing, the data of remote sensing images can be combined with past historical data information when processing, so as to realize future prediction and processing, facilitate the prediction of pasture grass growth status, and facilitate the control and regulation of pasture grazing.

[0028] Furthermore, when calculating the SARIMA model, exogenous variables are introduced to calculate the time series data, and the calculation formula is as follows: ; (14) In formula (14), is an exogenous variable, and is the coefficient of the exogenous variable, and the exogenous variable is used as an explanatory variable to eliminate the sudden changes of external factors on remote sensing images; In order to eliminate the impact of large-scale harvesting and storage of forage in autumn, the forage harvesting situation is calculated to facilitate accurate control of grazing time, give the pasture a certain amount of recovery time, improve the production efficiency of the pasture, maintain the ecological balance of the grassland, and enable the sustainable use of the grassland.

[0029] Furthermore, the specific steps of dynamically updating parameters of the SARIMA model through the Bayesian method are as follows: Assuming model parameters The prior distribution of ; Given observation data and model parameters , calculate the likelihood function , the likelihood function of the SARIMA model usually involves the calculation of the residual of the time series: , (15) In formula (15), Based on the current parameters The model prediction value of is the variance of the error; Combining the prior distribution and the likelihood function, the posterior distribution is calculated using Bayes' theorem ; , (16) In formula (16), is the marginal likelihood; With new data Arrived, update the posterior distribution of the parameters to the new prior distribution, and repeat the above steps: ,(17) In formula (17), is the posterior distribution based on the first t observations; In order to update the remote sensing image data, the dynamically displayed pasture conditions can be updated in real time to facilitate clear observation of pasture changes. The parameters are dynamically updated through the Bayesian method to improve the effect and accuracy of the dynamic display.

[0030] In this embodiment, preferably, the data visualization module displays the changes of the remote sensing image through dynamic display, that is, the changes and dynamic processes of the grass in the grassland are displayed through time series animation. The dynamic display steps of the data visualization module are as follows: Data loading: Import or connect processed remote sensing image data to ensure the accuracy and completeness of remote sensing image data; Configure display: set chart type, style and dynamic features, and configure real-time updates and interactivity; Realize dynamic functions: configure real-time update and interactive operation of remote sensing image data, and realize real-time update, interactivity and dynamic display through time series data; Run and monitor: Start the display module to monitor the update of remote sensing image data and user interaction in real time; Adjustment and optimization: Adjust according to actual usage to optimize performance and user experience; In order to dynamically display the changes in pasture grass and realize the optional adjustment parameters, the effect and accuracy of dynamic display are improved, and users can adjust the parameters according to their actual usage experience.

[0031] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A data analysis system based on integrated remote sensing data acquisition, characterized in that: It includes a data analysis system, which includes a remote sensing data acquisition module, a data preprocessing module, a data integration and fusion module, a data feature extraction module, a data time series analysis module and a data visualization module; The data analysis system adopts distributed computing and parallel and stream processing to improve the rapid processing of integrated massive remote sensing data; The distributed computing distributes remote sensing data and computing tasks to multiple computers through Apache Spark for collaborative processing; The parallel and streaming processing simultaneously processes multiple remote sensing data streams or tasks through Apache Kafka to improve processing efficiency; decomposes the computing task into multiple parallel tasks to improve computing efficiency; The remote sensing data acquisition module is used to receive and collect remote sensing data monitored by a remote sensing platform, wherein the remote sensing platform includes satellites, aircraft and drones, and the remote sensing platform monitors and collects data through sensors carried thereon, and the sensors include optical sensors, radar sensors and laser radars; Optical sensors capture information in the visible and near-infrared spectrum; Radar sensors obtain surface information by actively emitting electromagnetic waves and receiving reflected signals; LiDAR obtains high-precision terrain data and three-dimensional information by scanning the surface with lasers; The data preprocessing module preprocesses the collected remote sensing data, and the preprocessing includes geometric correction, atmospheric correction, radiation correction, noise removal, image enhancement and image fusion; Geometric correction is the process of converting pixel coordinates in remote sensing images into actual geographic coordinates on the surface; Atmospheric correction is to adjust the image radiation value caused by atmospheric scattering and absorption effects in remote sensing images in order to extract real surface information from the images; Radiometric correctionRadiometric correction adjusts the raw radiometric values ​​of remote sensing data to eliminate errors introduced by the sensor and the data collection process; Noise removal can improve the data quality by removing noise or pseudo-images in remote sensing data; Image enhancement helps to better extract information by improving the visualization of images; Image fusion combines image data from different sensors or at different times to obtain richer information; The data integration and fusion module is used to integrate data from different sensors and remote sensing platforms to ensure the consistency of data from different sources in space and time; The data feature extraction module is used to extract features from the fused remote sensing data, so as to facilitate subsequent analysis and processing based on the extracted features; The data time series analysis module applies statistical models to model and predict time series data, combines remote sensing data acquired at different times, and performs time series analysis and change monitoring; The data visualization module is used to display the prediction and monitoring data of the data time series analysis module in a visualized form to facilitate understanding and analysis; it includes integrated maps and dynamic displays. The integrated map overlays the analysis results of different data sets on the map for spatial visualization; the dynamic display displays surface changes and dynamic processes through time series animation.

2. The data analysis system based on integrated remote sensing data acquisition according to claim 1, characterized in that: The noise removal is performed by calculating and processing the remote sensing data through Gaussian filtering, and the calculation formula of Gaussian filtering is as follows: （1） In formula (1), It's on point The Gaussian function value of is the standard deviation of the Gaussian function, controlling the strength of smoothing; and are the coordinates relative to the center of the Gaussian function; Standard Deviation The calculation formula is as follows: （2） In formula (2), is the value of each remote sensing data point, and the value is The two-dimensional value of is the mean of the remote sensing data points, is the mean error of remote sensing data points, N is the number of remote sensing data points; The mean error The calculation formula is as follows: （3） In formula (3), is the actual calculated mean, is the predicted mean.

3. The data analysis system based on integrated remote sensing data acquisition according to claim 2, characterized in that: The separation of the Gaussian filter separates the two-dimensional Gaussian function into a one-dimensional Gaussian function to improve the efficiency of calculation. The separation formula of the Gaussian filter is as follows: （4） In formula (4), is a one-dimensional Gaussian function on the X-axis, is a one-dimensional Gaussian function on the Y axis; Then the one-dimensional Gaussian function on the X-axis is as follows: （5） The one-dimensional Gaussian function on the Y axis is as follows: （6） In formulas (5) and (6), x and y are the Gaussian function values, respectively. The values ​​on the X-axis and Y-axis; Then convolve the one-dimensional Gaussian function on the X-axis and Y-axis. The processing process is as follows: (7) (8) In formulas (7) and (8), and They are the convolution results of remote sensing data points on the X-axis and Y-axis respectively. and are the values ​​of the remote sensing data points on the X-axis and Y-axis respectively, and are the convolution kernels on the X-axis and Y-axis respectively, and L is the length of the convolution kernel.

4. The data analysis system based on integrated remote sensing data acquisition according to claim 1, characterized in that: The data integration fusion module performs fusion processing on remote sensing images collected by different remote sensing platforms or different sensors, and the calculation process of the fusion processing is as follows: （9） In formula (9), is the pixel value of the fused image, , are the pixel values ​​of the original images that need to be fused, is the weight coefficient, and its value is between 0 and 1.

5. The data analysis system based on integrated remote sensing data acquisition according to claim 1, characterized in that: The data feature extraction module accelerates the construction of the integral graph through a fast algorithm, and the construction process is calculated as follows: （10） In formula (10), is the image at position The original pixel value of is in position The integral image value of the image is the sum of all pixel values ​​in the area from the upper left corner (0,0) to (x,y); and the formula quickly calculates the integral image value of the current position by using the calculated integral image value; and the formula , , Represent the integral image values ​​of the left, top, and upper-left corners respectively.

6. The data analysis system based on integrated remote sensing data acquisition according to claim 5, characterized in that: The integral graph detects key points through the determinant of the Hessian matrix; and the determinant calculation formula of the Hessian matrix is ​​as follows: （11） In formula (11), It is the second-order derivative of the image in the X-axis direction, which represents the curvature of the image; It is the second-order derivative of the image in the Y-axis direction, which represents the curvature of the image; It is the mixed second-order derivative of the image in the X-axis and Y-axis directions, describing the interactive influence of the image in the two directions; The Haar wavelet response is used to generate feature descriptors for the local area of ​​each key point; and the calculation formula of the Haar wavelet response is as follows: （12） In formula (12), d is the size of the Haar feature, which is used to calculate the response in the local area; is the integral map value, which is used to speed up the calculation of the response of the local area.

7. The data analysis system based on integrated remote sensing data acquisition according to claim 1, characterized in that: The data time series analysis module performs time series analysis on the collected and processed remote sensing images through the SARIMA model, calculates the features of the remote sensing images through the integral graph established by the data feature extraction module, and performs time series analysis on the remote sensing images to improve the prediction of remote sensing impacts; The calculation formula for the SARIMA model to process seasonal time series data is as follows: （13） In formula (13), p is the order of the autoregressive term, which indicates how many lagged values ​​are used in the SARIMA model to predict the current value, q is the order of the moving average term, which indicates how many lagged error terms are used in the SARIMA model to predict the current value, P is the order of the seasonal autoregressive term, which indicates how many seasonal lagged values ​​are used in the SARIMA model, Q is the order of the seasonal moving average term, which indicates how many seasonal lagged error terms are used in the SARIMA model, s is the seasonal cycle, that is, the length of a complete season, where s is selected as 12, indicating that there are 12 months in a year, Φ1, Φ2, …, Φ p is the coefficient of the seasonal autoregressive term, which represents the relationship between the current value and the previous P seasonal lagged values, Θ1, Θ2, …, Θ q is the coefficient of the seasonal moving average term, which represents the relationship between the current value and the previous Q seasonal lagged error terms.

8. The data analysis system based on integrated remote sensing data acquisition according to claim 7, characterized in that: When calculating the SARIMA model, the time series data is calculated by introducing exogenous variables, and the calculation formula is as follows: （14） In formula (14), is an exogenous variable, and is the coefficient of the exogenous variable, and the exogenous variable is used as an explanatory variable to eliminate the sudden changes of external factors on remote sensing images.

9. The data analysis system based on integrated remote sensing data acquisition according to claim 8, characterized in that: The specific steps of dynamically updating parameters of the SARIMA model by Bayesian method are as follows: Assuming model parameters The prior distribution of ; Given observation data and model parameters , calculate the likelihood function , the likelihood function of the SARIMA model usually involves the calculation of the residual of the time series: （15） In formula (15), Based on the current parameters The model prediction value of is the variance of the error; Combining the prior distribution and the likelihood function, the posterior distribution is calculated using Bayes' theorem ; （16） In formula (16), is the marginal likelihood; With new data Arrived, update the posterior distribution of the parameters to the new prior distribution, and repeat the above steps: （17） In formula (17), is the posterior distribution based on the first t observations.

10. The data analysis system based on integrated remote sensing data acquisition according to claim 1, characterized in that: The data visualization module displays the changes of remote sensing images through dynamic display, that is, displays the changes and dynamic processes of the grass in the grassland through time series animation. The dynamic display steps of the data visualization module are as follows: Data loading: Import or connect processed remote sensing image data to ensure the accuracy and completeness of remote sensing image data; Configure display: set chart type, style and dynamic features, and configure real-time updates and interactivity; Realize dynamic functions: configure real-time update and interactive operation of remote sensing image data, and realize real-time update, interactivity and dynamic display through time series data; Run and monitor: Start the display module to monitor the update of remote sensing image data and user interaction in real time; Adjustment and optimization: Adjustments are made based on actual usage to optimize performance and user experience.

Images