Conductivity detection and evaluation method for soil element detection

Abstract

The invention relates to the technical field of soil detection, and discloses a conductivity detection and evaluation method for soil element detection, comprising the following steps: acquiring conductivity data and environmental characteristic data of target soil; constructing a space-time matrix and an evolution sequence of the target soil conductivity; extracting conductance response characteristics of the target soil; distributing an influence factor for each conductance response feature; obtaining a reference response feature of each content change mode of the target element, and distributing the credibility of each conductance response feature for each content change mode; calculating the confidence degree of each content change mode based on the confidence degree of the conductance response characteristics and the influence factors; judging a content change mode of a target element in the target soil based on the confidence coefficient; according to the scheme, the element content change mode in the soil can be accurately captured, the detection process can be simplified, and the recognition capability of the target element can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of soil detection, and in particular to a soil element detection conductivity detection evaluation method. BACKGROUND

[0002] Current soil element detection technology can utilize the correlation between ion conductivity and element content in soil solution to achieve rapid detection of target element content, and can be applied to farmland nutrient management, saline-alkali land treatment and other scenarios. Existing technologies focus on single-point instantaneous measurement, and obtain soil surface conductivity data through portable devices for rough assessment of soil fertility. Although soil conductivity detection technology has made some progress, there are still some problems in practical application.

[0003] Existing technologies focus on single time point or single depth conductivity data, lack of systematic characterization of spatial and temporal distribution of soil conductivity, and are difficult to reflect the migration path and time dynamic characteristics of elements in the vertical section of soil. Soil conductivity is significantly affected by environmental factors such as water content and pH value, but existing methods have not established a correlation mechanism between environmental factors and conductivity response characteristics, resulting in a lack of comparability of detection results under different environmental conditions. Different elements have significantly different migration characteristics in soil, but existing methods mostly use a unified analysis model, do not construct a differentiated recognition mechanism for element specificity, and are difficult to distinguish the conductivity contribution of target elements and non-target elements, resulting in a high misjudgment rate of element content change pattern.

[0004] A soil element online detection device and method are disclosed in Chinese patent with authorization announcement No. CN103185708B, which includes a soil detection component and a laser emission and detection system. The soil detection component is connected to the laser emission and detection system through an optical fiber. The soil detection component includes a housing and a focusing lens arranged in the housing. The laser emission and detection system includes a laser emission system, a laser collection system and a control system. The laser emission system is connected to the focusing lens and the control system respectively. The laser collection system is connected to the focusing lens and the control system respectively. This technical solution can realize online and in-situ detection of soil in a field environment.

[0005] A patent application with publication number CN119779981A discloses an optical detection device, a soil element detector and a soil element detection method. The device comprises: a laser beam expansion and collimation system that inputs the laser emitted by the laser into a multiplexing and folding optical path system after beam expansion and collimation; the multiplexing and folding optical path system folds the expanded and collimated laser beam to obtain a folded system exit laser input face scanning optical path system; the face scanning optical path system focuses the folded system exit laser onto a face scanning target window to generate plasma; a spectral signal acquisition optical path system acquires the signal light radiated by the plasma; and an imaging system images the window of the face scanning target window on a CMOS camera. The device modularizes functions such as laser beam expansion, collimation, folding, focusing and signal acquisition through the design of multiple optical systems, optimizes each module independently to improve the stability of the overall device, thereby improving the stability of the data, achieving rapid and accurate scanning detection of the measured surface, and improving the anti-disturbance ability.

[0006] The above technical solutions all have the problems raised in the background art: no element-specific differential recognition mechanism is constructed, and the misjudgment rate of element content change patterns is high.

[0007] The information disclosed in this background section is intended only to increase an understanding of the general context of the present application, and it should not be taken as an acknowledgement or any form of suggestion that this information forms prior art that is already widely known in the art before the filing date of the present application. SUMMARY

[0008] The technical problem to be solved by the present application is to overcome the defects of the prior art, provide a soil element detection conductivity detection evaluation method, accurately capture the element content change pattern in the soil, simplify the detection process, and improve the recognition ability of the target element.

[0009] To solve the above technical problems, the present application provides the following technical solutions:

[0010] A soil element detection conductivity detection evaluation method comprises the following steps:

[0011] Obtain the conductivity data and environmental characteristic data of the target soil;

[0012] Construct a space-time matrix and evolution sequence of the conductivity of the target soil based on the conductivity data;

[0013] Extract the conductivity response characteristics of the target soil based on the space-time matrix and evolution sequence; and assign an influence factor to each conductivity response characteristic based on the environmental characteristic data;

[0014] Obtain the reference response characteristics of each content change pattern of the target element; and assign a credibility of each conductivity response characteristic to each content change pattern based on the reference response characteristics;

[0015] The confidence level of each content change pattern is calculated based on the reliability of the conductivity response characteristics and the influencing factors; the content change pattern of the target element in the target soil is determined based on the confidence level.

[0016] As a preferred embodiment of the electrical conductivity detection and evaluation method for soil element detection described in this application, the method for obtaining electrical conductivity data of the target soil specifically includes: setting a sampling period; at the beginning of each sampling period, sampling the target soil in N layers in the vertical direction, and detecting the electrical conductivity of each soil sample; N is a positive integer; at the beginning of each sampling period, environmental characteristic data of the target soil are collected synchronously.

[0017] The construction of the spatiotemporal matrix for the target soil electrical conductivity specifically includes: initializing an M-row, N-column spatiotemporal matrix; M being a positive integer; acquiring electrical conductivity data from the most recent M sampling periods and filling the spatiotemporal matrix with elements; wherein, the element in the i-th row and j-th column of the spatiotemporal matrix represents the electrical conductivity of the j-th layer of soil sample in the i-th sampling period among the most recent M sampling periods; the value range of i is 1, 2, ..., M; the value range of j is 1, 2, ..., N;

[0018] The evolution sequence has N elements, where each element is the mean of all elements in the corresponding column of the spatiotemporal matrix.

[0019] As a preferred embodiment of the soil element detection and evaluation method for electrical conductivity as described in this application, the electrical conductivity response characteristics include time-series response characteristics, spatial response characteristics, and spatiotemporal correlation characteristics.

[0020] The time-series response features include at least one of the conductivity rise rate and conductivity peak shape ratio; the time-series response features are extracted based on the evolution sequence of the target soil conductivity; wherein, the method for extracting the conductivity rise rate is as follows:

[0021] The electrical conductivity of the target soil was fitted with a two-dimensional curve based on the evolution sequence, and the average slope of the fitted curve was extracted as the rate of increase of electrical conductivity.

[0022] The spatial response characteristics include at least the electrical conductivity gradient between any two adjacent soil samples.

[0023] The conductivity gradient is extracted based on the spatiotemporal matrix of the target soil conductivity, and the method is as follows: the mean conductivity of each soil sample layer is calculated based on the spatiotemporal matrix; wherein, the mean conductivity of any soil sample layer is the mean of all elements in the corresponding row of the spatiotemporal matrix; for any two adjacent soil samples, the mean conductivity of the upper soil sample is subtracted from the mean conductivity of the lower soil sample to obtain the corresponding conductivity gradient.

[0024] As a preferred embodiment of the conductivity detection and evaluation method for soil element detection described in this application, the method for extracting the conductivity peak shape ratio is as follows:

[0025] Based on the fitted conductivity curve, identify the peaks of conductivity variation; record the time corresponding to the peak value of each peak.

[0026] In the coordinate system of the conductivity curve, each peak is intercepted by a line parallel to the x-axis to obtain the start and end points of each peak.

[0027] For any peak, subtract the starting point from the time corresponding to the peak to obtain the first change time of the peak; subtract the peak from the ending point to obtain the second change time of the peak.

[0028] Calculate the peak shape ratio of each variation peak; the peak shape ratio of any variation peak is the ratio of the first variation time to the second variation time;

[0029] The average peak shape ratio of all the varying peaks is calculated to obtain the conductivity peak shape ratio.

[0030] As a preferred embodiment of the electrical conductivity detection and evaluation method for soil element detection described in this application, the spatiotemporal correlation features include at least the response delay sequences of electrical conductivity from different soil layers; the response delay sequences are extracted based on the spatiotemporal matrix of the target soil electrical conductivity, as follows:

[0031] Each row of the spatiotemporal matrix is ​​extracted as the electrical conductivity sequence of the corresponding soil sample layer; curve fitting is performed based on each electrical conductivity sequence to obtain the electrical conductivity variation curve of each soil sample layer.

[0032] Select a soil sample layer as a reference sample; extract the conductivity variation curve of the reference sample as a reference conductivity curve;

[0033] The peak values ​​were detected from the conductivity variation curves of each soil sample layer, and the timestamp corresponding to the peak value of each peak was marked.

[0034] According to the order of appearance of the change peaks from early to late, each change peak in the reference conductivity curve is used as a reference peak to classify all change peaks into different change peak families.

[0035] Based on the timestamp corresponding to the peak value of each variation peak, each family of variation peaks is encoded into a feature vector;

[0036] The response delay sequence is constructed by using the feature vector of each family of changing peaks as an element.

[0037] As a preferred embodiment of the electrical conductivity detection and evaluation method for soil element detection described in this application, each reference peak corresponds to a family of variation peaks; any family of variation peaks contains one or zero variation peaks from the electrical conductivity variation curve of any soil layer sample, and any two families of variation peaks have no overlap; the method for dividing the family of variation peaks using any variation peak in the reference electrical conductivity curve as a reference peak is as follows:

[0038] Set the time window with the timestamp corresponding to the peak value of the reference peak as the midpoint;

[0039] Each conductivity change curve is extracted through the time window, and the change peak whose timestamp falls into the time window is marked as a change peak to be determined.

[0040] Based on the conductivity sequence, a reference peak and a sequence segment corresponding to each undetermined variation peak are extracted;

[0041] The correlation between each undetermined variation peak and the reference peak is calculated based on the sequence fragments.

[0042] If the correlation between any undetermined variation peak and the reference peak is greater than the preset correlation threshold, then the corresponding undetermined variation peak will be assigned to the variation peak family corresponding to the reference peak.

[0043] If a family of variation peaks contains at least two variation peaks from any conductivity variation curve, then the variation peaks in the corresponding conductivity variation curves are removed from the family of variation peaks, and only the variation peaks in the corresponding conductivity variation curves whose peak values ​​are closest to the peak values ​​of the reference peaks in time are retained.

[0044] As a preferred embodiment of the electrical conductivity detection and evaluation method for soil element detection described in this application, wherein encoding each family of variation peaks into a feature vector specifically includes:

[0045] The order in which the peak values ​​of each variable peak appear in the variable peak family is encoded into a time-series vector of the variable peak family.

[0046] The time difference between the peak values ​​of the variation peaks in any two adjacent soil samples within the variation peak family is calculated based on timestamps.

[0047] The time difference of the occurrence of all the peaks is encoded as a hysteresis vector of a family of varying peaks;

[0048] The time-series vector and lag vector of the variable peak family are concatenated to form the feature vector of the variable peak family.

[0049] As a preferred embodiment of the electrical conductivity detection and evaluation method for soil element detection described in this application, the method involves assigning an influencing factor to each electrical conductivity response characteristic based on the environmental characteristic data, specifically including:

[0050] Calculate the mean of each environmental characteristic data in the most recent M sampling periods, and form the current environmental sequence of the target soil;

[0051] Acquire historical data of the target soil; each historical data point contains a set of environmental characteristic data and a corresponding set of electrical conductivity response characteristics; combine the environmental characteristic data of each historical data point into an environmental sequence;

[0052] Calculate the similarity between the environmental sequence of each historical data point and the current environmental sequence; extract historical data with a similarity greater than a preset similarity threshold as reference data;

[0053] Based on the conductivity response characteristics in the reference data, the instability of each conductivity response characteristic is calculated;

[0054] The influence factor for each conductivity response feature is assigned a value based on the instability; the greater the instability of any conductivity response feature, the smaller the influence factor.

[0055] As a preferred embodiment of the conductivity detection and evaluation method for soil element detection described in this application, wherein: the reference response characteristics for any content change pattern include reference values ​​for each reference response characteristic; and the method for assigning a confidence level to each conductivity response characteristic for any content change pattern includes:

[0056] For any time-series response feature or spatial response feature, calculate the absolute value of the difference between the conductivity response feature and the corresponding reference value, which is taken as the deviation value of the corresponding conductivity response feature; standardize the deviation value and take its reciprocal to obtain the confidence level of the corresponding conductivity response feature.

[0057] The confidence level for the response delay sequence is calculated as follows:

[0058] The similarity between the response delay sequence and the corresponding reference value is calculated using the dynamic time warping algorithm, and a confidence value is assigned. The greater the similarity with the reference value, the greater the confidence of the response delay sequence.

[0059] As a preferred embodiment of the conductivity detection and evaluation method for soil element detection described in this application, the method for calculating the confidence level of any content change pattern is as follows:

[0060] The confidence levels of each conductivity response feature are weighted and summed to obtain the confidence level of the corresponding content change pattern; in the weighted summation, the weight value of the confidence level of any conductivity response feature is the corresponding influence factor.

[0061] Based on the confidence level, the pattern of change in the content of the target element in the target soil is determined, specifically including:

[0062] The confidence levels of each content change pattern are sorted by magnitude; if the maximum confidence level among all content change patterns is greater than the preset confidence level threshold, then the content change pattern of the target element in the target soil is the corresponding content change pattern.

[0063] Compared with the prior art, the beneficial effects achieved by this application are as follows:

[0064] This application employs a combination of vertical stratified sampling and multi-period monitoring to construct a spatiotemporal matrix and evolution sequence of electrical conductivity, which can capture the distribution of target element content in soil and its changing trend over time. By considering environmental characteristic data to assign influencing factors to electrical conductivity response characteristics, the interference of environmental factors on electrical conductivity detection is reduced, the evaluation results remain stable under different environmental conditions, and the accuracy of the detection results is improved.

[0065] This proposed solution eliminates the need for real-time detection of specific element concentrations. Instead, it identifies typical patterns of element content changes using conductivity data, simplifying the soil nutrient monitoring process, providing a basis for agricultural production decisions, and reducing detection complexity. Attached Figure Description

[0066] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0067] Figure 1 A flowchart of a method for evaluating the electrical conductivity of soil elements provided in this application;

[0068] Figure 2 A conductivity curve and a schematic diagram of the variation peaks provided in this application. Detailed Implementation

[0069] The technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments and specific features in the embodiments are detailed descriptions of the technical solution of this application, rather than limitations thereof. In the absence of conflict, the embodiments and technical features in the embodiments can be combined with each other.

[0070] This embodiment introduces a method for evaluating the electrical conductivity of soil elements, referring to... Figure 1 The method includes the following steps:

[0071] Obtain electrical conductivity and environmental characteristic data of the target soil;

[0072] The conductivity data includes the conductivity of each soil sample in each sampling period;

[0073] The acquisition of electrical conductivity data of the target soil specifically includes: setting a sampling period; at the beginning of each sampling period, sampling the target soil in N layers vertically, and detecting the electrical conductivity of each soil sample; N is a positive integer; for example, optionally, the target soil is divided into upper, middle, and lower layers, and the electrical conductivity is collected separately. By performing vertical stratified sampling of the soil in each sampling period and measuring the electrical conductivity using sensors, the distribution and variation trend of the target element at different depths can be obtained, facilitating the identification of its vertical migration behavior;

[0074] At the start of each sampling cycle, environmental characteristic data of the target soil are collected synchronously; the environmental characteristic data includes at least one of water content and pH value. By collecting environmental characteristic data, a basis is provided for subsequent adjustment of the reliability of conductivity characteristics, thereby improving the accuracy and environmental adaptability of the identification results of target element content change patterns.

[0075] Based on the conductivity data, a spatiotemporal matrix and evolution sequence of the target soil conductivity are constructed;

[0076] The construction of the spatiotemporal matrix for the target soil electrical conductivity specifically includes: initializing an M-row, N-column spatiotemporal matrix; M being a positive integer; acquiring electrical conductivity data from the most recent M sampling periods and filling the spatiotemporal matrix with elements; wherein, the element in the i-th row and j-th column of the spatiotemporal matrix represents the electrical conductivity of the j-th layer of soil sample in the i-th sampling period among the most recent M sampling periods; the value range of i is 1, 2, ..., M; the value range of j is 1, 2, ..., N;

[0077] Constructing an evolution sequence of the target soil electrical conductivity specifically includes: initializing an evolution sequence of length M; filling the evolution sequence with elements based on the spatiotemporal matrix, wherein the evolution sequence has N elements, and any element is the mean of all elements in the corresponding column of the spatiotemporal matrix.

[0078] Based on the spatiotemporal matrix and evolution sequence, the electrical conductivity response characteristics of the target soil are extracted; based on the environmental characteristic data, an influencing factor is assigned to each electrical conductivity response characteristic;

[0079] The conductivity response characteristics include time-series response characteristics, spatial response characteristics, and spatiotemporal correlation characteristics;

[0080] The time-series response features include at least one of the conductivity rise rate and conductivity peak shape ratio; the time-series response features are extracted based on the evolution sequence of the target soil conductivity; wherein, the method for extracting the conductivity rise rate is as follows:

[0081] Based on the aforementioned evolution sequence, a two-dimensional curve is fitted to the electrical conductivity of the target soil, and the average slope of the fitted curve is extracted as the rate of increase in electrical conductivity. The average slope of the fitted curve can be positive or negative; a positive slope indicates an increase in electrical conductivity, and a negative slope indicates a decrease in electrical conductivity. Electrical conductivity directly reflects ion concentration. The sign and absolute value of the slope of increase in electrical conductivity reflect the intensity of ion activity in the soil, are related to the release and migration characteristics of specific elements, and are an indicator of the intensity of element application or loss processes. For example, after applying fertilizer containing a specific element to the soil, the ion concentration of that element in the soil increases, resulting in an increase in electrical conductivity, and the rate of increase in electrical conductivity is positive.

[0082] The method for extracting the conductivity peak shape ratio is as follows:

[0083] Based on the fitted conductivity curve, the peaks of conductivity variation are identified; the time corresponding to the peak value of each variation peak is recorded; in this embodiment, the places where the conductivity curve shows an initial rise followed by a fall or a fall followed by a rise are all variation peaks, that is, both peaks and valleys in the curve are defined as variation peaks.

[0084] In the coordinate system of the conductivity curve, each peak is intercepted by a line parallel to the x-axis to obtain the start and end points of each peak.

[0085] Reference Figure 2 The curve in the coordinate system represents the change of conductivity over time. The dashed line is a parallel line to the x-axis. The portion of the curve intercepted by this parallel line, i.e. the portion above the parallel line, is a peak. In this peak, the intersection of the parallel line and the curve are the starting and ending points of the peak, respectively.

[0086] For any peak, subtract the starting point from the time corresponding to the peak to obtain the first change time of the peak; subtract the peak from the ending point to obtain the second change time of the peak.

[0087] Calculate the peak shape ratio of each variation peak; the peak shape ratio of any variation peak is the ratio of the first variation time to the second variation time;

[0088] The average peak shape ratio of all the varying peaks is calculated to obtain the conductivity peak shape ratio.

[0089] The conductivity peak shape ratio calculated in this embodiment describes the skewness and symmetry of the changing peaks. When the content of a specific element changes, its release rate and decay rate are usually asymmetrical. If the release is fast and the decay is slow, that is, the element diffuses quickly in the soil and is absorbed and consumed slowly by plant roots, it is manifested as a rapid increase in ion concentration, followed by a slow decrease in ion concentration after reaching the peak. These characteristics are reflected in the changing peaks as the peak time being closer to the start time, the first change time being shorter than the second change time, the conductivity peak shape ratio being less than 1, and the changing peak skewed to the left. If the release is slow and the decay is fast, the conductivity peak shape ratio is greater than 1, and the changing peak skewed to the right. Conductivity is a mapping of the total ion activity in the soil, not a linear function of a single element, but a composite signal with hysteresis and mixed perturbations. Even if the content of the target element changes monotonically, the conductivity curve will still show non-monotonic changing peaks. This application can better restore the true content change characteristics of the target element under background interference by observing the shape characteristics of these changing peaks.

[0090] The spatial response characteristics include at least the electrical conductivity gradient between any two adjacent soil samples.

[0091] The conductivity gradient is extracted based on the spatiotemporal matrix of the target soil conductivity, and the method is as follows: the mean conductivity of each soil sample layer is calculated based on the spatiotemporal matrix; wherein, the mean conductivity of any soil sample layer is the mean of all elements in the corresponding row of the spatiotemporal matrix; for any two adjacent soil samples, the mean conductivity of the upper soil sample is subtracted from the mean conductivity of the lower soil sample to obtain the corresponding conductivity gradient.

[0092] In this embodiment, a total of N soil samples are used, so the spatial response characteristics of the target soil include at least N-1 conductivity gradients. The conductivity gradient between any two adjacent soil samples can be positive or negative, reflecting the magnitude and direction of the conductivity gradient between different soil samples. The spatial response characteristics constructed in this application can help determine the spatial distribution characteristics of element content changes and help eliminate the influence of non-target elements. Different target elements have different migration characteristics, and conductivity will form different gradient patterns. For example, the leaching of nitrate ions containing nitrogen will form a profile structure with low conductivity at the surface and high conductivity at the bottom; potassium ions are often adsorbed at the soil surface, and the conductivity change at the bottom is not obvious.

[0093] The spatiotemporal correlation features include at least the response delay sequences of electrical conductivity from different soil layers; the response delay sequences are extracted based on the spatiotemporal matrix of the target soil electrical conductivity, using the following method:

[0094] Each row of the spatiotemporal matrix is ​​extracted as the electrical conductivity sequence of the corresponding soil sample layer; curve fitting is performed based on each electrical conductivity sequence to obtain the electrical conductivity variation curve of each soil sample layer.

[0095] Select a soil sample layer as a reference sample; extract the conductivity variation curve of the reference sample as a reference conductivity curve;

[0096] The peak values ​​were detected from the conductivity variation curves of each soil sample layer, and the timestamp corresponding to the peak value of each peak was marked.

[0097] According to the order of appearance of the change peaks from early to late, each change peak in the reference conductivity curve is used as a reference peak to classify all change peaks into different change peak families.

[0098] Each reference peak corresponds to a family of variation peaks; any family of variation peaks contains one or zero variation peaks from the conductivity variation curve of any soil sample layer, and no two families of variation peaks overlap; the method for dividing the families of variation peaks using any variation peak in the reference conductivity curve as the reference peak is as follows:

[0099] Set a time window with the timestamp corresponding to the peak value of the reference peak as the midpoint; for example, optionally, the length of the time window is set to 2 days, then the starting point of the time window is one day earlier than the timestamp corresponding to the peak value of the reference peak, and the ending point of the time window is one day later than the timestamp corresponding to the peak value of the reference peak.

[0100] Each conductivity change curve is extracted through the time window, and the change peak whose timestamp falls into the time window is marked as a change peak to be determined.

[0101] Based on the conductivity sequence, a reference peak and a sequence segment corresponding to each undetermined variation peak are extracted;

[0102] The correlation between each undetermined change peak and the reference peak is calculated based on the sequence fragments. In this embodiment, it is preferable to calculate the cross-correlation coefficient between the sequence fragments corresponding to the undetermined change peak and the sequence fragments corresponding to the reference peak, and select the largest cross-correlation coefficient as the correlation between the undetermined change peak and the reference peak.

[0103] If the correlation between any undetermined variation peak and the reference peak is greater than the preset correlation threshold, then the corresponding undetermined variation peak will be assigned to the variation peak family corresponding to the reference peak.

[0104] If a family of variation peaks contains at least two variation peaks from any conductivity variation curve, then the variation peaks in the corresponding conductivity variation curves are removed from the family of variation peaks, and only the variation peaks in the corresponding conductivity variation curves whose peak values ​​are closest to the peak values ​​of the reference peaks in time are retained.

[0105] Based on the timestamp corresponding to the peak value of each variation peak, each family of variation peaks is encoded into a feature vector; specifically including:

[0106] The order in which the peak values ​​of each variable peak appear in the variable peak family is encoded into a time-series vector of the variable peak family.

[0107] The time difference between the peak values ​​of any two adjacent soil samples in the family of change peaks is calculated based on the timestamp. Optionally, if there is no change peak in a certain soil sample in the family of change peaks, the moment when the peak value of the corresponding soil sample appears is determined by interpolation.

[0108] The time difference of the occurrence of all the peaks is encoded as a hysteresis vector of a family of varying peaks;

[0109] The time-series vector and lag vector of the variable peak family are concatenated to form the feature vector of the variable peak family;

[0110] The response delay sequence is constructed by using the feature vector of each family of changing peaks as an element.

[0111] In this embodiment, the response delay sequence identifies the migration path and velocity characteristics of target elements in the vertical soil profile by analyzing the response time difference and response sequence of different soil layers to changes in electrical conductivity. Different elements exhibit significant differences in their migration patterns in the soil, meaning that the peak electrical conductivity responses induced by different soil layers have different temporal sequences and delay patterns. By constructing a structured sequence composed of the interlayer peak response sequence and time difference, the ability to identify target elements is enhanced, which helps improve the accuracy of determining the change patterns in target element content.

[0112] The aforementioned conductivity response characteristics are all obtained from the spatiotemporal characteristics of conductivity itself, and are highly correlated with the actual content variation characteristics of the target element. They can effectively eliminate interference from non-target elements and are important evidence factors for improving the identification of target element content variations.

[0113] Based on the aforementioned environmental characteristic data, an influencing factor is assigned to each conductivity response feature, specifically including:

[0114] Calculate the mean of each environmental characteristic data in the most recent M sampling periods, and form the current environmental sequence of the target soil;

[0115] Acquire historical data of the target soil; each historical data point contains a set of environmental characteristic data and a corresponding set of electrical conductivity response characteristics; combine the environmental characteristic data of each historical data point into an environmental sequence;

[0116] Calculate the similarity between the environmental sequence of each historical data point and the current environmental sequence; optionally, calculate the Euclidean distance and take its reciprocal as the similarity; extract historical data with similarity greater than a preset similarity threshold as reference data;

[0117] Based on the conductivity response characteristics in the reference data, the instability of each conductivity response characteristic is calculated;

[0118] The influence factor for each conductivity response feature is assigned a value based on the instability; the greater the instability of any conductivity response feature, the smaller the influence factor. For example, a value range is set for the influence factor of each conductivity response feature, and a specific value is assigned to the influence factor within the value range according to the corresponding instability, with the larger the instability, the smaller the value assigned to the influence factor.

[0119] Optionally, the instability of each conductivity response characteristic is calculated in this embodiment as follows:

[0120] For any time-series or spatial response characteristic, calculate the standard deviation of each conductivity response characteristic in all reference data as its instability.

[0121] For spatiotemporal correlation features, the response delay sequence is extracted from each reference data point, and the similarity between any two reference data response delay sequences is calculated using the DTW (Dynamic Time Warping) algorithm. The average similarity between the response delay sequences of all reference data is calculated and assigned to the instability of the spatiotemporal correlation feature; the lower the average similarity, the higher the instability of the spatiotemporal correlation feature. For example, the average similarity is normalized, and 1 minus the normalized average similarity is used as the instability of the spatiotemporal correlation feature.

[0122] In this embodiment, the aforementioned environmental characteristic data does not directly participate in the calculation of conductivity response characteristics, but it does affect the influence of each conductivity response characteristic as evidence of changes in the target element. By statistically evaluating the instability of conductivity response characteristics under similar environmental conditions in historical data, it is determined whether the performance of each characteristic is stable in the current environment. Based on the calculation logic of instability, the lower the instability, the more stable the conductivity response characteristics in the reference data, indicating that the corresponding conductivity response characteristics are less affected by the current environment and better reflect the true trend of changes in the content of the target element in the current environment.

[0123] Obtain reference response features for each content variation pattern of the target element; assign confidence level of each conductivity response feature to each content variation pattern based on the reference response features;

[0124] The content change pattern refers to a pattern in which the content of the target element changes significantly, such as the target element being absorbed in large quantities, the target element being lost, or the target element increasing significantly after fertilization. Understanding the content change pattern of the target element helps to determine whether intervention to regulate its content is necessary.

[0125] The reference response characteristics for any content change pattern include the reference value for each reference response characteristic.

[0126] Methods for assigning confidence levels to each conductivity response feature for any given content variation pattern include:

[0127] For any time-series response feature or spatial response feature, the absolute value of the difference between the conductivity response feature and the corresponding reference value is calculated as the deviation value of the corresponding conductivity response feature; the deviation value is standardized and its reciprocal is taken to obtain the confidence level of the corresponding conductivity response feature; optionally, the standardization process is to scale the deviation value to between 1 and 10; furthermore, those skilled in the art can adjust the confidence level range of each conductivity response feature by scaling the corresponding deviation value to different value ranges.

[0128] The confidence level for the response delay sequence is calculated as follows:

[0129] The similarity between the response delay sequence and its corresponding reference value is calculated using a dynamic time warping algorithm, and a confidence score is assigned accordingly. The greater the similarity with the reference value, the greater the confidence score of the response delay sequence. For example, the similarity between the response delay sequence and its corresponding reference value is normalized, and the normalized reference value is used as the confidence score of the corresponding response delay sequence.

[0130] The confidence level of each content change pattern is calculated based on the reliability of the conductivity response characteristics and the influencing factors; the content change pattern of the target element in the target soil is determined based on the confidence level.

[0131] The method for calculating the confidence level of any content change pattern is as follows:

[0132] The confidence levels of each conductivity response feature are weighted and summed to obtain the confidence level of the corresponding content change pattern; in the weighted summation, the weight value of the confidence level of any conductivity response feature is the corresponding influence factor.

[0133] Based on the confidence level, the pattern of change in the content of the target element in the target soil is determined, specifically including:

[0134] The confidence levels of each content change pattern are sorted by magnitude; if the maximum confidence level among all content change patterns is greater than the preset confidence level threshold, then the content change pattern of the target element in the target soil is the corresponding content change pattern.

[0135] This application provides a mechanism for screening and identifying anomalies in soil nutrient content for agricultural or environmental applications. In scenarios where real-time monitoring of specific element concentrations is not required, by uniformly collecting conductivity data, the changes in the content of various target elements can be assessed separately. This simplifies soil nutrient monitoring procedures, reduces detection complexity, and is particularly suitable for scenarios requiring the evaluation of significant changes in one or more key elements over a specific period. Typical applications include: assessing changes in target element content after fertilization to determine fertilization effectiveness and whether supplemental fertilization is necessary; and monitoring the loss of essential elements to enable timely intervention decisions.

[0136] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0137] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of protection of this application, and these forms are all within the protection scope of this application.

Claims

1. A method for detecting and evaluating the electrical conductivity of soil elements, characterized in that, Includes the following steps: Obtain electrical conductivity and environmental characteristic data of the target soil; Based on the conductivity data, a spatiotemporal matrix and evolution sequence of the target soil conductivity are constructed; Based on the spatiotemporal matrix and evolution sequence, the electrical conductivity response characteristics of the target soil are extracted; Based on the environmental characteristic data, an influence factor is assigned to each conductivity response feature; Obtain reference response features for each content variation pattern of the target element; assign confidence level of each conductivity response feature to each content variation pattern based on the reference response features; The confidence level of each content change pattern is calculated based on the reliability of the conductivity response characteristics and the influencing factors; the content change pattern of the target element in the target soil is determined based on the confidence level.

2. The method for detecting and evaluating the electrical conductivity of soil elements as described in claim 1, characterized in that: The acquisition of electrical conductivity data of the target soil specifically includes: setting a sampling period; at the beginning of each sampling period, sampling the target soil in N layers in the vertical direction, and detecting the electrical conductivity of each soil sample; N is a positive integer; at the beginning of each sampling period, environmental characteristic data of the target soil are collected synchronously. The construction of the spatiotemporal matrix for the target soil electrical conductivity specifically includes: initializing an M-row, N-column spatiotemporal matrix; M being a positive integer; acquiring electrical conductivity data from the most recent M sampling periods and filling the spatiotemporal matrix with elements; wherein, the element in the i-th row and j-th column of the spatiotemporal matrix represents the electrical conductivity of the j-th layer of soil sample in the i-th sampling period among the most recent M sampling periods; the value range of i is 1, 2, ..., M; the value range of j is 1, 2, ..., N; The evolution sequence has N elements, where each element is the mean of all elements in the corresponding column of the spatiotemporal matrix.

3. The method for detecting and evaluating the electrical conductivity of soil elements as described in claim 2, characterized in that: The conductivity response characteristics include time-series response characteristics, spatial response characteristics, and spatiotemporal correlation characteristics; The time-series response features include at least one of the conductivity rise rate and conductivity peak shape ratio; the time-series response features are extracted based on the evolution sequence of the target soil conductivity; wherein, the method for extracting the conductivity rise rate is as follows: The electrical conductivity of the target soil was fitted with a two-dimensional curve based on the evolution sequence, and the average slope of the fitted curve was extracted as the rate of increase of electrical conductivity. The spatial response characteristics include at least the electrical conductivity gradient between any two adjacent soil samples. The conductivity gradient is extracted based on the spatiotemporal matrix of the target soil conductivity, and the method is as follows: the mean conductivity of each soil sample layer is calculated based on the spatiotemporal matrix; wherein, the mean conductivity of any soil sample layer is the mean of all elements in the corresponding row of the spatiotemporal matrix; for any two adjacent soil samples, the mean conductivity of the upper soil sample is subtracted from the mean conductivity of the lower soil sample to obtain the corresponding conductivity gradient.

4. The method for detecting and evaluating the electrical conductivity of soil elements as described in claim 3, characterized in that: The method for extracting the conductivity peak shape ratio is as follows: Based on the fitted conductivity curve, identify the peaks of conductivity variation; record the time corresponding to the peak value of each peak. In the coordinate system of the conductivity curve, each peak is intercepted by a line parallel to the x-axis to obtain the start and end points of each peak. For any peak, subtract the starting point from the time corresponding to the peak to obtain the first change time of the peak; subtract the peak from the ending point to obtain the second change time of the peak. Calculate the peak shape ratio of each variation peak; the peak shape ratio of any variation peak is the ratio of the first variation time to the second variation time; The average peak shape ratio of all the varying peaks is calculated to obtain the conductivity peak shape ratio.

5. The method for detecting and evaluating the electrical conductivity of soil elements as described in claim 4, characterized in that: The spatiotemporal correlation features include at least the response delay sequences of electrical conductivity from different soil layers; the response delay sequences are extracted based on the spatiotemporal matrix of the target soil electrical conductivity, using the following method: Each row of the spatiotemporal matrix is ​​extracted as the electrical conductivity sequence of the corresponding soil sample layer; curve fitting is performed based on each electrical conductivity sequence to obtain the electrical conductivity variation curve of each soil sample layer. Select one layer of soil sample as a reference sample; The conductivity variation curve of the reference sample is extracted as the reference conductivity curve; The peak values ​​were detected from the conductivity variation curves of each soil sample layer, and the timestamp corresponding to the peak value of each peak was marked. According to the order of appearance of the change peaks from early to late, each change peak in the reference conductivity curve is used as a reference peak to classify all change peaks into different change peak families. Based on the timestamp corresponding to the peak value of each variation peak, each family of variation peaks is encoded into a feature vector; The response delay sequence is constructed by using the feature vector of each family of changing peaks as an element.

6. The method for detecting and evaluating the electrical conductivity of soil elements as described in claim 5, characterized in that: Each reference peak corresponds to a family of variation peaks; any family of variation peaks contains at most one variation peak from the conductivity variation curve of any soil sample layer, and no two families of variation peaks overlap; the method for dividing the families of variation peaks using any variation peak in the reference conductivity curve as the reference peak is as follows: Set the time window with the timestamp corresponding to the peak value of the reference peak as the midpoint; Each conductivity change curve is extracted through the time window, and the change peak whose timestamp falls into the time window is marked as a change peak to be determined. Based on the conductivity sequence, a reference peak and a sequence segment corresponding to each undetermined variation peak are extracted; The correlation between each undetermined variation peak and the reference peak is calculated based on the sequence fragments. If the correlation between any undetermined variation peak and the reference peak is greater than the preset correlation threshold, then the corresponding undetermined variation peak will be assigned to the variation peak family corresponding to the reference peak. If a family of variation peaks contains at least two variation peaks from any conductivity variation curve, then the variation peaks in the corresponding conductivity variation curves are removed from the family of variation peaks, and only the variation peaks in the corresponding conductivity variation curves whose peak values ​​are closest to the reference peaks in time are retained.

7. The method for detecting and evaluating the electrical conductivity of soil elements as described in claim 6, characterized in that: Encoding each family of varying peaks into a feature vector specifically includes: The order in which the peak values ​​of each variable peak appear in the variable peak family is encoded into a time-series vector of the variable peak family. The time difference between the peak values ​​of the variation peaks in any two adjacent soil samples within the variation peak family is calculated based on timestamps. The time difference of the occurrence of all the peaks is encoded as a hysteresis vector of a family of varying peaks; The time-series vector and lag vector of the variable peak family are concatenated to form the feature vector of the variable peak family.

8. The method for detecting and evaluating the electrical conductivity of soil elements as described in claim 7, characterized in that: Based on the aforementioned environmental characteristic data, an influencing factor is assigned to each conductivity response feature, specifically including: Calculate the mean of each environmental characteristic data in the most recent M sampling periods, and form the current environmental sequence of the target soil; Acquire historical data of the target soil; each historical data point contains a set of environmental characteristic data and a corresponding set of electrical conductivity response characteristics; combine the environmental characteristic data of each historical data point into an environmental sequence; Calculate the similarity between the environmental sequence of each historical data point and the current environmental sequence; extract historical data with a similarity greater than a preset similarity threshold as reference data; Based on the conductivity response characteristics in the reference data, the instability of each conductivity response characteristic is calculated; The influence factor for each conductivity response feature is assigned a value based on the instability; the greater the instability of any conductivity response feature, the smaller the influence factor.

9. The method for detecting and evaluating the electrical conductivity of soil elements as described in claim 8, characterized in that: The reference response characteristics for any content change pattern include reference values ​​for each reference response characteristic; Methods for assigning confidence levels to each conductivity response feature for any given content variation pattern include: For any time-series response feature or spatial response feature, calculate the absolute value of the difference between the conductivity response feature and the corresponding reference value, and use it as the deviation value of the corresponding conductivity response feature. The deviation value is standardized and its reciprocal is taken to obtain the confidence level of the corresponding conductivity response characteristic; The confidence level for the response delay sequence is calculated as follows: The similarity between the response delay sequence and the corresponding reference value is calculated using the dynamic time warping algorithm, and a confidence value is assigned. The greater the similarity with the reference value, the greater the confidence of the response delay sequence.

10. The method for detecting and evaluating the electrical conductivity of soil elements as described in claim 9, characterized in that: The method for calculating the confidence level of any content change pattern is as follows: The confidence levels of each conductivity response feature are weighted and summed to obtain the confidence level of the corresponding content change pattern; in the weighted summation, the weight value of the confidence level of any conductivity response feature is the corresponding influence factor. Based on the confidence level, the pattern of change in the content of the target element in the target soil is determined, specifically including: The reliability of each content variation pattern was ranked by magnitude; If the maximum confidence level among all content change patterns is greater than the preset confidence level threshold, then the content change pattern of the target element in the target soil is the corresponding content change pattern.

Images