Intelligent monitoring and dynamic early warning system for arable land quality for food security

Abstract

The application relates to the technical field of farmland monitoring, in particular to a farmland quality intelligent monitoring and dynamic early warning system for food safety, which comprises a memory and a processor, the processor executes a computer program stored in the memory to realize the following steps: land element monitoring data and land humidity monitoring data are acquired; according to the difference between the true value and the predicted value of the land element monitoring data at each time, the difference influence of the true value and the predicted value of the land humidity monitoring data is combined to determine element correction data; according to the difference between the predicted results of the land element monitoring data in all non-irrigation time periods and a preset monitoring period, data prediction evaluation indexes are analyzed; the global element predicted value at the current time is determined by using the data prediction evaluation indexes, and then the farmland quality is monitored. The application avoids the influence of irrigation operation and improves the accuracy of the monitoring data corresponding to the predicted results.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cultivated land monitoring, in particular to a cultivated land quality intelligent monitoring and dynamic early warning system for food safety. BACKGROUND

[0002] As the core carrier of food production, the quality of cultivated land directly determines the yield and quality of food. Intelligent monitoring technology helps farmers to reasonably fertilize and irrigate, and promotes the development of agriculture towards green and efficient and sustainable direction. Reasonable fertilization can improve soil fertility and avoid excessive soil nutrients; precise irrigation can save water resources and reduce water waste. At the same time, green agricultural production methods can help to improve the quality of food and other issues, improve the level of food quality and safety, and meet the needs of consumers for high-quality food.

[0003] With the development of remote sensing images, it has become a common method to obtain relevant chemical and physical parameters of food cultivated land quality through remote sensing inversion, which is widely used in the monitoring of cultivated land quality of large-scale food planting, and the required elements in the growth and development process of food are predicted, that is, the cultivated land quality is predicted, and the current safety state of food is evaluated.

[0004] However, during the initial planting of food, frequent irrigation and fertilization operations are required, and the elements in the cultivated land change frequently. The existing cultivated land quality monitoring method uses various cultivated soil elements collected by sensors in real time to evaluate the cultivated land quality by using a comprehensive index. However, in the process of real-time monitoring of cultivated soil element data by sensors, irrigation operation causes water to flow in the soil, which may cause fluctuations in the monitoring data of soil elements, making the evaluation results of real-time monitoring of cultivated land quality less accurate. SUMMARY

[0005] In order to solve the technical problem that the existing cultivated land quality monitoring method is affected by irrigation and other environmental factors, resulting in fluctuations in the monitoring data, and making the evaluation results of cultivated land quality less accurate, the purpose of the present application is to provide a cultivated land quality intelligent monitoring and dynamic early warning system for food safety, and the technical scheme adopted is as follows:

[0006] The cultivated land quality intelligent monitoring and dynamic early warning system for food safety comprises a storage and a processor, and the processor executes the computer program stored in the storage to realize the following steps:

[0007] Obtain the land element monitoring data and the land humidity monitoring data of the food cultivated land at each time within a preset monitoring period, wherein the preset monitoring period includes an irrigation period and a non-irrigation period;

[0008] According to the difference between the true value and the predicted value of the land element monitoring data at each time point, and combined with the difference between the true value and the predicted value of the land humidity monitoring data, element correction data at each time point in each irrigation time period is determined;

[0009] According to the difference between the predicted result of the land element monitoring data in the non-irrigation time period before each irrigation time period and the predicted result of the land element monitoring data in all non-irrigation time periods in the preset monitoring period, a data prediction evaluation index of each irrigation time period is analyzed;

[0010] Using the data prediction evaluation index, the element prediction data at the current time point is weighted based on the element correction data at each time point in each irrigation time period, and a global element prediction value at the current time point is determined, and the global element prediction value at the current time point is used to monitor the quality of the grain farmland.

[0011] Preferably, the element correction data at each time point in each irrigation time period is determined according to the difference between the true value and the predicted value of the land element monitoring data at each time point, and combined with the difference between the true value and the predicted value of the land humidity monitoring data, and specifically includes:

[0012] Based on the land element monitoring data at each time point before all time points, a land element fitting value at each time point is determined, and based on the difference between the land element monitoring data and the land element fitting value at each time point, an element difference factor at each time point is determined;

[0013] Based on the land humidity monitoring data in all non-irrigation time periods before each time point, a land humidity fitting value at each time point is determined, and based on the difference between the land humidity monitoring data and the land humidity fitting value at each time point, a humidity difference factor at each time point is determined;

[0014] According to the element difference factor and the humidity difference factor at each time point in each irrigation time period, the influence of irrigation operation is analyzed, the land element fitting value is corrected, and the element correction data at each time point in each irrigation time period is obtained.

[0015] Preferably, the element correction data at each time point in each irrigation time period is determined according to the element difference factor and the humidity difference factor at each time point in each irrigation time period, the influence of irrigation operation is analyzed, and the land element fitting value is corrected, and specifically includes:

[0016] For the first time point of any irrigation time period, the humidity difference factor and the element difference factor are used as weights, and the land element fitting value and the land element monitoring value are corrected to determine the element correction data at the first time point;

[0017] From the second time of the irrigation time period, the element correction data corrected at the previous time is input as new monitoring data, the data fitting operation is re-executed to generate updated land element fitting values at each time, and the element correction data at each time from the second time is iteratively calculated based on the updated land element fitting values and the humidity difference factor at the corresponding time.

[0018] Preferably, the element correction data at the first time is determined by correcting the land element fitting values and the land element monitoring values with the humidity difference factor and the element difference factor as weights, and specifically includes:

[0019] The product of the humidity difference factor and the element difference factor at the first time is calculated to obtain a prediction weight at the first time, and a preset weight corresponding to the land element monitoring value at each time is obtained.

[0020] A first cumulative sum of the prediction weight and the preset weight is calculated, a second cumulative sum is obtained by weighting and summing the land element fitting value at the first time and the land element monitoring value at the first time with the prediction weight corresponding to the land element fitting value and the preset weight corresponding to the land element monitoring value, respectively, and the element correction data at the first time is determined as the ratio of the second cumulative sum to the first cumulative sum.

[0021] Preferably, the data prediction evaluation index of each irrigation time period is analyzed according to the difference between the prediction result of the land element monitoring data in the non-irrigation time period before each irrigation time period and the prediction result of the land element monitoring data in all non-irrigation time periods in the preset monitoring period, and specifically includes:

[0022] The local fitting result at each time corresponding to each non-irrigation time period is determined by data fitting based on all land element monitoring data in the non-irrigation time period before each irrigation time period.

[0023] The global monitoring fitting data at each time corresponding to the global fitting result of the preset monitoring period is determined by data fitting based on the local monitoring fitting data at all times in all non-irrigation time periods in the preset monitoring period.

[0024] The first similarity characteristic value of each irrigation time period is obtained according to the difference between the local monitoring fitting data at the same time in each non-irrigation time period and other non-irrigation time periods in each irrigation time period, combined with the humidity difference factor at each time in each irrigation time period.

[0025] The second similarity characteristic value of each irrigation time period is obtained according to the difference between the local monitoring fitting data and the global monitoring fitting data at the same time in each non-irrigation time period and the preset monitoring period in each irrigation time period, combined with the humidity difference factor at each time in each irrigation time period.

[0026] The cumulative sum of the first similarity feature value and the second similarity feature value of each irrigation time period is determined as the data prediction evaluation index of each irrigation time period.

[0027] Preferably, the first similarity feature value of each irrigation time period is obtained according to the difference between the local monitoring fitting data of each non-irrigation time period and other non-irrigation time periods at the same time in each irrigation time period, in combination with the humidity difference factor of each time in each irrigation time period, and specifically includes:

[0028] Any one of the irrigation time periods is taken as an irrigation target time period, and the non-irrigation time period before the irrigation target time period is taken as a non-irrigation reference time period;

[0029] The mean value of the difference between the local monitoring fitting data of the non-irrigation reference time period and each other non-irrigation time period at the same time in the irrigation target time period is taken as the local difference factor of each other non-irrigation time period;

[0030] The mean value of the humidity difference factors of all times in each irrigation time period is taken as the irrigation influence degree of each irrigation time period;

[0031] The third cumulative sum of the fitting data amount of the non-irrigation reference time period and the fitting data amount of each other non-irrigation time period, and the fourth cumulative sum between the irrigation influence degrees of the irrigation target time period and the adjacent irrigation time period of each other non-irrigation time period are calculated, and the ratio of the third cumulative sum to the fourth cumulative sum is the evaluation coefficient of each other non-irrigation time period;

[0032] The mean value of the product of the negative correlation coefficient of the local difference factor of each other non-irrigation time period and the evaluation coefficient is calculated to obtain the first similarity feature value of the irrigation target time period.

[0033] Preferably, the second similarity feature value of each irrigation time period is obtained according to the difference between the local monitoring fitting data and the global monitoring fitting data of each non-irrigation time period and a preset monitoring period at the same time in each irrigation time period, in combination with the humidity difference factor of each time in each irrigation time period, and specifically includes:

[0034] The ratio of the fitting data amount of the non-irrigation reference time period to the irrigation influence degree of the irrigation target time period is taken as the first feature factor;

[0035] The mean value of the absolute value of the difference between the local monitoring fitting data and the global monitoring fitting data of the non-irrigation reference time period at the same time in the irrigation target time period is taken as the global difference factor of the non-irrigation reference time period;

[0036] The average value of the negative correlation coefficient of the global difference factor of the non-irrigation reference time period and the product of the first characteristic factor is obtained as a second similar characteristic value of the irrigation target time period.

[0037] Preferably, the evaluation index is predicted by using the data, the element correction data at each time in each irrigation time period is corrected, the element prediction data at the current time is weighted, and the global element prediction value at the current time is determined, and specifically includes the following steps.

[0038] The element prediction data at the current time is obtained by respectively performing data fitting according to the element correction data at each time in each irrigation time period.

[0039] The time decay coefficient of each irrigation time period is obtained based on the time interval between each irrigation time period and the current time.

[0040] The weight is determined by using the data prediction evaluation index and the time decay coefficient of each irrigation time period, the element prediction data at the current time corresponding to each irrigation time period is weighted and summed, and the global element prediction value at the current time is obtained.

[0041] Preferably, the time decay coefficient of each irrigation time period is obtained based on the time interval between each irrigation time period and the current time, and specifically includes the following steps.

[0042] The time decay coefficient of each irrigation time period is obtained by performing negative correlation processing on the time interval between the last time in each irrigation time period and the current time.

[0043] Preferably, the weight is determined by using the data prediction evaluation index and the time decay coefficient of each irrigation time period, and specifically includes the following steps.

[0044] The sum value between the standardization result of the data prediction evaluation index of each irrigation time period and the standardization result of the time decay coefficient is taken as the weight of the corresponding irrigation time period.

[0045] The embodiment of the application has at least the following beneficial effects:

[0046] The application firstly collects soil element monitoring data and soil humidity monitoring data, divides the monitoring period into irrigation time periods and non-irrigation time periods, provides a data basis for the influence of data changes on subsequent irrigation operation. Then, the difference between the predicted value and the actual value of the land element, the difference between the predicted value and the actual value of the land humidity, the difference existing before and after irrigation are analyzed, the monitoring data is corrected to obtain element correction data; the actual value of the soil element content at the end of irrigation is corrected, which can be used for prediction operation of future monitoring data, to a certain extent, the influence of irrigation operation on data is avoided. Further, the accuracy of the data prediction result is considered, the difference between the prediction results is determined, and the data prediction evaluation index is used to represent the accuracy of the data prediction result of the corresponding irrigation time period. Finally, the weight of the predicted value of each irrigation time period is determined based on the accuracy of the predicted value, so that the current time prediction result avoids the influence of irrigation operation, improves the accuracy of the corresponding prediction result of the monitoring data, and makes the final grain land quality evaluation result more accurate. BRIEF DESCRIPTION OF DRAWINGS

[0047] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the drawings needed in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0048] Figure 1 is the method step flow chart of the method executed by the grain safety-oriented cultivated land quality intelligent monitoring and dynamic early warning system provided by the present application;

[0049] Figure 2 is the step flow chart of the method for obtaining element correction data provided by the present application;

[0050] Figure 3 is the step flow chart of the method for obtaining data prediction evaluation index provided by the present application;

[0051] Figure 4 is the step flow chart of the method for obtaining the first similar characteristic value provided by the present application;

[0052] Figure 5 is the step flow chart of the method for obtaining the global element prediction value of the current time provided by the present application. DETAILED DESCRIPTION

[0053] In order to further illustrate the technical means and effects taken by the present application to achieve the predetermined inventive purpose, the specific implementation, structure, features and effects of the cultivated land quality intelligent monitoring and dynamic early warning system for food security according to the present application are described in detail below in combination with the drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0054] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0055] The specific scheme of the cultivated land quality intelligent monitoring and dynamic early warning system for food security provided by the present application is described below in combination with the drawings.

[0056] Please refer to Figure 1 which shows the step flowchart of the cultivated land quality intelligent monitoring and dynamic early warning method provided by one embodiment of the present application, which includes the following steps:

[0057] Step S100, acquiring land element monitoring data and land humidity monitoring data of the food cultivated land at each time point within a preset monitoring period, wherein the preset monitoring period includes an irrigation time period and a non-irrigation time period.

[0058] When monitoring the food cultivated land, the land for food cultivation is monitored through high-spectral data of satellites or unmanned aerial vehicles, and soil organic matter content, total nitrogen, available phosphorus and other cultivated land chemical parameters are obtained through remote sensing inversion; soil humidity and temperature related data are obtained through fusion inversion of thermal infrared remote sensing and microwave remote sensing, thereby obtaining monitoring data of multiple monitoring points. According to remote sensing inversion, the cultivated land is divided into multiple regions (spatial grid division, such as 50m*50m, which is determined according to the clarity of the inversion image and the cultivated area). NDVI of each region is obtained through inversion.

[0059] Based on this, the content of various elements and humidity at each time point within the same growth period of food are acquired as land element monitoring data and land humidity monitoring data respectively by performing monitoring data collection operation on different positions in the land for food cultivation. At the same time, the starting time and the ending time of each irrigation are marked within a preset monitoring period, which is used to divide the monitoring period time points, and the time period in which irrigation operation exists is the irrigation time period, i.e. the time period between the starting time and the ending time of each irrigation, and the time period in which irrigation operation does not exist is the non-irrigation time period.

[0060] It can be understood that the preset monitoring period of the embodiment is one growth stage, and one production stage of grain cultivation corresponds to a period, for example, germination stage, tillering stage, heading stage, etc. The time interval between adjacent data collection time points in a preset monitoring period is equal. Considering that the time length of different growth periods is different, the implementer determines according to the specific implementation scene. Different types of soil elements include the pH value, organic matter content, total nitrogen, available phosphorus, available potassium, etc. of the soil. It can be understood that the monitoring data of different types of soil elements represents the monitoring content change of the corresponding type of elements, for example, the change of phosphorus concentration before and after irrigation operation.

[0061] It should be noted that, in order to avoid the influence of different dimensions on the analysis process of data characteristics, standardization processing is required after collecting various data. The method of standardization processing can use Z-score standardization, which will not be introduced in detail here.

[0062] Meanwhile, in the process of monitoring the growth stage of the grain cultivated land, irrigation operation is needed to supplement the water content of the cultivated land. During the irrigation operation, the change of the water content of the cultivated land will reduce the concentration of the fertilizer of the land to a certain extent, affecting the accuracy of the soil element content monitored. The analysis of the irrigation influence of each sensor at the monitoring position is the same. In the embodiment, the correction process of any type of element content data is taken as an example for description.

[0063] Step S200, according to the difference between the true value and the predicted value of the land element monitoring data at each time, and the difference between the true value and the predicted value of the land humidity monitoring data, determine the element correction data at each time in each irrigation time period.

[0064] When the grain is just planted, in order to ensure the normal growth of the grain, more frequent fertilization, irrigation and other operations are needed, which leads to frequent changes of soil chemical parameters of grain planting, affecting the prediction of the quality of the cultivated land. In addition, due to the lack of independent (i.e. without fertilization and irrigation time period) data, the prediction accuracy of the soil cultivated land of grain planting is low, and the safety of the grain is evaluated incorrectly.

[0065] Based on this aspect, during the non-irrigation period, the change trend of the land humidity monitoring data reflects the change trend of the normal consumption of land moisture without water replenishment, that is, reflects the moisture change in the natural state in the grain growth stage. During the irrigation period, the actual change of the land element monitoring data in the grain field deviates from the normal growth process of the grain due to the influence of irrigation operation. This change trend is affected by the temporary change of soil moisture and belongs to the normal change in the actual monitoring process. Directly using these data for the evaluation operation of the cultivated land quality makes it impossible to accurately reflect the actual condition of the cultivated land quality, and the cultivated land quality is evaluated and predicted incorrectly, affecting the protection and decision of the cultivated land.

[0066] Based on this feature, first, the soil humidity monitoring data and the soil element monitoring data under the normal growth state without the influence of irrigation need to be predicted. The data difference between the prediction result and the actual collection result can realize the correction operation of the soil element data.

[0067] In this embodiment, as shown in Figure 2 The element correction data acquisition method for each irrigation time period and each time can be realized by steps S201 to S203, taking the data acquired by one monitoring point as an example.

[0068] Step S201, data fitting is performed based on the land element monitoring data of all time periods before each time period to determine the land element fitting value of each time period; and the element difference factor of each time period is determined based on the difference between the land element monitoring data and the land element fitting value of each time period.

[0069] In the first aspect, first, the data prediction operation is performed on the soil element monitoring data, and then the difference between the data prediction result and the actual collection result is analyzed in the soil element dimension. It can be understood that the land element monitoring data will only have normal fluctuations during the irrigation period, so the purpose of data prediction is to predict the change trend of the soil element monitoring data during the irrigation period under the normal change trend by using the change trend of the data during the non-irrigation period under the normal state.

[0070] Specifically, the land element monitoring data of all time periods before each time period is fitted by using the logistic model to obtain the land element fitting value of each time period, for example, for the first irrigation time of the first irrigation time period, the soil element monitoring data of all time periods before the irrigation time is taken as the fitting data set for the data prediction operation. The construction method of the logistic model is a known technology, and will not be described in detail here.

[0071] Further, the absolute value of the difference between the soil element monitoring data and the soil element fitting value at the same time is normalized to obtain an element difference factor at each time, which reflects the difference between the normal prediction result and the actual collection result of the soil element at each time.

[0072] In step S202, the soil humidity fitting value at each time is determined based on the data fitting of the soil humidity monitoring data in all non-irrigation time periods before each time. The humidity difference factor at each time is determined based on the difference between the soil humidity monitoring data and the soil humidity fitting value at each time.

[0073] In a second aspect, the soil humidity monitoring data is subjected to data prediction, and then the difference between the data prediction result and the actual collection result is analyzed in the humidity dimension. The soil humidity fitting value at each time is obtained by using a time series prediction model, such as an ARIMA model, to perform data fitting on the soil humidity monitoring data at all times in all non-irrigation time periods before each time.

[0074] Further, the absolute value of the difference between the soil humidity monitoring data and the soil humidity fitting value at the same time is normalized to obtain a humidity difference factor at each time. The soil humidity monitoring data represents the actual monitored cultivated land moisture after irrigation, the soil humidity fitting value represents the theoretical cultivated land moisture determined by the prediction result of the actual monitored cultivated land moisture, and the humidity difference factor represents the change degree of the cultivated land moisture between the predicted moisture data without irrigation and the actual moisture data after irrigation, reflecting the influence degree of irrigation on the change of the cultivated land moisture. The normalization method is a known technology, and the maximum-minimum normalization method can be used, which will not be described in detail here.

[0075] In step S203, the influence of the irrigation operation is analyzed according to the element difference factor and the humidity difference factor at each time in each irrigation time period, and the soil element fitting value is corrected to obtain the element correction data at each time in each irrigation time period.

[0076] Each irrigation operation corresponds to an irrigation time period, and the time distribution before each irrigation time period is a non-irrigation time period. The data monitoring process in the non-irrigation time period can simulate the normal growth state of the grain. Therefore, in order to ensure the accuracy of the data prediction result in the irrigation time period, it is necessary to ensure that the data set for data prediction at each time in the irrigation time period contains the data collection result in the non-irrigation time period or the data after correction of the data collection result in the irrigation time. It can be understood that the present embodiment mainly corrects the soil element monitoring value in each irrigation time period, and therefore the data prediction operation is mainly performed on each irrigation time period.

[0077] In a first step, for a first time point of any irrigation time period, the humidity difference factor and the element difference factor are taken as weights, and the land element fitting value and the land element monitoring value are combined to correct the element correction data of the first time point.

[0078] As a specific example, the first time point in the first irrigation time period in the preset monitoring period is taken as an example for illustration. At this time, since all time points before the first time point are non-irrigation time periods, the historical data set for prediction operation does not need to be updated when the land element fitting value is obtained. The humidity difference factor and the element difference factor of the first time point are obtained according to the methods of steps S201 and S202. Further, the humidity difference factor and the element difference factor are taken as weights, and the land element fitting value and the land element monitoring value are combined to correct the element correction data of the first time point.

[0079] In a second step, from the second time point of the irrigation time period, the element correction data corrected at the previous time point is taken as new monitoring data input, and the data fitting operation is re-executed to generate the updated land element fitting value of each time point. Based on the updated land element fitting value and the humidity difference factor of the corresponding time point, the element correction data of each time point from the second time point is iteratively calculated.

[0080] More specifically, according to the chronological order of the irrigation time period, the first time point has completed the correction operation of the land element monitoring data. All time points before the second time point include the first time point, that is, all time points before the second time point include time points that do not belong to non-irrigation time periods. Therefore, when obtaining the land element fitting value at the second time point, the historical data set for prediction operation needs to be updated. That is, the land element monitoring data of all time points in the non-irrigation time period before the second time point and the element correction data of all time points in the irrigation time period are taken as the historical data set of the second time point. The historical data set is used for data prediction of the second time point to obtain the land element fitting value of the second time point.

[0081] Further, the humidity difference factor and the element difference factor of the second time point are obtained according to the same methods as steps S201 and S202, and then the element correction data of the second time point is calculated according to the same method as the first time point.

[0082] It should be noted that the element correction data of each time point is iteratively calculated in chronological order. The calculation process is the same each time, except that when performing the prediction operation of the land element monitoring data, there are time points in the irrigation time period before the certain time point, and the data set needs to be updated with the element correction data of the corresponding time point.

[0083] In the embodiment, the first time in the first irrigation time period is taken as an example for illustration, and the method for obtaining the element correction data is specifically as follows: a product of the humidity difference factor and the element difference factor at the first time is calculated to obtain a prediction weight at the first time, and a preset weight corresponding to the land element monitoring value at each time is obtained; a first accumulation sum of the prediction weight and the preset weight is calculated, and a second accumulation sum is obtained by weighting and summing the land element fitting value at the first time and the land element monitoring value at the first time respectively by using the prediction weight corresponding to the land element fitting value and the preset weight corresponding to the land element monitoring value; and the element correction data at the first time is determined as a ratio of the second accumulation sum to the first accumulation sum.

[0084] As a specific example, the calculation formula of the element correction data at the first time can be represented as:

[0085]

[0086] wherein, represents the element correction data at the first time in the irrigation time period, represents the first time, represents the element difference factor at the first time, represents the humidity difference factor at the first time, represents the land element fitting value at the first time, represents the land element monitoring data at the first time.

[0087] represents the land element fitting value at the first time corresponding to the prediction weight, and the land element monitoring data at the first time corresponding to the preset weight is 1, which represents the default baseline data level of the actual value. The prediction weight reflects the influence of the irrigation operation on the data monitoring at the corresponding time through the difference between the prediction data and the actual data at the first time, and the greater the value of the prediction weight, the more reliable the prediction result at the corresponding time, that is, the greater the weight in data correction.

[0088] is the second accumulation sum, and the first accumulation sum represents the sum of the weights, which is used to ensure that the correction result of the weighted average is mapped to a reasonable range. When the prediction weight has a large value, it indicates that the influence of the irrigation operation on the data monitoring is large, and at this time, the actual value is diluted or migrated and distorted, and the prediction value is introduced to correct the data. When the prediction weight is small, it indicates that the influence of the irrigation operation on the data monitoring is small, and at this time, the actual value can still reflect part of the data characteristics, so the actual value is given a higher weight for calculation.

[0089] In some embodiments, after obtaining the element correction data, before analyzing the data prediction evaluation index of each irrigation time period, the method further comprises: performing a second correction on the element correction data of each monitoring point at each time in each irrigation time period, to obtain element adjustment data of each monitoring point at each time in each irrigation time period.

[0090] When only relying on the area where the single monitoring point is located to analyze the element characteristics of the cultivated land, the characteristics of the individual case are easily affected, that is, the monitoring result is easily less accurate due to the influence of the growth state of the grain. In order to make the characteristic analysis result of the growth state parameter of the grain more accurate, the characteristics of the cultivated land corresponding to the remaining grains in good growth state also need to be referred to.

[0091] As a specific example, the method for obtaining the element adjustment data can be implemented by steps S204 to S207.

[0092] Step S204: Obtain the cultivated land state index of each monitoring point at each time.

[0093] Specifically, each monitoring point corresponds to a monitoring area. In step S100, the NDVI index of each monitoring area at each time is obtained by remote sensing inversion, which is used as the cultivated land state index of each monitoring point at each time, for representing the growth state of the cultivated grain. Since the NDVI indexes of the grains corresponding to different cultivation time periods are different under normal circumstances, the distribution of the NDVI indexes of the other monitoring points is also referred to for the area where each monitoring point is located.

[0094] Step S205: Obtain the correlation adjustment weight corresponding between each monitoring point and the other monitoring points according to the difference between the element correction data of each monitoring point and the other monitoring points at each time in each irrigation time period, and the difference of the cultivated land state index.

[0095] Firstly, for any time, the monitoring points corresponding to each cultivated land state index are sequentially taken as the monitoring points to be analyzed in the order of the cultivated land state index from large to small, and all the monitoring points other than the monitoring points to be analyzed are taken as the reference monitoring points of the growth state of the cultivated grain.

[0096] It should be noted that the second adjustment process at each time in each irrigation time period is completely the same in the order of time, and the first time in any time period is taken as an example for description.

[0097] Secondly, the absolute value of the difference between the element correction data of the monitoring point to be analyzed and each reference monitoring point at the first time is obtained, which is recorded as the element correction amplitude between the monitoring point to be analyzed and each reference monitoring point at the first time.

[0098] If the correction amplitude is higher, it indicates that the reference value of the NDVI index of each reference monitoring point relative to the monitoring point to be detected is smaller, so the weight of the further correction of the correction element data of the monitoring point to be analyzed is smaller. In view of this, in a third step, based on the negative correlation coefficient between the difference between the element correction data of the monitoring point to be analyzed and each reference monitoring point at the first time in the irrigation period, and the ratio of the farmland state indicators between each reference monitoring point and the monitoring point to be analyzed, the correlation adjustment weight between the monitoring point to be analyzed and each reference monitoring point at the first time is determined.

[0099] As a specific example, the calculation formula of the correlation adjustment weight between the monitoring point to be analyzed and the kth reference monitoring point at the first time in the irrigation period can be expressed as:

[0100]

[0101] wherein, represents the correlation adjustment weight between the monitoring point to be analyzed and the kth reference monitoring point at the first time in the irrigation period, represents the first time, represents the farmland state indicator of the monitoring point to be analyzed at the first time in the irrigation period, represents the farmland state indicator of the kth reference monitoring point at the first time in the irrigation period, represents the element correction data of the monitoring point to be analyzed at the first time in the irrigation period, represents the element correction data of the kth reference monitoring point at the first time in the irrigation period, and Norm is a normalization function.

[0102] In step S206, based on the corresponding correlation adjustment weight between each monitoring point and other monitoring points, the local Moran index is used to obtain the correlation coefficient of each monitoring point.

[0103] At each time in each irrigation period, the corresponding correlation adjustment weight between the monitoring point to be analyzed and each reference monitoring point is normalized to serve as a spatial weight, and then the local Moran index is used to obtain the Moran index between the monitoring point to be analyzed and each reference monitoring point, as the correlation coefficient of the monitoring point to be analyzed at each time in each irrigation period.

[0104] The Morant index reflects the strength of the correlation between the land elements of the monitoring point to be analyzed and the land elements of the reference monitoring point. The greater the value, the greater the correlation between the two, and the stronger the data correlation between the monitoring point to be analyzed and the surrounding monitoring points. The smaller the value, the smaller the correlation between the land elements of the monitoring point to be analyzed and the surrounding monitoring points, and the greater the data isolation of the monitoring point to be analyzed. Therefore, the data of the monitoring point to be analyzed cannot be corrected by using the element change of other monitoring points as a reference.

[0105] In step S207, the element adjustment data of each monitoring point is obtained by combining the correlation coefficient with the element correction data of the other monitoring points and the correlation adjustment weight corresponding to each monitoring point.

[0106] Taking the monitoring point to be analyzed and each reference monitoring point corresponding thereto as an example, in the first step, at the first time in any irrigation time period, when the correlation coefficient of the monitoring point to be analyzed is greater than 0, the element correction data of each reference monitoring point is weighted and averaged based on the correlation adjustment weight corresponding to each reference monitoring point to obtain the element adjustment data of the monitoring point to be analyzed. At this time, the Morant index of the monitoring point to be analyzed is greater than 0, indicating that the element correlation between the monitoring point to be analyzed and other monitoring points is strong, so the second correction data correction result of the monitoring point to be analyzed at the current time is obtained by using the weighted method of the corresponding monitoring points in other areas.

[0107] In the second step, at the first time in any irrigation time period, when the correlation coefficient of the monitoring point to be analyzed is less than or equal to 0, the element correction data of the monitoring point to be analyzed at the corresponding time is not secondly corrected. At this time, the Morant index of the monitoring point to be analyzed is less than or equal to 0, indicating that the element correlation between the monitoring point to be analyzed and other monitoring points is weak, so no second correction operation is needed for the data of the monitoring point to be analyzed at this time, that is, the element correction data at this time is the corresponding element adjustment data.

[0108] It should be noted that steps S204 to S207 can be regarded as a process of secondly updating the element correction data of each monitoring point at each time in each irrigation time period. In order to facilitate description, the data updating process is introduced in the above steps by taking the "element adjustment data" as the object. It should be understood that the element correction data mentioned after step S300 can be the element correction data after the feature analysis in step S203, and as a second embodiment, it can also be the element adjustment data after the second correction.

[0109] At step S300, the data prediction evaluation index of each irrigation time period is analyzed according to the prediction result of the land element monitoring data in the non-irrigation time period before each irrigation time period and the difference between the prediction results of the land element monitoring data in all non-irrigation time periods in the preset monitoring period.

[0110] The soil element content in each irrigation time period is subjected to data correction operation, which can correct the influence of irrigation operation to a certain extent. The multiple soil element content prediction values at the current time are obtained by comprehensively predicting the data in different irrigation time periods in a weighted manner, so that the soil element content at the current time can be more accurately obtained, and the cultivated land quality evaluation result is more accurate. First, the data accuracy of each irrigation time period is measured, and the higher the accuracy of the data prediction result in each irrigation time period, the higher the data reliability.

[0111] In this embodiment, as shown in Figure 3 The data prediction evaluation index can be obtained by steps S301 to S305.

[0112] At step S301, data fitting is performed on all land element monitoring data in each non-irrigation time period before each irrigation time period, and the local fitting result of each non-irrigation time period is determined as the local monitoring fitting data at each time.

[0113] It can be understood that each irrigation time period corresponds to an adjacent non-irrigation time period, and the land element monitoring data at all times of each non-irrigation time period is subjected to data prediction operation by using the logistic model, so as to obtain the prediction result of the local data corresponding to each non-irrigation time period. However, it should be noted that each local data prediction result covers the entire monitoring period, that is, the corresponding local monitoring fitting data can be determined in each local fitting result at each time.

[0114] At step S302, data fitting is performed on the local monitoring fitting data at all times in all non-irrigation time periods in the preset monitoring period, and the global fitting result of the preset monitoring period is determined as the global monitoring fitting data at each time.

[0115] The irrigation time period and the non-irrigation time period correspond to different time segments in a monitoring period, and the local monitoring fitting data at all times in all non-irrigation time periods in the preset monitoring period is subjected to data prediction operation by using the logistic model, so as to obtain the data prediction result of the entire monitoring period.

[0116] It should be noted that all the land element monitoring data in the non-irrigation period is used in steps S301 and S302 to perform the prediction operation, so that the prediction evaluation result is more accurate, that is, the influence of the irrigation data needs to be excluded.

[0117] In step S303, according to the difference between the local monitoring fitting data of each non-irrigation period and other non-irrigation periods at the same time in each irrigation period, and in combination with the humidity difference factor of each time in each irrigation period, a first similarity characteristic value of each irrigation period is obtained.

[0118] The first similarity characteristic value represents the prediction data difference and similarity of the corresponding local prediction result of each non-irrigation period at the same time, and the greater the similarity between them, the higher the accuracy and reliability of the corresponding local prediction result.

[0119] More specifically, as shown in Figure 4 The first similarity characteristic value can be obtained by steps S3031 to S3035.

[0120] In step S3031, any irrigation period is taken as an irrigation target period, and the non-irrigation period before the irrigation target period is taken as a non-irrigation reference period.

[0121] The irrigation target period and the non-irrigation reference period are in a one-to-one correspondence, and thus the accuracy of the data prediction and correction result of the irrigation target period needs to be evaluated based on the non-irrigation reference period, and the data difference between the non-irrigation reference period and other non-irrigation periods in the preset monitoring period except the non-irrigation reference period needs to be analyzed.

[0122] In step S3032, the mean value of the difference between the local monitoring fitting data of the non-irrigation reference period and each other non-irrigation period at the same time in the irrigation target period is taken as the local difference factor of each other non-irrigation period.

[0123] In the local prediction result of each non-irrigation period, there is a corresponding local monitoring fitting data at each time in the irrigation target period. Any time in the irrigation target period is taken as a target time, and first, the absolute value of the difference between the local monitoring fitting data of the non-irrigation reference period and each other non-irrigation period at the target time is calculated, which is taken as the local prediction difference value of each other non-irrigation period at the target time, and then the mean value of the local prediction difference value of each other non-irrigation period at all times in the irrigation target period is taken as the local difference factor of each other non-irrigation period.

[0124] Step S3033, taking the mean value of the humidity difference factors of all time points in each irrigation time period as the irrigation influence degree of each irrigation time period.

[0125] Step S3034, calculating the third cumulative sum of the fitting data amount of the non-irrigation reference time period and the fitting data amount of each other non-irrigation time period, and the fourth cumulative sum between the irrigation influence degree of the irrigation target time period and the irrigation time period adjacent to each other non-irrigation time period, and the ratio of the third cumulative sum to the fourth cumulative sum is the evaluation coefficient of each other non-irrigation time period.

[0126] The fitting data amount represents the total amount of data participating in the data prediction operation, and in this embodiment, the total amount of all time points in each non-irrigation time period is the corresponding fitting data amount. The more the fitting data amount, the higher the prediction accuracy.

[0127] The third cumulative sum reflects the performance degree of the fitting data amount of the non-irrigation reference time period and the other non-irrigation time period, and the fourth cumulative sum reflects the summation result between the irrigation influence degree of the non-irrigation reference time period and the corresponding irrigation time period of the other non-irrigation time period. The greater the value of the fourth cumulative sum, the greater the influence degree of irrigation, and the lower the data prediction accuracy; if the influence degree of irrigation is small, such as a small amount of water is supplemented to the farmland, at this time, the accuracy is higher.

[0128] The evaluation coefficient of each other non-irrigation time period reflects the preliminary evaluation result of the accuracy of data prediction in each other non-irrigation time period.

[0129] Step S3035 calculates the average value of the product of the negative correlation coefficient of the local difference factor of each other non-irrigation time period and the evaluation coefficient, to obtain the first similar feature value of the irrigation target time period.

[0130] As a specific example, the negative correlation coefficient of the local difference factor of each other non-irrigation time period is represented by represents the local difference factor of any other non-irrigation time period, and exp represents the exponential function with the natural constant e as the base.

[0131] The smaller the value of the local difference factor, the smaller the difference between the data prediction results of the non-irrigation time period corresponding to the irrigation target time period and the other non-irrigation time periods, and the greater the similarity between the two, that is, the greater the value of the first similar feature value, and the product of the accuracy evaluation result and the similarity evaluation result by preliminary analysis comprehensively reflects the similarity degree of the non-irrigation time period corresponding to the irrigation target time period relative to the other non-irrigation time periods.

[0132] ​In step S304, according to the difference between the local monitoring fitting data and the global monitoring fitting data at the same time in each irrigation time period and the preset monitoring period, combined with the humidity difference factor at each time in each irrigation time period, the second similarity characteristic value of each irrigation time period is obtained.

[0133] The second similarity characteristic value reflects the difference and similarity between the prediction data of the prediction result of the irrigation time period and the prediction result of the whole monitoring period at the same time. The greater the similarity between the two, the higher the accuracy of the corresponding local prediction result and the greater the reliability.

[0134] More specifically, the method for obtaining the second similarity characteristic value is as follows: taking the ratio of the fitting data amount of the non-irrigation reference time period and the irrigation influence degree of the irrigation target time period as the first characteristic factor; taking the average of the absolute value of the difference between the local monitoring fitting data and the global monitoring fitting data of the non-irrigation reference time period at the same time in the irrigation target time period as the global difference factor of the non-irrigation reference time period; calculating the average of the product of the negative correlation coefficient of the global difference factor of the non-irrigation reference time period and the first characteristic factor to obtain the second similarity characteristic value of the irrigation target time period.

[0135] By analogy with the acquisition process of the first similarity characteristic value, the first characteristic factor only uses the irrigation influence degree of the irrigation target time period and the fitting data amount of the corresponding non-irrigation time period to preliminarily evaluate the accuracy of the prediction result of the irrigation target time period. Further, the global difference factor reflects the difference between the prediction result of the irrigation target time period and the prediction result of the whole monitoring period. The greater the difference, the smaller the similarity between the two, that is, the smaller the value of the second similarity characteristic factor. According to the same method, the function can be used for negative correlation processing, where x represents the index.

[0136] In step S305, the cumulative sum of the first similarity characteristic value and the second similarity characteristic value of each irrigation time period is determined as the data prediction evaluation index of each irrigation time period.

[0137] In a first aspect, the first similarity feature value represents the similarity of the local prediction result of the irrigation target time period, and is characterized by the difference and similarity of the prediction data at the same time between the non-irrigation time period corresponding to the irrigation target time period and each other non-irrigation time period. In a second aspect, the second similarity feature value represents the similarity of the overall prediction result of the irrigation target time period, and is characterized by the difference and similarity of the prediction data at the same time between the irrigation target time period and the overall monitoring period. Then, the data prediction evaluation index of the irrigation target time period is obtained by accumulation, that is, the greater the value of the first similarity feature value and the greater the value of the second similarity feature value, the greater the accuracy of the data prediction in the irrigation target time period, and the more reliable the data prediction operation.

[0138] In step S400, the data prediction evaluation index is used to weight the element correction data based on each time in each irrigation time period and the element prediction data at the current time, to determine the global element prediction value at the current time, and to monitor the grain farmland quality by using the global element prediction value at the current time.

[0139] By performing data prediction on the element correction data in each irrigation time period, the soil element content after correction can be obtained, which avoids the influence of irrigation operation. When the different prediction results are weighted, the prediction result of each irrigation time period is more accurate, the corresponding prediction data weight is greater, the time of each irrigation time period from the current time is closer, the corresponding prediction data reference degree is greater, that is, the weight is greater.

[0140] Based on this feature, as shown in Figure 5 The method for obtaining the global element prediction value at the current time can be implemented by steps S401 to S403.

[0141] In step S401, data fitting is performed on the element correction data at each time in each irrigation time period respectively, to obtain the element prediction data of the correction fitting result of each irrigation time period at the current time.

[0142] It should be understood that the current time is the time at which soil element prediction is currently needed, and a plurality of irrigation time periods are included in the same preset monitoring period before the current time. For each irrigation time period, the logistic model is used to perform data prediction operation on the element correction data at all times of the irrigation time period, to obtain the element prediction data of the correction fitting result of each irrigation time period at the current time, and the data prediction evaluation index of each irrigation time period can represent the accuracy of the correction fitting result in the corresponding time period.

[0143] In step S402, the time decay coefficient of each irrigation time period is obtained based on the time interval between each irrigation time period and the current time.

[0144] The time interval between the last time in each irrigation time period and the current time is negatively correlated to obtain the time decay coefficient of each irrigation time period. The function can be used The time interval is negatively correlated, which will not be introduced here. The smaller the value of the time decay coefficient, the farther the time interval of the irrigation time period from the current time, and the lower the reference value of the modified fitting result corresponding to the irrigation time period.

[0145] Step S403, using the data prediction evaluation index and the time decay coefficient of each irrigation time period to determine the weight, and performing weighted summation on the element prediction data corresponding to the current time of each irrigation time period to obtain the global element prediction value at the current time.

[0146] Specifically, the sum of the standardized results of the data prediction evaluation index and the time decay coefficient of each irrigation time period is used as the weight of the corresponding irrigation time period.

[0147] The greater the value of the data prediction evaluation index, the higher the accuracy of the modified fitting result of the corresponding irrigation time period, the higher the reference value of the element prediction data at the current time, and the greater the value of the time decay coefficient, the closer the time distance between the corresponding irrigation time period and the current time, and the higher the reference value of the element prediction data at the current time of the modified fitting result of the irrigation time period.

[0148] Further, each irrigation time period corresponds to a weight and an element prediction data, and the weighted summation result of the element prediction data corresponding to all irrigation time periods by using the weight is the global element prediction value at the current time. The finally determined global element prediction value can effectively exclude the influence of irrigation operation on soil element content, so that the monitoring data of soil content is more accurate, and further based on the accurate soil element content, more accurate farmland quality evaluation results can be obtained.

[0149] Finally, it should be noted that the data analysis process of the sulfur content of one kind of soil element is used as an example in this embodiment, and the global element prediction value of each kind of soil element content at the current time can be obtained according to the same method. The global element prediction values corresponding to multiple soil element contents form an element content vector, which is used as the input of the evaluation model to input the evaluation result of farmland quality.

[0150] The evaluation model can be a neural network model trained based on historical data, the input of the training process of the model is an element content vector corresponding to historical data at each time point, and the output is a cultivated land quality evaluation coefficient corresponding to each time point. The cultivated land quality evaluation coefficient can be manually marked by relevant staff based on the numerical value of different soil element contents at each time point, and the value range is (0, 1). The greater the value, the higher the cultivated land quality, and the smaller the value, the lower the cultivated land quality. The cultivated land quality is evaluated by various soil element contents, which is a known technology and will not be described in detail here.

[0151] It can be understood that the main purpose of the embodiment is to perform data prediction and other processing based on historical monitoring data of grain cultivated land, and the influence of irrigation operation is excluded in the data prediction process, so that the prediction data used for cultivated land quality evaluation is more accurate. The cultivated land quality evaluation method is not limited, and the implementer can select according to the specific implementation scene. Under the condition of ensuring the consistency of the evaluation method, it can be known that since the data used for cultivated land quality evaluation excludes the influence of irrigation operation to a certain extent, the cultivated land quality evaluation result of the embodiment is more accurate compared with the result of cultivated land quality evaluation by data without shielding the influence of irrigation operation.

[0152] In summary, in the embodiment of the application, the influence of irrigation on monitoring data is determined at the analysis time, and the difference between the predicted value and the actual value of the soil element content and the difference between the predicted value and the actual value of the soil humidity are analyzed. The actual value of the soil element content at the end of irrigation is corrected so that it can be used for future monitoring data prediction. To a certain extent, the influence of irrigation operation on data is avoided. Further, the accuracy of data prediction is considered, the weight of the predicted value of each irrigation time period is determined based on the accuracy of the predicted value and the time decay, the accuracy of future prediction data is improved, and the accuracy and timeliness of the cultivated land quality judgment and early warning are improved.

[0153] The above-described embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. An intelligent monitoring system for arable land quality aimed at food security, characterized in that, The system includes a memory and a processor, the processor executing a computer program stored in the memory to perform the following steps: Acquire land element monitoring data and land moisture monitoring data of farmland at each moment within a preset monitoring period, wherein the preset monitoring period includes irrigation time period and non-irrigation time period; Based on the differences between the actual and predicted values ​​of land element monitoring data at each moment, and considering the impact of the differences between the actual and predicted values ​​of land moisture monitoring data, the element correction data for each moment within each irrigation period is determined. Specifically, this includes: fitting data to land element monitoring data from all moments prior to each moment to determine the fitted value of land elements at each moment; determining the element difference factor for each moment based on the difference between the land element monitoring data and the fitted value of land elements at each moment; fitting data to land moisture monitoring data from all non-irrigation periods prior to each moment to determine the fitted value of land moisture at each moment; determining the moisture difference factor for each moment based on the difference between the land moisture monitoring data and the fitted value of land moisture at each moment; and analyzing the impact of irrigation operations based on the element difference factor and moisture difference factor for each moment within each irrigation period, and correcting the fitted values ​​of land elements to obtain the element correction data for each moment within each irrigation period. Based on the predicted results of land element monitoring data in non-irrigation periods before each irrigation period, and the differences between the predicted results of land element monitoring data in all non-irrigation periods within the preset monitoring cycle, the data prediction evaluation index for each irrigation period is analyzed. Using the aforementioned data prediction evaluation index, the element correction data based on each moment within each irrigation period is weighted and the element prediction data at the current moment is weighted to determine the global element prediction value at the current moment. The global element prediction value at the current moment is then used to monitor the quality of farmland for grain production.

2. The intelligent monitoring system for arable land quality for food security according to claim 1, characterized in that, The step involves analyzing the impact of irrigation operations based on the elemental and humidity difference factors at each moment within each irrigation period, and correcting the fitted values ​​of the land elements to obtain corrected elemental data at each moment within each irrigation period. Specifically, this includes: For the first moment of any irrigation period, the humidity difference factor and element difference factor are used as weights, and the soil element fitting value and soil element monitoring value are combined to make corrections to determine the element correction data for the first moment. Starting from the second moment of the irrigation period, the element correction data corrected at the previous moment is used as the new monitoring data input. The data fitting operation is re-executed to generate the updated land element fitting values ​​at each moment. Based on the updated land element fitting values ​​and the humidity difference factor at the corresponding moment, the element correction data at each moment starting from the second moment is iteratively calculated.

3. The intelligent monitoring system for arable land quality for food security according to claim 2, characterized in that, The step of using the humidity difference factor and element difference factor as weights, combined with the land element fitted value and land element monitoring value, to correct and determine the element correction data for the first time moment specifically includes: The predicted weight for the first time moment is obtained by multiplying the humidity difference factor and the element difference factor at the first time moment, and the preset weight corresponding to the land element monitoring value at each time moment is obtained. Calculate the first cumulative sum of the predicted weight and the preset weight. Using the predicted weight corresponding to the land element fitting value and the preset weight corresponding to the land element monitoring value, respectively, perform a weighted summation on the land element fitting value and the land element monitoring value at the first time point to obtain the second cumulative sum. The ratio of the second cumulative sum to the first cumulative sum is determined as the element correction data at the first time point.

4. The intelligent monitoring system for arable land quality for food security according to claim 1, characterized in that, The method involves analyzing the data prediction evaluation indicators for each irrigation period based on the prediction results of land element monitoring data during non-irrigation periods prior to each irrigation period, and the differences between the prediction results of land element monitoring data during all non-irrigation periods within a preset monitoring period. Specifically, this includes: Data fitting is performed based on all land element monitoring data during the non-irrigation period before each irrigation period to determine the local fitting result of each non-irrigation period corresponding to the local monitoring fitting data at each time point. Based on the local monitoring fitting data of all times in all non-irrigation time periods within the preset monitoring period, the global fitting result of the preset monitoring period is determined by the global monitoring fitting data of each time. Based on the differences between the local monitoring fitted data of each non-irrigation period and other non-irrigation periods at the same time within each irrigation period, and combined with the humidity difference factor at each time within each irrigation period, the first similarity feature value of each irrigation period is obtained. Based on the differences between the local monitoring fitting data and the global monitoring fitting data at the same moment within each irrigation period and the preset monitoring cycle for each non-irrigation period, and combined with the humidity difference factor at each moment within each irrigation period, the second similarity feature value for each irrigation period is obtained. The sum of the first and second similarity feature values ​​for each irrigation period is determined as the data prediction and evaluation index for each irrigation period.

5. The intelligent monitoring system for arable land quality for food security according to claim 4, characterized in that, The method involves determining the first similarity feature value for each irrigation time period based on the differences between local monitoring and fitting data from different non-irrigation time periods and other non-irrigation time periods at the same time within each irrigation time period, combined with the humidity difference factor at each time within each irrigation time period. This process specifically includes: Take any irrigation period as the target irrigation period, and the non-irrigation period before the target irrigation period as the non-irrigation reference period. The mean of the differences between the local monitoring fitted data at the same time within the irrigation target time period and the non-irrigation reference time period and each other non-irrigation time period is used as the local difference factor for each other non-irrigation time period. The mean of the humidity difference factor at all times within each irrigation period is taken as the degree of irrigation impact for each irrigation period. The third sum of the fitted data for the non-irrigation reference time period and the fitted data for each other non-irrigation time period is calculated, and the fourth sum of the irrigation impact between the irrigation target time period and the irrigation time periods adjacent to each other non-irrigation time period is calculated. The ratio of the third sum to the fourth sum is the evaluation coefficient for each other non-irrigation time period. The average of the product of the negative correlation coefficient and the evaluation coefficient of the local difference factor for each other non-irrigation time period is calculated to obtain the first similarity feature value of the irrigation target time period.

6. The intelligent monitoring system for arable land quality for food security according to claim 5, characterized in that, The method involves analyzing the differences between local and global monitoring fitted data at the same time within each irrigation period, based on the non-irrigation time period and the preset monitoring cycle, and combining this with the humidity difference factor at each time within each irrigation period to obtain a second similarity feature value for each irrigation period. Specifically, this includes: The ratio of the amount of fitted data during the non-irrigation reference period to the degree of irrigation impact during the irrigation target period was used as the first feature factor. The mean of the absolute values ​​of the differences between the local monitoring fitted data and the global monitoring fitted data at the same time within the irrigation target time period during the non-irrigation reference time period is used as the global difference factor for the non-irrigation reference time period. The average of the negative correlation coefficient of the global difference factor and the product of the first characteristic factor for the non-irrigated reference time period is calculated to obtain the second similarity characteristic value for the irrigation target time period.

7. The intelligent monitoring system for arable land quality for food security according to claim 1, characterized in that, The step of using the data prediction evaluation index to weight the element prediction data at the current moment based on the element correction data at each moment within each irrigation period, and determining the global element prediction value at the current moment, specifically includes: Data fitting is performed on the element correction data at each moment within each irrigation period to obtain the element prediction data at the current moment based on the correction fitting results of each irrigation period. The time decay coefficient for each irrigation period is obtained based on the time interval between each irrigation period and the current time. The weights are determined by using the data prediction evaluation index and time decay coefficient for each irrigation period. The element prediction data corresponding to each irrigation period at the current moment are weighted and summed to obtain the global element prediction value at the current moment.

8. The intelligent monitoring system for arable land quality for food security according to claim 7, characterized in that, The process of obtaining the time decay coefficient for each irrigation time period based on the time interval between each irrigation time period and the current time specifically includes: The time decay coefficient for each irrigation period is obtained by negatively correlating the time interval between the last moment and the current moment in each irrigation period.

9. The intelligent monitoring system for arable land quality for food security according to claim 7, characterized in that, The method of using data from each irrigation period to predict evaluation indicators and determine weights specifically includes: The sum of the standardized results of the data prediction evaluation indicators and the standardized results of the time decay coefficient for each irrigation period is used as the weight of the corresponding irrigation period.

Images