Smart rural comprehensive management method and system

Abstract

The invention discloses a smart rural comprehensive management method and system, and particularly relates to the technical field of data management. Monitoring areas are divided according to the geographic distribution density of sensors in a farmland, the accuracy weight of sensor data in each area is calculated by using the data deviation characteristics between adjacent sensors and the fluctuation characteristics of crop growth environment data, and a comprehensive accuracy index reflecting the environment condition is generated. Therefore, the problems of data blind areas and data distortion caused by uneven sensor distribution are solved, a monitoring area is divided into a high-accuracy area, a medium-accuracy area and a low-accuracy area based on the accuracy index, the abnormal condition of the medium-accuracy area is predicted, and the data acquisition frequency and the distribution density are dynamically adjusted according to the prediction result. According to the method, the accuracy and reliability of a data analysis system are effectively improved, managers can master the crop growth environment more accurately, farmland management decisions are optimized, data prejudice and misleading decisions are avoided, and the controllability and efficiency of management are improved.

Description

Technical Field

[0001] The present invention relates to the field of data management technology, and in particular to a smart village comprehensive management method and system. Background Art

[0002] Smart village integrated management is a model that uses information, digitalization and intelligent means to improve the management efficiency and service level of rural areas. Through technologies such as the Internet of Things, artificial intelligence, big data analysis, and cloud computing, the smart village integrated management system can monitor and manage rural resources, infrastructure, ecological environment, agricultural production, etc. in real time, thereby achieving scientific decision-making and efficient governance. In the management methods of existing technologies, smart villages mainly rely on sensor networks and data management systems. For example, when managing farmland, data is collected in real time and transmitted to the cloud platform through soil sensors, meteorological stations, crop monitoring instruments and other equipment, which facilitates managers to analyze and make decisions remotely. At the same time, the big data platform is used to analyze and integrate rural geographic information, crop planting data, villagers' needs, etc., to formulate precise agricultural production plans.

[0003] The prior art has the following deficiencies: When managing farmland, due to the low density of geographical distribution of sensors, some locations may be overlooked (i.e., "sensor black spots"), resulting in incomplete data. At the same time, the data analysis system may generate prediction and decision models based on limited sensor data, making it difficult for the system to accurately reflect the actual environmental conditions, which in turn leads to data bias and misleading decisions. For example, soil sensors in a certain area may show moderate humidity, but due to uneven distribution, severe drought in the black spot area cannot be monitored. In addition, long-term data distortion will cause the algorithm model of the management system to deviate more, and the decision-making will lose its reference value. Managers may gradually lose accurate judgment of crop growth conditions, resulting in reduced controllability of management, and even a chain of management out of control. Summary of the invention

[0004] The purpose of the present invention is to provide a smart village comprehensive management method and system to solve the deficiencies in the background technology.

[0005] In order to achieve the above object, the present invention provides the following technical solution: a smart village comprehensive management method, comprising the following steps: S1: According to the geographical density of sensors in the farmland, the farmland is divided into several monitoring areas, and the monitoring data deviation characteristics between adjacent sensor pairs in different time periods in each monitoring area are obtained, as well as the fluctuation characteristics of crop growth environment data monitored by the same type of sensors in different time periods; S2: Determine the accuracy weight assignment of the sensor monitoring data in different time periods in each monitoring area according to the acquired monitoring data deviation characteristics and crop growth environment data fluctuation characteristics, and calculate the weighted average of the accuracy weight assignment of the sensor monitoring data in different time periods in each monitoring area to obtain the accuracy index of the data analysis system reflecting the actual environmental conditions; S3: Divide the monitoring area into high-accuracy monitoring area, medium-accuracy monitoring area and low-accuracy monitoring area according to the calculated accuracy index of the data analysis system reflecting the actual environmental conditions, and manage them accordingly; S4: For the medium-accuracy monitoring area, further prediction is made on the abnormal degree of accuracy of the data analysis system in reflecting the actual environmental conditions, and the data collection frequency and distribution density of the sensors in the medium-accuracy monitoring area are dynamically adjusted according to the prediction results.

[0006] Preferably, in S2, the acquired monitoring data deviation characteristics are analyzed to generate a monitoring value deviation abnormality index, and the monitoring value deviation abnormality index is obtained by: Obtain monitoring data from each pair of adjacent sensors, calculate the deviation between each pair of adjacent sensors, arrange the deviation values ​​of each pair of adjacent sensors at each time point in chronological order, and generate time series data , where t represents the time point in the time series; select the ARIMA model parameters. The ARIMA model is represented by ARIMA(p,d,q), where: p represents the number of autoregressive terms, d represents the number of differences, and q represents the number of moving average terms. If the time series is not stable, perform a difference operation according to the parameter d to obtain a stable time series. , the expression is: ; Use the optimized parameters p, d, and q to train the ARIMA model, use historical data to fit the ARIMA model, and predict the future monitoring value deviation based on the trained ARIMA model. Set the prediction value to , then the prediction formula of the model at time t is: ;in: is the coefficient of the autoregressive term, is the coefficient of the moving average term, is the residual at the previous time point, calculate the residual , the residual represents the difference between the actual deviation value and the predicted value, and the calculation formula is: ; Standardize the residuals and convert them into standardized residuals , the expression is: ;in is the standard deviation of the residual. The threshold range of the residual under normal conditions is set through historical data. When the standardized residual exceeds the threshold range, the deviation is judged to be abnormal. The frequency of abnormal occurrence in the T time period is calculated. Let Indicates the total number of time points, The number of abnormal time points is used to calculate the abnormal index of the monitoring value deviation. The expression is: ; Where AE is the abnormal index of monitoring value deviation.

[0007] Preferably, in S2, the environmental data fluctuation index is generated after analyzing the fluctuation characteristics of the acquired crop growth environment data. The method for obtaining the environmental data fluctuation index is: Get the time series of environmental data within a period of time, set the data to ,in Represents the observed value at time n, selects the observation time interval, and divides the data sequence X into several segments , each segment is an independent subsequence and obeys a normal distribution with different mean and variance. For each segment of data, set the mean and variance Normal distribution ,for and For the prior , choose the conjugate priors of the normal distribution and the inverse gamma distribution: ;in, is a priori parameter. Based on the observed data X, the posterior probability of each possible change point position is calculated. For a given set of change points , the posterior probability Represents the probability of change, and uses the Bayesian formula to update the posterior probability: ;in: is the likelihood of the data at a given change point, and calculates the probability that each segment of data follows the corresponding normal distribution. P(C) is the prior probability, which represents the prior distribution of the change point position. For each segment of data , using its standard deviation Represents the fluctuation range, and the expression is: ;in, It is a section The mean of Represents the number of data points in a segment. Based on the fluctuation range of each segment, the environmental data fluctuation index is defined. The calculation expression is: ; GH is the environmental data fluctuation index.

[0008] Preferably, in S2, the monitoring value deviation anomaly index and the environmental data fluctuation index are converted into a comprehensive feature vector, and the comprehensive feature vector is used as the input of the machine learning model. The machine learning model uses each group of comprehensive feature vectors to predict the accuracy weighted assignment labels of the sensor monitoring data in each monitoring area in different time periods as the prediction target, and takes minimizing the sum of the prediction errors of the accuracy weighted assignment labels of the sensor monitoring data in all monitoring areas in different time periods as the training target. The machine learning model is trained until the sum of the prediction errors converges and the model training is stopped. The accuracy weight assignment of the sensor monitoring data in different time periods in each monitoring area is determined according to the model output results, wherein the machine learning model is a polynomial regression model, and the accuracy index of the data analysis system reflecting the actual environmental conditions is obtained by weighted average calculation of the accuracy weight assignment of the sensor monitoring data in different time periods in each monitoring area.

[0009] Preferably, in S3, the monitoring area is divided according to the calculated accuracy index of the data analysis system reflecting the actual environmental conditions, and the monitoring area is divided into a high-accuracy monitoring area, a medium-accuracy monitoring area and a low-accuracy monitoring area, specifically: The acquired accuracy index is compared with a gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and the accuracy index is compared with the first standard threshold and the second standard threshold respectively; If the accuracy index is greater than the second standard threshold, it means that the data analysis system reflects the actual environmental conditions with high accuracy, and a high accuracy signal is generated, and the corresponding monitoring area is divided into a high accuracy monitoring area; If the accuracy index is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the accuracy of the data analysis system in reflecting the actual environmental conditions is average, and a medium accuracy signal is generated, and the corresponding monitoring area is divided into a medium accuracy monitoring area; If the accuracy index is less than the first standard threshold, it means that the data analysis system has low accuracy in reflecting the actual environmental conditions. In this case, a low accuracy signal is generated and the corresponding monitoring area is divided into a low accuracy monitoring area.

[0010] Preferably, in S4, for the medium accuracy monitoring area, the abnormal degree of the accuracy of the data analysis system in reflecting the actual environmental conditions is further predicted, and the data collection frequency and distribution density of the sensors in the medium accuracy monitoring area are dynamically adjusted according to the prediction results, specifically: For the medium accuracy monitoring area, that is, the accuracy index generated within a fixed time period is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the accuracy index generated in the subsequent fixed time period that is greater than or equal to the first standard threshold and less than or equal to the second standard threshold is collected, and the corresponding data set is established, and the standard deviation of the data set is calculated as the accuracy anomaly index, which is used as the basis for dynamically adjusting the sensor data collection frequency and point density.

[0011] Preferably, the frequency of data collection is adjusted based on the accuracy anomaly index ADI, and the initial data collection frequency is defined as , and adjust the acquisition frequency according to ADI, the expression is: ; Where: f is the data acquisition frequency after dynamic adjustment, is the initial data collection frequency, is the acquisition frequency adjustment coefficient; Set the distribution density adjustment formula, and calculate the target distribution density. The expression is: ; Where: ρ is the density of points after dynamic adjustment, is the initial point density, is the point density adjustment coefficient; By calculating the abnormal degree of ADI real-time prediction monitoring data, the acquisition frequency f and point density ρ in the medium accuracy area are dynamically adjusted. , that is, if the abnormality exceeds the preset threshold α, then f and ρ will be increased; if , then maintain or restore the initial value to reduce resource consumption.

[0012] The present invention also provides a smart village integrated management system, including a data collection module, an accuracy assessment module, a regional division management module and an optimization module; Data acquisition module: According to the geographical density of sensors in the farmland, it is divided into several monitoring areas, and the monitoring data deviation characteristics between adjacent sensor pairs in different time periods in each monitoring area are obtained, as well as the fluctuation characteristics of crop growth environment data monitored by the same type of sensors in different time periods; Accuracy evaluation module: Determine the accuracy weight of the sensor monitoring data in different time periods in each monitoring area according to the acquired monitoring data deviation characteristics and crop growth environment data fluctuation characteristics, and calculate the weighted average of the accuracy weight of the sensor monitoring data in different time periods in each monitoring area to obtain the accuracy index of the data analysis system reflecting the actual environmental conditions; Area division management module: divides the monitoring area according to the accuracy index of the calculated data analysis system reflecting the actual environmental conditions, and divides the monitoring area into high-accuracy monitoring area, medium-accuracy monitoring area and low-accuracy monitoring area, and manages them accordingly; Optimization module: For medium-accuracy monitoring areas, further prediction is made on the degree of abnormality in the accuracy of the data analysis system in reflecting the actual environmental conditions, and the data collection frequency and distribution density of sensors in the medium-accuracy monitoring areas are dynamically adjusted based on the prediction results.

[0013] In the above technical solution, the technical effects and advantages provided by the present invention are: 1. The present invention divides the monitoring area according to the density of sensor points and obtains the monitoring data deviation and environmental data fluctuation characteristics of each area, and calculates the accuracy weight and comprehensive accuracy index of the sensor monitoring data. This method effectively overcomes the data blind spot problem in the prior art, eliminates the data deviation caused by the "black spot" area of ​​the sensor, and improves the data analysis system's accurate reflection of the actual environmental conditions by dynamically adjusting the acquisition frequency and point density of the monitoring area. Monitoring areas with different accuracy levels are subject to targeted management measures to further ensure the reliability of the system and the accuracy of decision support, and reduce the risk of misleading decisions.

[0014] 2. The present invention predicts the accuracy weight assignment of monitoring data through a machine learning model, and converts the monitoring value deviation anomaly index and the environmental data fluctuation index into a comprehensive feature vector, which can dynamically predict and adjust the degree of data accuracy anomaly in the medium-accuracy monitoring area, and achieve real-time optimization. This dynamic adjustment mechanism of frequency and point density based on the accuracy anomaly index not only reduces resource consumption while ensuring high data quality, but also improves the controllability and refinement of farmland management, thereby effectively enhancing the intelligent level of rural environmental management. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0016] Figure 1 The figure is a flow chart of the method of the present invention.

[0017] Figure 2 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are 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.

[0019] Example 1, please refer to Figure 1 and Figure 2 As shown, the smart village comprehensive management method described in this embodiment includes the following steps: S1: According to the geographical density of sensors in the farmland, the farmland is divided into several monitoring areas, and the monitoring data deviation characteristics between adjacent sensor pairs in different time periods in each monitoring area are obtained, as well as the fluctuation characteristics of crop growth environment data monitored by the same type of sensors in different time periods; S2: Determine the accuracy weight assignment of the sensor monitoring data in different time periods in each monitoring area according to the acquired monitoring data deviation characteristics and crop growth environment data fluctuation characteristics, and calculate the weighted average of the accuracy weight assignment of the sensor monitoring data in different time periods in each monitoring area to obtain the accuracy index of the data analysis system reflecting the actual environmental conditions; S3: Divide the monitoring area into high-accuracy monitoring area, medium-accuracy monitoring area and low-accuracy monitoring area according to the calculated accuracy index of the data analysis system reflecting the actual environmental conditions, and manage them accordingly; S4: For the medium-accuracy monitoring area, further prediction is made on the abnormal degree of accuracy of the data analysis system in reflecting the actual environmental conditions, and the data collection frequency and distribution density of the sensors in the medium-accuracy monitoring area are dynamically adjusted according to the prediction results.

[0020] In S1, the farmland is divided into several monitoring areas according to the geographical density of sensors in the farmland, and the monitoring data deviation characteristics between adjacent sensor pairs in different time periods in each monitoring area are obtained, as well as the fluctuation characteristics of crop growth environment data monitored by the same type of sensors in different time periods. Specifically, Conduct a preliminary analysis based on the density of sensor points to understand the distribution of sensors in different areas. If the sensors are densely distributed in a certain area, the monitoring reliability of the area is higher; on the contrary, there may be monitoring blind spots in areas with sparse distribution. Consider factors such as farmland topography, soil type, irrigation method, and crop type to divide the monitoring area more accurately. For example, low-lying areas may be more prone to waterlogging, and different crops have different water and nutrient requirements. Confirm the signal coverage of the data transmission equipment to ensure that the data in each area can be stably transmitted to the system.

[0021] According to the density and geographical location of the sensors, the farmland is initially divided into several candidate monitoring areas, including: Densely distributed area: The area with densely distributed sensors is divided into an independent monitoring area, which is usually located in the key area of ​​the farmland, such as the main crop planting belt. Sparsely distributed area: For areas with fewer sensors or sparsely distributed sensors, adjacent areas are merged to form a larger monitoring area to reduce monitoring blind spots.

[0022] Adjust the boundaries of the initially divided monitoring areas so that the sensors in the area are relatively evenly distributed and ensure that the data coverage of each monitoring area is reasonable. For example, low-lying areas should be distinguished from highlands in order to capture microclimate differences. In each monitoring area, sensors that are close to each other are paired and marked as adjacent sensor pairs. The data deviation characteristics of these adjacent sensor pairs will be used to judge the consistency of data in the area. Usually, sensor pairs that are close to each other (such as less than 50 meters) are marked as adjacent pairs to improve the accuracy of the analysis. Confirm the coverage of each area so that it can reflect relatively independent climate, soil and crop growth conditions. The final monitoring area should have a high internal consistency, and the data in each area can represent the overall situation of the area.

[0023] In each monitoring area, data is collected from each pair of adjacent sensors at a predetermined frequency (e.g., every hour, every day). Soil moisture, temperature, air pressure, etc. are key collection parameters. Deviation calculation is performed on the monitoring data of each pair of adjacent sensors, and the formula is: ;in, and It is the monitoring value of adjacent sensors at the same time point. The calculation result is used to quantify the consistency of data within the area. The deviation within a period of time is counted to obtain the overall characteristics of the sensor data deviation in the monitoring area. If the deviation value is large, it indicates that the monitoring data consistency in the area is poor, and it may be necessary to increase the sensor density or calibrate the data.

[0024] Obtain monitoring data from sensors of the same type (such as all soil moisture sensors) at different time periods to monitor changes in the crop growth environment. The data collection time period can be set to daytime and nighttime, sunny and rainy days, or planting and growing seasons, etc. For sensors of the same type in each monitoring area, calculate their data fluctuation characteristics. The fluctuation characteristic calculation formula is: ;in, and is the monitoring data of two consecutive time points, and N is the total number of time points. This calculation can reflect the degree of dynamic change of the crop growth environment. Statistics on data fluctuations over a period of time and analyze the fluctuation characteristics of different monitoring areas. If the data fluctuations in a certain area are large, it means that the environment changes frequently, and intensive data collection may be required to accurately capture the growth needs of crops.

[0025] S2: Determine the accuracy weight assignment of the sensor monitoring data in different time periods in each monitoring area based on the acquired monitoring data deviation characteristics and crop growth environment data fluctuation characteristics, and obtain the accuracy index of the data analysis system reflecting the actual environmental conditions by performing weighted average calculation on the accuracy weight assignment of the sensor monitoring data in different time periods in each monitoring area.

[0026] After analyzing the deviation characteristics of the acquired monitoring data, the monitoring value deviation anomaly index is generated. The method for obtaining the monitoring value deviation anomaly index is as follows: Obtain monitoring data (such as soil moisture, temperature, etc.) from each pair of adjacent sensors, calculate the deviation between each pair of adjacent sensors, and arrange the deviation values ​​of each pair of adjacent sensors at each time point in chronological order to generate time series data , where t represents the time point in the time series; select the ARIMA model parameters. The ARIMA model is represented by ARIMA (p, d, q), where: p represents the number of autoregressive terms, that is, how many past deviation values ​​the current value is related to, and d represents the number of differences, which is used to eliminate non-stationarity and obtain a stable time series through differences. q represents the number of moving average terms, that is, how many past residuals the error term is related to. Determine the p, d, q parameters: usually use the autocorrelation function (ACF) and partial autocorrelation function (PACF) to preliminarily select parameters, or use information criteria (such as AIC or BIC) for parameter optimization. If the time series is not stable, perform a difference operation based on the parameter d to obtain a stable time series , the expression is: ; Use the optimized parameters p, d, q to train the ARIMA model so that the model can capture the trend and cyclical fluctuations in the time series. Use historical data to fit the ARIMA model and determine the coefficients of the model. Based on the trained ARIMA model, predict the deviation of the future monitoring value and set the prediction value to , then the prediction formula of the model at time t is: ;in: is the coefficient of the autoregressive term, is the coefficient of the moving average term, is the residual at the previous time point (i.e., prediction error), and calculate the residual , the residual represents the difference between the actual deviation value and the predicted value, and the calculation formula is: ; To facilitate the detection of anomalies, the residuals are standardized and converted into standardized residuals , the expression is: ;in is the standard deviation of the residual. The threshold range of the residual under normal conditions is set through historical data. When the standardized residual exceeds the threshold range, the deviation is judged to be abnormal. The frequency of abnormal occurrence in the T time period is calculated. Let Indicates the total number of time points, The number of abnormal time points is used to calculate the abnormal index of the monitoring value deviation. The expression is: ; Where AE is the abnormal index of monitoring value deviation.

[0027] The larger the monitoring value deviation anomaly index is, the poorer the accuracy of the sensor monitoring data in each monitoring area in different time periods is, and more abnormal deviations have occurred. This means that the data collected by the sensor is unstable and may be frequently affected by factors such as environmental interference, sensor failure, or data drift, resulting in the inability to accurately reflect the actual situation. In addition, a high monitoring value deviation anomaly index may also indicate that the sensor points are uneven, and the sensor data in some areas is unreliable and prone to deviations. If managers rely on these data to make decisions, they will face a high risk of misjudgment. Therefore, when the monitoring value deviation anomaly index is large, it may be necessary to increase the density of sensor points, adjust the collection frequency, or calibrate the equipment to improve the stability and credibility of the data.

[0028] On the contrary, the smaller the monitoring value deviation anomaly index is, the higher the accuracy of the sensor monitoring data in each monitoring area in different time periods is, the smaller the deviation is, and the data is relatively consistent and stable. This shows that the sensor can stably collect data under various environmental conditions, reflecting the changes in the actual situation, thus providing a reliable reference for managers. A low monitoring value deviation anomaly index also means that the data consistency and distribution rationality in the system are high, and the monitoring information in each area can represent the real environmental conditions. For the integrated management of smart villages, a smaller monitoring value deviation anomaly index will greatly improve the quality of data-driven decision-making, reduce the possibility of misjudgment, and contribute to the efficient operation of the management system and the rational allocation of resources.

[0029] After analyzing the fluctuation characteristics of the acquired crop growth environment data, an environmental data fluctuation index is generated. The method for obtaining the environmental data fluctuation index is as follows: Get the time series of environmental data (such as temperature, humidity, soil moisture, etc.) over a period of time. Set the data to ,in represents the observed value at time n, and the observation time interval (such as hours or days) is selected to ensure that the amount of data is sufficient but not too much noise is introduced. The observation interval of the data should be related to the typical period of environmental fluctuations. Split the data sequence X into several segments , each segment is an independent subsequence and follows a normal distribution with different mean and variance. The location of the change point can be assumed to follow a uniform distribution, that is, any time point may be a change point. For each segment of data, set the mean and variance Normal distribution .for and For the prior, we can choose the conjugate prior of the normal distribution and the inverse gamma distribution: ;in, is a priori parameter, set based on experience or historical data. Based on the observed data X, the posterior probability of each possible change point position is calculated. For a given set of change points , the posterior probability Represents the probability of changes at these locations, and uses the Bayesian formula to update the posterior probability: ;in: is the likelihood of the data at a given change point, and calculates the probability that each segment of data follows the corresponding normal distribution. P(C) is the prior probability, which represents the prior distribution of the change point position. For each segment of data , using its standard deviation Represents the fluctuation range, and the expression is: ;in, It is a section The mean of Represents the number of data points in a segment. Based on the fluctuation range of each segment, the environmental data fluctuation index is defined. The calculation expression is: ; GH is the environmental data fluctuation index.

[0030] The larger the environmental data volatility index, the more dramatic the fluctuations in environmental variables (such as temperature, humidity, etc.) in each monitoring area. Such dramatic fluctuations may reflect reduced data accuracy, because a more volatile environment usually means that sensor data is more susceptible to external influences, resulting in outliers or noise. In addition, a higher volatility index may indicate that the density of sensor points is insufficient, and that small changes in the environment cannot be captured, resulting in large deviations in the data. If managers make decisions based on data with a high volatility index, large errors may be introduced, leading to inaccurate management measures such as irrigation and fertilization. Therefore, when the environmental data volatility index is large, it may be necessary to increase the monitoring frequency or increase the density of sensor points to obtain more accurate data.

[0031] On the contrary, the smaller the environmental data fluctuation index, the more stable the environmental changes in each monitoring area, the lower the fluctuation of the data collected by the sensors, and the higher the accuracy of the data. A stable environment means that the monitoring data is more consistent and less susceptible to short-term fluctuations. A low environmental data fluctuation index also indicates that the existing sensor layout and data collection frequency are reasonable, which can fully capture environmental changes without introducing significant errors. Therefore, a smaller fluctuation index reflects the high credibility of the data, which helps managers make accurate agricultural management decisions and ensure the stable implementation of various measures.

[0032] The monitoring value deviation anomaly index and the environmental data fluctuation index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model uses each group of comprehensive feature vectors to predict the accuracy weighted assignment labels of the sensor monitoring data in each monitoring area in different time periods as the prediction target, and takes minimizing the sum of the prediction errors of the accuracy weighted assignment labels of the sensor monitoring data in all monitoring areas in different time periods as the training target. The machine learning model is trained until the sum of the prediction errors converges and the model training is stopped. The accuracy weight assignment of the sensor monitoring data in different time periods in each monitoring area is determined according to the model output results. Among them, the machine learning model is a polynomial regression model, and the accuracy index of the data analysis system reflecting the actual environmental conditions is obtained by weighted average calculation of the accuracy weight assignment of the sensor monitoring data in different time periods in each monitoring area.

[0033] The method for obtaining the accuracy weight assignment of the sensor monitoring data in different time periods in each monitoring area is as follows: from the comprehensive feature vector training data of the trained machine learning model, the corresponding function expression is obtained: ; In the formula, is the output function of the model, AE is the monitoring value deviation anomaly index, GH is the environmental data fluctuation index, Assign accuracy weights to the sensor monitoring data in each monitoring area at different time periods.

[0034] S3: Divide the monitoring area into high-accuracy monitoring area, medium-accuracy monitoring area and low-accuracy monitoring area according to the calculated accuracy index of the data analysis system reflecting the actual environmental conditions, and carry out corresponding management.

[0035] The acquired accuracy index is compared with a gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and the accuracy index is compared with the first standard threshold and the second standard threshold respectively; If the accuracy index is greater than the second standard threshold, it means that the data analysis system reflects the actual environmental conditions with high accuracy. At this time, a high-accuracy signal is generated and the corresponding monitoring area is divided into a high-accuracy monitoring area. The data in the high-accuracy area can be directly used for agricultural management decisions, such as irrigation, fertilization and other daily operations. Relying on such regional data can improve the reliability of decision-making.

[0036] If the accuracy index is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the accuracy of the data analysis system in reflecting the actual environmental conditions is average. In this case, a medium accuracy signal is generated, and the corresponding monitoring area is divided into a medium accuracy monitoring area. It is necessary to further analyze the accuracy of the data analysis system in reflecting the actual environmental conditions within a fixed time period. If the accuracy index is less than the first standard threshold, it means that the data analysis system reflects the actual environmental conditions with low accuracy. In this case, a low accuracy signal is generated and the corresponding monitoring area is divided into a low accuracy monitoring area. More sensors are added in the low accuracy area to increase the density of points, reduce monitoring blind spots, and capture more complete environmental data. Frequently calibrate sensor data and regularly check data deviations and environmental fluctuation indexes to identify and exclude unstable data.

[0037] S4: For the medium-accuracy monitoring area, further prediction is made on the abnormal degree of accuracy of the data analysis system in reflecting the actual environmental conditions, and the data collection frequency and distribution density of the sensors in the medium-accuracy monitoring area are dynamically adjusted according to the prediction results.

[0038] For the medium accuracy monitoring area, that is, the accuracy index generated within a fixed time period is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the accuracy index generated in the subsequent fixed time period that is greater than or equal to the first standard threshold and less than or equal to the second standard threshold is collected, and the corresponding data set is established, and the standard deviation of the data set is calculated as the accuracy anomaly index, which is used as the basis for dynamically adjusting the sensor data collection frequency and point density.

[0039] Set a data collection frequency adjustment formula, adjust the frequency of data collection based on the accuracy anomaly index ADI, and define the initial data collection frequency as (Unit: times / hour), and adjust the acquisition frequency f according to ADI, the expression is: ; Where: f is the data collection frequency after dynamic adjustment, in times / hour. is the initial data collection frequency, in times / hour. is the acquisition frequency adjustment coefficient, which reflects the influence of ADI on the acquisition frequency. Experience values ​​or historical data can be used to determine .

[0040] When ADI is large, the acquisition frequency f increases, allowing the system to collect data more frequently and capture more environmental changes; conversely, when ADI is small, the data acquisition frequency f will be close to the initial frequency , avoid unnecessary increase in collection frequency and save resources.

[0041] Determine whether it is necessary to increase the density of sensor points based on the ADI index. Set the point density adjustment formula and calculate the target point density. The expression is: ; Where: ρ is the density of points after dynamic adjustment, the unit is the number of sensors / area, is the initial point density, the unit is the number of sensors / area, It is the adjustment coefficient of the point density, which reflects the influence of ADI on the point density and is usually set based on historical experience data. When ADI is high, it means that the current point density cannot fully capture data fluctuations, so the point density ρ is increased to improve monitoring coverage; when ADI is low, the point density is close to the initial value. , maintain the existing points. By calculating the abnormality of ADI real-time prediction monitoring data, the system can dynamically adjust the collection frequency f and point density ρ in the medium accuracy area. , that is, if the abnormality exceeds the preset threshold α, then f and ρ will be increased; if , then maintain or restore the initial value to reduce resource consumption.

[0042] In this embodiment, first, the farmland is divided into multiple monitoring areas according to the density of the distribution of farmland sensors, and the deviation characteristics of the monitoring data of adjacent sensor pairs in each area and the fluctuation characteristics of the crop growth environment are obtained; then, the accuracy weights of the sensor monitoring data in different time periods in each monitoring area are calculated based on these deviations and fluctuation characteristics, and the weights of each area are weighted averaged to obtain a comprehensive accuracy index of the data analysis system for the actual environmental conditions; then, the monitoring area is divided into high, medium and low accuracy monitoring areas according to the accuracy index, and corresponding management measures are implemented; finally, the degree of abnormality of the data accuracy in the medium accuracy monitoring area is further predicted, and the sensor data acquisition frequency and distribution density in the area are dynamically adjusted according to the prediction results, thereby improving the monitoring accuracy and decision reliability of the system.

[0043] Embodiment 2, a smart village integrated management system described in this embodiment includes a data collection module, an accuracy assessment module, a regional division management module and an optimization module; Data acquisition module: According to the geographical density of sensors in the farmland, it is divided into several monitoring areas, and the monitoring data deviation characteristics between adjacent sensor pairs in different time periods in each monitoring area are obtained, as well as the fluctuation characteristics of crop growth environment data monitored by the same type of sensors in different time periods; Accuracy evaluation module: Determine the accuracy weight of the sensor monitoring data in different time periods in each monitoring area according to the acquired monitoring data deviation characteristics and crop growth environment data fluctuation characteristics, and calculate the weighted average of the accuracy weight of the sensor monitoring data in different time periods in each monitoring area to obtain the accuracy index of the data analysis system reflecting the actual environmental conditions; Area division management module: divides the monitoring area according to the accuracy index of the calculated data analysis system reflecting the actual environmental conditions, and divides the monitoring area into high-accuracy monitoring area, medium-accuracy monitoring area and low-accuracy monitoring area, and manages them accordingly; Optimization module: For medium-accuracy monitoring areas, further prediction is made on the degree of abnormality in the accuracy of the data analysis system in reflecting the actual environmental conditions, and the data collection frequency and distribution density of sensors in the medium-accuracy monitoring areas are dynamically adjusted based on the prediction results.

[0044] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0045] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0046] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0047] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A smart village comprehensive management method, characterized by: The steps include: S1: According to the geographical density of sensors in the farmland, the farmland is divided into several monitoring areas, and the monitoring data deviation characteristics between adjacent sensor pairs in different time periods in each monitoring area are obtained, as well as the fluctuation characteristics of crop growth environment data monitored by the same type of sensors in different time periods; S2: Determine the accuracy weight assignment of the sensor monitoring data in different time periods in each monitoring area according to the acquired monitoring data deviation characteristics and crop growth environment data fluctuation characteristics, and calculate the weighted average of the accuracy weight assignment of the sensor monitoring data in different time periods in each monitoring area to obtain the accuracy index of the data analysis system reflecting the actual environmental conditions; S3: Divide the monitoring area into high-accuracy monitoring area, medium-accuracy monitoring area and low-accuracy monitoring area according to the calculated accuracy index of the data analysis system reflecting the actual environmental conditions, and manage them accordingly; S4: For the medium-accuracy monitoring area, further prediction is made on the abnormal degree of accuracy of the data analysis system in reflecting the actual environmental conditions, and the data collection frequency and distribution density of the sensors in the medium-accuracy monitoring area are dynamically adjusted according to the prediction results.

2. According to claim 1, a smart village integrated management method is characterized by: In S2, the deviation characteristics of the acquired monitoring data are analyzed to generate a monitoring value deviation anomaly index. The method for obtaining the monitoring value deviation anomaly index is as follows: Obtain monitoring data from each pair of adjacent sensors, calculate the deviation between each pair of adjacent sensors, arrange the deviation values ​​of each pair of adjacent sensors at each time point in chronological order, and generate time series data , where t represents the time point in the time series; select the ARIMA model parameters. The ARIMA model is represented by ARIMA(p,d,q), where: p represents the number of autoregressive terms, d represents the number of differences, and q represents the number of moving average terms. If the time series is not stable, perform a difference operation according to the parameter d to obtain a stable time series. , the expression is: ; Use the optimized parameters p, d, and q to train the ARIMA model, use historical data to fit the ARIMA model, and predict the future monitoring value deviation based on the trained ARIMA model. Set the prediction value to , then the prediction formula of the model at time t is: ;in: is the coefficient of the autoregressive term, is the coefficient of the moving average term, is the residual at the previous time point, calculate the residual , the residual represents the difference between the actual deviation value and the predicted value, and the calculation formula is: ; Standardize the residuals and convert them into standardized residuals , the expression is: ;in is the standard deviation of the residual. The threshold range of the residual under normal conditions is set through historical data. When the standardized residual exceeds the threshold range, the deviation is judged to be abnormal. The frequency of abnormal occurrence in the T time period is calculated. Let Indicates the total number of time points, The number of abnormal time points is used to calculate the abnormal index of the monitoring value deviation. The expression is: ; Where AE is the abnormal index of monitoring value deviation.

3. A smart village integrated management method according to claim 2, characterized in that: In S2, the obtained crop growth environment data fluctuation characteristics are analyzed to generate an environmental data fluctuation index. The environmental data fluctuation index is obtained by: Get the time series of environmental data within a period of time, set the data to ,in Represents the observed value at time n, selects the observation time interval, and divides the data sequence X into several segments , each segment is an independent subsequence and obeys a normal distribution with different mean and variance. For each segment of data, set the mean and variance Normal distribution ,for and For the prior , choose the conjugate priors of the normal distribution and the inverse gamma distribution: ;in, is a priori parameter. Based on the observed data X, the posterior probability of each possible change point position is calculated. For a given set of change points , the posterior probability Represents the probability of change, and uses the Bayesian formula to update the posterior probability: ;in: is the likelihood of the data at a given change point, and calculates the probability that each segment of data follows the corresponding normal distribution. P(C) is the prior probability, which represents the prior distribution of the change point position. For each segment of data , using its standard deviation Represents the fluctuation range, and the expression is: ;in, It is a section The mean of Represents the number of data points in a segment. Based on the fluctuation range of each segment, the environmental data fluctuation index is defined. The calculation expression is: ; GH is the environmental data fluctuation index.

4. The method for integrated management of smart villages according to claim 3, characterized in that: In S2, the monitoring value deviation anomaly index and the environmental data fluctuation index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as inputs of the machine learning model. The machine learning model predicts the accuracy weighted labels of the sensor monitoring data in each monitoring area in different time periods with each group of comprehensive feature vectors as the prediction target, and minimizes the sum of the prediction errors of the accuracy weighted labels of the sensor monitoring data in different time periods in all monitoring areas as the training target. The machine learning model is trained until the sum of the prediction errors converges and the model training is stopped. The accuracy weighted values ​​of the sensor monitoring data in different time periods in each monitoring area are determined according to the output results of the model. The machine learning model is a polynomial regression model, and the accuracy index of the data analysis system reflecting the actual environmental conditions is obtained by weighted average calculation of the accuracy weighted values ​​of the sensor monitoring data in different time periods in each monitoring area.

5. A smart village integrated management method according to claim 4, characterized in that: In S3, the monitoring area is divided according to the calculated accuracy index of the data analysis system reflecting the actual environmental conditions, and the monitoring area is divided into a high-accuracy monitoring area, a medium-accuracy monitoring area and a low-accuracy monitoring area, specifically: The acquired accuracy index is compared with a gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and the accuracy index is compared with the first standard threshold and the second standard threshold respectively; If the accuracy index is greater than the second standard threshold, it means that the data analysis system reflects the actual environmental conditions with high accuracy, and a high accuracy signal is generated, and the corresponding monitoring area is divided into a high accuracy monitoring area; If the accuracy index is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the accuracy of the data analysis system in reflecting the actual environmental conditions is average, and a medium accuracy signal is generated, and the corresponding monitoring area is divided into a medium accuracy monitoring area; If the accuracy index is less than the first standard threshold, it means that the data analysis system has low accuracy in reflecting the actual environmental conditions. In this case, a low accuracy signal is generated and the corresponding monitoring area is divided into a low accuracy monitoring area.

6. The method for integrated management of smart villages according to claim 1, characterized in that: In S4, for the medium accuracy monitoring area, the abnormal degree of accuracy of the data analysis system in reflecting the actual environmental conditions is further predicted, and the data collection frequency and point density of the sensors in the medium accuracy monitoring area are dynamically adjusted according to the prediction results, specifically: For the medium accuracy monitoring area, that is, the accuracy index generated within a fixed time period is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the accuracy index generated in the subsequent fixed time period that is greater than or equal to the first standard threshold and less than or equal to the second standard threshold is collected, and the corresponding data set is established, and the standard deviation of the data set is calculated as the accuracy anomaly index, which is used as the basis for dynamically adjusting the sensor data collection frequency and point density.

7. A smart village integrated management method according to claim 6, characterized in that: The frequency of data collection is adjusted based on the accuracy anomaly index ADI, and the initial data collection frequency is defined as , and adjust the acquisition frequency according to ADI, the expression is: ; Where: f is the data acquisition frequency after dynamic adjustment, is the initial data collection frequency, is the acquisition frequency adjustment coefficient; Set the distribution density adjustment formula, and calculate the target distribution density. The expression is: ; Where: ρ is the density of points after dynamic adjustment, is the initial point density, is the point density adjustment coefficient; By calculating the abnormal degree of ADI real-time prediction monitoring data, the acquisition frequency f and point density ρ in the medium accuracy area are dynamically adjusted. , that is, if the abnormality exceeds the preset threshold α, then f and ρ will be increased; if , then maintain or restore the initial value to reduce resource consumption.

8. A smart village integrated management system, used to implement a smart village integrated management method according to any one of claims 1 to 7, characterized in that: It includes data collection module, accuracy assessment module, area division management module and optimization module; Data acquisition module: According to the geographical density of sensors in the farmland, it is divided into several monitoring areas, and the monitoring data deviation characteristics between adjacent sensor pairs in different time periods in each monitoring area are obtained, as well as the fluctuation characteristics of crop growth environment data monitored by the same type of sensors in different time periods; Accuracy evaluation module: Determine the accuracy weight of the sensor monitoring data in different time periods in each monitoring area according to the acquired monitoring data deviation characteristics and crop growth environment data fluctuation characteristics, and calculate the weighted average of the accuracy weight of the sensor monitoring data in different time periods in each monitoring area to obtain the accuracy index of the data analysis system reflecting the actual environmental conditions; Area division management module: divides the monitoring area according to the accuracy index of the calculated data analysis system reflecting the actual environmental conditions, and divides the monitoring area into high-accuracy monitoring area, medium-accuracy monitoring area and low-accuracy monitoring area, and manages them accordingly; Optimization module: For medium-accuracy monitoring areas, further prediction is made on the degree of abnormality in the accuracy of the data analysis system in reflecting the actual environmental conditions, and the data collection frequency and distribution density of sensors in the medium-accuracy monitoring areas are dynamically adjusted based on the prediction results.

Images