Soil state-based backwashing water and fertilizer integrated machine irrigation method

Abstract

The present application relates to the technical field of smart farmland, and particularly relates to a backwashing water and fertilizer integrated machine irrigation method based on soil state. According to the similarity of recent fluctuations and the spatial distribution difference of the same soil indexes between sampling points, the reference degree of abnormal analysis between sampling points is obtained, and then combined with the local fluctuation characteristic difference of the same soil index at the same time, the abnormal characteristic value of each monitoring parameter is obtained. According to the abnormal characteristic value of each monitoring parameter and the fluctuation characteristic of the corresponding soil index, the isolated abnormality degree of each monitoring parameter is obtained. According to the isolated abnormality degree of the monitoring parameter, the optimal interpolation curve of each soil parameter sequence is screened out, and then the soil state at the current time is evaluated by fusing all soil index parameters to regulate the irrigation parameter. The present application combines the isolated abnormality degree of each monitoring parameter to evaluate the interpolation error, reduces the influence of noise on interpolation fitting, and thus accurately evaluates the soil state at each time.

Description

Technical Field

[0001] This invention relates to the field of soil data analysis and processing technology, specifically to a backwashing fertigation method based on soil conditions. Background Technology

[0002] Currently, poor water-fertilizer ratios are common in agricultural planting, leading to low fertilizer absorption rates in crops, resulting in significant waste and reduced yields. A backwashing fertigation system, on the other hand, integrates a water pump, sprinkler system, and fertilizer supply system, enabling intelligent management such as water and fertilizer fusion, precise allocation, and automatic fertilization and irrigation. This can effectively improve agricultural production and management efficiency. However, the use of a fertigation system requires consideration of the soil conditions in the planting area to precisely adjust irrigation parameters.

[0003] When collecting soil index parameters to assess the current soil condition, due to limitations in sensor configuration and the inherent properties of soil indexes, the sampling frequencies of various soil index parameters differ. This results in some soil index monitoring parameters not being collected at the same time, making it impossible to match them one-to-one. Consequently, it is impossible to comprehensively assess the soil condition at the current time based on the monitoring parameters of all soil indexes. Therefore, interpolation processing of various monitoring parameters is often required. However, due to the possibility of noise interference in the collected soil index parameters, the final interpolation results are poor, leading to inaccurate interpolated monitoring parameters at each time point. This affects the assessment of soil condition and makes it impossible to accurately adjust irrigation parameters. Summary of the Invention

[0004] To address the technical problem of poor accuracy in soil condition assessment caused by noise interference in existing interpolation fitting methods, the present invention aims to provide a backwashing fertigation method based on soil condition. The specific technical solution adopted is as follows:

[0005] Obtain the soil parameter sequence for each soil index at each sampling point in the planting area. The soil parameter sequence includes all monitoring parameters for the corresponding soil index at the corresponding sampling point.

[0006] Taking any sampling point as the target point, based on the recent fluctuation similarity and spatial distribution difference of the monitoring parameters of the same soil index at the target point and other sampling points within a preset neighboring time period, the reference degree of each other sampling point for the target point under each soil index is obtained; based on the fluctuation characteristics difference of the monitoring parameters of the same soil index at the same sampling time between the target point and other sampling points within a preset range in the soil parameter sequence and the corresponding reference degree, the abnormal characteristic value of the monitoring parameters of each soil index at the target point at each sampling time is obtained; combining the abnormal characteristic value and the fluctuation characteristics of the corresponding monitoring parameters of each soil index at the target point, the isolation anomaly degree of the monitoring parameters of each soil index at the target point at each sampling time is obtained.

[0007] Obtain at least two interpolation curves for each soil parameter sequence; based on the difference between the soil parameter sequence and the monitoring parameters at the same sampling time in the corresponding interpolation curve and the degree of isolation anomaly, select the optimal interpolation curve for each soil index at each sampling point.

[0008] Furthermore, the method for obtaining the recent fluctuation similarity includes:

[0009] In the soil parameter sequence, recent parameter sequences corresponding to preset adjacent time periods are obtained; DTW matching is performed on the recent parameter sequences of the same soil index at the target point and each sampling point to obtain all DTW matching points between the recent parameter sequences of the target point and the corresponding sampling point; the recent fluctuation similarity is obtained according to the recent fluctuation similarity calculation formula; the recent fluctuation similarity calculation formula is:

[0010] Among them, D ihj d represents the similarity of recent fluctuations of the j-th soil index at the i-th target point and the h-th sampling point; ihjf Let f be the parameter difference between the recent parameter sequence of the i-th target point and the recent parameter sequence of the h-th sampling point under the j-th soil index, and f be the DTW matching point number between the recent parameter sequence of the i-th target point and the recent parameter sequence of the h-th sampling point under the j-th soil index; F be the total number of DTW matching points between the recent parameter sequence of the i-th target point and the recent parameter sequence of the h-th sampling point under the j-th soil index; Δt ihjf Let f be the time interval between the recent parameter sequence of the i-th target point and the recent parameter sequence of the h-th sampling point under the j-th soil index; e is the natural constant; norm{} is the standard normalization function.

[0011] Furthermore, the method for obtaining the reference degree includes:

[0012] The spatial distance between each sampling point and the target point is obtained. The spatial distance is negatively correlated and normalized to obtain the distance weight of the anomaly analysis of the target point for the corresponding sampling point. The recent fluctuation similarity of the same soil index at the target point and each of the other sampling points is multiplied by the corresponding distance weight to obtain the reference degree of the anomaly analysis of the target point for each sampling point under each soil index.

[0013] Furthermore, the method for obtaining the abnormal feature values ​​includes:

[0014] The abnormal feature value is obtained according to the calculation formula of the abnormal feature value; the calculation formula of the abnormal feature value is:

[0015] Where, β ijg Let ε be the abnormal characteristic value of the monitoring parameter at time g in the soil parameter sequence of the j-th soil index at the i-th target point; I is the number of all sampling points in the planting area; h is the sequence number of the sampling point; ε ijg ε is the preset range of the monitoring parameters at time g in the soil parameter sequence of the j-th soil index at the i-th target point, and the standard deviation of all monitoring parameters; hjg c is the standard deviation of all monitoring parameters within a preset range of the monitoring parameters at time g in the soil parameter sequence of the j-th soil index at the h-th sampling point; ihj This serves as the reference level for the anomaly analysis of the h-th sampling point under the j-th soil index on the i-th target point.

[0016] Furthermore, the method for obtaining the degree of isolated anomalies includes:

[0017] Based on the fluctuation characteristics of each monitoring parameter of each soil index at the target point within the corresponding preset range, the amplitude constant characteristic value of the corresponding soil index at the target point is obtained; the amplitude constant characteristic value of each soil index at the target point is negatively correlated and a preset normal number is added, and then multiplied by the abnormal characteristic value of the monitoring parameter of the corresponding soil index at each sampling time. The product is normalized to obtain the isolation anomaly degree of the monitoring parameter of each soil index at the target point at each sampling time.

[0018] Furthermore, the method for obtaining the amplitude constant eigenvalue includes:

[0019] The amplitude constant feature is obtained according to the calculation formula of the amplitude constant feature value; the calculation formula of the amplitude constant feature value is:

[0020] Among them, f ij Let be the amplitude constant characteristic value of the j-th soil index at the i-th target point; N is the total number of monitoring parameters in the soil parameter sequence of the j-th soil index at the i-th target point; n is the sequence number of the monitoring parameter in the soil parameter sequence of the j-th soil index at the i-th target point; a ijn ε represents the parameter difference between the nth and (n-1)th monitoring parameters in the soil parameter sequence for the j-th soil index at the i-th target point; ijn ε is the standard deviation of all monitoring parameters within a preset range of the nth monitoring parameter in the soil parameter sequence of the jth soil index at the i-th target point; m is the type number of the soil index; J is the total number of soil indices; ε imnWithin a time period of the preset range of the nth monitoring parameter in the soil parameter sequence of the jth soil index at the i-th target point, the standard deviation of all monitoring parameters corresponding to the m-th soil index at the i-th target point; norm{} is the standard normalization function, and δ is the preset positive parameter.

[0021] Furthermore, the method for obtaining the optimal interpolation curve includes:

[0022] The fitting error of each interpolation curve is obtained based on the isolation anomaly degree of each monitoring parameter. The interpolation curve with the smallest fitting error among all the interpolation curves corresponding to each soil index at each sampling point is taken as the optimal interpolation curve for the corresponding soil index within the corresponding sampling point.

[0023] Furthermore, the method for obtaining the fitting error includes:

[0024] The parameter differences between the soil parameter sequence and the monitoring parameters at the same sampling time in the corresponding interpolation curve are obtained. The isolation anomalies of the monitoring parameters at the corresponding sampling time are negatively correlated, normalized, and then multiplied by the parameter differences to obtain the fitting deviation of the monitoring parameters at each sampling time. The fitting deviations of the monitoring parameters at all sampling times in the corresponding interpolation curve are summed to obtain the fitting error of the corresponding interpolation curve.

[0025] Furthermore, the sampling points are obtained using grid sampling.

[0026] Furthermore, the method for obtaining the recent parameter sequence includes:

[0027] The data segment between the second endpoint and the first endpoint in the soil parameter sequence is taken as the recent parameter sequence, with the end time of the soil parameter sequence as the first endpoint and a preset time earlier than the end time as the second endpoint.

[0028] The present invention has the following beneficial effects:

[0029] This invention obtains the reference degree for anomaly analysis between sampling points based on the recent fluctuation similarity and spatial distribution differences between sampling points. The reference degree combines recent fluctuation information and spatial information to reflect the similarity and anomaly analysis reference value between sampling points. Simultaneously, based on the characteristic that monitoring parameters at the same sampling time corresponding to similar sampling points should have similar fluctuation characteristics, the invention obtains the isolation anomaly degree of each monitoring parameter by combining the fluctuation characteristics of each monitoring parameter in the soil parameter sequence of each soil index at each sampling point. Multiple interpolation fittings are performed on the soil parameter sequence, and the optimal interpolation curve among all interpolation curves corresponding to the monitoring parameter sequence is evaluated based on the isolation anomaly degree of the monitoring parameter. The larger the isolation anomaly degree, the greater the possibility that the monitoring parameter at that time is noise, and the lower the reference value of its corresponding fitted parameter for the final fitting result. The optimal interpolation curve selected based on the isolation anomaly degree best reflects the true changes of the monitoring parameters. Furthermore, the monitoring parameters of all soil indices are integrated at each time point to evaluate the soil state at that time to regulate irrigation parameters. This invention combines the isolation anomaly degree of each monitoring parameter to evaluate interpolation fitting error, reducing the impact of abnormal noise on interpolation fitting, thereby accurately evaluating the soil state at each time point. Attached Figure Description

[0030] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0031] Figure 1 This is a flowchart illustrating a soil-condition-based backwashing fertigation method according to an embodiment of the present invention. Detailed Implementation

[0032] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the backwashing fertigation method based on soil conditions proposed in this invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0033] 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 invention pertains.

[0034] The following description, in conjunction with the accompanying drawings, details the specific scheme of the backwashing fertigation method based on soil conditions provided by this invention.

[0035] Please see Figure 1 The diagram illustrates a method flowchart of a backwashing fertigation system based on soil conditions, according to an embodiment of the present invention. The method includes the following steps:

[0036] Step S1: Obtain the soil parameter sequence for each soil index at each sampling point in the planting area. The soil parameter sequence includes all monitoring parameters for the corresponding soil index at the corresponding sampling point.

[0037] To accurately assess the soil condition of the planting area, this invention samples and monitors soil data within the planting area, then performs anomaly analysis on the sampled and monitored soil data, assesses the degree of isolation anomaly of each monitoring parameter, and adjusts the interpolation deviation of each sampled and monitored parameter in subsequent interpolation fitting based on the degree of isolation anomaly to obtain more accurate interpolation results. Finally, at each time point, the soil condition at the current time point is assessed by comprehensively considering the parameters of all soil indicators, and irrigation is controlled accordingly.

[0038] In one embodiment of the present invention, soil in the planting area is sampled using a grid sampling method. Various sensors are set up at each sampling point, and monitoring data of various soil indicators are collected at corresponding sampling frequencies. The monitoring data includes the numerical values ​​of all soil indicators at each sampling point, including soil temperature, soil moisture, soil nitrogen content, soil phosphorus content, soil potassium content, and soil pH, which are key information for assessing soil condition. Since there are differences in magnitude and units between the monitoring data of each soil indicator, to facilitate subsequent isolated anomaly analysis of each monitoring data point, the monitoring data needs to be standardized and preprocessed, mapping each monitoring data point to a range of 0-10 to obtain corresponding monitoring parameters. The monitoring parameters of the soil indicators obtained by each sensor at each sampling point in the planting area at the corresponding sampling time are mapped into a sequence according to the collection time sequence to obtain a soil parameter sequence. Each element in the soil parameter sequence is the monitoring parameter of the corresponding soil indicator at the corresponding sampling point at the corresponding sampling time.

[0039] It should be noted that the sampling frequency for each soil indicator is determined based on the properties of the soil indicator and the model and configuration of the sensor. Different soil indicators have different sampling frequency settings, resulting in different numbers of monitoring parameters collected within the same time period for different soil indicators; that is, different soil parameter sequence lengths correspond to different soil indicators. Meanwhile, grid sampling is a commonly used analytical method in the field of science and will not be elaborated upon here. In one embodiment of the present invention, the sample point spacing for grid sampling is set to 5 meters, and monitoring parameters within one month are collected to construct a soil parameter sequence. In other embodiments of the present invention, implementers can also use other sample spacings or sampling methods to sample and monitor the soil and construct soil parameter sequences of other lengths according to actual needs. They can also set the number or types of soil indicators themselves, and set the specific sampling frequency and deployment location according to the properties of the soil indicators and the model and configuration of the sensor.

[0040] Step S2: Taking any sampling point as the target point, based on the similarity of recent fluctuations and spatial distribution differences of the monitoring parameters of the same soil index at the target point and other sampling points within a preset adjacent time period, obtain the reference degree of the anomaly analysis of each other sampling point for the target point under each soil index; based on the fluctuation characteristics of the monitoring parameters of the same soil index at the same sampling time between the target point and other sampling points within a preset range in the soil parameter sequence and the corresponding reference degree, obtain the abnormal characteristic value of the monitoring parameters of each soil index at the target point at each sampling time; combine the abnormal characteristic value and the fluctuation characteristics of each soil index at the target point to obtain the isolation anomaly degree of the monitoring parameters of each soil index at the target point at each sampling time.

[0041] Within a planting area, monitoring parameters corresponding to the same soil index tend to exhibit more similar variation characteristics among sampling points that are closer together. When an anomaly occurs in the soil condition at a target point, the likelihood of anomalies in the soil condition at neighboring sampling points is relatively higher. Therefore, the probability of anomalies at a sampling point can be determined by analyzing the similarity of fluctuations at different sampling points and the location information between sampling points. However, considering that analyzing the similarity of fluctuations of the same soil index at different sampling points based on all collected monitoring parameters not only involves a large amount of computation but also excessively smooths out some important fluctuation characteristics due to the long time span, making the analysis of the similarity of fluctuations of monitoring parameters of the same soil index at different sampling points inaccurate, this embodiment of the invention first uses any one sampling point as the target point for analysis among all sampling points. Then, based on the recent fluctuation similarity and spatial distribution differences between the corresponding soil parameter sequences of the same soil index at the target point and other sampling points within a preset adjacent time period, the reference degree of each sampling point for the anomaly analysis of the target point under each soil index is obtained.

[0042] Preferably, in one embodiment of the present invention, considering that the Dynamic Time Warping (DTW) algorithm typically calculates the shortest path between matching points in two sequences to obtain the final DTW distance, and then evaluates the similarity between sequences; if the difference between each pair of matching points is smaller, and the time span corresponding to the matching points in the corresponding sequences is smaller, it indicates that the two matching points not only have high similarity in amplitude, but also that the corresponding matching points are more aligned in time, and the pair of matching points has greater reference value for obtaining the similarity between the two sequences. Based on this, in the soil parameter sequence, the recent parameter sequence corresponding to the preset adjacent time period is obtained; DTW matching is performed on the recent parameter sequences of the same soil index at the target point and other sampling points to obtain all DTW matching points between the recent parameter sequences of the target point and the corresponding sampling points; then, the recent fluctuation similarity is obtained according to the recent fluctuation similarity calculation formula. The DTW algorithm for obtaining matching points is a prior art well known to those skilled in the art, and will not be described in detail here.

[0043] It should be noted that, in one embodiment of the present invention, the method for obtaining the recent parameter sequence is as follows: taking the end time of the soil parameter sequence as the first endpoint and a preset time earlier than the end time as the second endpoint, the data segment between the second endpoint and the first endpoint in the soil parameter sequence is taken as the recent parameter sequence. The time period corresponding to the data segment between the second endpoint and the first endpoint in each soil parameter sequence is 3 days, meaning the recent parameter sequence includes monitoring parameters collected in the last 3 days, thereby analyzing the fluctuation similarity of monitoring parameters corresponding to the same soil index within a preset adjacent time period. In other embodiments of the present invention, parameter sequences can also be constructed based on data segments corresponding to preset time periods of other durations according to actual needs, i.e., arbitrarily selecting the first endpoint and the second endpoint to construct the parameter sequence. Parameter sequences constructed from different endpoints corresponding to different time periods may not yield consistent results for analyzing fluctuation similarity. In this embodiment of the present invention, to more accurately assess the current abnormal fluctuations of parameters, a preset adjacent time period is selected to construct the recent parameter sequence.

[0044] The formula for calculating recent volatility similarity is:

[0045]

[0046] Among them, D ihj d represents the similarity of recent fluctuations of the j-th soil index at the i-th target point and the h-th sampling point; ihjfLet f be the parameter difference between the recent parameter sequence of the i-th target point and the recent parameter sequence of the h-th sampling point under the j-th soil index, and f be the DTW matching point number between the recent parameter sequence of the i-th target point and the recent parameter sequence of the h-th sampling point under the j-th soil index; F be the total number of DTW matching points between the recent parameter sequence of the i-th target point and the recent parameter sequence of the h-th sampling point under the j-th soil index; Δt ihjf Let f be the time interval between the recent parameter sequence of the i-th target point and the recent parameter sequence of the h-th sampling point under the j-th soil index; e is the natural constant; norm{} is the standard normalization function.

[0047] In the formula for calculating recent fluctuation similarity, the negative correlation of time intervals is mapped to an exponential function for normalization. This results in a higher time alignment, i.e., a larger time interval, corresponding to a larger time series scaling reference weight. The fluctuation trends of two recent parameter sequences at similar times are more similar. Then, the normalized value of the corresponding time interval is subtracted from 1 to adjust the logical relationship. If the parameter difference between the corresponding matching points is also smaller, the comprehensive parameter difference of the matching point is smaller, and the reference value for measuring the similarity between recent parameter sequences is greater. If the comprehensive difference corresponding to all matching points between recent reference sequences is very small, the difference between recent parameter sequences is smaller. After summing the comprehensive differences corresponding to all matching points and normalizing, the corresponding recent fluctuation similarity is greater by subtracting the corresponding normalized value from 1.

[0048] After obtaining the recent fluctuation similarity of monitoring parameters between the target point and each sampling point within a preset adjacent time period, and combining their spatial distribution differences, the reference degree of each sampling point for the anomaly analysis of the target point under each soil index is obtained.

[0049] Preferably, in one embodiment of the present invention, the method for obtaining the reference degree includes: obtaining the spatial distance between each sampling point and the target point; performing negative correlation mapping and normalization on the spatial distance to obtain the distance weight of the anomaly analysis of the corresponding sampling point to the target point; multiplying the recent fluctuation similarity of the same soil index at the target point and each of the other sampling points by the corresponding distance weight to obtain the reference degree of the anomaly analysis of each sampling point to the target point for each soil index. The formula for calculating the reference degree is:

[0050]

[0051] Among them, c ihj D serves as the reference level for anomaly analysis of the h-th sampling point under the j-th soil index relative to the i-th target point; ihj Let l represent the similarity of recent fluctuations of the j-th soil index at the i-th target point and the h-th sampling point; ihLet be the spatial distance between the i-th target point and the h-th sampling point; e is the natural constant.

[0052] It should be noted that the spatial distance between the target point and the sampling point is calculated using Euclidean distance, and the calculation method is a well-known existing technology in the field of science, and will not be described in detail here.

[0053] In the formula for calculating the reference degree, the negative correlation between the spatial distance between the sampling point and the target point is mapped to an exponential function for normalization. Then, the normalized distance weight is combined with the recent fluctuation similarity through multiplication, so that both the distance weight and the recent fluctuation similarity are positively correlated with the reference degree. In other embodiments of the present invention, the spatial distance can also be normalized by other normalization methods, or the distance weight can be combined with the recent fluctuation similarity by addition or a power function.

[0054] Within a planting area, sampling points that are closer together and have similar recent fluctuations (i.e., have higher reference value) should exhibit similar fluctuation trends in the monitoring parameters of the same soil index at the same sampling time within their corresponding local ranges. If there are significant differences in the fluctuations of the monitoring parameters of the same soil index at the same sampling time, it indicates that the monitoring parameters of the corresponding soil index at that sampling time are highly likely to be noise. Therefore, this embodiment of the invention obtains the abnormal characteristic values ​​of the monitoring parameters of each soil index at the target point at each sampling time based on the differences in fluctuation characteristics and corresponding reference value between the target point and other sampling points of the same soil index at the same sampling time within their corresponding preset ranges.

[0055] Preferably, in one embodiment of the present invention, considering that the fluctuation difference of the monitoring parameters of the same soil index at the target point and all other sampling points at the same sampling time reflects the possibility that there is noise in the monitoring parameters of the corresponding soil index at the target point at that sampling time; at the same time, considering that the reference degree of each sampling point to the target point combines recent fluctuation information and spatial information to reflect the similarity and reference value between the sampling point and the target point, that is, the local fluctuation characteristics of the monitoring parameters of the same soil index at the same sampling time of sampling points with high reference degree to the target point should also be similar; based on this, abnormal feature values ​​can be obtained.

[0056] It should be noted that, in this embodiment of the invention, the least common multiple of the sampling frequencies of all soil indicators is used as the preset range corresponding to the time duration. Preset ranges of the same duration but different sizes are constructed for each parameter of each soil indicator to accurately analyze the fluctuations of the monitoring parameters corresponding to each soil indicator within the same time period. The preset range size of the monitoring parameter for the corresponding soil indicator can be obtained by dividing the time duration by the sampling frequency of each soil indicator. In other embodiments of the invention, preset ranges of other sizes can also be determined according to actual needs.

[0057] The formula for calculating abnormal feature values ​​is:

[0058]

[0059] Where, β ijg Let ε be the abnormal characteristic value of the monitoring parameter at time g in the soil parameter sequence of the j-th soil index at the i-th target point; I is the number of all sampling points in the planting area; h is the sequence number of the sampling point; ε ijg ε is the standard deviation of all monitoring parameters within a preset range of the monitoring parameters at time g in the soil parameter sequence of the j-th soil index at the i-th target point; hjg The preset range of the monitoring parameters at time g in the soil parameter sequence of the j-th soil index at the h-th sampling point is the standard deviation of all monitoring parameters; c ihj This serves as the reference level for the anomaly analysis of the h-th sampling point under the j-th soil index on the i-th target point.

[0060] In the formula for calculating the abnormal characteristic value, the reference degree of the sampling point to the target point for anomaly analysis is used as the reference weight for the local fluctuation difference of the monitoring parameters of the same soil index at the same time. The greater the reference degree between the sampling point and the target point, the smaller the corresponding local fluctuation difference should be. However, if the local fluctuation difference is larger at this time, it indicates that the monitoring parameters of the same soil index at the same time are more likely to be abnormal, and the abnormal characteristic value is larger.

[0061] Considering that when the planting and irrigation conditions in the planting area are not interfered with by humans, the monitoring parameters of various soil indicators will fluctuate accordingly with the growth and absorption of crops. However, the absorption rates of water and various nutrients by crops are not the same, which means that the fluctuation degree of the monitoring parameters corresponding to different soil indicators may not be consistent. Under normal circumstances, the monitoring parameters of some soil indicators may tend to fluctuate drastically, that is, the monitoring parameters show isolated fluctuation characteristics. Therefore, when analyzing the degree of anomaly of such soil indicators, it is necessary to combine their own fluctuation tendency, that is, the constant amplitude characteristics, to appropriately reduce the sensitivity of anomaly degree analysis. Therefore, the embodiments of the present invention combine the anomaly characteristic values ​​and the fluctuation characteristics of each soil indicator at the target point to obtain a more accurate degree of isolated anomaly of the monitoring parameters of each soil indicator at each sampling time within the target point.

[0062] Preferably, in one embodiment of the present invention, the method for obtaining the isolation anomaly degree includes: obtaining the amplitude constant characteristic value of the corresponding soil index at the target point based on the fluctuation characteristics of each monitoring parameter of each soil index at the target point within a corresponding preset range; performing negative correlation mapping on the amplitude constant characteristic value of each soil index at the target point, adding a preset normal number, multiplying by the anomaly characteristic value of the monitoring parameter of the corresponding soil index at each sampling time, and normalizing the product to obtain the isolation anomaly degree of the monitoring parameter of each soil index at the target point at each sampling time. The formula for calculating the isolation anomaly degree is:

[0063] k ijg =norm{(γ-f ij )×β ijg}

[0064] Where, k ijg f is the isolation anomaly degree of the monitoring parameter corresponding to the j-th soil index in the soil parameter sequence at the i-th target point at the g-th time step; ij β is the constant characteristic value of the amplitude of the j-th soil index at the i-th target point; ijg Let f be the abnormal characteristic value of the monitoring parameter at the g-th time in the soil parameter sequence of the j-th soil index at the i-th target point; norm{} is the standard normalization function; γ is the preset normal number. In one embodiment of the present invention, since f ij The value range of f is 0-1, so γ is 1 to change f. ij With β ijg The corresponding logical relationships between them can be set by the implementer according to the specific implementation situation.

[0065] In the formula for calculating isolation anomaly, the amplitude constant characteristic value reflects the amplitude fluctuation tendency of the corresponding soil index at the target point. The larger this value, the more isolated and discrete the soil index exhibits under normal conditions. By subtracting the amplitude constant characteristic value from 1 to change the logical relationship, when analyzing the isolation anomaly of the monitoring parameter at each sampling time, the larger the amplitude constant characteristic value, the more the monitoring parameter of the corresponding soil index tends to have isolated and outlier fluctuation characteristics, and the lower the sensitivity should be in the isolation anomaly analysis of the monitoring parameter. Then, the two are combined by multiplication, and (1-f ij Using these as reference weights for anomalous feature values, a more accurate degree of isolated anomalies can be obtained.

[0066] Preferably, in one embodiment of the present invention, considering that the amplitude difference between adjacent parameters reflects the change of the parameters, and the standard deviation of the parameter amplitude within a local range reflects the degree of fluctuation of the parameters; and considering that noise will also exhibit certain fluctuation characteristics within the corresponding local range, but the possibility of noise interference causing all soil indicators to fluctuate simultaneously is extremely low, the reliability of the amplitude fluctuation of the monitoring parameters corresponding to the local range of the soil indicator can be evaluated by the fluctuation of the monitoring parameters of all soil indicators within the same time period within the local range. Furthermore, the amplitude constant characteristic value of the soil indicator corresponding to the target point can be evaluated based on the fluctuation of all monitoring parameters in the soil parameter sequence and the fluctuation reliability.

[0067] The constant amplitude characteristic is obtained according to the formula for calculating the constant amplitude characteristic value; the formula for calculating the constant amplitude characteristic value is:

[0068]

[0069] Among them, f ij Let be the amplitude constant characteristic value of the j-th soil index at the i-th target point; N is the total number of monitoring parameters in the soil parameter sequence of the j-th soil index at the i-th target point; n is the sequence number of the monitoring parameter in the soil parameter sequence of the j-th soil index at the i-th target point; a ijn ε represents the parameter difference between the nth monitored parameter and the previous adjacent monitored parameter in the soil parameter sequence of the j-th soil index at the i-th target point; ijn ε is the standard deviation of all monitoring parameters within a preset range of the nth monitoring parameter in the soil parameter sequence of the jth soil index at the i-th target point; m is the type number of the soil index; J is the total number of soil indices; ε imn Within a time period of the preset range of the nth monitoring parameter in the soil parameter sequence of the jth soil index at the i-th target point, the standard deviation of all monitoring parameters corresponding to the mth soil index at the i-th target point; δ is a preset positive parameter. In one embodiment of the present invention, δ is taken as 0.01 to prevent the denominator from being 0. The implementer may also set other values ​​according to the specific implementation situation.

[0070] In the formula for calculating the eigenvalue of the constant amplitude The fluctuation information of all soil indicators at the i-th target point within the same time period reflects the reliability of the fluctuation analysis of the n-th monitoring parameter of the j-th soil indicator at the i-th target point within the corresponding preset range. When the fluctuation characteristics of the monitoring parameters of each other soil indicator within the same time period are relatively small compared with the fluctuation characteristics of the monitoring parameters of the j-th soil indicator, it indicates that the fluctuation characteristics of the j-th soil indicator are more reliable. At the same time, the greater the parameter difference between adjacent monitoring parameters and the larger the standard deviation of the monitoring parameter in the local range, the more likely the monitoring parameter is to be isolated, discrete, and fluctuate violently. Then, by combining the fluctuation characteristics and fluctuation reliability of all monitoring parameters in the soil parameter sequence, the amplitude constant characteristic value of the corresponding soil indicator is obtained. The larger the amplitude constant characteristic value, the more violent the fluctuation of the soil indicator is under normal conditions, and all monitoring parameters show relatively more isolated and discrete characteristics. By changing the target point and obtaining the isolation anomaly degree corresponding to each monitoring parameter in each soil monitoring sequence according to the above-mentioned method of obtaining isolation anomaly degree, the interpolation results can be adjusted according to the isolation anomaly degree of each monitoring parameter when performing interpolation fitting based on the monitoring parameters in the future, so as to obtain more accurate parameter monitoring results and thus accurately assess the soil condition to regulate irrigation.

[0071] Step S3: Obtain at least two interpolation curves for each soil parameter sequence; based on the differences and isolation anomalies of the monitoring parameters at the same sampling time in the soil parameter sequence and the corresponding interpolation curve, select the optimal interpolation curve for each soil index at each sampling point, and then automatically configure the water and fertilizer ratio, and send a drip irrigation signal from the control system to precisely irrigate the soil.

[0072] Since monitoring data for different soil indicators are collected at different sampling frequencies, it may be impossible to obtain the monitoring parameters corresponding to each soil indicator at the same time. Interpolation fitting can perform smooth function fitting between existing monitoring parameters, thereby providing continuous estimates and obtaining the monitoring parameters of all soil indicators at each same time. However, there may be noisy monitoring parameters among the existing monitoring parameters, which will seriously affect the accuracy of the final interpolation fitting. Therefore, in this embodiment of the invention, the monitoring parameters in the soil detection sequence of each soil indicator at each sampling point are repeatedly interpolated and fitted. At least two interpolation curves are obtained for each soil parameter sequence, and then the optimal interpolation curve is obtained as the final interpolation fitting result for the corresponding soil parameter sequence. In one embodiment of the invention, the interpolation curve of each soil parameter sequence is repeatedly fitted 10 times using the spline interpolation method. After each fitting, the spline interpolation parameters are adjusted using the gradient descent method, and the next spline interpolation is performed based on the adjusted spline interpolation parameters to obtain the interpolation curve. The implementer can also set the number of fittings and other spline interpolation parameter adjustment methods according to actual needs. Spline interpolation and gradient descent are techniques well-known to those skilled in the art, and will not be elaborated upon here.

[0073] After obtaining all interpolation curves, the difference between the monitoring parameters at the same sampling time in the soil parameter sequence and the corresponding interpolation curve, as well as the degree of isolation anomaly, are used. Specifically, the difference between the monitoring parameters in the soil parameter sequence at the same time and the fitted parameters obtained by interpolation, and the degree of isolation anomaly of the monitoring parameters at that time, are used. The greater the degree of isolation anomaly, the greater the possibility that the monitoring parameters at that time are abnormal noise, and the lower the reference value of the corresponding fitted parameters for the final fitting result. Then, a more accurate fitting error for each interpolation curve can be obtained, and the optimal interpolation curve can be selected from all the interpolation curves of each soil parameter sequence.

[0074] In a preferred embodiment of the present invention, the fitting error of each interpolation curve is obtained based on the isolation anomaly degree of each monitoring parameter, and then the interpolation curve with the smallest fitting error among all interpolation curves for each soil index at each sampling point is taken as the optimal interpolation curve for the corresponding soil index within the corresponding sampling point.

[0075] Preferably, in one embodiment of the present invention, the method for obtaining the fitting error includes: obtaining the parameter differences between the soil parameter sequence and the monitoring parameters at the same sampling time in the corresponding interpolation curve; performing negative correlation mapping on the isolation anomalies of the monitoring parameters at the corresponding sampling time, normalizing them, and then multiplying them by the parameter differences to obtain the fitting deviation of the monitoring parameters at each sampling time; and summing the fitting deviations of the monitoring parameters at all sampling times in the corresponding interpolation curve to obtain the fitting error of the corresponding interpolation curve. The specific formula for calculating the fitting deviation is as follows:

[0076] B ijg =Δb ijg ×(1-k ijg )

[0077] Among them, B ijg Δb represents the fitting bias at time g in the interpolation curve corresponding to the soil parameter sequence of the j-th soil index at the i-th sampling point; ijg k is the parameter deviation between the monitored parameter at time g and the fitted parameter in the corresponding interpolation curve in the soil parameter sequence of the j-th soil index at the i-th sampling point; ijg Let be the isolation anomaly degree of the monitoring parameter corresponding to the j-th soil index in the soil parameter sequence at the i-th target point at the g-th time step.

[0078] In the formula for calculating the fitting bias, 1-k ijg For isolated outlier degree k ijg Perform negative correlation mapping and normalization to adjust their logical relationships, i.e., the isolation anomaly degree k. ijg The larger the value, the more likely it is to correspond to 1-k. ijg The smaller the value, the lower the reference value for the fitting deviation; the fitting deviation Δb is multiplied.ijg with (1-k) ijg The fitting deviation of each monitoring parameter is combined and comprehensively evaluated. If the fitting deviation is large and the corresponding isolation anomaly degree is small, the fitting effect of the monitoring parameter is poor.

[0079] By summing the fitting deviations of all known monitoring parameters and fitting parameters on each interpolation curve, the fitting error of the corresponding interpolation curve can be obtained. Then, based on the fitting errors of all interpolation curves for each soil index at each sample point, the optimal interpolation curve is selected as the final monitoring curve for the corresponding soil index at that sample point.

[0080] After obtaining the optimal interpolation curve for each soil index at each sampling point, the implementer can transmit the optimal interpolation curve to the backwashing fertigation machine. The machine automatically assesses the soil condition at each sampling point at the current moment, and then automatically configures the water and fertilizer ratio. The control system sends a drip irrigation signal to the drip irrigation device at each sampling point to precisely irrigate and fertilize the soil in the corresponding area of ​​the sampling point. Alternatively, it can combine the optimal interpolation curves of all soil indices at each sampling point to assess the soil quality of the current planting area, and then formulate targeted soil use policies and restoration and protection measures based on the current soil condition to reasonably protect soil resources and maintain the balance of the ecosystem.

[0081] In summary, this invention obtains the reference degree for anomaly analysis between sampling points based on the recent fluctuation similarity and spatial distribution differences between sampling points; based on the characteristic that monitoring parameters corresponding to similar sampling points at the same sampling time should have similar fluctuation characteristics, it obtains the isolation anomaly degree of each monitoring parameter by combining the fluctuation characteristics of each monitoring parameter in the soil parameter sequence of each soil index at each sampling point; it performs multiple interpolation fittings on each soil parameter sequence and selects the optimal interpolation curve among all interpolation curves corresponding to the soil parameter sequence based on the isolation anomaly degree of the monitoring parameter. The optimal interpolation curve selected based on the isolation anomaly degree best reflects the true changes of the monitoring parameters; and then, at each time point, it integrates the monitoring parameters of all soil indices to evaluate the soil state at the current time to regulate irrigation parameters. This invention combines the isolation anomaly degree of each monitoring parameter to evaluate the interpolation fitting error, reduces the influence of abnormal noise on the interpolation fitting, and thus accurately evaluates the soil state at each time point.

[0082] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0083] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A backwashing water and fertilizer integrated machine irrigation method based on soil state, characterized in that, The method comprises: Obtaining a soil parameter sequence of each sampling point of each soil index in the planting area, wherein the soil parameter sequence comprises all monitoring parameters of the corresponding soil index at the corresponding sampling point; Taking any sampling point as a target point, obtaining a reference degree of abnormal analysis of each sampling point on the target point according to a recent fluctuation similarity and a spatial distribution difference of monitoring parameters of the same soil index within a preset adjacent period at the target point and the remaining sampling points, obtaining an abnormal feature value of the monitoring parameters of each soil index at each sampling time at the target point according to a fluctuation feature difference of the monitoring parameters of the same soil index at the same sampling time between the target point and the remaining sampling points within a preset range in the soil parameter sequence and the reference degree corresponding to the reference degree, and obtaining an isolated abnormal degree of the monitoring parameters of each soil index at each sampling time at the target point by combining the abnormal feature value and the fluctuation feature of the corresponding monitoring parameters of each soil index at the target point. Obtaining at least two interpolation curves of each soil parameter sequence, and screening an optimal interpolation curve of each soil index at each sampling point according to a difference of monitoring parameters at the same sampling time between the soil parameter sequence and the interpolation curve and the isolated abnormal degree. The isolated abnormal degree comprises: Obtaining an amplitude constant feature value of the corresponding soil index at the target point according to a fluctuation feature of each monitoring parameter of each soil index within a corresponding preset range at the target point, performing negative correlation mapping on the amplitude constant feature value of each soil index at the target point, adding a preset normal number, multiplying the abnormal feature value of the monitoring parameters of the corresponding soil index at each sampling time, normalizing the product, and obtaining the isolated abnormal degree of the monitoring parameters of each soil index at each sampling time at the target point. The amplitude constant feature value comprises: Obtaining the amplitude constant feature value according to a calculation formula of the amplitude constant feature value, wherein the calculation formula of the amplitude constant feature value is: ; wherein, is a constant amplitude eigenvalue of the soil index of the th target point; th soil index; is a total number of monitoring parameters in the soil parameter sequence of the th soil index of the th target point; th monitoring parameter; th monitoring parameter; th monitoring parameter of the soil parameter sequence of the th soil index of the th target point; th monitoring parameter; th monitoring parameter of the soil parameter sequence of the th target point; th monitoring parameter of the soil parameter sequence of the th target point; th monitoring parameter; th monitoring parameter of the soil parameter sequence of the is a category number of the soil index; is a total number of categories of the soil index; th monitoring parameter of the soil parameter sequence of the th target point; th monitoring parameter; th target point; th monitoring parameter of the soil parameter sequence of the th target point; is a standard normalization function, is a preset positive parameter.

2. The soil condition-based backwash water and fertilizer integrated machine irrigation method according to claim 1, characterized in that, The recent fluctuation similarity comprises: In the soil parameter sequence, obtaining a recent parameter sequence corresponding to a preset adjacent period, performing DTW matching on the recent parameter sequence of the same soil index at the target point and each sampling point, obtaining all DTW matching points between the recent parameter sequence of the target point and the corresponding sampling point, and obtaining the recent fluctuation similarity according to a calculation formula of the recent fluctuation similarity, wherein the calculation formula of the recent fluctuation similarity is: ;in, For the first Soil index in the first The target point and the first Similarity of recent fluctuations at each sampling point; For the first Soil index under the first The recent parameter sequence of the target point and the... The recent parameter sequence of the sampling point is between the th sampling point and the th sampling point. The differences in parameters between the corresponding monitoring parameters at the matching points; For the first Soil index under the first The recent parameter sequence of the target point and the... The sequence number of DTW matching points between recent parameter sequences of each sampling point; For the first Soil index under the first The recent parameter sequence of the target point and the... The total number of DTW matching points between recent parameter sequences of each sampling point; For the first Soil index under the first The recent parameter sequence of the target point and the... The recent parameter sequence of the sampling point is between the th sampling point and the th sampling point. The time interval between matching points; It is a natural constant; This is the standard normalization function.

3. The soil condition-based backwash water and fertilizer integrated machine irrigation method according to claim 1, characterized in that, The reference degree comprises: Obtaining a spatial distance between each sampling point and the target point, performing negative correlation mapping and normalization on the spatial distance to obtain a distance weight of abnormal analysis of the target point by the corresponding sampling point, multiplying the recent fluctuation similarity of the same soil index at the target point and the remaining sampling points by the distance weight corresponding to the distance weight, and obtaining the reference degree of abnormal analysis of the target point by each sampling point under each soil index.

4. The soil condition-based backwash water and fertilizer integrated machine irrigation method according to claim 1, characterized by, The abnormal feature value comprises: Obtaining the abnormal feature value according to a calculation formula of the abnormal feature value, wherein the calculation formula of the abnormal feature value is: ; wherein, is the abnormal feature value of the monitoring parameter corresponding to the i-th moment in the soil parameter sequence of the j-th soil index of the i-th target point; is the number of all sampling points in the planting area; is the serial number of the sampling point; <000010 5. The soil condition-based backwash water and fertilizer integrated machine irrigation method according to claim 1, characterized in that, The method for obtaining the optimal interpolation curve comprises the following steps: According to the isolated abnormality of each monitoring parameter, a fitting error of each interpolation curve is obtained, and the interpolation curve with the minimum fitting error of each soil index of each sampling point is taken as the optimal interpolation curve of the corresponding soil index in the corresponding sampling point. 6.The soil condition-based backwash water and fertilizer integrated machine irrigation method according to claim 5, characterized in that, The method for obtaining the fitting error comprises the following steps: A parameter difference between the soil parameter sequence and the monitoring parameter at the same sampling time in the corresponding interpolation curve is obtained, the isolated abnormality of the monitoring parameter at the corresponding sampling time is negatively correlated and normalized, and then multiplied by the parameter difference to obtain a fitting deviation of the monitoring parameter at each sampling time; the fitting deviations of the monitoring parameters at all sampling times in the corresponding interpolation curve are summed to obtain the fitting error of the corresponding interpolation curve.

7. The soil condition-based backwash water and fertilizer integrated machine irrigation method according to claim 1, characterized in that, The method for obtaining the sampling point is grid sampling. 8.The soil condition-based backwash water and fertilizer integrated machine irrigation method according to claim 2, wherein, The method for obtaining the recent parameter sequence comprises the following steps: The end time of the soil parameter sequence is taken as a first endpoint, a preset time earlier than the end time is taken as a second endpoint, and a data segment between the second endpoint and the first endpoint in the soil parameter sequence is taken as the recent parameter sequence.

Images