Soil environment management method and system based on data analysis

Abstract

The invention relates to the field of soil environment management, in particular to a soil environment management method and system based on data analysis. A soil environment management system based on data analysis comprises a data monitoring module, a soil characteristic distinguishing module, a soil quality evaluation module and a soil risk prediction module. According to the method, an environment variable-driven dynamic adjustment mechanism is constructed, the weight coefficient of the comprehensive environment suitable value and the soil body data affected by the environment is introduced, the ideal range of each soil parameter is dynamically adjusted, the method does not depend on a fixed threshold value, and the flexibility and scientificity of soil quality evaluation are remarkably improved; the mechanism not only solves the problem that the static evaluation standard is disjointed with the dynamic change of the environment, but also realizes the transparency and interpretability of the evaluation process through mathematical quantification, and provides a reliable basis for precise agricultural management.

Description

Technical Field

[0001] The present invention relates to the field of soil environmental management, and particularly relates to a soil environmental management method and system based on data analysis. Background Art

[0002] With the rapid development of modern agriculture and ecological environment management, the precise and dynamic management of soil environmental quality has become a key requirement for the sustainable development of agriculture and ecological protection; traditional soil management methods mostly rely on manual experience or static threshold evaluation, and have the following limitations: defects of static evaluation, traditional methods usually use fixed soil parameter thresholds for quality evaluation and cannot flexibly adjust the standards according to the dynamic changes of environmental conditions; for example, under extreme climates such as continuous drought or heavy rain, the tolerance of the soil may change significantly, but the existing methods are difficult to reflect this adaptive demand in a timely manner, resulting in the disconnection between the evaluation results and the actual risks; ignoring regional differences: there are significant differences in the response degrees of soil characteristics in different regions to the same environmental variable, while the existing technologies often adopt globally unified weights or models and lack personalized analysis for local regions, resulting in insufficient prediction accuracy; lagging risk warning: traditional methods mostly evaluate based on current or historical data and lack the ability to predict future environmental changes; for example, it is impossible to combine future rainfall prediction to judge the soil erosion risk in advance, resulting in insufficient warning timeliness and being difficult to support proactive intervention.

[0003] In view of the above problems, a soil environmental management solution that can integrate multi-source time-series data, dynamically adapt to environmental changes and has regional personalization is needed. Therefore, the present invention proposes a soil environmental management method and system based on data analysis. Through innovative data analysis architecture and algorithm design, the limitations of traditional methods are broken through, and the precise evaluation of soil quality and risk warning are realized, providing technical support for smart agriculture and ecological governance. Summary of the Invention

[0004] The present invention constructs a dynamic adjustment mechanism driven by environmental variables, introduces a comprehensive environmental suitability value and a weight coefficient of soil body data affected by the environment, and dynamically adjusts the ideal range of each soil parameter, rather than relying on fixed thresholds, significantly improving the flexibility and scientificity of soil quality evaluation; this mechanism not only solves the problem of the disconnection between static evaluation criteria and dynamic environmental changes, but also realizes the transparency and interpretability of the evaluation process through mathematical quantification, providing a reliable basis for precise agricultural management.

[0005] A soil environmental management method based on data analysis includes: Set a number of local management areas, and collect soil body data and environmental variable data of each local management area at the current monitoring time point. The soil body data includes soil humidity, soil compaction, pH value, organic matter content, conductivity, and heavy metal content; the environmental variable data includes environmental temperature, environmental humidity, light intensity, precipitation, and wind speed. Set an alternation period. At the beginning of any alternation period, based on the environmental variable data and soil body data collected from each local management area in the recent period, calculate the influence degree of the soil body data of each local management area by the environmental variable data respectively. At the current monitoring time point, based on the environmental variable data collected from each local management area in the recent period, calculate the comprehensive environmental suitability value of all environmental variable data respectively; use the calculated comprehensive environmental suitability value and the influence degree of the soil body data of each local management area by the environmental variable data to calculate the ideal range of the soil body data of each local management area, and then calculate the comprehensive soil quality score of each local management area; judge whether the soil quality score of each local management area meets the preset compliance conditions. If so, no operation is performed; if not, execute the soil problem treatment measures. At the beginning of any alternation period, for any local management area, obtain the predicted environmental variable data of the current local management area in the next period; based on the influence degree of the soil body data of the current local management area obtained in the current alternation period by the environmental variable data, calculate the contribution coefficient of each predicted environmental variable data to the current local management area respectively; construct a soil risk prediction model, and use the soil body data sorted by time, the predicted environmental variable data sorted by time, and the contribution coefficient of each predicted environmental variable data obtained in the recent period as the input of the soil risk prediction model, and output the risk probability of the current local management area having soil problems in the next period; if the risk probability is higher than the preset risk threshold, execute the soil problem warning measures.

[0006] Preferably, based on the environmental variable data and soil body data collected from each local management area in the recent period, calculate the influence degree of the soil body data of each local management area by the environmental variable data respectively. The specific operations are as follows: For any local management area, record the soil body data of the current local management area as , i = 1, 2,..., 6; to respectively represent soil humidity, soil compaction, pH value, organic matter content, conductivity, and heavy metal content; record the environmental variable data of the current local management area as , = 1, 2, …, 5; to respectively represent ambient temperature, ambient humidity, light intensity, precipitation, and wind speed; Set an alternation period. At the start of any alternation period, based on the soil body data and environmental variable data , = 1, 2, …, ; respectively calculate the mean value of the soil body data and the mean value of the and environmental variable data ; use to calculate and obtain the degree of influence of the soil body data of the current local management area by the environmental variable data . .

[0007] Preferably, based on the environmental variable data collected and obtained in each local management area in the recent period, calculate and obtain the comprehensive environmental suitability value of all environmental variable data. The specific operations are as follows: For any local management area, according to the crop type of the current local management area, obtain the environmental suitability range corresponding to each environmental variable data to ; for any environmental variable data , if or , then the single - time environmental suitability value of this environmental variable data is 0; if , then use the formula to calculate and obtain the single - time environmental suitability value , where is the intermediate value between and ; if , then use the formula to calculate and obtain the single - time environmental suitability value ; based on the environmental variable data obtained in the recent period , calculate and obtain single - time environmental suitability values , and then calculate single - time environmental suitability values ​ The average value of is obtained to get the comprehensive environmental suitability value. From to

[0008] preferably, using the calculated comprehensive environmental suitability value and the influence degree of the soil body data of each local management area by the environmental variable data, calculate the ideal range of the soil body data of each local management area, and then calculate and obtain the comprehensive soil quality score of each local management area. The specific operations are as follows: For any local management area, use the formula to calculate and obtain the soil body data of the 5 weight coefficients ; then use the formula to calculate and obtain the soil suitability value of the soil body data ; From to correspond to soil humidity, soil compactness, pH value, organic matter content, conductivity and heavy metal content in sequence; Set the reference range of the soil body data as to and use the formula to calculate and obtain the ideal minimum value of the soil body data , where is the intermediate value of and , is the adjustment coefficient, ; use the formula to calculate and obtain the ideal maximum value of the soil body data ; Based on the currently obtained soil body data , if or , then the quality score corresponding to the currently obtained soil body data is 0; if , then use the formula to calculate and obtain the quality score ; if , then use the formula to calculate and obtain the quality score ; finally use the formula ​Calculate and obtain the comprehensive soil quality score of the current local management area 。

[0009] Preferably, based on the degree of influence of the soil ontology data of the current local management area obtained in the current replacement cycle by the environmental variable data, calculate and obtain the contribution coefficients of each predicted environmental variable data to the current local management area respectively. The specific operations are as follows: Based on the obtained soil ontology data of the current local management area affected by the environmental variable data degree of influence for any environmental variable data select the value with the largest absolute value from the six degrees of influence corresponding to this environmental variable data ; Use the formula to calculate and obtain the contribution coefficients of each predicted environmental variable data to the current local management area 。 。

[0010] Preferably, the soil risk prediction model is established based on LSTM, including an input layer, a feature extraction layer, an environmental feature weighting layer, a feature splicing layer, a fully connected layer and an output layer; The input layer is used to receive the soil ontology data sorted by time, the predicted environmental variable data sorted by time, and the contribution coefficients of each predicted environmental variable data obtained in the recent period of time; The feature extraction layer includes a parallel first LSTM layer and a second LSTM layer. The first LSTM layer is used to extract the soil time series feature vector of the soil ontology data sorted by time; the second LSTM layer is used to extract the environmental time series feature vector of the predicted environmental variable data sorted by time; The environmental feature weighting layer is used to apply the contribution coefficients of each predicted environmental variable data to the current local management area to weight and adjust the environmental time series feature vector; The feature splicing layer is used to splice the soil time series feature vector and the weighted and adjusted environmental time series feature vector to obtain a comprehensive time series feature vector; The fully connected layer is used to further process the comprehensive time series feature vector to extract non-linear features; The output layer is used to convert the output result of the fully connected layer into the final risk probability.

[0011] Preferably, for the training of the soil risk prediction model, the specific operations are as follows: Obtain a number of soil risk prediction training samples with marked risk probabilities. Each soil risk prediction training sample contains a segment of soil body data sorted by time, prediction environmental variable data sorted by time, and contribution coefficients of each prediction environmental variable data; divide all the obtained soil risk prediction training samples into a training set and a validation set, input the training set into the soil risk prediction model with initialized parameters for training, and use the validation set to validate the soil risk prediction model to obtain a validation result. Set training conditions and determine whether the validation result meets the training conditions. If so, output the trained soil risk prediction model; if not, continue to train the soil risk prediction model using the training set.

[0012] A soil environmental management system based on data analysis, comprising: A data monitoring module for collecting soil body data and environmental variable data of each local management area at the current monitoring time point; A soil characteristic differentiation module for calculating, at the start of any replacement cycle, the degree of influence of the environmental variable data on the soil body data of each local management area based on the environmental variable data and soil body data collected from each local management area in the recent period; A soil quality assessment module, including an environmental suitability value calculation unit, an ideal range adjustment unit, and a quality score calculation unit; the environmental suitability value calculation unit is used to calculate the comprehensive environmental suitability value of all environmental variable data respectively by using the environmental variable data collected from each local management area in the recent period; the ideal range adjustment unit is used to calculate the ideal range of the soil body data of each local management area by using the calculated comprehensive environmental suitability value and the degree of influence of the environmental variable data on the soil body data of each local management area; the quality score calculation unit is used to calculate the comprehensive soil quality score of each local management area and determine whether the soil quality score of each local management area meets the preset compliance conditions. If so, no operation is performed; if not, soil problem handling measures are executed; A soil risk prediction module is used to, at the beginning of any replacement cycle, for any local management area, obtain the predicted environmental variable data of the current local management area in a future period of time; calculate and obtain the contribution coefficients of each predicted environmental variable data to the current local management area respectively based on the influence degree of the soil ontology data of the current local management area obtained in the current replacement cycle by the environmental variable data; construct a soil risk prediction model, take the soil ontology data sorted by time, the predicted environmental variable data sorted by time, and the contribution coefficients of each predicted environmental variable data obtained in the recent period as the input of the soil risk prediction model, and output the risk probability of soil problems occurring in the current local management area in a future period of time; if the risk probability is higher than the preset risk threshold, then execute the soil problem warning measure.

[0013] The present invention has the following advantages: 1. By constructing a dynamically adjustable mechanism driven by environmental variables, introducing the comprehensive environmental suitability value and the weight coefficient of the influence of soil ontology data by the environment, and dynamically adjusting the ideal range of each soil parameter, rather than relying on fixed thresholds, the flexibility and scientificity of soil quality assessment are significantly improved; this mechanism not only solves the problem of the disconnection between static assessment criteria and dynamic environmental changes, but also realizes the transparency and interpretability of the assessment process through mathematical quantification, providing a reliable basis for precision agriculture management.

[0014] 2. By constructing a soil risk prediction model, the present invention can predict the occurrence risk of soil problems in advance and issue warnings in a timely manner, enabling soil management to change from passive response to active prevention, greatly reducing the occurrence probability of soil problems and the treatment cost; in agricultural production, predicting soil problems in advance can timely adjust the planting plan or take preventive measures, avoiding the reduction or even failure of crops caused by soil problems, ensuring the quality and safety of agricultural products, and at the same time reducing the subsequent economic investment in soil remediation. Description of the Drawings

[0015] Figure 1 It is a schematic structural diagram of a soil environment management system based on data analysis adopted in an embodiment of the present invention. Detailed Embodiments

[0016] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0017] Embodiment 1, a soil environment management method based on data analysis, includes: Set several local management areas. At the current monitoring time point, collect the soil body data and environmental variable data of each local management area. The soil body data includes soil humidity, soil compaction, pH value, organic matter content, electrical conductivity, and heavy metal content, which directly reflect the physical and chemical properties of the soil; the environmental variable data includes environmental temperature, environmental humidity, light intensity, precipitation, and wind speed, and these factors will affect the soil condition; the role of collecting data is to provide basic data support for subsequent soil environmental management; by dividing the local management areas, the refined management of large-area soil can be realized, and corresponding management strategies can be formulated according to the characteristics of different areas; at the same time, collecting the soil body data and environmental variable data helps to analyze the relationship between the soil condition and environmental factors, so as to more accurately evaluate the soil quality, predict the occurrence risk of soil problems, and take effective management measures; Set a replacement cycle. At the beginning of any replacement cycle, based on the environmental variable data and soil body data collected in each local management area in the recent period, calculate the influence degree of the soil body data of each local management area by the environmental variable data respectively; the core of setting the replacement cycle is to dynamically capture the interaction relationship between the environment and the soil by regularly updating data and model parameters, ensuring the scientificity and timeliness of the management strategy; at the beginning of each cycle, the system first collects the recent environmental variables (such as temperature, precipitation) and soil body data (such as humidity, pH value) of each area, and uses the Pearson correlation coefficient to quantify the correlation between the two; for example, through formula calculation, it is found that the correlation coefficient between precipitation and soil humidity in a sandy soil area is as high as 0.85, indicating that rainfall has a significant impact on humidity, while this coefficient may drop to 0.6 after continuous drought in the same area, reflecting the change of soil permeability; this dynamic calculation mechanism enables the system to adjust the environmental variable weight in real time, avoiding the evaluation deviation caused by environmental mutations in traditional fixed models; for example, during the rainy season, the system automatically reduces the precipitation weight to prevent over-warning of saturated soil; in the industrial pollution area, if the correlation coefficient between wind speed and heavy metal migration is monitored to increase from 0.2 to 0.6, the model will give priority to including wind speed in the risk prediction and trigger a pollution diffusion alarm in advance; the periodic update not only solves the lag problem of static models, but also reduces manual intervention through data driving, making the soil quality assessment and risk warning always fit the actual environmental evolution, and providing adaptive and high-precision decision support for smart agriculture; At the current monitoring time point, based on the environmental variable data collected and obtained in each local management area in the recent period, the comprehensive environmental suitability values ​​of all environmental variable data are calculated respectively; the ideal range of soil ontology data in each local management area is calculated by using the calculated comprehensive environmental suitability values ​​and the degree of influence of environmental variable data on the soil ontology data of each local management area. This ideal range is determined according to the current environmental conditions and the response of the soil to the environment, and can reflect the optimal state of the soil ontology data under the current environment. Then, the comprehensive soil quality score of each local management area is calculated, which can intuitively reflect the soil quality status of each local management area; it is judged whether the soil quality score of each local management area meets the preset standard conditions. If so, no operation is performed; if not, soil problem treatment measures are implemented; then, the system dynamically adjusts the ideal range of soil ontology data based on the environmental suitability value and the degree of influence of soil parameters on the environment; for example, when the environmental suitability value is high (0.9), the soil moisture baseline range [20%, 60%] will be narrowed to [30%, 50%], which is more stringent; if the environmental suitability value is low ( 0.4), then relax to [15%, 65%] to tolerate greater fluctuations; the purpose is to avoid misjudgment of static thresholds in extreme environments: when environmental conditions are excellent, slight abnormalities in soil parameters may cause problems, and the standards need to be raised; conversely, in harsh environments, the standards are appropriately relaxed to reduce false alarms; finally, based on the adjusted ideal range, the system calculates the comprehensive soil quality score: if the current soil moisture is 45% (ideal range 30%-50%), the moisture score is 1; if the pH value is 6.8 (ideal range 6.5-7.0), the pH score is 0.8; each parameter The scores are averaged to get a comprehensive score (such as 0.85). If it is lower than the threshold (such as 0.7), an early warning is triggered. The purpose is to provide an intuitive soil health indicator to help managers quickly identify problem areas (such as excessive heavy metals or insufficient organic matter) and give priority to high-risk areas. For example, if the soil moisture score in a certain area drops sharply to 0.3 due to continuous heavy rains, the system will immediately prompt drainage measures to prevent crop root rot. This dynamic closed-loop mechanism deeply couples environmental variables with soil characteristics, breaking through the static limitations of traditional methods and providing real-time, adaptive decision support for precision agriculture. At the beginning of any replacement cycle, for any local management area, obtain the predicted environmental variable data of the current local management area in the next period of time; based on the degree of influence of the soil ontology data of the current local management area obtained in the current replacement cycle by the environmental variable data, calculate the contribution coefficients of each predicted environmental variable data to the current local management area respectively; construct a soil risk prediction model, and use the soil ontology data sorted by time, the predicted environmental variable data sorted by time, and the contribution coefficients of each predicted environmental variable data obtained in the recent period as the input of the soil risk prediction model, and output the risk probability of soil problems occurring in the current local management area in the next period of time; if the risk probability is higher than the preset risk threshold, execute the soil problem warning measure.

[0018] Based on the environmental variable data and soil ontology data collected in each local management area in the recent period, calculate the degree of influence of the soil ontology data of each local management area by the environmental variable data respectively. The specific operations are as follows: For any local management area, denote the soil ontology data of the current local management area as , = 1, 2, …, 6; to represent soil moisture, soil compaction, pH value, organic matter content, electrical conductivity, and heavy metal content in sequence; denote the environmental variable data of the current local management area as , = 1, 2, …, 5; to represent environmental temperature, environmental humidity, light intensity, precipitation, and wind speed in sequence; Set the replacement cycle. At the beginning of any replacement cycle, based on the soil ontology data and environmental variable data obtained in the current local management area in the recent period, = 1, 2, …, ; calculate the mean value of the soil ontology data and the mean value of the environmental variable data respectively; use to calculate the degree of influence of the soil ontology data of the current local management area by the environmental variable data .

[0019] Based on the environmental variable data collected in each local management area in the recent period, calculate the comprehensive environmental suitability values of all environmental variable data respectively. The specific operations are as follows: For any local management area, obtain the environmental suitability ranges corresponding to each environmental variable data according to the crop type of the current local management area to ; for any environmental variable data , if or , then the single - time environmental suitability value of this environmental variable data is 0; if , then use the formula to calculate and obtain the single - time environmental suitability value , where is and the median of; if , then use the formula to calculate and obtain the single - time environmental suitability value ; based on the environmental variable data obtained in the recent period , calculate and obtain single - time environmental suitability values , and then calculate the average value of the single - time environmental suitability values to obtain the comprehensive environmental suitability value , to correspond to environmental temperature, environmental humidity, light intensity, precipitation, and wind speed in sequence.

[0020] Using the calculated comprehensive environmental suitability values and the influence degrees of the soil body data of each local management area by the environmental variable data, calculate the ideal ranges of the soil body data of each local management area, and then calculate the comprehensive soil quality scores of each local management area. The specific operations are as follows: For any local management area, use the formula to calculate and obtain the 5 weight coefficients of the soil body data ; then use the formula to calculate and obtain the soil suitability value of the soil body data , to correspond to soil humidity, soil compactness, pH value, organic matter content, conductivity, and heavy metal content in sequence; Set the soil body data The reference range of to is used to calculate and obtain the soil body data by using the formula The ideal minimum value of is obtained, where is the mid-value of and and is the adjustment coefficient, ; the ideal maximum value of the soil body data is calculated and obtained by using the formula ; Based on the currently obtained soil body data if or then the quality score corresponding to the currently obtained soil body data is 0; if then the quality score is calculated and obtained by using the formula ; if then the quality score is calculated and obtained by using the formula ; finally, the comprehensive soil quality score of the current local management area is calculated by using the formula

[0021] Based on the influence degree of the soil body data of the current local management area obtained in the current replacement cycle by the environmental variable data, the contribution coefficients of each predicted environmental variable data to the current local management area are calculated respectively. The specific operations are as follows: Based on the obtained soil body data of the current local management area affected by the environmental variable data the influence degree for any environmental variable data select the value with the largest absolute value from the six influence degrees corresponding to the environmental variable data ; the contribution coefficients of each predicted environmental variable data to the current local management area are calculated by using the formula

[0022] The soil risk prediction model is established based on LSTM and includes an input layer, a feature extraction layer, an environmental feature weighting layer, a feature splicing layer, a fully connected layer and an output layer; The input layer is used to receive the soil ontology data sorted by time, the predicted environmental variable data sorted by time, and the contribution coefficients of each predicted environmental variable data obtained in the recent period. The feature extraction layer includes a parallel first LSTM layer and a second LSTM layer. The first LSTM layer is used to extract the soil time-series feature vectors of the soil ontology data sorted by time; the second LSTM layer is used to extract the environmental time-series feature vectors of the predicted environmental variable data sorted by time. The environmental feature weighting layer is used to apply the contribution coefficients of each predicted environmental variable data to the current local management area to weight and adjust the environmental time-series feature vectors. The feature concatenation layer is used to concatenate the soil time-series feature vectors and the weighted and adjusted environmental time-series feature vectors to obtain comprehensive time-series feature vectors. The fully connected layer is used to further process the comprehensive time-series feature vectors to extract non-linear features. The output layer is used to convert the output result of the fully connected layer into the final risk probability.

[0023] For the training of the soil risk prediction model, the specific operations are as follows: Obtain a number of soil risk prediction training samples with marked risk probabilities. Each soil risk prediction training sample contains a section of soil ontology data sorted by time, predicted environmental variable data sorted by time, and the contribution coefficients of each predicted environmental variable data; divide all the obtained soil risk prediction training samples into a training set and a validation set, input the training set into the soil risk prediction model with initialized parameters for training, and use the validation set to verify the soil risk prediction model to obtain a verification result. Set training conditions and determine whether the verification result meets the training conditions. If so, output the trained soil risk prediction model; if not, continue to train the soil risk prediction model using the training set.

[0024] Embodiment 2, a soil environmental management system based on data analysis, as Figure 1 shown, includes: A data monitoring module, used to collect soil ontology data and environmental variable data of each local management area at the current monitoring time point. A soil property differentiation module, used to calculate and obtain the influence degree of the soil ontology data of each local management area by the environmental variable data respectively based on the environmental variable data and soil ontology data collected in the recent period of each local management area at the beginning of any replacement cycle. Soil quality assessment module, including an environmental suitability value calculation unit, an ideal range adjustment unit, and a quality score calculation unit; the environmental suitability value calculation unit is used to calculate and obtain the comprehensive environmental suitability value of all environmental variable data respectively by using the environmental variable data collected in each local management area in the recent period; the ideal range adjustment unit is used to calculate the ideal range of the soil body data of each local management area by using the calculated comprehensive environmental suitability value and the influence degree of the soil body data of each local management area by the environmental variable data; the quality score calculation unit is used to calculate and obtain the comprehensive soil quality score of each local management area, and judge whether the soil quality score of each local management area meets the preset compliance conditions. If so, no operation is performed; if not, soil problem treatment measures are executed. Soil risk prediction module, which is used to obtain the predicted environmental variable data of the current local management area in the future for any local management area at the beginning of any replacement cycle; based on the influence degree of the soil body data of the current local management area obtained in the current replacement cycle by the environmental variable data, calculate and obtain the contribution coefficient of each predicted environmental variable data to the current local management area respectively; construct a soil risk prediction model, and use the soil body data sorted by time, the predicted environmental variable data sorted by time, and the contribution coefficient of each predicted environmental variable data obtained in the recent period as the input of the soil risk prediction model, and output the risk probability of soil problems occurring in the current local management area in the future; if the risk probability is higher than the preset risk threshold, soil problem warning measures are executed.

[0025] It should be understood that those of ordinary skill in the art can make improvements or transformations according to the above description, and all such improvements and transformations should fall within the protection scope of the appended claims of the present invention. The parts not described in detail in this specification belong to the prior art well known to those of ordinary skill in the art.

Claims

1. A soil environment management method based on data analysis, characterized in that: include: Set up several local management areas, and collect soil entity data and environmental variable data of each local management area at the current monitoring time point. Soil entity data includes soil moisture, soil compaction, pH value, organic matter content, electrical conductivity and heavy metal content; environmental variable data includes ambient temperature, ambient humidity, light intensity, precipitation and wind speed; Set a replacement cycle. At the beginning of any replacement cycle, based on the environmental variable data and soil ontology data collected in each local management area in the recent period, calculate the influence of the environmental variable data on the soil ontology data of each local management area; At the current monitoring time point, based on the environmental variable data collected and obtained in the recent period of time in each local management area, the comprehensive environmental suitability values ​​of all environmental variable data are calculated respectively; using the calculated comprehensive environmental suitability values ​​and the degree to which the soil ontology data of each local management area is affected by the environmental variable data, the ideal range of the soil ontology data of each local management area is calculated, and then the comprehensive soil quality score of each local management area is calculated; it is judged whether the soil quality score of each local management area meets the preset compliance conditions, if so, no operation is performed; if not, soil problem treatment measures are implemented; At the beginning of any replacement cycle, for any local management area, the predicted environmental variable data of the current local management area in the future period is obtained; based on the degree of influence of the soil ontology data of the current local management area obtained in the current replacement cycle on the environmental variable data, the contribution coefficient of each predicted environmental variable data to the current local management area is calculated respectively; a soil risk prediction model is constructed, and the soil ontology data sorted by time, the predicted environmental variable data sorted by time and the contribution coefficient of each predicted environmental variable data obtained in the recent period are used as the input of the soil risk prediction model, and the risk probability of soil problems occurring in the current local management area in the future period is output; If the risk probability is higher than the preset risk threshold, soil problem early warning measures will be implemented.

2. A soil environment management method based on data analysis according to claim 1, characterized in that: Based on the environmental variable data and soil ontology data collected in each local management area in the recent period, the degree to which the soil ontology data of each local management area is affected by the environmental variable data is calculated respectively. The specific operations are as follows: For any local management area, the soil ontology data of the current local management area is recorded as , =1, 2, …, 6; to represents soil moisture, soil compactness, pH value, organic matter content, electrical conductivity and heavy metal content respectively; the environmental variable data of the current local management area is recorded as , =1, 2, …, 5; to It represents the ambient temperature, ambient humidity, light intensity, precipitation and wind speed respectively; Set the replacement cycle. At the beginning of any replacement cycle, based on the data obtained in the recent period of the current local management area, Soil ontology data And environment variable data , =1, 2, …, ; Calculate separately Soil ontology data The mean as well as Environment variable data The mean ;use Calculate and obtain the soil ontology data of the current local management area Affected by environment variable data The degree of influence .

3. A soil environment management method based on data analysis according to claim 2, characterized in that: Based on the environmental variable data collected in each local management area in the recent period, the comprehensive environmental suitability values ​​of all environmental variable data are calculated and obtained. The specific operations are as follows: For any local management area, according to the crop type of the current local management area, obtain the environmental suitability range corresponding to each environmental variable data to ; For any environment variable data ,like or , then the environment variable data Single environmental suitability value is 0; if , then use the formula Calculate and obtain a single environmental suitability value ,in, for and The middle value of , then use the formula Calculate and obtain a single environmental suitability value ; Based on the most recent Environment variable data , calculate and obtain A single environmental suitability value , then calculate A single environmental suitability value The average value of the comprehensive environmental suitability is obtained , to They correspond to ambient temperature, ambient humidity, light intensity, precipitation and wind speed respectively.

4. A soil environment management method based on data analysis according to claim 3, characterized in that: Using the calculated comprehensive environmental suitability value and the degree to which the soil ontology data of each local management area is affected by the environmental variable data, the ideal range of the soil ontology data of each local management area is calculated, and then the comprehensive soil quality score of each local management area is calculated. The specific operations are as follows: For any local management area, use the formula Calculate and obtain soil ontology data The five weight coefficients ; Then use the formula Calculate and obtain soil ontology data Soil suitability value , to They correspond to soil moisture, soil compactness, pH value, organic matter content, electrical conductivity and heavy metal content; Setting soil data The reference range is to , using the formula Calculate and obtain soil ontology data The ideal minimum value ,in, for and The middle value of is the adjustment coefficient, ; Using the formula Calculate and obtain soil ontology data The ideal maximum value ; Based on the currently acquired soil ontology data ,like or , then the soil ontology data currently obtained The corresponding quality score is 0; if , then use the formula Calculate the quality score ;like , then use the formula Calculate the quality score ; Finally, use the formula Calculate the comprehensive soil quality score for the current local management area .

5. A soil environment management method based on data analysis according to claim 4, characterized in that: Based on the degree of influence of the environmental variable data on the soil ontology data of the current local management area obtained in the current replacement cycle, the contribution coefficient of each predicted environmental variable data to the current local management area is calculated respectively. The specific operations are as follows: Based on the soil ontology data of the current local management area Affected by environment variable data The degree of influence , for any environment variable data , from the environment variable data The corresponding six levels of impact Select the value with the largest absolute value ; Using the formula Calculate the contribution coefficient of each predicted environmental variable data to the current local management area .

6. A soil environment management method based on data analysis according to claim 5, characterized in that: The soil risk prediction model is built based on LSTM, including input layer, feature extraction layer, environmental feature weighting layer, feature concatenation layer, fully connected layer and output layer; The input layer is used to receive the soil ontology data sorted by time, the predicted environmental variable data sorted by time, and the contribution coefficient of each predicted environmental variable data obtained in the recent period; The feature extraction layer includes a first LSTM layer and a second LSTM layer in parallel, wherein the first LSTM layer is used to extract soil time series feature vectors of soil ontology data sorted by time; and the second LSTM layer is used to extract environmental time series feature vectors of predicted environmental variable data sorted by time. The environmental characteristic weighted layer is used to apply the contribution coefficient of each predicted environmental variable data to the current local management area to make weighted adjustments to the environmental time series characteristic vector; The feature splicing layer is used to splice the soil time series feature vector and the weighted adjusted environment time series feature vector to obtain a comprehensive time series feature vector; The fully connected layer is used to further process the comprehensive time series feature vector and extract nonlinear features; The output layer is used to convert the output results of the fully connected layer into the final risk probability.

7. A soil environment management method based on data analysis according to claim 6, characterized in that: The specific operations for training the soil risk prediction model are as follows: A number of soil risk prediction training samples with labeled risk probabilities are obtained, each of which contains a period of soil ontology data sorted by time, predicted environmental variable data sorted by time, and contribution coefficients of each predicted environmental variable data; all the obtained soil risk prediction training samples are divided into a training set and a validation set, the training set is input into the parameter-initialized soil risk prediction model for training, and the validation set is used to validate the soil risk prediction model, the validation result is obtained, the training condition is set, and it is judged whether the validation result meets the training condition. If so, the trained soil risk prediction model is output; if not, the training set is continued to be used to train the soil risk prediction model.

8. A soil environment management system based on data analysis, characterized in that: The system is applied to a soil environment management method based on data analysis as described in any one of claims 1 to 7, comprising: The data monitoring module is used to collect soil ontology data and environmental variable data of each local management area at the current monitoring time point; The soil characteristic differentiation module is used to calculate the influence degree of the soil ontology data of each local management area on the environmental variable data and soil ontology data collected and obtained in the recent period at the beginning of any replacement cycle; The soil quality assessment module includes an environmental suitability value calculation unit, an ideal range adjustment unit and a quality score calculation unit; the environmental suitability value calculation unit is used to use the environmental variable data collected and obtained in each local management area in the recent period to calculate the comprehensive environmental suitability value of all environmental variable data respectively; the ideal range adjustment unit is used to use the calculated comprehensive environmental suitability value and the degree of influence of the soil ontology data of each local management area by the environmental variable data to calculate the ideal range of the soil ontology data of each local management area; the quality score calculation unit is used to calculate the comprehensive soil quality score of each local management area, and judge whether the soil quality score of each local management area meets the preset standard conditions. If so, no operation is performed; if not, soil problem treatment measures are implemented; The soil risk prediction module is used to obtain the predicted environmental variable data of the current local management area in the future period for any local management area at the beginning of any replacement cycle; based on the degree of influence of the environmental variable data on the soil ontology data of the current local management area obtained in the current replacement cycle, calculate and obtain the contribution coefficient of each predicted environmental variable data to the current local management area; construct a soil risk prediction model, use the soil ontology data sorted by time, the predicted environmental variable data sorted by time and the contribution coefficient of each predicted environmental variable data obtained in the recent period as the input of the soil risk prediction model, and output the risk probability of soil problems occurring in the current local management area in the future period; if the risk probability is higher than the preset risk threshold, execute soil problem early warning measures.

Images