Intelligent query method and system for water and soil conservation measures based on multidimensional parameterization

Abstract

The invention relates to the technical field of intelligent environment recognition, and particularly discloses a water and soil conservation measure intelligent query method and system based on multi-dimensional parameterization. Temperature and humidity data are collected in real time through a multi-source sensor deployed in water and soil, and a temperature stability characteristic value is calculated by adopting zero-sequence processing and fast Fourier transform; a humidity change rate characteristic value is extracted by combining first-order difference processing and Haar wavelet transform; the feature values are constructed into a comprehensive environment state feature vector, and the comprehensive environment state feature vector is input into a trained random forest model for environment recognition result evaluation; according to the method, the sensing technology, signal analysis and machine learning prediction are fused, the sensing ability, the regulation and control precision and the self-adaptability of the system are improved, the method is suitable for complex and changeable environment scenes, and the method has the advantages of being high in adaptability, high in adaptability and the like. Good application prospects are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent environment recognition, and particularly relates to a multi-dimensional parameter-based intelligent query method and system for soil and water conservation measures. BACKGROUND

[0002] Soil and water conservation measures refer to a series of measures taken to protect water resources, maintain and improve the integrity and fertility of cultivated land, reduce water and soil loss and environmental pollution, and improve the sustainable utilization of land. According to different goals and needs, soil and water conservation measures can be divided into various types. Different soil and water conservation measures should be suitable for different environmental characteristics, and therefore, corresponding query measures need to be generated, so that when environmental parameters are input, the appropriate soil and water conservation measures under the environment can be obtained.

[0003] The prior art has the following disadvantages:

[0004] Although the current soil and water conservation measure query system has a certain level of intelligence in terms of hardware configuration and basic control logic, there are still many problems in actual application. First, most systems rely only on simple threshold judgment for regulation and control, lack in-depth analysis of environmental change trends, and are difficult to accurately identify temperature fluctuation stability and humidity abnormal change characteristics, resulting in frequent response lags or misadjustment phenomena. Second, traditional methods often ignore the time series characteristics of environmental data, fail to effectively extract key feature information, and limit the improvement of regulation and control precision and prediction ability. In addition, the existing query system generally lacks a dynamic adaptive mechanism and cannot flexibly adjust the operation strategy according to seasonal changes, weather mutations and other factors, resulting in insufficient query accuracy. Therefore, there is an urgent need for a new regulation and control method that integrates multi-source sensor data, deep feature extraction and intelligent prediction model to achieve efficient, stable and intelligent query of soil and water conservation measures. SUMMARY

[0005] The present application relates to the technical field of intelligent environment recognition, and particularly relates to a multi-dimensional parameter-based intelligent query method and system for soil and water conservation measures.

[0006] The object of the present application can be achieved by the following technical solutions:

[0007] The multi-dimensional parameter-based intelligent query method for soil and water conservation measures comprises the following steps:

[0008] S1: Real-time acquisition of temperature data and humidity data through multi-source sensors deployed in the soil and water environment;

[0009] S2: Zero sequence processing of the acquired temperature data, construction of a zero sequence temperature sequence, calculation of a temperature stability feature value for evaluating the stability of the current environmental temperature fluctuation;

[0010] S3: first-order difference processing is performed on the humidity data, a differential humidity sequence is constructed, a humidity change rate characteristic value is calculated, and whether the environmental humidity has a potential anomaly is identified;

[0011] S4: the temperature stability characteristic value and the humidity change rate characteristic value are constructed into a comprehensive environmental state characteristic vector, and input into a trained machine learning model for environmental recognition result evaluation;

[0012] S5: the query control parameters are dynamically adjusted according to the evaluation result, and adaptive query of water and soil environmental protection measures is realized.

[0013] As a further scheme of the present application, the evaluation of the stability of the current environmental temperature fluctuation specifically comprises:

[0014] Through the temperature sensor deployed in the water and soil environment, temperature data is collected in real time according to the time sequence, the collected temperature data is subjected to zero sequence processing, a zero sequence temperature sequence is constructed, the temperature stability characteristic value is calculated according to the fluctuation degree of the zero sequence temperature sequence, and whether the temperature stability characteristic value is greater than or equal to a preset threshold value is judged, if yes, the current environmental temperature fluctuation is unstable, and if not, the current environmental temperature fluctuation is stable.

[0015] As a further scheme of the present application, the acquisition process of the temperature stability characteristic value is:

[0016] The real-time collected temperature data is obtained and integrated into a temperature time sequence, a linear regression model is used to fit the trend item of the temperature time sequence, specifically, the least square method is used to fit a straight line between the time point and the corresponding temperature value, and the best fitting straight line equation is obtained; then the trend value corresponding to each time point is calculated according to the fitting straight line equation, and the corresponding trend value is subtracted from the original temperature time sequence point by point, so as to obtain the temperature fluctuation sequence after removing the linear trend, that is, the zero sequence temperature sequence;

[0017] The temperature sequence after zero sequence processing is subjected to fast Fourier transform to obtain a frequency domain signal sequence, the frequency resolution is determined and the frequency interval is divided according to the sampling frequency and the number of data points, and the target frequency range is set;

[0018] The sum of the power spectral density of each frequency component in the target frequency range is calculated to obtain the temperature stability characteristic value.

[0019] As a further scheme of the present application, the identification of whether the environmental humidity has a potential anomaly specifically comprises:

[0020] The humidity sensor is arranged in the water and soil environment, humidity data is collected in time sequence in real time, the humidity data is processed by first-order difference, a differential humidity sequence is constructed, a humidity change rate characteristic value is calculated according to the change degree of the differential humidity sequence, whether the humidity change rate characteristic value is greater than or equal to a preset threshold is judged, if yes, the environmental humidity has potential abnormality, and if no, the environmental humidity has no potential abnormality.

[0021] As a further scheme of the present application, the process of obtaining the humidity change rate characteristic value is:

[0022] The humidity data collected in real time is obtained and integrated into a humidity time sequence, the humidity time sequence is processed by first-order difference, specifically, each humidity value in the humidity time sequence is subtracted by the humidity value at the previous time point, that is, the humidity difference value of any two adjacent time points is integrated to obtain a differential humidity sequence.

[0023] The differential humidity sequence is analyzed by using Haar wavelet transform, Haar wavelet is selected as the wavelet base function, and the differential humidity sequence is decomposed by multiple scales to generate a group of wavelet coefficients under different scales.

[0024] According to the result of Haar wavelet transform, the square sum of all wavelet coefficients under each scale is accumulated to obtain the humidity change rate characteristic value.

[0025] As a further scheme of the present application, the temperature stability characteristic value and the humidity change rate characteristic value are constructed into a comprehensive environmental state characteristic vector, and input into a trained machine learning model, specifically including:

[0026] The temperature stability characteristic value and the humidity change rate characteristic value of the water and soil are obtained, the temperature stability characteristic value and the humidity change rate characteristic value are constructed into a comprehensive environmental state characteristic vector, which is input into a machine learning model, the error between the predicted environmental recognition result score and the actual environmental recognition result score is minimized as the training target, the machine learning model is trained, and the environmental recognition result score is output according to the trained machine learning model, and the machine learning model is a random forest model.

[0027] As a further scheme of the present application, the training process of the machine learning model is:

[0028] The comprehensive environmental state feature vectors extracted within multiple sampling periods are used as input samples, and the actual environmental recognition result scores within the corresponding time period are used as target output labels to construct a training data set. The random forest model is used as the prediction model. The model consists of multiple decision trees. The training samples of each tree are generated from the training data set through the replacement sampling method, and the out-of-bag error is used to evaluate the generalization ability of the model. During the training process, the objective function is to minimize the mean square error between the environmental recognition result scores predicted by the model and the actual scores. By optimizing the decision tree node splitting criterion and the number of decision trees in the forest, the model prediction accuracy is gradually improved. Finally, the random forest model after training can receive the new comprehensive environmental state feature vector as input and output the corresponding environmental recognition result score prediction value.

[0029] As a further solution of the present invention: the environment recognition result evaluation specifically includes:

[0030] Determine whether the environment recognition result score of the environment recognition is greater than or equal to the preset threshold. If so, the environment recognition result is normal; if not, the environment recognition result is abnormal.

[0031] As a further solution of the present invention, the method of dynamically adjusting the query control parameters according to the evaluation results to realize the adaptive query of the water and soil environmental protection measures specifically includes:

[0032] If the result of environmental identification is judged to be normal, the query control parameter will not be adjusted. If the result of environmental identification is judged to be abnormal, the query control parameter will be adjusted, and the first average change rate of the temperature stability characteristic value over time will be calculated, and the second average change rate of the humidity change rate characteristic value over time will be calculated. The weight of the humidity in the query control parameter is set to the second average change rate, and the weight of the temperature in the query control parameter is set to the first average change rate. The adjusted query control parameters are used to query soil and water conservation measures.

[0033] The intelligent query system for soil and water conservation measures based on multi-dimensional parameterization includes:

[0034] A data acquisition module, which collects temperature and humidity data in real time through multi-source sensors deployed in the water and soil environment;

[0035] An ambient temperature assessment module, which performs zero-sequence processing on the collected temperature data, constructs a zero-sequence temperature sequence, and calculates a temperature stability characteristic value for evaluating the stability of the current ambient temperature fluctuation;

[0036] An environmental humidity evaluation module, which performs first-order difference processing on the humidity data, constructs a differential humidity sequence, calculates a humidity change rate characteristic value, and is used to identify whether there is a potential abnormality in the environmental humidity;

[0037] An environmental recognition result evaluation module, which constructs a comprehensive environmental state characteristic vector from the temperature stability characteristic value and the humidity change rate characteristic value, and inputs the vector into a trained machine learning model for environmental recognition result evaluation;

[0038] A query adjustment module, which dynamically adjusts query control parameters according to the evaluation result, and realizes adaptive query of water and soil environmental protection measures.

[0039] Beneficial results of the present application:

[0040] The present application uses high-precision and strong anti-interference temperature and humidity sensors to collect temperature and humidity data in the water and soil environment in real time, and performs deep feature extraction on the original data based on advanced multi-dimensional signal processing technology. Specifically, for temperature data, first, a linear regression model is used to fit and deduct the trend item of the time series to construct a zero-order temperature sequence, and then a fast Fourier transform is used to extract frequency domain features and calculate a temperature stability characteristic value, thereby realizing quantitative evaluation of temperature fluctuation stability. For humidity data, first-order difference processing is performed on the time series to remove the long-term trend, and a Haar wavelet transform is introduced for multi-scale decomposition to further extract energy features at each scale, and finally a humidity change rate characteristic value reflecting the degree of short-term humidity change is obtained. The above two key characteristic values are fused into a representative comprehensive environmental state characteristic vector, which is input into a trained random forest model as an input variable to predict the current environmental recognition result score. The model takes the minimization of the mean square error between the predicted score and the actual score as the objective function, has good generalization ability and prediction accuracy. By comparing the predicted result with the set threshold, the system can dynamically adjust the query control parameters to realize closed-loop adaptive environmental recognition. Compared with traditional methods, the present application not only improves the perception ability and response speed of complex environmental changes, but also significantly enhances the intelligent level and decision reliability of the control system, truly realizes the leap from "passive adjustment" to "active perception-precise prediction-automatic optimization", and has significant technological progress and application value. BRIEF DESCRIPTION OF DRAWINGS

[0041] The present application will be further described below with reference to the accompanying drawings.

[0042] Figure 1 is a flowchart of the intelligent query method of the present application based on multi-dimensional parameterization of water and soil conservation measures;

[0043] Figure 2 is a flow chart of the intelligent query system for soil and water conservation measures based on multi-dimensional parameterization in the present application. DETAILED DESCRIPTION

[0044] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0045] Please refer to Figure 1 The present application is an intelligent query method for soil and water conservation measures based on multi-dimensional parameterization, which comprises the following steps:

[0046] S1: Real-time collection of temperature data and humidity data through multi-source sensors deployed in the soil and water environment;

[0047] S2: Zero sequence processing of the collected temperature data to construct a zero sequence temperature sequence and calculate a temperature stability eigenvalue for evaluating the stability of the current environmental temperature fluctuation;

[0048] S3: First-order difference processing of the humidity data to construct a difference humidity sequence and calculate a humidity change rate eigenvalue for identifying whether there is a potential abnormality in the environmental humidity;

[0049] S4: Construction of a comprehensive environmental state feature vector from the temperature stability eigenvalue and the humidity change rate eigenvalue, and inputting the trained machine learning model for environmental recognition result evaluation;

[0050] S5: Dynamic adjustment of the query control parameters according to the evaluation results to realize adaptive query of the soil and water environment protection measures.

[0051] In S1, real-time collection of temperature data and humidity data through multi-source sensors deployed in the soil and water environment, specifically including:

[0052] The multi-source sensors include but are not limited to temperature sensors and humidity sensors, wherein the temperature sensors are used to collect air temperature information in the soil and water environment, and the humidity sensors are used to synchronously obtain the relative humidity value in the air. Each sensor is distributed in different areas according to a preset spatial layout to ensure that the collected data is representative and consistent, thereby fully reflecting the current environmental state.

[0053] Each sensor establishes a connection with the central control module through an Internet of Things communication protocol, enabling periodic uploading and centralized processing of temperature and humidity data. The sensor device has high precision and anti-interference characteristics, and the sampling frequency can be dynamically adjusted according to actual needs to adapt to different application scenarios' data update rate requirements. The collected raw temperature and humidity data form time series respectively, and are transmitted to the subsequent data processing module, providing basic data support for constructing zero sequence temperature sequence, differential humidity sequence, and extracting temperature stability characteristic value and humidity change rate characteristic value.

[0054] In S2, the collected temperature data is subjected to zero sequence processing to construct a zero sequence temperature sequence and calculate a temperature stability characteristic value for evaluating the stability of the current environmental temperature fluctuation, specifically including:

[0055] By deploying temperature sensors in the water and soil environment, real-time temperature data is collected in time series, the collected temperature data is subjected to zero sequence processing to construct a zero sequence temperature sequence, and a temperature stability characteristic value is calculated based on the fluctuation degree of the zero sequence temperature sequence to determine whether the temperature stability characteristic value is greater than or equal to a preset threshold. If yes, the current environmental temperature fluctuation is unstable, and if no, the current environmental temperature fluctuation is stable.

[0056] The process of obtaining the temperature stability characteristic value is as follows:

[0057] The real-time collected temperature data is obtained and integrated into a temperature time series. A linear regression model is used to fit the trend item of the temperature time series. Specifically, the least squares method is used to fit a straight line between the time points and the corresponding temperature values to obtain the best fitting straight line equation. Then, the trend value corresponding to each time point is calculated based on the fitting straight line equation, and the corresponding trend value is subtracted from the original temperature time series point by point to obtain the temperature fluctuation sequence after removing the linear trend, i.e., the zero sequence temperature sequence.

[0058] Finally, the zero sequence temperature sequence is subjected to stationarity test. If it meets the preset trendless standard, it is output as the basis data for subsequent frequency domain analysis and temperature stability evaluation. Otherwise, the fitting model parameters are adjusted or a higher order trend fitting method is used for reprocessing until the required zero sequence temperature sequence is obtained.

[0059] The zero sequence processed temperature sequence is subjected to fast Fourier transform to obtain a frequency domain signal sequence. Based on the sampling frequency and the number of data points, the frequency resolution is determined, and the frequency interval is divided, where n represents the number of data points, f s represents the sampling frequency, and the target frequency range [f low , f high ] is set, where f low represents the low frequency cutoff frequency, and f highdenotes the high frequency cutoff frequency, corresponding to the index range [k,m] in the frequency domain signal, wherein,

[0060] The sum of the power spectral densities of each frequency component in the target frequency range is calculated to obtain a temperature stability characteristic value, and the calculation expression is:

[0061]

[0062] In the formula, S represents the temperature stability characteristic value, i represents the number of frequency components, |f i | represents the absolute value of the amplitude of the i-th frequency component.

[0063] It should be noted that: the temperature sensor deployed in the water and soil environment collects temperature data in real time, and a linear regression model is used to fit the trend item of the original temperature time series, and the least squares method is used to obtain the best fitting straight line, and then the trend item is deducted point by point, so as to build the zero sequence temperature sequence which removes the long-term trend influence; Further, the zero sequence temperature sequence is subjected to fast Fourier transform to obtain its frequency domain signal distribution, and a target frequency band in a specific frequency range is set, and the sum of the power spectral densities of each frequency component in the frequency band is calculated to calculate the quantitative index temperature stability characteristic value reflecting the temperature fluctuation stability. By comparing the temperature stability characteristic value with the preset threshold value, it can be judged whether the current environmental temperature is in a stable state, so as to provide a scientific basis for subsequent regulation and decision-making. The technical scheme effectively solves the technical defects that the traditional method cannot distinguish between short-term fluctuations and long-term trends and cannot quantitatively evaluate the temperature stability, and has the advantages of high data processing accuracy, objective and reliable evaluation results, strong adaptability and the like.

[0064] In S3, the humidity data is subjected to first-order difference processing to build a difference humidity sequence, and a humidity change rate characteristic value is calculated to identify whether there is a potential abnormality in the environmental humidity, specifically including:

[0065] Through the humidity sensor deployed in the water and soil environment, humidity data is collected in real time according to the time sequence, the humidity data is subjected to first-order difference processing to build a difference humidity sequence, and a humidity change rate characteristic value is calculated according to the change degree of the difference humidity sequence, to judge whether the humidity change rate characteristic value is greater than or equal to a preset threshold value, if yes, the environmental humidity has a potential abnormality, and if no, the environmental humidity does not have a potential abnormality;

[0066] The humidity data collected in real time is obtained and integrated into a humidity time sequence, and the humidity time sequence is subjected to first-order difference processing, specifically, each humidity value in the humidity time sequence is subtracted by the humidity value at the previous time point, that is, for the humidity difference value of any two adjacent time points, all humidity difference values are integrated to obtain a difference humidity sequence;

[0067] The differential humidity sequence is analyzed by using Haar wavelet transform, Haar wavelet is selected as the wavelet base function, and the differential humidity sequence is multi-scale decomposed to generate a group of wavelet coefficients at different scales. The Haar wavelet transform is a simple form of discrete wavelet transform, which realizes multi-level signal representation by recursively decomposing the signal into approximation and detail parts;

[0068] According to the results of the Haar wavelet transform, the total energy sum of the wavelet coefficients in the selected scale range is calculated as the humidity change rate characteristic value. Specifically, the square sum of all wavelet coefficients at each scale is accumulated to obtain the humidity change rate characteristic value. For example, in the first layer decomposition, the energy sum of the detail coefficients is calculated, and the approximation coefficients are further decomposed and energy calculated until the predetermined decomposition layer is reached.

[0069] Finally, the calculated humidity change rate characteristic value is output for subsequent quantitative analysis and monitoring of environmental humidity changes. This process not only effectively captures the short-term fluctuation characteristics of humidity data, but also focuses on different periodic change characteristics by adjusting the decomposition layer number of Haar wavelet transform, thereby providing more accurate evaluation of humidity change rate characteristic value.

[0070] It should be noted that: the humidity sensor deployed in the water and soil environment collects humidity data in real time, and constructs a humidity time series; then it is first-order differential processed to eliminate the long-term trend influence and highlight the short-term fluctuation characteristics, forming a differential humidity sequence; further, the Haar wavelet transform is used to multi-scale decompose the differential humidity sequence, extract wavelet coefficients at different frequency scales, and calculate the humidity change rate characteristic value by accumulating the square sum of wavelet coefficients at each scale, thereby realizing quantitative evaluation of the humidity change rate. By comparing the humidity change rate characteristic value with the preset threshold, it can effectively judge whether there is humidity abnormal fluctuation in the current environment, and then provide basis for subsequent control strategy. This technical solution not only improves the sensitivity and accuracy of humidity anomaly identification, but also uses the multi-resolution analysis capability of Haar wavelet transform to realize fine description of different periodic fluctuation characteristics, has the advantages of fast response speed, low computational complexity, strong adaptability, etc., significantly improves the identification ability of existing environmental monitoring system to humidity abnormal state.

[0071] In S4, the temperature stability characteristic value and the humidity change rate characteristic value are constructed into a comprehensive environmental state characteristic vector, which is input into the trained machine learning model for environmental recognition result evaluation, specifically including:

[0072] The temperature stability characteristic value and the humidity change rate characteristic value are obtained, the temperature stability characteristic value and the humidity change rate characteristic value are constructed into a comprehensive environmental state characteristic vector, which is used as an input of a machine learning model, and a minimum error between a predicted environmental recognition result score and an actual environmental recognition result score is taken as a training target to train the machine learning model, and an environmental recognition result score is output according to the trained machine learning model, and the machine learning model is a random forest model;

[0073] The comprehensive environmental state characteristic vectors extracted in multiple sampling periods are taken as input samples, and the actual environmental recognition result scores in the corresponding time periods are taken as target output labels to construct a training data set, a random forest model is used as a prediction model, the model is composed of multiple decision trees, training samples of each tree are generated from the training data set by a sampling method with replacement, and the generalization ability of the model is evaluated by using out-of-bag error, in the training process, a mean square error between the environmental recognition result score predicted by the model and the actual score is taken as an objective function, the prediction accuracy of the model is gradually improved by optimizing the decision tree node splitting criterion and the number of decision trees in the forest, and finally, the trained random forest model can receive a new comprehensive environmental state characteristic vector as an input and output a corresponding environmental recognition result score prediction value.

[0074] The environmental recognition result evaluation specifically includes:

[0075] It is judged whether the environmental recognition result score of the environmental recognition is greater than or equal to a preset threshold, if yes, the environmental recognition result is normal, and if no, the environmental recognition result is abnormal.

[0076] In S5, the query control parameters are dynamically adjusted according to the evaluation result to realize adaptive query of the water and soil environmental protection measures, and specifically includes:

[0077] If the result of the environmental recognition is determined to be normal, the query control parameters are not adjusted, if the result of the environmental recognition is determined to be abnormal, the query control parameters are adjusted, a first average change rate of the temperature stability characteristic value changing with time is calculated, a second average change rate of the humidity change rate characteristic value changing with time is calculated, the weight of humidity in the query control parameters is set to the second average change rate, the weight of temperature in the query control parameters is set to the first average change rate, and the query control parameters are used to query the water and soil conservation measures.

[0078] Referring to Figure 2 The intelligent query system of water and soil conservation measures based on multi-dimensional parameterization includes:

[0079] The data acquisition module acquires temperature data and humidity data in real time through the multi-source sensors deployed in the water and soil environment.

[0080] An ambient temperature evaluation module, which carries out zero sequence processing on the collected temperature data, constructs a zero sequence temperature sequence, calculates a temperature stability characteristic value, and is used for evaluating the stability of the current ambient temperature fluctuation;

[0081] An ambient humidity evaluation module, which carries out first-order difference processing on the humidity data, constructs a difference humidity sequence, calculates a humidity change rate characteristic value, and is used for identifying whether there is a potential abnormality in the ambient humidity;

[0082] An ambient recognition result evaluation module, which constructs a comprehensive ambient state characteristic vector from the temperature stability characteristic value and the humidity change rate characteristic value, inputs the trained machine learning model, and carries out ambient recognition result evaluation;

[0083] A query adjustment module, which dynamically adjusts the query control parameters according to the evaluation result, and realizes adaptive query of the water and soil environment protection measures.

[0084] The working principle of the present application: the present application collects water and soil temperature and humidity data through multi-source sensors, and realizes real-time perception of the environment state by combining signal processing and machine learning technology. The method includes five core steps: through the deployment of temperature sensors and humidity sensors in water and soil, real-time collection of environmental parameters according to the preset spatial layout and sampling frequency, forming representative temperature and humidity time series; the temperature data is processed by zero sequence, the trend item is removed by using linear regression model, the zero sequence temperature sequence is constructed, the frequency domain features are extracted by fast Fourier transform, the temperature stability characteristic value is calculated, and whether the current environment temperature is stable is judged; the humidity data is processed by first-order difference, the difference humidity sequence is constructed, and the humidity change rate characteristic value is calculated by using Haar wavelet transform for multi-scale decomposition, whether there is humidity abnormal fluctuation is identified; the above two characteristic values are constructed into a comprehensive ambient state characteristic vector, input into the trained random forest model, the environment recognition result score is predicted, and the effectiveness of the current control strategy is evaluated; the query control parameters are dynamically adjusted according to the evaluation result, and adaptive query of the water and soil environment protection measures is realized. The present application introduces high-precision data acquisition, time-frequency analysis and machine learning prediction mechanism, improves the intelligent level and response ability of the query system, solves the problem of insufficient query precision caused by the general lack of dynamic adaptive mechanism in the existing query system, and has good application prospect and popularization value.

[0085] The above formulas are dimensionless values, and the formulas are obtained by software simulation of a large amount of data to obtain a formula of the latest real situation, and the preset parameters in the formula are set by the person skilled in the art according to the actual situation.

[0086] The above-described embodiments can be implemented in part or in whole through software, hardware, firmware or any combination thereof. When implemented in software, the above-described embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded into a computer, all or part of the processes described in the embodiments of the present application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions can be transferred from one website, computer, server or data center to another website, computer, server or data center through wired or wireless (such as infrared, wireless, microwave, etc.) manner. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. containing one or more available medium collections. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid state disk.

[0087] It should be understood that the term "and / or" herein only describes the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B, which can represent three cases of A alone, A and B together, and B alone, where A and B can be singular or plural. In addition, the character " / " herein generally represents that the front and rear associated objects are in an "or" relationship, but can also represent an "and / or" relationship, which can be understood in the context.

[0088] It should be understood that in various embodiments of the present application, the size of the sequence number of the above-mentioned processes does not mean the order of execution, and the execution order of the processes should be determined by its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0089] The above describes one embodiment of the present application in detail, but the content described is only the preferred embodiment of the present application, and cannot be considered as limiting the scope of the implementation of the present application. Any equivalent changes and improvements made within the scope of the present application are still within the scope of the patent coverage of the present application.

Claims

1. An intelligent query method for soil and water conservation measures based on multidimensional parameterization, characterized by: The following steps are involved: S1: Real-time temperature and humidity data are collected through multi-source sensors deployed in the water and soil environment; S2: Perform zero-sequence processing on the collected temperature data, construct a zero-sequence temperature sequence, and calculate the temperature stability characteristic value to evaluate the stability of the current ambient temperature fluctuation; S3: Perform first-order difference processing on the humidity data, construct a differential humidity series, and calculate the characteristic value of the humidity change rate to identify whether there is a potential anomaly in the ambient humidity; S4: The temperature stability eigenvalue and the humidity change rate eigenvalue are constructed into a comprehensive environmental state eigenvector, which is input into the trained machine learning model to evaluate the environmental recognition results. S5: Dynamically adjust query control parameters according to the evaluation results to achieve adaptive query of water and soil environmental protection measures.

2. The intelligent query method for soil and water conservation measures based on multidimensional parameterization according to claim 1 is characterized in that: The evaluation of the stability of the current ambient temperature fluctuation specifically includes: By deploying temperature sensors in the water and soil environment, temperature data is collected in real time according to the time series, and the collected temperature data is processed by zero sequence to construct a zero sequence temperature sequence. According to the fluctuation degree of the zero sequence temperature sequence, the temperature stability characteristic value is calculated to determine whether the temperature stability characteristic value is greater than or equal to the preset threshold. If so, the current ambient temperature fluctuation is unstable; if not, the current ambient temperature fluctuation is stable.

3. The intelligent query method for soil and water conservation measures based on multidimensional parameterization according to claim 2 is characterized in that: The process of obtaining the temperature stability characteristic value is as follows: Real-time temperature data is acquired and integrated into a temperature time series. A linear regression model is used to fit the trend term of the temperature time series. Specifically, the relationship between the time point and the corresponding temperature value is fitted using the least squares method to obtain the best fitting linear equation. The trend value corresponding to each time point is then calculated based on the fitted linear equation, and the corresponding trend value is subtracted point by point from the original temperature time series to obtain the temperature fluctuation series after removing the linear trend, namely the zero-sequence temperature series. Perform fast Fourier transform on the temperature sequence after zero-sequence processing to obtain a frequency domain signal sequence. According to the sampling frequency and the number of data points, the frequency resolution is determined and the frequency interval is divided to set the target frequency range. The sum of the power spectral density of each frequency component within the target frequency range is calculated to obtain the temperature stability characteristic value.

4. The intelligent query method for soil and water conservation measures based on multidimensional parameterization according to claim 1 is characterized in that: The identifying whether there is a potential abnormality in the ambient humidity specifically includes: By deploying humidity sensors in the water and soil environment, humidity data is collected in real time according to the time series, and the humidity data is processed by first-order difference to construct a differential humidity sequence. According to the degree of change of the differential humidity sequence, the characteristic value of the humidity change rate is calculated, and it is judged whether the characteristic value of the humidity change rate is greater than or equal to the preset threshold. If so, there is a potential anomaly in the environmental humidity. If not, there is no potential anomaly in the environmental humidity.

5. The intelligent query method for soil and water conservation measures based on multidimensional parameterization according to claim 4 is characterized in that: The process of obtaining the characteristic value of the humidity change rate is as follows: Obtain real-time humidity data and integrate it into a humidity time series. Perform first-order difference processing on the humidity time series. Specifically, subtract the humidity value of the previous time point from each humidity value in the humidity time series. That is, for the humidity difference between any two adjacent time points, all humidity differences are integrated to obtain a differential humidity series. The differential humidity series is analyzed by using Haar wavelet transform, and Haar wavelet is selected as the wavelet basis function. The differential humidity series is decomposed into multiple scales to generate a set of wavelet coefficients at different scales. According to the results of Haar wavelet transform, the square sum of all wavelet coefficients at each scale is accumulated to obtain the characteristic value of humidity change rate.

6. The intelligent query method for soil and water conservation measures based on multidimensional parameterization according to claim 1 is characterized in that: The temperature stability characteristic value and the humidity change rate characteristic value are constructed into a comprehensive environmental state characteristic vector and input into the trained machine learning model, specifically including: The temperature stability characteristic value and humidity change rate characteristic value of water and soil are obtained, and the temperature stability characteristic value and the humidity change rate characteristic value are constructed into a comprehensive environmental state characteristic vector as the input of the machine learning model. The machine learning model is trained with minimizing the error between the predicted environmental recognition result score and the actual environmental recognition result score as the training goal, and the environmental recognition result score is output according to the trained machine learning model. The machine learning model is a random forest model.

7. The intelligent query method for soil and water conservation measures based on multidimensional parameterization according to claim 6 is characterized in that: The training process of the machine learning model is: The comprehensive environmental state feature vectors extracted within multiple sampling periods are used as input samples, and the actual environmental recognition result scores within the corresponding time period are used as target output labels to construct a training data set. The random forest model is used as the prediction model. The model consists of multiple decision trees. The training samples of each tree are generated from the training data set through the replacement sampling method, and the out-of-bag error is used to evaluate the generalization ability of the model. During the training process, the objective function is to minimize the mean square error between the environmental recognition result scores predicted by the model and the actual scores. By optimizing the decision tree node splitting criterion and the number of decision trees in the forest, the model prediction accuracy is gradually improved. Finally, the random forest model after training can receive the new comprehensive environmental state feature vector as input and output the corresponding environmental recognition result score prediction value.

8. The intelligent query method for soil and water conservation measures based on multidimensional parameterization according to claim 1 is characterized in that: The environmental recognition result evaluation specifically includes: Determine whether the environment recognition result score of the environment recognition is greater than or equal to the preset threshold. If so, the environment recognition result is normal; if not, the environment recognition result is abnormal.

9. The intelligent query method for soil and water conservation measures based on multidimensional parameterization according to claim 1 is characterized in that: The method of dynamically adjusting the query control parameters according to the evaluation results to realize the adaptive query of the water and soil environmental protection measures specifically includes: If the result of environmental identification is judged to be normal, the query control parameter will not be adjusted. If the result of environmental identification is judged to be abnormal, the query control parameter will be adjusted, and the first average change rate of the temperature stability characteristic value over time will be calculated, and the second average change rate of the humidity change rate characteristic value over time will be calculated. The weight of the humidity in the query control parameter is set to the second average change rate, and the weight of the temperature in the query control parameter is set to the first average change rate. The adjusted query control parameters are used to query soil and water conservation measures.

10. The intelligent query system for soil and water conservation measures based on multi-dimensional parameterization is characterized by: The intelligent query method for soil and water conservation measures based on multidimensional parameterization according to any one of claims 1 to 9 comprises: A data acquisition module, which collects temperature and humidity data in real time through multi-source sensors deployed in the water and soil environment; An ambient temperature assessment module, which performs zero-sequence processing on the collected temperature data, constructs a zero-sequence temperature sequence, and calculates a temperature stability characteristic value for evaluating the stability of the current ambient temperature fluctuation; An environmental humidity assessment module, which performs first-order difference processing on humidity data, constructs a differential humidity sequence, and calculates a characteristic value of the humidity change rate to identify whether there is a potential anomaly in the environmental humidity; An environment recognition result evaluation module, which constructs a comprehensive environment state feature vector from the temperature stability feature value and the humidity change rate feature value, and inputs the vector into a trained machine learning model to evaluate the environment recognition result; A query adjustment module dynamically adjusts query control parameters according to the evaluation results to achieve adaptive query of water and soil environmental protection measures.

Images