A cultivated land quality evaluation method based on soil productivity ecological multi-element fusion

Abstract

The present application relates to the field of agricultural data processing, and more particularly to a cultivated land quality evaluation method based on soil productivity ecological multi-element fusion, which comprises: collecting and preprocessing multi-source soil element time series data to obtain a standardized multi-element feature matrix; cross-collaborative mapping of soil volume moisture content and apparent soil conductivity to obtain an instantaneous cross-collaborative factor; conditional integration combined with the evolution trajectory of soil volume moisture content to obtain a stress fatigue integral factor; correction of the basic worst solution distance using the stress fatigue integral factor to obtain a corrected worst solution distance, and calculation of the cultivated land quality comprehensive evaluation result combined with the basic optimal solution distance, thereby improving the accuracy and stability of the dynamic evaluation of cultivated land quality under the condition of water-salt collaborative degradation.

Description

Technical Field

[0001] This invention relates to the field of agricultural data processing technology, and in particular to a method for assessing arable land quality based on the integration of multiple factors related to soil productivity and ecology. Background Technology

[0002] With the development of agricultural IoT, soil sensing and monitoring, and agricultural information technology, dynamic assessment of arable land quality using multi-source soil element data has become an important technical means for arable land protection and agricultural production management. Existing arable land quality assessment methods typically involve deploying agricultural IoT monitoring equipment to continuously collect time-series data on multiple elements such as soil volumetric moisture content, soil apparent electrical conductivity, soil organic matter content, soil porosity, and soil pH. An entropy-weighted approximation of the ideal solution ranking method is then used to objectively assign weights to different indicators. By calculating the distance between the current soil state and the optimal and worst solutions, as well as the corresponding relative proximity, a comprehensive quantitative assessment of arable land quality is achieved.

[0003] Existing multi-factor assessment methods based on EWM-TOPSIS typically use weighted aggregation of the independent deviations of each indicator dimension from the ideal solution during distance measurement. While this reflects the degree of change in individual soil indicators, it lacks characterization of the cross-interactions between different indicator dimensions. In actual arable land environments, soil moisture deficit and salinity accumulation often exhibit significant synergistic degradation characteristics. When soil volumetric water content decreases and is accompanied by an increase in apparent soil electrical conductivity, the impact on the ecological state of arable land is not a simple summation of the degradation degrees of individual indicators. Existing distance measurement methods struggle to effectively extract these synergistic water-salt degradation characteristics, easily leading to an underestimation of the true degree of arable land degradation.

[0004] Furthermore, under high-frequency agricultural IoT monitoring conditions, soil volumetric moisture content and apparent electrical conductivity are easily affected by short-term environmental changes such as diurnal evaporation, condensation, rainfall, and irrigation, resulting in periodic fluctuations. Evaluating solely based on instantaneous water and salt conditions can easily lead to significant oscillations in the assessment results, making it difficult to distinguish between short-term recoverable fluctuations and persistent water-salt co-deterioration. Therefore, accurately extracting the co-deterioration characteristics of water deficit and salt accumulation during multi-source soil element fusion assessment, and combining this with their temporal evolution to suppress short-term fluctuation interference, thereby improving the accuracy and stability of dynamic farmland quality assessment results, has become a pressing issue that existing farmland quality assessment technologies need to address. Summary of the Invention

[0005] In view of this, the present invention aims to propose a method for assessing arable land quality based on the integration of multiple factors of soil productivity and ecology, in order to solve the problem that existing multi-factor assessment methods are unable to accurately characterize the synergistic deterioration characteristics of water deficit and salt accumulation.

[0006] To achieve the above objectives, the technical solution of the present invention is implemented as follows:

[0007] A method for assessing arable land quality based on the integration of multiple soil productivity and ecological factors, the method comprising:

[0008] Step S1: Obtain a standardized multi-element feature matrix by collecting and preprocessing time-series data of multi-source soil elements;

[0009] Step S2: Obtain the instantaneous cross-cooperative factor by performing cross-cooperative mapping processing on soil volumetric water content and soil apparent electrical conductivity;

[0010] Step S3: Obtain the stress fatigue integral factor by conditionally integrating the instantaneous cross-synergistic factor and the evolution trajectory of soil volumetric water content;

[0011] Step S4: Obtain the corrected worst solution distance by performing distance correction processing on the basic worst solution distance and the stress fatigue integral factor;

[0012] Step S5: Obtain the comprehensive assessment result of cultivated land quality by calculating the relative closeness between the distance of the modified worst solution and the distance of the basic best solution.

[0013] Furthermore, the process of acquiring and preprocessing time-series data of multi-source soil elements to obtain a standardized multi-element feature matrix includes:

[0014] Agricultural IoT monitoring equipment is deployed in the cultivated land area to be assessed. The agricultural IoT monitoring equipment is equipped with soil moisture sensor for collecting soil volumetric moisture content data, soil conductivity sensor for collecting soil apparent conductivity data, soil organic matter monitoring sensor for collecting soil organic matter content data, soil porosity monitoring sensor for collecting soil porosity data, and soil pH sensor for collecting soil pH data.

[0015] A fixed sampling time step is set, and continuous monitoring and data collection are carried out on the cultivated land area to be evaluated according to the fixed sampling time step to obtain time series data of multi-source soil elements within the evaluation period. The time series data of multi-source soil elements include time series data of soil volumetric water content, soil apparent electrical conductivity, soil organic matter content, soil porosity, and soil pH.

[0016] The time series data of multi-source soil elements are time-stamped according to the standard timestamp corresponding to the fixed sampling time step. The missing data after time-stamp alignment is completed by linear interpolation. The completed time series data of multi-source soil elements is continuously extracted according to the preset sliding time window to obtain the time series data fragments of multi-source soil elements corresponding to each sliding time window.

[0017] For any target sliding time window, the historical maximum value, historical minimum value, historical maximum value, and historical minimum value of soil volumetric moisture content, soil apparent electrical conductivity, and soil apparent electrical conductivity are statistically obtained from the multi-source soil element time series data segments corresponding to the target sliding time window. Extreme value standardization processing is then performed on the soil volumetric moisture content time series data, soil apparent electrical conductivity time series data, soil organic matter content time series data, soil porosity time series data, and soil pH time series data, respectively. The corresponding standardized data are arranged according to the sampling time and soil element dimensions to construct a standardized multi-element feature matrix.

[0018] Furthermore, the instantaneous cross-cooperative factor is obtained by performing cross-cooperative mapping processing on soil volumetric water content and soil apparent electrical conductivity, including:

[0019] By performing local normalization on soil volumetric water content data, soil apparent electrical conductivity data and corresponding historical boundary data, the water deficit degree and salinity accumulation degree were obtained.

[0020] Instantaneous cross-coupling and nonlinear mapping of water deficit and salinity accumulation are used to obtain the cross-cooperative factor.

[0021] Furthermore, the step of obtaining water deficit and salinity accumulation by locally normalizing soil volumetric water content data, soil apparent electrical conductivity data, and corresponding historical boundary data includes:

[0022] For any target sampling time within any target sliding time window, extract the soil volumetric moisture content data and soil apparent electrical conductivity data corresponding to the target sampling time from the multi-source soil element time series data segment corresponding to the target sliding time window, and obtain the historical maximum value, historical minimum value, historical maximum value, and historical minimum value of soil volumetric moisture content, corresponding to the target sliding time window.

[0023] The difference between the historical maximum value of soil volumetric moisture content and the soil volumetric moisture content data corresponding to the target sampling time is used as the water deficit assessment corresponding to the target sampling time, and the difference between the historical maximum value of soil volumetric moisture content and the historical minimum value of soil volumetric moisture content is used as the water historical range assessment. The water deficit assessment is used as the numerator, the water historical range assessment is used as the denominator, and the corresponding fraction is used as the water deficit degree corresponding to the target sampling time.

[0024] The difference between the soil apparent electrical conductivity data at the target sampling time and the historical minimum value of soil apparent electrical conductivity is used as the salt accumulation assessment at the target sampling time, and the difference between the historical maximum value and the historical minimum value of soil apparent electrical conductivity is used as the salt historical range assessment. The salt accumulation assessment is used as the numerator, the salt historical range assessment is used as the denominator, and the corresponding fraction is used as the salt accumulation degree at the target sampling time.

[0025] Furthermore, the instantaneous cross-coupling and nonlinear mapping process for water deficit and salinity accumulation to obtain the cross-cooperative factor includes:

[0026] For any target sampling time within any target sliding time window, obtain the water deficit and salinity accumulation corresponding to the target sampling time;

[0027] Multiply the water deficit and salinity accumulation at the target sampling time, and use the resulting product as the cross-cooperative product evaluation at the target sampling time.

[0028] Add the water deficit and salinity accumulation at the target sampling time, and use the sum as the assessment of the synergistic degradation intensity at the target sampling time. Perform an exponential mapping on the synergistic degradation intensity assessment with the natural constant as the base, and use the mapping result as the synergistic degradation exponential amplification assessment at the target sampling time.

[0029] Multiply the cross-cooperative product evaluation corresponding to the target sampling time with the cooperative degradation index amplification evaluation, and add the corresponding product to the constant 1. Use the corresponding calculation result as the instantaneous cooperative mapping evaluation corresponding to the target sampling time.

[0030] The instantaneous collaborative mapping evaluation corresponding to the target sampling time is subjected to natural logarithmic mapping, and the corresponding mapping result is used as the instantaneous cross-collaborative factor corresponding to the target sampling time.

[0031] Furthermore, the step of obtaining the stress fatigue integral factor by conditionally integrating the instantaneous cross-synergy factor with the soil volumetric water content evolution trajectory includes:

[0032] By extracting and normalizing the unidirectional change data of soil volumetric water content evolution trajectory, an assessment of unidirectional water loss is obtained.

[0033] By dynamically coupling the instantaneous cross-synergistic factor with the unidirectional moisture loss assessment, a single-step fatigue penalty assessment is obtained.

[0034] By performing cumulative and nonlinear mapping processing on the single-step fatigue penalty assessment within the sliding time window, the stress fatigue integral factor is obtained.

[0035] Furthermore, the step of extracting and normalizing unidirectional changes in soil volumetric water content evolution trajectory data to obtain an assessment of unidirectional water loss includes:

[0036] For any target sampling time other than the first sampling time within any target sliding time window, obtain the soil volumetric moisture content data corresponding to the target sampling time and the soil volumetric moisture content data corresponding to the previous sampling time of the target sampling time, and obtain the historical maximum value and historical minimum value of soil volumetric moisture content corresponding to the target sliding time window.

[0037] The difference between the soil volumetric moisture content data corresponding to the previous sampling time and the soil volumetric moisture content data corresponding to the target sampling time is used as the moisture change assessment corresponding to the target sampling time; when the moisture change assessment is less than or equal to a constant 0, the one-way water loss assessment corresponding to the target sampling time is set to a constant 0.

[0038] When the moisture change assessment is greater than the constant 0, the difference between the historical maximum and the historical minimum of soil volumetric moisture content is used as the moisture historical range assessment. The moisture change assessment corresponding to the target sampling time is used as the numerator, the moisture historical range assessment is used as the denominator, and the corresponding fraction is used as the unidirectional moisture loss assessment corresponding to the target sampling time.

[0039] Furthermore, the process of dynamically coupling the instantaneous cross-coordination factor with the unidirectional moisture loss assessment to obtain a single-step fatigue penalty assessment includes:

[0040] For any target sampling time other than the first sampling time within any target sliding time window, obtain the instantaneous cross-coordination factor corresponding to the target sampling time and the unidirectional moisture loss assessment corresponding to the target sampling time;

[0041] Multiply the instantaneous cross-coordination factor corresponding to the target sampling time with the unidirectional moisture loss assessment, and use the resulting product as the single-step fatigue penalty assessment corresponding to the target sampling time.

[0042] Furthermore, the step of obtaining the stress fatigue integral factor by accumulating and nonlinearly mapping the single-step fatigue penalty assessment within the sliding time window includes:

[0043] For any target sliding time window, the single-step fatigue penalty assessments corresponding to all sampling times except the first sampling time within the target sliding time window are accumulated, and the corresponding accumulated result is used as the cumulative fatigue penalty assessment corresponding to the target sliding time window.

[0044] The negative number of the cumulative fatigue penalty assessment corresponding to the target sliding time window is subjected to an exponential mapping with the natural constant as the base, and the corresponding mapping result is used as the cumulative fatigue decay assessment corresponding to the target sliding time window.

[0045] The difference between constant 1 and the fatigue cumulative decay assessment corresponding to the target sliding time window is used as the stress fatigue integral factor corresponding to the target sliding time window.

[0046] Furthermore, the step of obtaining the corrected worst-case solution distance by performing distance correction processing on the basic worst-case solution distance and the stress fatigue integral factor includes:

[0047] Based on the standardized multi-element feature matrix, the entropy weight method is used to calculate the index weights corresponding to soil volumetric water content, soil apparent electrical conductivity, soil organic matter content, soil porosity and soil pH, and the worst solution data corresponding to each index dimension is determined based on the standardized soil element data.

[0048] For any target sampling time corresponding to a target sliding time window, calculate the difference between the standardized soil element data and the worst solution data of the corresponding index dimension at the target sampling time. Multiply the square of each difference by the corresponding index weight and use the result as the weighted worst solution distance component of the corresponding index dimension. Sum the weighted worst solution distance components corresponding to all index dimensions and calculate the square root of the sum. Use the result as the basic worst solution distance at the target sampling time.

[0049] Obtain the stress fatigue integral factor corresponding to the target sliding time window to which the target sampling time belongs, add the stress fatigue integral factor to the constant 1, and use the corresponding calculation result as the distance correction denominator evaluation corresponding to the target sampling time;

[0050] The distance to the worst solution at the target sampling time is used as the numerator, the distance correction denominator evaluation is used as the denominator, and the resulting fraction is used as the corrected worst solution distance at the target sampling time.

[0051] Compared with the prior art, the present invention has the following advantages:

[0052] This invention presents a method for assessing arable land quality based on the integration of multiple soil productivity and ecological factors. By constructing a cross-synergistic characterization mechanism between soil volumetric water content and soil apparent electrical conductivity, it can distinguish between single water anomalies, single salinity anomalies, and complex deterioration states where low water and high salinity occur simultaneously. Furthermore, it utilizes nonlinear mapping to enhance the response to severe water-salt synergistic stress. In actual farmland monitoring scenarios, it can reduce the weakening effect of traditional multi-indicator independent distance aggregation on the superposition of water and salt, making the arable land quality assessment results more consistent with the true ecological state under the combined effects of salt enrichment and water deficit. Simultaneously, this invention combines the continuous evolution direction of soil volumetric water content to perform temporal screening and accumulation of instantaneous synergistic features, which can suppress recoverable high-frequency fluctuations caused by factors such as diurnal evaporation, condensation, short-term rainfall, or irrigation, and highlight the gradually accumulating water-salt stress characteristics under continuous dehydration conditions. Based on this, the worst-case solution distance of the EWM-TOPSIS model is dynamically corrected, which can reduce short-term oscillations in arable land quality scores, improve the stability and accuracy of assessment results under continuous IoT monitoring conditions, and provide more reliable data for arable land quality change identification and agricultural production management. Attached Figure Description

[0053] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0054] Figure 1 This is a flowchart illustrating a method for assessing arable land quality based on the integration of multiple soil productivity and ecological factors, as described in an embodiment of the present invention.

[0055] Figure 2 This is a simulation verification diagram of the effectiveness of the optimized model of the arable land quality assessment method based on the integration of multiple soil productivity and ecological factors, as described in an embodiment of the present invention. Detailed Implementation

[0056] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0057] See Figure 1 This is a flowchart of a method for assessing arable land quality based on the integration of multiple soil productivity and ecological factors, as provided in Embodiment 1 of the present invention. Figure 1 As shown, a method for assessing arable land quality based on the integration of multiple soil productivity and ecological factors may include:

[0058] Step S1: Obtain a standardized multi-element feature matrix by collecting and preprocessing time-series data of multi-source soil elements.

[0059] First, agricultural IoT monitoring equipment is deployed in the cultivated land area to be assessed. The agricultural IoT monitoring equipment is equipped with soil moisture sensors for collecting soil volumetric moisture content data, soil conductivity sensors for collecting soil apparent conductivity data, soil organic matter monitoring sensors for collecting soil organic matter content data, soil porosity monitoring sensors for collecting soil porosity data, and soil pH sensors for collecting soil pH data.

[0060] A fixed sampling time step is set, and continuous monitoring and data collection are carried out on the cultivated land area to be evaluated according to the fixed sampling time step to obtain time series data of multi-source soil elements within the evaluation period. The time series data of multi-source soil elements includes time series data of soil volumetric water content, soil apparent electrical conductivity, soil organic matter content, soil porosity, and soil pH. It should be noted that in this embodiment of the invention, the sampling time step is set to 1 hour, that is, data is collected once per hour.

[0061] The time series data of multi-source soil elements are time-stamped according to the standard timestamp corresponding to the fixed sampling time step. The missing data after time-stamp alignment is completed by linear interpolation. The completed time series data of multi-source soil elements is continuously extracted according to the preset sliding time window to obtain the time series data fragments of multi-source soil elements corresponding to each sliding time window. In this embodiment of the invention, the length of the sliding time window is set to 168, that is, 168 hours of data collected is a window.

[0062] For any target sliding time window, the historical maximum value, historical minimum value, historical maximum value, and historical minimum value of soil volumetric moisture content, soil apparent electrical conductivity, and soil apparent electrical conductivity are statistically obtained from the multi-source soil element time series data segments corresponding to the target sliding time window. Extreme value standardization processing is then performed on the soil volumetric moisture content time series data, soil apparent electrical conductivity time series data, soil organic matter content time series data, soil porosity time series data, and soil pH time series data, respectively. The corresponding standardized data are arranged according to the sampling time and soil element dimensions to construct a standardized multi-element feature matrix.

[0063] Thus, the standardized multi-element feature matrix was obtained by collecting and preprocessing time-series data of multi-source soil elements.

[0064] Step S2: Obtain the instantaneous cross-cooperative factor by performing cross-cooperative mapping processing on soil volumetric water content and soil apparent electrical conductivity.

[0065] Existing multi-factor assessment models based on the EWM-TOPSIS algorithm measure the distance to land parcel status by independently squaring and summing the distances of each indicator dimension using Euclidean distance. This calculation method is based on the underlying assumption that the data dimensions are orthogonal and independent. However, in actual soil environmental monitoring data, the decrease in volumetric water content and the increase in apparent conductivity often have a mutually reinforcing synergistic relationship. When extremely low moisture and high salinity data occur simultaneously, they have a cumulative effect, leading to an accelerated decline in the overall soil environmental quality. The existing linear summation structure of Euclidean distance cannot identify this cross-synergistic state of multi-dimensional data; it can only calculate the independent degradation values ​​of moisture and salinity and perform simple addition, thus masking the accelerated degradation phenomenon caused by the superposition of indicators at the data level. This results in assessment results that cannot accurately reflect the true degree of soil degradation. To solve the problem of the failure of synergistic feature extraction caused by the linear summation of Euclidean distance, a cross-dimensional correlation mapping mechanism must be introduced into the assessment model. This mechanism requires the use of real-time data boundaries of water and salinity to construct a nonlinear feature that is activated only when both deteriorate simultaneously. By amplifying the response value of this extreme crossover state at the numerical level, it can compensate for the loss of data features in the basic algorithm when dealing with water and salinity co-deterioration segments.

[0066] In summary, this invention first performs local normalization on soil volumetric moisture content data, soil apparent electrical conductivity data, and corresponding historical boundary data to obtain water deficit and salinity accumulation. Specifically, for any target sampling time within any target sliding time window, soil volumetric moisture content data and soil apparent electrical conductivity data corresponding to the target sampling time are extracted from the multi-source soil element time-series data segment corresponding to the target sliding time window. The historical maximum, minimum, maximum, and minimum values ​​of soil volumetric moisture content, soil apparent electrical conductivity, and soil apparent electrical conductivity corresponding to the target sliding time window are then obtained. The difference between the historical maximum soil volumetric moisture content and the soil volumetric moisture content data corresponding to the target sampling time is used as the water deficit assessment for the target sampling time, and the difference between the historical maximum and minimum soil volumetric moisture content is used as the historical water deficit range assessment. The water deficit assessment is used as the numerator, the historical water deficit range assessment is used as the denominator, and the resulting fraction is used as the water deficit degree corresponding to the target sampling time. The difference between the soil apparent electrical conductivity data at the target sampling time and the historical minimum value of soil apparent electrical conductivity is used as the salt accumulation assessment at the target sampling time, and the difference between the historical maximum value and the historical minimum value of soil apparent electrical conductivity is used as the salt historical range assessment. The salt accumulation assessment is used as the numerator, the salt historical range assessment is used as the denominator, and the corresponding fraction is used as the salt accumulation degree at the target sampling time.

[0067] After obtaining the water deficit and salinity accumulation, the instantaneous cross-coupling and nonlinear mapping processes are applied to these values ​​to obtain the instantaneous cross-coupling factor. Specifically, for any target sampling time within any target sliding time window, the water deficit and salinity accumulation corresponding to that sampling time are obtained. The water deficit and salinity accumulation at the target sampling time are multiplied, and the resulting product is used as the cross-coupling product evaluation for that target sampling time. The water deficit and salinity accumulation at the target sampling time are added together, and the sum is used as the co-deterioration intensity evaluation for that target sampling time. An exponential mapping with a natural constant as the base is applied to the co-deterioration intensity evaluation, and the resulting mapping is used as the co-deterioration exponential amplification evaluation for that target sampling time. The cross-coupling product evaluation and the co-deterioration exponential amplification evaluation at the target sampling time are multiplied together, and the resulting product is added to a constant 1. The calculated result is used as the instantaneous co-coupling mapping evaluation for that target sampling time. The instantaneous collaborative mapping evaluation corresponding to the target sampling time is subjected to natural logarithmic mapping, and the corresponding mapping result is used as the instantaneous cross-collaborative factor corresponding to the target sampling time.

[0068] In one implementation, assume the first The water deficit at each sampling time was ;No. The salt accumulation at each sampling time is Then the first The expression for calculating the instantaneous cross-cooperation factor corresponding to each sampling time is:

[0069]

[0070] in, Indicates the first The instantaneous cross-cooperation factor corresponding to each sampling moment; This represents the logarithmic function with the natural constant e as the base. Indicates the first Moisture deficit at each sampling time; Indicates the first Salt accumulation at each sampling time; This represents an exponential function with the natural constant e as the base.

[0071] It should be noted that, to address the issue that the Euclidean distance formula cannot recognize the overlapping effects of water and salinity, this embodiment constructs a cross-product structure consisting of the product of water deficit and salinity accumulation. In actual data acquisition, if only a single indicator deteriorates, such as water deficit while salinity is normal, the cumulative degree corresponding to the normal indicator will approach 0, and the result of the cross-product structure will also be constrained and decay to approach 0. This design ensures that the algorithm only triggers the cooperative penalty mechanism in segments where low water and high salinity occur simultaneously, effectively avoiding misjudgment of regular single fluctuation data. Furthermore, since the linear superposition property of existing distance formulas cannot characterize the accelerated decay characteristics when multiple indicators are cross-superimposed, this embodiment introduces an exponential amplification structure based on the cross-product structure. When the soil environment enters a severe state of dual water and salt degradation, the sum of the two deficit values ​​increases. The exponential function can nonlinearly amplify this extreme cross-degradation state, thus accurately reproducing the true data evolution law of accelerated decline in arable land quality under multiple environmental pressures in a numerical dimension. Furthermore, to prevent the exponential function from outputting excessively large response values ​​under extremely harsh data, thereby disrupting the weight balance of other conventional indicator dimensions in the basic EWM-TOPSIS model, this formula... While preserving the nonlinear amplification trend of the internal exponential term, it can smoothly converge the final generated collaborative factor output value to a reasonable mathematical range, thereby ensuring the data fusion stability when the optimization factor is incorporated into the original distance calculation formula.

[0072] Thus, the instantaneous cross-cooperative factor was obtained by cross-cooperative mapping of soil volumetric water content and soil apparent electrical conductivity.

[0073] Step S3: By performing conditional integration on the evolution trajectory of instantaneous cross-synergistic factors and soil volumetric water content, the stress fatigue integral factor is obtained.

[0074] After constructing the instantaneous cross-coordination factor in step S2, the model possesses the ability to identify the water-salt superposition state at a specific moment. However, directly applying this instantaneous feature to the final evaluation faces data interference issues caused by diurnal variation fluctuations in high-frequency IoT monitoring scenarios. In actual high-frequency farmland monitoring, especially with sensors deployed in the shallow tillage layer, the collected data is driven by surface water heat exchange, exhibiting significant diurnal periodic extreme fluctuations. For example, during the midday period of strong radiation, capillary water in the surface soil is rapidly consumed due to intense evaporation. The strong evaporation pull simultaneously draws soluble salts from the lower layers to the surface for a brief period of enrichment. This diurnal micro-salinization and dehydration phenomenon leads to a large characteristic value output by the instantaneous cross-coordination factor during this period. However, as the surface temperature decreases at night and evaporation ceases, coupled with trace amounts of surface condensation and the slow replenishment of deep capillary water, the surface water-salt concentration experiences a periodic decline and dilution, thus dissolving the aforementioned brief cross-superposition state. This high-frequency periodic fluctuation driven by diurnal meteorological variations does not cause irreversible long-term damage to the soil's physical structure. If the algorithm directly adopts this unfiltered instantaneous state for distance measurement, the final farmland assessment score will oscillate frequently with the diurnal cycle of the environment, losing its significance in assessing the stable productivity of farmland. To solve the assessment oscillation problem caused by high-frequency data instantaneous fluctuations, isolated instantaneous features must be considered within the evolutionary trajectory of the time series. This mechanism needs to identify the macroscopic evolution trend of the data, and only when the water-salt cross-state is accompanied by a monotonous and continuous dehydration path of the soil is it identified as substantial environmental damage, thereby extracting a fatigue damage feature based on the data evolution path and filtering out the noise of conventional diurnal environmental fluctuations.

[0075] In summary, this invention first extracts and normalizes the unidirectional change data of soil volumetric moisture content evolution trajectory to obtain a unidirectional water loss assessment. Specifically, for any target sampling time within any target sliding time window (excluding the first sampling time), the soil volumetric moisture content data corresponding to the target sampling time and the soil volumetric moisture content data corresponding to the previous sampling time are obtained. The historical maximum and minimum values ​​of soil volumetric moisture content corresponding to the target sliding time window are also obtained. The difference between the soil volumetric moisture content data corresponding to the previous sampling time and the soil volumetric moisture content data corresponding to the target sampling time is used as the water change assessment corresponding to the target sampling time. When the water change assessment is less than or equal to a constant 0, the unidirectional water loss assessment corresponding to the target sampling time is set to a constant 0. When the water change assessment is greater than a constant 0, the difference between the historical maximum and historical minimum values ​​of soil volumetric moisture content is used as the historical water range assessment. The water change assessment corresponding to the target sampling time is used as the numerator, the historical water range assessment is used as the denominator, and the resulting fraction is used as the unidirectional water loss assessment corresponding to the target sampling time.

[0076] After obtaining the unidirectional moisture loss assessment, the instantaneous cross-coordination factor and the unidirectional moisture loss assessment are dynamically coupled to obtain the single-step fatigue penalty assessment. Specifically, for any target sampling time within any target sliding time window (excluding the first sampling time), the instantaneous cross-coordination factor and the unidirectional moisture loss assessment corresponding to the target sampling time are obtained. The instantaneous cross-coordination factor and the unidirectional moisture loss assessment corresponding to the target sampling time are multiplied, and the resulting product is used as the single-step fatigue penalty assessment corresponding to the target sampling time.

[0077] After obtaining the single-step fatigue penalty assessment, the stress fatigue integral factor is obtained by accumulating and nonlinearly mapping the single-step fatigue penalty assessment within the sliding time window. Specifically, for any target sliding time window, the single-step fatigue penalty assessments corresponding to all sampling times within the target sliding time window (excluding the first sampling time) are accumulated, and the accumulated result is used as the cumulative fatigue penalty assessment corresponding to the target sliding time window. The negative of the cumulative fatigue penalty assessment corresponding to the target sliding time window is exponentially mapped to the natural constant, and the resulting mapping result is used as the cumulative fatigue decay assessment corresponding to the target sliding time window. The difference between the constant 1 and the cumulative fatigue decay assessment corresponding to the target sliding time window is used as the stress fatigue integral factor corresponding to the target sliding time window.

[0078] In one implementation, assume the first The soil volumetric moisture content data at each time point are: , No. The soil volumetric moisture content data at each time point are: The historical maximum value of soil volumetric moisture content corresponding to the sliding time window is The historical minimum value of soil volumetric moisture content corresponding to the sliding time window is... The length of the sliding time window is Then the first The expression for calculating the stress fatigue integral factor corresponding to the sliding time window at each moment is:

[0079]

[0080] in, Indicates the first The stress fatigue integral factor corresponding to the sliding time window at each moment. Indicates the first Instantaneous cross-cooperation factor at each moment; Represents the maximum value function; Indicates the first Soil volumetric moisture content data at each time point; Indicates the first Soil volumetric moisture content data at each time point; This represents the historical maximum soil volumetric moisture content corresponding to the sliding time window. This represents the historical minimum soil volumetric moisture content corresponding to the sliding time window.

[0081] It should be noted that, to address the issue of spurious degradation feature accumulation caused by diurnal periodic fluctuations, this formula internally constructs a unidirectional evolution triggering structure based on temporal difference. This structure determines the true direction of soil environmental evolution by calculating the difference in moisture content between adjacent time steps. When encountering temporary backwatering caused by nighttime condensation or normal irrigation and rainfall, the moisture content shows an upward trend, the difference result is negative, and the maximum value function will directly output 0, thus forcibly cutting off the instantaneous cross-coefficient factor values ​​accumulated at that moment and before in the overall damage at the algorithm level. Only when the soil moisture content shows a strictly monotonically decreasing state does the structure output a positive water loss. This data filtering mechanism based on evolution direction can accurately remove daytime cycle noise that has not formed a continuous environmental degradation from high-frequency time series signals. At the same time, this formula extracts the unidirectional water loss by dividing it by the historical range within the window. Local normalization is performed, and this normalized value is multiplied as the dynamic loss gradient with the instantaneous cross-coordination factor at the current moment. This cross-product structure couples the static cross-forced state of the environment at that moment with its actual dynamic loss displacement. Only when extreme instantaneous environmental degradation is accompanied by a high data loss rate will a meaningful single-step fatigue penalty weight be generated. This dynamic weighting mechanism realistically recreates the damage accumulation process of the farmland environment along the continuous deterioration trajectory at the pure data level. In addition, after completing the data accumulation and summation within the time window, in order to smoothly integrate this fatigue damage feature into the subsequent multi-factor distance measurement, this formula applies a negative exponential truncation structure on the outer side. This structure can smoothly and strictly monotonically shrink the internal accumulated sum characteristic value, which is not constrained by the upper bound, to a value. Within the standardized interval. When no continuous dehydration degradation meeting the conditions occurs within the time window, the internal sum is zero. This mapping structure ensures that the stress fatigue integral factor output is zero, avoiding invalid interference to the basic evaluation distance. However, as continuous dehydration and high-salt stress accumulate and amplify, the stress fatigue integral factor will gradually and steadily approach 1, thereby extracting a stable penalty feature set for the algorithm backend that is resistant to high-frequency oscillations and represents long-term environmental trends.

[0082] Thus, the stress fatigue integral factor was obtained by conditionally integrating the instantaneous cross-synergistic factor and the evolution trajectory of soil volumetric water content.

[0083] Step S4: Obtain the corrected worst solution distance by performing distance correction processing on the basic worst solution distance and the stress fatigue integral factor.

[0084] After extracting the features of the stress fatigue integral factor, the optimization result needs to be fed back into the basic EWM-TOPSIS evaluation model to compensate for the deficiency of its Euclidean distance metric formula in handling water-salt synergistic features. The existing EWM-TOPSIS model uses the square root of the sum of squared distances in each independent dimension when calculating the distance from the evaluated object to the ideal worst solution. To utilize the extracted damage features based on evolutionary trajectories to correct the error caused by this linear superposition, this embodiment adopts a denominator correction mechanism based on nonlinear compression. The stress fatigue integral factor obtained in step S3 is used as a distance attenuation weight and directly applied to the distance to the basic negative ideal solution calculated by the original model. Since the integral factor approaches 1 when water and salt continuously deteriorate, the final corrected distance output by the algorithm is significantly reduced through division correction. This forcibly removes the false safe distance obtained by the linear independence assumption for this plot at the algorithm's fusion calculation level, accurately bringing it closer to the worst solution state of extreme deterioration of multiple factors in the feature space.

[0085] Specifically, based on a standardized multi-element feature matrix, the entropy weight method is used to calculate the index weights corresponding to soil volumetric water content, soil apparent electrical conductivity, soil organic matter content, soil porosity, and soil pH, respectively. The worst-case solution data for each index dimension is determined based on the standardized soil element data. For any target sampling time corresponding to a target sliding time window, the difference between the standardized soil element data and the worst-case solution data for the corresponding index dimension is calculated. The square of each difference multiplied by the corresponding index weight is used as the weighted worst-case solution distance component for the corresponding index dimension. The weighted worst-case solution distance components for all index dimensions are summed, and the square root of the sum is calculated. The result is used as the basic worst-case solution distance for the target sampling time. The stress fatigue integral factor corresponding to the target sliding time window to which the target sampling time belongs is obtained. The stress fatigue integral factor is added to a constant 1, and the result is used as the distance correction denominator for the target sampling time. The distance to the worst solution at the target sampling time is used as the numerator, the distance correction denominator evaluation is used as the denominator, and the resulting fraction is used as the corrected worst solution distance at the target sampling time.

[0086] It should be noted that when the cultivated land within the assessment window does not experience sustained water-salt co-deterioration, the stress fatigue integral factor is equal to zero, and the distance to the corrected worst solution remains consistent with the distance to the basic worst solution. The conventional dimensional feature fusion process of the original assessment model is unaffected. When sensor data reflects that the plot has experienced severe and sustained water-salt cross-dehydration and deterioration during the assessment period, the stress fatigue integral factor outputs a larger feature weight. In this case, the distance correction denominator is significantly greater than 1, causing the distance to the basic worst solution calculated by the original model to be forcibly reduced. This correction process breaks the feature masking effect caused by the linear accumulation structure of the original model, anchoring plots with ecological damage in the distance metric space to a position close to the worst solution, thus providing an accurate distance benchmark for subsequent calculation of the comprehensive relative proximity, eliminating the co-superposition error.

[0087] Thus, the distance to the corrected worst solution is obtained by performing distance correction processing on the distance to the basic worst solution and the stress fatigue integral factor.

[0088] Step S5: Calculate the relative proximity between the distance to the modified worst solution and the distance to the basic best solution to obtain the comprehensive assessment result of farmland quality.

[0089] Based on the standardized multi-factor feature matrix, the optimal solution data corresponding to each indicator dimension is determined according to the EWM-TOPSIS algorithm. Combined with the entropy weight corresponding to each indicator, the weighted Euclidean distance between the standardized soil element data and the corresponding optimal solution data at the target sampling time is calculated. The obtained distance is used as the basic optimal solution distance at the target sampling time.

[0090] For any target sampling time, obtain the corrected worst-case solution distance obtained in step S4. Add the basic optimal solution distance to the corrected worst-case solution distance to obtain the total distance span corresponding to the target sampling time. Use the corrected worst-case solution distance as the numerator and the total distance span as the denominator, and use the resulting fraction as the comprehensive relative proximity score corresponding to the target sampling time. The larger the comprehensive relative proximity score, the closer the current soil multi-element state is to the ideal optimal state; the smaller the comprehensive relative proximity score, the closer the current soil state is to the worst state. Arrange the comprehensive relative proximity scores corresponding to each target sampling time according to the time sequence of sampling times to form a comprehensive relative proximity score sequence. Based on the comprehensive relative proximity score sequence, dynamically assess the farmland quality, productivity, and ecological health status of the farmland area to be evaluated to obtain the comprehensive farmland quality assessment result.

[0091] See Figure 2 , Figure 2 The figure shows the simulation verification results of the optimization model provided in the embodiment of the present invention. Figure 2 (a) is a time series diagram of simulated monitoring data for sites with synergistic water and salt degradation. Figure 2 (b) is a comparison chart of the assessment scores and actual quality of land parcels with synergistic water and salt degradation. Figure 2 (c) is a graph showing the evolution of intermediate feature factors in the optimization model of this invention. Figure 2 (d) is a comparison chart of the assessment errors of three typical land parcels.

[0092] Specifically, Figure 2 In (a), the horizontal axis represents time (unit: days), the left vertical axis represents soil volumetric water content (unit: m³ / m³, blue curve in the figure), and the right vertical axis represents soil apparent electrical conductivity (unit: dS / m, red curve in the figure). The shaded area in the figure represents the period of water-salt synergistic degradation starting from day 18. It can be seen that before day 18, the soil volumetric water content exhibits normal periodic fluctuations under the action of irrigation pulses, and the apparent electrical conductivity remains stable at a low level, indicating that the plot is in a healthy state. From day 18 onwards, the soil volumetric water content continuously decreases along a monotonous path, while the apparent electrical conductivity simultaneously and continuously increases. The two form a synergistic evolution process of water deficit and salt accumulation, simulating a typical scenario in actual farmland where water and salt stress leads to accelerated decline in arable land productivity.

[0093] Figure 2In (b), the gray dashed line is the preset real farmland quality curve, the blue curve is the comprehensive relative closeness output by the existing entropy weight approximation ideal solution ranking method based on time-series sliding window (traditional entropy weight TOPSIS), and the red curve is the comprehensive relative closeness output by the optimization model of this invention. It can be seen that before the degradation begins on day 18, the three curves basically overlap around 0.9, indicating that the optimization model of this invention does not interfere with the basic assessment process under normal conditions. After degradation begins, the actual farmland quality exhibits a precipitous decline due to the combined effects of water and salt. The traditional entropy weight TOPSIS, due to its Euclidean distance metric formula treating each indicator dimension as independent and linearly accumulating them, masks the synergistic effect of simultaneous degradation of water and salt, and its output score stagnates at around 0.5, significantly overestimating the actual farmland quality. In contrast, the optimization model of this invention identifies the cross-state of simultaneous degradation of water and salt through instantaneous cross-synergistic factors, and after accumulating and amplifying the continuous dehydration trajectory through stress fatigue integral factors, it corrects the distance of the worst solution. Its output score drops to around 0.35, which is highly consistent with the accelerated degradation pattern of the actual farmland quality curve, indicating that this invention can accurately restore the actual farmland state under the combined influence of multiple factors.

[0094] Figure 2 In (c), the orange curve represents the instantaneous cross-cooperation factor C(t), and the green curve represents the stress fatigue integral factor F(t). It can be seen that the instantaneous cross-cooperation factor is influenced by the diurnal variation cycle driven by surface water and heat exchange, exhibiting obvious high-frequency oscillation characteristics. In contrast, the stress fatigue integral factor, after conditional integration using a unidirectional evolution triggering structure based on temporal difference, exhibits a smooth, monotonically increasing pattern that gradually approaches saturation. This verifies that the unidirectional evolution triggering structure described in this invention can eliminate diurnal cycle noise that does not lead to continuous environmental degradation from high-frequency time-series signals, accumulating substantial damage characteristics only when water-salt cross-states are accompanied by a monotonous continuous dehydration path, thereby outputting stable penalty characteristics that resist high-frequency oscillations and represent long-term environmental trends.

[0095] Figure 2In (d), the blue bars represent the evaluation error (RMSE) of the traditional entropy-weighted TOPSIS, and the red bars represent the evaluation error (RMSE) of the optimized model of this invention. The error is represented by the root mean square error between the output score of each method and the actual farmland quality. The lower the value, the more accurate the evaluation. It can be seen that for healthy plots, the evaluation errors of the two methods are 0.010 and 0.011, respectively, which are basically consistent. This indicates that the synergistic penalty mechanism introduced in this invention is approximately inactive when water and salt synergistic degradation does not occur, and will not produce misjudgments. For plots with single salinity degradation, the evaluation errors of the two methods are 0.046 and 0.040, respectively. This indicates that when a single index deteriorates, the cross-product structure is automatically decayed due to the constraint of the normal index, and the model of this invention does not produce an over-response. For plots with water and salt synergistic degradation, the evaluation error of the traditional method is as high as 0.145, while the evaluation error of the optimized model of this invention is reduced to 0.072, a reduction of about 50.1%.

[0096] In summary, this invention significantly improves the assessment accuracy of water-salt synergistic degradation scenarios by constructing an instantaneous cross-synergistic factor and an evolutionary trajectory-based stress fatigue integral factor, and integrating them into the distance measurement stage of the basic assessment model. This is achieved while ensuring the stability of assessments in conventional scenarios and preventing misjudgments in single degradation scenarios. It overcomes the shortcomings of existing linear Euclidean distance accumulation structures, which cannot extract multi-factor synergistic change characteristics and cause assessment results to deviate from the actual state of cultivated land. This verifies the effectiveness of the technical solution of this invention.

[0097] Thus, the comprehensive assessment results of arable land quality were obtained by calculating the relative closeness between the distance of the modified worst solution and the distance of the basic best solution.

[0098] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for assessing arable land quality based on the integration of multiple factors including soil productivity and ecology, characterized in that, The method includes: Step S1: Obtain a standardized multi-element feature matrix by collecting and preprocessing time-series data of multi-source soil elements; Step S2: Obtain the instantaneous cross-cooperative factor by performing cross-cooperative mapping processing on soil volumetric water content and soil apparent electrical conductivity; Step S3: Obtain the stress fatigue integral factor by conditionally integrating the instantaneous cross-synergistic factor and the evolution trajectory of soil volumetric water content; Step S4: Obtain the corrected worst solution distance by performing distance correction processing on the basic worst solution distance and the stress fatigue integral factor; Step S5: Obtain the comprehensive assessment result of cultivated land quality by calculating the relative closeness between the distance of the modified worst solution and the distance of the basic best solution.

2. The method for assessing arable land quality based on the integration of multiple soil productivity and ecological factors according to claim 1, characterized in that, The process involves collecting and preprocessing time-series data of multi-source soil elements to obtain a standardized multi-element feature matrix, including: Agricultural IoT monitoring equipment is deployed in the cultivated land area to be assessed. The agricultural IoT monitoring equipment is equipped with soil moisture sensor for collecting soil volumetric moisture content data, soil conductivity sensor for collecting soil apparent conductivity data, soil organic matter monitoring sensor for collecting soil organic matter content data, soil porosity monitoring sensor for collecting soil porosity data, and soil pH sensor for collecting soil pH data. A fixed sampling time step is set, and continuous monitoring and data collection are carried out on the cultivated land area to be evaluated according to the fixed sampling time step to obtain time series data of multi-source soil elements within the evaluation period. The time series data of multi-source soil elements include time series data of soil volumetric water content, soil apparent electrical conductivity, soil organic matter content, soil porosity, and soil pH. The time series data of multi-source soil elements are time-stamped according to the standard timestamp corresponding to the fixed sampling time step. The missing data after time-stamp alignment is completed by linear interpolation. The completed time series data of multi-source soil elements is continuously extracted according to the preset sliding time window to obtain the time series data fragments of multi-source soil elements corresponding to each sliding time window. For any target sliding time window, the historical maximum value, historical minimum value, historical maximum value, and historical minimum value of soil volumetric moisture content, soil apparent electrical conductivity, and soil apparent electrical conductivity are statistically obtained from the multi-source soil element time series data segments corresponding to the target sliding time window. Extreme value standardization processing is then performed on the soil volumetric moisture content time series data, soil apparent electrical conductivity time series data, soil organic matter content time series data, soil porosity time series data, and soil pH time series data, respectively. The corresponding standardized data are arranged according to the sampling time and soil element dimensions to construct a standardized multi-element feature matrix.

3. The method for assessing arable land quality based on the integration of multiple soil productivity and ecological factors according to claim 1, characterized in that, The instantaneous cross-cooperative factor is obtained by performing cross-cooperative mapping processing on soil volumetric water content and soil apparent electrical conductivity, including: By performing local normalization on soil volumetric water content data, soil apparent electrical conductivity data and corresponding historical boundary data, the water deficit degree and salinity accumulation degree were obtained. Instantaneous cross-coupling and nonlinear mapping of water deficit and salinity accumulation are used to obtain the cross-cooperative factor.

4. The method for assessing arable land quality based on the integration of multiple soil productivity and ecological factors according to claim 3, characterized in that, The process involves locally normalizing soil volumetric water content data, soil apparent electrical conductivity data, and corresponding historical boundary data to obtain water deficit and salinity accumulation, including: For any target sampling time within any target sliding time window, extract the soil volumetric moisture content data and soil apparent electrical conductivity data corresponding to the target sampling time from the multi-source soil element time series data segment corresponding to the target sliding time window, and obtain the historical maximum value, historical minimum value, historical maximum value, and historical minimum value of soil volumetric moisture content, corresponding to the target sliding time window. The difference between the historical maximum value of soil volumetric moisture content and the soil volumetric moisture content data corresponding to the target sampling time is used as the water deficit assessment corresponding to the target sampling time, and the difference between the historical maximum value of soil volumetric moisture content and the historical minimum value of soil volumetric moisture content is used as the water historical range assessment. The water deficit assessment is used as the numerator, the water historical range assessment is used as the denominator, and the corresponding fraction is used as the water deficit degree corresponding to the target sampling time. The difference between the soil apparent electrical conductivity data at the target sampling time and the historical minimum value of soil apparent electrical conductivity is used as the salt accumulation assessment at the target sampling time, and the difference between the historical maximum value and the historical minimum value of soil apparent electrical conductivity is used as the salt historical range assessment. The salt accumulation assessment is used as the numerator, the salt historical range assessment is used as the denominator, and the corresponding fraction is used as the salt accumulation degree at the target sampling time.

5. The method for assessing arable land quality based on the integration of multiple soil productivity and ecological factors according to claim 3, characterized in that, The instantaneous cross-coupling and nonlinear mapping process for water deficit and salinity accumulation is used to obtain the cross-cooperative factor, including: For any target sampling time within any target sliding time window, obtain the water deficit and salinity accumulation corresponding to the target sampling time; Multiply the water deficit and salinity accumulation at the target sampling time, and use the resulting product as the cross-cooperative product evaluation at the target sampling time. Add the water deficit and salinity accumulation at the target sampling time, and use the sum as the assessment of the synergistic degradation intensity at the target sampling time. Perform an exponential mapping on the synergistic degradation intensity assessment with the natural constant as the base, and use the mapping result as the synergistic degradation exponential amplification assessment at the target sampling time. Multiply the cross-cooperative product evaluation corresponding to the target sampling time with the cooperative degradation index amplification evaluation, and add the corresponding product to the constant 1. Use the corresponding calculation result as the instantaneous cooperative mapping evaluation corresponding to the target sampling time. The instantaneous collaborative mapping evaluation corresponding to the target sampling time is subjected to natural logarithmic mapping, and the corresponding mapping result is used as the instantaneous cross-collaborative factor corresponding to the target sampling time.

6. The method for assessing arable land quality based on the integration of multiple soil productivity and ecological factors according to claim 1, characterized in that, The stress fatigue integral factor is obtained by conditionally integrating the instantaneous cross-synergy factor with the soil volumetric water content evolution trajectory, including: By extracting and normalizing the unidirectional change data of soil volumetric water content evolution trajectory, an assessment of unidirectional water loss is obtained. By dynamically coupling the instantaneous cross-synergistic factor with the unidirectional moisture loss assessment, a single-step fatigue penalty assessment is obtained. By performing cumulative and nonlinear mapping processing on the single-step fatigue penalty assessment within the sliding time window, the stress fatigue integral factor is obtained.

7. The method for assessing arable land quality based on the integration of multiple soil productivity and ecological factors according to claim 6, characterized in that, The method involves extracting and normalizing unidirectional changes in soil volumetric water content evolution trajectory data to obtain an assessment of unidirectional water loss, including: For any target sampling time other than the first sampling time within any target sliding time window, obtain the soil volumetric moisture content data corresponding to the target sampling time and the soil volumetric moisture content data corresponding to the previous sampling time of the target sampling time, and obtain the historical maximum value and historical minimum value of soil volumetric moisture content corresponding to the target sliding time window. The difference between the soil volumetric moisture content data corresponding to the previous sampling time and the soil volumetric moisture content data corresponding to the target sampling time is used as the moisture change assessment corresponding to the target sampling time; when the moisture change assessment is less than or equal to a constant 0, the one-way water loss assessment corresponding to the target sampling time is set to a constant 0. When the moisture change assessment is greater than the constant 0, the difference between the historical maximum and the historical minimum of soil volumetric moisture content is used as the moisture historical range assessment. The moisture change assessment corresponding to the target sampling time is used as the numerator, the moisture historical range assessment is used as the denominator, and the corresponding fraction is used as the unidirectional moisture loss assessment corresponding to the target sampling time.

8. The method for assessing arable land quality based on the integration of multiple soil productivity and ecological factors according to claim 6, characterized in that, The process of dynamically coupling the instantaneous cross-coordination factor with the unidirectional moisture loss assessment to obtain a single-step fatigue penalty assessment includes: For any target sampling time other than the first sampling time within any target sliding time window, obtain the instantaneous cross-coordination factor corresponding to the target sampling time and the unidirectional moisture loss assessment corresponding to the target sampling time; Multiply the instantaneous cross-coordination factor corresponding to the target sampling time with the unidirectional moisture loss assessment, and use the resulting product as the single-step fatigue penalty assessment corresponding to the target sampling time.

9. A method for assessing arable land quality based on the integration of multiple soil productivity and ecological factors according to claim 6, characterized in that, The step of obtaining the stress fatigue integral factor by accumulating and nonlinearly mapping the single-step fatigue penalty assessment within the sliding time window includes: For any target sliding time window, the single-step fatigue penalty assessments corresponding to all sampling times except the first sampling time within the target sliding time window are accumulated, and the corresponding accumulated result is used as the cumulative fatigue penalty assessment corresponding to the target sliding time window. The negative number of the cumulative fatigue penalty assessment corresponding to the target sliding time window is subjected to an exponential mapping with the natural constant as the base, and the corresponding mapping result is used as the cumulative fatigue decay assessment corresponding to the target sliding time window. The difference between constant 1 and the fatigue cumulative decay assessment corresponding to the target sliding time window is used as the stress fatigue integral factor corresponding to the target sliding time window.

10. The method for assessing arable land quality based on the integration of multiple soil productivity and ecological factors according to claim 1, characterized in that, The process of obtaining the corrected worst-case solution distance by performing distance correction processing on the basic worst-case solution distance and the stress fatigue integral factor includes: Based on the standardized multi-element feature matrix, the entropy weight method is used to calculate the index weights corresponding to soil volumetric water content, soil apparent electrical conductivity, soil organic matter content, soil porosity and soil pH, and the worst solution data corresponding to each index dimension is determined based on the standardized soil element data. For any target sampling time corresponding to a target sliding time window, calculate the difference between the standardized soil element data and the worst solution data of the corresponding index dimension at the target sampling time. Multiply the square of each difference by the corresponding index weight and use the result as the weighted worst solution distance component of the corresponding index dimension. Sum the weighted worst solution distance components corresponding to all index dimensions and calculate the square root of the sum. Use the result as the basic worst solution distance at the target sampling time. Obtain the stress fatigue integral factor corresponding to the target sliding time window to which the target sampling time belongs, add the stress fatigue integral factor to the constant 1, and use the corresponding calculation result as the distance correction denominator evaluation corresponding to the target sampling time; The distance to the worst solution at the target sampling time is used as the numerator, the distance correction denominator evaluation is used as the denominator, and the resulting fraction is used as the corrected worst solution distance at the target sampling time.

Images