A rice field soil detection and adjustment system and method based on multi-source sensing data analysis

Abstract

The application relates to the technical field of paddy field soil improvement, and discloses a paddy field soil detection and adjustment system and method based on multi-source sensing data analysis, which comprises a region division module, a data acquisition module, a soil health assessment module, a rice health assessment module, a comprehensive health index calculation module, a soil quality grade division module and an implementation and simulation prediction module. The application comprehensively evaluates the soil and crop health states of different regions of a paddy field, formulates corresponding soil adjustment strategies, effectively predicts the evolution trend of soil health under different climate scenarios in the future by combining a fuzzy logic relationship group and an artificial wolf swarm optimization algorithm, and outputs a clear soil health change curve, thereby providing technical support for agricultural managers.

Description

Technical Field

[0001] This application relates to the technical field of paddy field soil improvement, and in particular to a paddy field soil detection and regulation system and method based on multi-source sensor data analysis. Background Technology

[0002] Rice is a major food crop belonging to the Poaceae family. It is typically grown in paddy fields and is the staple food source for more than half of the world's population. It not only occupies an important position in Asia but is also widely cultivated in parts of Africa and the Americas. Rice requires a relatively large amount of water to grow and has certain temperature and light requirements, making it suitable for cultivation in humid and warm climates. Depending on the cultivation method, it can be divided into early rice, mid-season rice, and late rice. In terms of varieties, there are japonica rice, indica rice, and glutinous rice. Rice has high food value and plays a vital role in ensuring global food security.

[0003] Paddy field soil testing mainly uses various sensors to measure changes in the dielectric constant of the soil, and then uses built-in models and algorithms to estimate the soil content, providing scientific data support for agricultural production. This helps to rationally plan irrigation, prevent water waste caused by excessive irrigation, or affect crop growth due to insufficient irrigation, and achieve the goal of efficient and economical agricultural water use.

[0004] However, in traditional agriculture, soil conditioning strategies are usually formulated based solely on the properties of the soil itself or the appearance of rice plants. There is a lack of stratified testing of soil parameters in paddy fields, making it impossible to effectively assess the current soil health score of the area through each soil layer. Furthermore, there is no synergistic analysis with the crop growth status. At the same time, there is a lack of corresponding regulatory mechanisms when facing the potential impact of climate change on soil health, and it is also difficult to combine the complex relationship between soil, crops and the environment, which brings certain challenges to agricultural managers. Summary of the Invention

[0005] To address the problem that soil conditioning strategies are often formulated based solely on soil properties or the appearance of rice plants, lacking a synergistic analysis of soil and crop conditions, this application provides a paddy field soil detection and conditioning system and method based on multi-source sensor data analysis.

[0006] In a first aspect, this application provides a paddy field soil detection and regulation system based on multi-source sensor data analysis, the paddy field soil detection and regulation system comprising:

[0007] The regional division module divides the paddy field soil into multiple regions and deploys sensing and monitoring equipment in each region;

[0008] The data acquisition module uses the sensor monitoring device to collect soil parameters of paddy fields at different depths in each area, and acquires image data including rice leaves, rice roots and stems, and rice plants; the different depths are divided into topsoil, subsoil and subsoil.

[0009] The soil health assessment module analyzes the soil health status scores of the topsoil, subsoil, and subsoil layers based on the screened paddy field soil parameters, and obtains the soil health score for each region based on the different depths of the paddy field gley layer distribution in the current region.

[0010] The comprehensive health index calculation module comprehensively assesses the comprehensive health index of paddy field soil in each region based on the health status score and the rice health status score.

[0011] The soil quality grading module classifies the quality of paddy field soil according to the comprehensive health index, dividing the quality into healthy growth, light pollution, moderate pollution and heavy pollution, and formulates soil conditioning strategies for paddy field soil in each region.

[0012] The implementation and simulation prediction module implements soil regulation strategies, simulates the changing trends of paddy field soil health under different climatic conditions, predicts the impact of future climate change on paddy field soil, and builds a visualization dashboard.

[0013] Alternatively, the expression for the soil health score is:

[0014]

[0015] In the formula, k represents the soil parameter;

[0016] K represents the total number of soil parameters;

[0017] This represents the weight of soil parameter k;

[0018] This represents the health score of soil parameter k in the Lth soil layer;

[0019] Represents the soil parameter k in the Lth soil layer;

[0020] This represents the mean value of soil parameter k in the L-th soil layer;

[0021] Let represent the standard deviation of soil parameter k in the L-th soil layer.

[0022] Optionally, the rice health assessment module includes:

[0023] The leaf health scoring module analyzes the color, shape, and lesions of rice leaves using image data to assess the health status of the leaves and obtain a leaf health score N.

[0024] The root health scoring module analyzes the length, distribution, and density of the root system using image data of rice roots and stems to assess the development of the rice root system and obtain a root health score M.

[0025] The plant growth scoring module measures the height of rice plants based on image data, analyzes the growth status of rice plants, and obtains a plant growth score Q.

[0026] The comprehensive health status scoring module integrates leaf health score, root health score, and plant growth score, assigns weights according to the importance of preset indicators, and obtains the rice health status score for each region.

[0027] Optionally, by comprehensively considering leaf health score, root health score, and plant growth score, and assigning weights according to the importance of preset indicators, the rice health status score for each region is obtained, including:

[0028] The scores for leaf health (N), root health (M), and plant height (Q) were obtained, with each score ranging from 0 to 100.

[0029] The ranges of leaf health score N, root health score M, and plant height score Q are scaled to the range of 0-1 to obtain leaf health score N′, root health score M′, and plant height score Q′.

[0030] Based on the importance of the preset indicators, the weights F of leaf health score N′, root health score M′, and plant height score Q′ are respectively assigned. N′ F M′ and F Q′ And substitute it into the following formula:

[0031]

[0032] The rice health status score was calculated. ;

[0033] In the formula, N′ represents the leaf health score; M′ represents the root health score; Q′ represents the plant height score; F N′ Indicates the weight of the leaf health score; F M′ Indicates the weight of the root health score; The weighting of the plant height score; This represents the exponential penalty function.

[0034] Optionally, based on the health status score and the rice health status score, a comprehensive health index of paddy field soil in each region is assessed, including:

[0035] Based on the different impacts of paddy soil and rice health on agricultural production, weights are assigned to soil health scores and rice health status scores, and then substituted into the following formula:

[0036]

[0037] The comprehensive health index of paddy field soil in each region was calculated. ;

[0038] In the formula, This indicates the health score of the L layer soil, where the paddy field's gley layer is located. This represents the weight of the health score in the Lth soil layer; The score represents the health status of the rice. The weights representing the health status scores of rice. This represents the exponential penalty function.

[0039] Optionally, the quality of paddy field soil can be graded according to a comprehensive health index, categorizing it into healthy growth, slightly polluted, moderately polluted, and heavily polluted, including:

[0040] Paddy field soil quality is divided into four levels based on a comprehensive health index;

[0041] The scores are as follows: Level 1: 80-100 points; Level 2: 60-80 points; Level 3: 30-60 points; Level 4: 0-30 points.

[0042] When the comprehensive health index When the soil quality is within the first level range, it is considered to be healthy for growth, indicating that the soil conditions and the rice growing environment are suitable.

[0043] When the comprehensive health index Within the second level, the soil quality is judged as slightly polluted, indicating that the soil has a slight nutrient imbalance or minor pollution.

[0044] When the comprehensive health index When the soil quality falls within the third level, it is classified as moderately polluted, indicating that the soil has begun to suffer serious damage.

[0045] When the comprehensive health index When the soil quality falls within the fourth level, it is classified as severely polluted, indicating that the soil health has been seriously degraded.

[0046] Optionally, soil conditioning strategies can be implemented to simulate the changing trends of paddy field soil health under different climatic conditions and predict the impacts of future climate change on paddy field soils, including:

[0047] The established soil conditioning strategies are matched with historical climate factors in different regions, the soil health indicators to be predicted are selected, and the domain of the target variables is determined based on the historical value range.

[0048] The universe of discourse of the target variable is divided into several fuzzy intervals, and the parameters of the wolf pack optimization algorithm are set. By randomly initializing the positions of artificial wolves in the universe of discourse, the corresponding fuzzy partitioning scheme is generated.

[0049] The artificial wolf constructs fuzzy logical relation groups based on a fuzzy partitioning scheme, evaluates fitness indicators using multi-source agricultural time-series data, and updates the fuzzy logical relation groups.

[0050] By inputting different future climate paths and combining them with soil regulation strategies for each region, the system uses updated fuzzy logic relation sets to predict the future trend of soil health changes in each region and outputs soil health change curves under each climate scenario.

[0051] Optionally, the artificial wolf constructs fuzzy logical relation groups based on a fuzzy partitioning scheme, and uses agricultural multi-source time-series data to evaluate fitness indicators, including:

[0052] Each artificial wolf, according to the corresponding fuzzy partitioning scheme, divides the target variable into a corresponding fuzzy set;

[0053] A fuzzy logic relation group containing fuzzy logic relations is constructed based on fuzzy sets, and historical agricultural multi-source time series data is input into the model for prediction to obtain a sequence of predicted values.

[0054] Calculate the root mean square error between the predicted value and the true value, and use the root mean square error as the fitness index of the fuzzy partitioning scheme.

[0055] The optimal artificial wolf is selected as the alpha wolf based on the fitness index, and the fuzzy logic relation group is updated based on the alpha wolf.

[0056] Optionally, different future climate paths are input, and combined with soil regulation strategies for each region, the updated fuzzy logic relation set is used to predict the soil health change trend of each region in the future, outputting soil health change curves under each climate scenario, including:

[0057] Input multiple future climate path scenarios to generate corresponding future climate time series data for each region;

[0058] By combining established soil regulation strategies in each region with future climate time series data, a joint input sequence for future environment and management can be constructed.

[0059] The optimal fuzzy partitioning scheme determined by the alpha wolf and the updated fuzzy logic relation group model are used to predict the joint input sequence of future environment and management, and generate the predicted value sequence of soil health indicators.

[0060] Based on the predicted value sequence, the soil health evolution trend of each region under different climate scenarios is output, generating a change curve in the time dimension.

[0061] Secondly, a method for detecting and regulating paddy field soil based on multi-source sensor data analysis is characterized in that the method is implemented based on the aforementioned paddy field soil detection and regulation system based on multi-source sensor data analysis.

[0062] In summary, this application includes at least one of the following beneficial technical effects: By integrating soil sensor data, rice image information, and fuzzy logic modeling technology, this application divides paddy field soil into layers, obtains soil health scores for each region based on core evaluation indicators in different layers, comprehensively assesses the soil and crop health status in different areas of the paddy field, and formulates corresponding soil regulation strategies accordingly. At the same time, by combining fuzzy logic relation groups and artificial wolf pack optimization algorithms, it effectively predicts the evolution trend of soil health under different future climate scenarios and outputs clear soil health change curves, enhancing the reliability of prediction and providing technical support for agricultural managers. Attached Figure Description

[0063] Figure 1 This is a system principle block diagram of this application. Detailed Implementation

[0064] The embodiments of this application are described in detail below, and examples of the embodiments are shown in the accompanying drawings.

[0065] In the description of this specification, the references to "certain embodiments," "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples" refer to specific features, structures, materials, or characteristics described in connection with the described embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0066] This application discloses a paddy field soil detection and regulation system based on multi-source sensor data analysis, referring to... Figure 1 The paddy field soil testing and conditioning system includes:

[0067] The regional division module divides the paddy field soil into multiple regions and deploys sensing and monitoring equipment in each region.

[0068] The data acquisition module uses the sensor monitoring device to collect soil parameters of paddy fields at different depths in each area, and acquires image data including rice leaves, rice roots and stems and rice plants; the different depths are divided into topsoil layer, subsoil layer and subsoil layer.

[0069] It should be explained that in a certain provincial agricultural science and technology experimental station, the research team divided the experimental field into 12 sub-regions, each with an area of ​​approximately 20m × 20m. The division was based on factors such as terrain elevation, irrigation and drainage conditions, and historical fertilization intensity to ensure that each sub-region has representative differences.

[0070] Multiple soil sensing nodes were deployed in each area, installed in the topsoil (0–10 cm), subsoil (10–20 cm), and subsoil (20–40 cm) layers to monitor in real time: soil temperature, soil moisture content, electrical conductivity, redox potential, and pH value.

[0071] Image data is collected periodically via a ground-based mobile platform, specifically including:

[0072] 1. Rice leaf images: used to identify nitrogen deficiency, changes in leaf color index (such as SPAD value estimation), and early symptoms of pests and diseases;

[0073] 2. Root and stem images: Root density and vitality images are collected through ground-level windows or root observation tubes;

[0074] 3. Whole rice plant image: Used for three-dimensional reconstruction of plant structure, and to help estimate growth traits such as biomass, canopy height, and canopy density.

[0075] The soil health assessment module analyzes the soil health status scores of the topsoil, subsoil, and subsoil layers based on the screened paddy field soil parameters, and obtains the soil health score for each region based on the different depths of the paddy field gley layer distribution in the current region.

[0076] Preferably, the expression for the soil health score is:

[0077]

[0078] In the formula, k represents the soil parameter;

[0079] K represents the total number of soil parameters;

[0080] This represents the weight of soil parameter k;

[0081] This represents the health score of soil parameter k in the Lth soil layer;

[0082] Represents the soil parameter k in the Lth soil layer;

[0083] This represents the mean value of soil parameter k in the L-th soil layer;

[0084] Let represent the standard deviation of soil parameter k in the L-th soil layer.

[0085] It should be explained that the purpose of the soil health assessment module is to transform the collected multi-parameter, multi-depth soil data into quantifiable and comparable soil health scores, thereby supporting agricultural managers in making scientific judgments about the condition of arable land; specifically, it mainly performs the following tasks:

[0086] 1. Stratified analysis: Based on soil monitoring data at different depths (such as topsoil, subsoil, and subsoil) within the region, assess the soil health status of each layer.

[0087] 2. Multi-parameter fusion: Standardize and weight multiple key parameters (such as organic matter, moisture, electrical conductivity, pH, etc.) in each soil layer to calculate the soil health score of that layer;

[0088] The paddy field gley layer is a soil-forming layer formed through alternating redox reactions under conditions of periodic rise and fall of groundwater or seasonal waterlogging. This soil layer is characterized by the migration and deposition of iron and manganese substances, forming new growths such as rust streaks, rust spots, iron and manganese nodules, or eel blood spots. The formation and function of the gley layer in soil pores are crucial for soil moisture regulation, nutrient supply, soil health, and crop growth. By regulating moisture, promoting the migration and deposition of iron and manganese substances, improving soil structure, and enhancing microbial activity, it can effectively support plant root growth and soil ecological balance. Therefore, screening the paddy field gley layer is a core evaluation indicator for soil quality in the current region.

[0089] Based on the location of the paddy field gley layer, if the paddy field gley layer is mainly located in the current division area and is mainly distributed in the subsoil layer, then the soil health status score of the subsoil layer is used.

[0090] If the paddy field gley layer is mainly distributed in the subsoil layer in the current division area, the soil health status score of the current area is calculated using the soil health status score of the subsoil layer.

[0091] If the paddy field gley layer is mainly distributed in the subsoil layer in the current division area, the soil health status score of the current area is calculated using the soil health status score of the subsoil layer.

[0092] Similarly, based on the location of the main distribution of the paddy field's gley layer, the current soil health score of the area, and the depth of the topsoil, subsoil, and subsoil layers, adjustments can be made according to the actual situation. This is a conventional technique and will not be explained further here.

[0093] Specifically, in soil layer L, organic matter, moisture, electrical conductivity, and pH were weighted at ratios of 0.2, 0.4, 0.2, and 0.2, respectively. These parameters were then standardized and mapped to the [0, 1] region. Soil health scores were then used to determine the optimal parameters. The calculation formula is used to obtain the health scores of organic matter, moisture, electrical conductivity and pH respectively. Then, the health scores of organic matter, moisture, electrical conductivity and pH are added together to obtain the soil health score of layer L in paddy field gley layer.

[0094] 3. Regional aggregation: The soil health scores of each layer are weighted according to their ecological importance (for example, the weight of the surface layer is 0.3, the weight of the middle layer is 0.5, and the weight of the deep layer is 0.2). The overall soil health score of the region is generated by multiplying the soil health scores of each layer by the corresponding weight ratio.

[0095] By quantitatively evaluating the soil health status of different soil layers, regions, and time periods, risk warnings can be achieved.

[0096] The rice health assessment module analyzes the health of rice leaves, the development of rice roots, and the height of the plant based on image data to obtain a rice health status score for each region.

[0097] Preferably, the rice health assessment module includes:

[0098] The leaf health scoring module analyzes the color, shape, and lesions of rice leaves using image data to assess the health status of the leaves and obtain a leaf health score N.

[0099] The root health scoring module analyzes the length, distribution, and density of the root system using image data of rice roots and stems to assess the development of the rice root system and obtain a root health score M.

[0100] The plant growth scoring module measures the height of rice plants based on image data, analyzes the growth status of rice plants, and obtains a plant growth score Q.

[0101] It should be noted that observation tubes or scanning imaging devices were set up at multiple paddy field locations; leaf images, root and stem images, and plant images were collected every few days. Leaf parameters such as leaf area, leaf color, leaf thickness, degree of damage, and leaf moisture content were extracted through the image recognition module; root and stem parameters such as taproot length, lateral root length, root density, and root and stem moisture were extracted; and plant parameters such as leaf area, plant nutrients, root and stem growth rate, and plant moisture were extracted.

[0102] All parameters below are standardized, that is, the standardized value is the current parameter minus the mean of the current parameter and divided by the standard deviation of the current parameter;

[0103] Among them, the leaf health score N=a1 Leaf area (standardized value) + a2 Leaf color (standardized value) + a3 Blade thickness (standardized value) + a4 Damage level (standardized value) + a5 Leaf moisture content (standardized value);

[0104] Root health score M=b1 Principal root length (standardized value) + b2 Lateral root length (standardized value) + b3 Root density (standardized value) + b4 Root hair vitality index (standardized value);

[0105] Plant growth score Q=c1 Leaf area (standardized value) + c2 Plant nutrients (standardized values) + c3 Rhizome growth rate (standardized value) + c4 Plant moisture (standardized value);

[0106] In the formula, a1, a2, a3, a4 and a5 are the weights of leaf area, leaf color, leaf thickness, degree of damage and leaf moisture content, respectively.

[0107] b1, b2, b3, and b4 are the weights of taproot length, lateral root length, root density, and root and stem moisture index, respectively.

[0108] c1, c2, c3, and c4 are the weights of leaf area, plant nutrients, root and stem growth rate, and plant water, respectively.

[0109] In the process of scoring plant growth, the root and stem growth rate is obtained from the length of the main root, the length of the lateral root, and the root density, while the plant moisture is obtained from the moisture content of the root and stem and the moisture content of the leaves. It can be concluded that the leaf health score will affect the plant growth score, and similarly, the root health score will also affect the plant growth score.

[0110] Taking a paddy field in a certain area as an example:

[0111] The leaf area is 8.5, the leaf color is 7.5, the leaf thickness is 6.5, the degree of damage is 4.5, and the leaf moisture content is 8.

[0112] a1, a2, a3, a4 and a5 are 0.3, 0.2, 0.1, 0.15 and 0.25 respectively. Substituting them into the formula for calculating the leaf health score N, the leaf health score N is found to be 7.375.

[0113] The taproot length is 8, the lateral root length is 6, the root density is 5, and the root hair vitality index is 6.

[0114] b1, b2, b3 and b4 are 0.4, 0.25, 0.15 and 0.2 respectively. Substituting them into the root health score N calculation formula, the root health score N is 6.65.

[0115] Leaf area is 8.5, plant nutrients are 10, root and stem growth rate is 19 (including main root length of 8, lateral root length of 6, and root density of 5), and plant water content is 20.

[0116] c1, c2, c3 and c4 are 0.1, 0.2, 0.4 and 0.3 respectively. Substituting these values ​​into the plant growth score Q, we get a plant growth score Q of 11.65.

[0117] The comprehensive health status scoring module integrates leaf health score, root health score, and plant growth score, assigns weights according to the importance of preset indicators, and obtains the rice health status score for each region.

[0118] The comprehensive health index calculation module comprehensively assesses the comprehensive health index of paddy field soil in each region based on the health status score and the rice health status score.

[0119] Preferably, the specific steps for obtaining the rice health status score for each region by comprehensively considering leaf health score, root health score, and plant growth score, and assigning weights according to the importance of preset indicators, are as follows:

[0120] The scores for leaf health (N), root health (M), and plant height (Q) were obtained, with each score ranging from 0 to 100.

[0121] The ranges of leaf health score N, root health score M, and plant height score Q are scaled to the range of 0-1 to obtain leaf health score N′, root health score M′, and plant height score Q′.

[0122] Based on the importance of the preset indicators, the weights F of leaf health score N′, root health score M′, and plant height score Q′ are respectively assigned. N′ F M′ and F Q′ And substitute it into the following formula:

[0123]

[0124] The rice health status score was calculated. ;

[0125] In the formula, N′ represents the leaf health score; M′ represents the root health score; Q′ represents the plant height score; F N′ Indicates the weight of the leaf health score; F M′ Indicates the weight of the root health score; The weighting of the plant height score; This represents the exponential penalty function.

[0126] It should be explained that the rice image data of each region was acquired and intelligently identified to extract three types of health scoring indicators: leaf health score N (e.g., 85 points), root health score M (e.g., 78 points), and plant height score Q (e.g., 92 points).

[0127] The initial rating range is 0–100, which is a uniform dimension, and it is then standardized.

[0128] =N / 100=0.85, M =M / 100=0.78, Q =Q / 100=0.92;

[0129] Subsequently, the weights of each score item are set according to the importance of the current growth stage of rice (which can be dynamically adjusted): F N′ =0.4, F M′ =0.3 and =0.3, the rice health status score is calculated. The rice health status score was obtained through calculation. Approximately 86.4;

[0130] Preferably, the specific steps for comprehensively evaluating the overall health index of paddy field soil in each region based on the health status score and the rice health status score are as follows:

[0131] Based on the different impacts of paddy soil and rice health on agricultural production, weights are assigned to soil health scores and rice health status scores, and then substituted into the following formula:

[0132]

[0133] The comprehensive health index of paddy field soil in each region was calculated. ;

[0134] In the formula, This indicates the health score of the L layer soil, where the paddy field's gley layer is located. This represents the weight of the health score in the Lth soil layer; The score represents the health status of the rice. The weights representing the health status scores of rice. This represents the exponential penalty function.

[0135] The obtained rice health status score Health score of the L layer soil where the paddy field gley layer is located Based on the different impacts of paddy soil and rice health on agricultural production, weights were set for soil health score and rice health status score (e.g., soil health score 0.5 and comprehensive health index 0.5), and the comprehensive health index of paddy soil in each region was calculated. If calculated If the value is approximately 88, the soil quality is considered healthy for growth. It should be noted that the main purpose of introducing the exponential penalty function is to penalize scores that deviate from the expected value, so that the impact of excessively high or low outliers on the final health score is gradually reduced, avoiding an excessive impact on the overall score, and resulting in a more stable comprehensive health index.

[0136] The soil quality grading module classifies the quality of paddy field soil according to the comprehensive health index, dividing the quality into healthy growth, slightly polluted, moderately polluted, and heavily polluted, and formulates soil conditioning strategies for paddy field soil in each region.

[0137] Preferably, the specific steps for classifying the quality of paddy field soil into grades based on a comprehensive health index, categorizing the quality into healthy growth, slightly polluted, moderately polluted, and heavily polluted, are as follows:

[0138] Paddy field soil quality is divided into four levels based on a comprehensive health index;

[0139] The scores are as follows: Level 1: 80-100 points; Level 2: 60-80 points; Level 3: 30-60 points; Level 4: 0-30 points.

[0140] When the comprehensive health index When the soil quality is within the first level range, it is considered to be healthy for growth, indicating that the soil conditions and the rice growing environment are suitable.

[0141] When the comprehensive health index Within the second level, the soil quality is judged as slightly polluted, indicating that the soil has a slight nutrient imbalance or minor pollution.

[0142] When the comprehensive health index When the soil quality falls within the third level, it is classified as moderately polluted, indicating that the soil has begun to suffer serious damage.

[0143] When the comprehensive health index When the soil quality falls within the fourth level, it is classified as severely polluted, indicating that the soil health has been seriously degraded.

[0144] It should be noted that the specific soil conditioning strategies are shown in the table below:

[0145] Soil Conditioning Strategy Formulation Table

[0146] Health Level score range Main problems core of adjustment strategy Management intensity First level 80–100 Normal, virtuous cycle Fine-tuning + prevention Low First level 60–80 Nutritional imbalance or acid-base imbalance Adjust pH, replenish organic matter and microbial activators middle Level 3 30–60 Compacted soil, insufficient organic matter, poor root system Multiple agents applied simultaneously + deep soil improvement high Fourth level 0–30 Severe pollution, heavy metals / pesticide residues Suspension of farming + closed restoration + joint governance Extremely high

[0147] The first level indicates that the soil is in an ideal state; the soil structure is good, the organic matter content is appropriate, and the core indicators such as moisture, pH, and electrical conductivity are stable. Crops grow vigorously in this environment without obvious growth inhibition or disease. Regulation strategies include: adopting maintenance fertilization strategies, such as using slow-release fertilizers to prevent nutrient loss; regularly adding biological agents to enhance soil microbial activity; planting green manure crops for rotation, such as milkvetch and alfalfa, to maintain soil fertility; and limiting the frequency of tillage to protect soil aggregate structure and surface microhabitat.

[0148] The second level indicates that the soil has initial problems, such as acid-base imbalance, local nutrient imbalance, or slight structural degradation; it manifests as yellowish rice leaves and decreased root vitality, but is generally still reversible; adjustment strategies include: using acid-base regulators such as lime or sulfur according to the actual pH conditions; supplementing organic matter and trace elements, and improving buffering capacity through organic fertilizers or humic acid; carrying out appropriate deep loosening operations to improve aeration and permeability; and scientifically controlling irrigation methods and optimizing water management.

[0149] Level 3 indicates that the soil has been significantly damaged; soil compaction, decreased organic matter, salt accumulation, and root obstruction may occur, seriously affecting the normal growth and development of crops. Remedial strategies include: applying large amounts of well-rotted organic fertilizer or farmyard manure to rebuild the soil organic matter pool; using multifunctional conditioners, such as biochar, humic acid, and silicon-calcium-magnesium amendments; conducting deep tillage or breaking up the plow pan to clear the rhizosphere; targeted supplementation of micronutrients, such as iron, zinc, and boron; and implementing zoned field management for precise restoration.

[0150] Level 4 indicates that the soil has been severely degraded or polluted, possibly due to destructive factors such as heavy metals, pesticide residues, and salinity, leading to crop death or extreme stunting, and is no longer suitable for cultivation. Remedial strategies include: suspending cultivation and implementing closed-loop management of the fields to prevent the spread of pollution; implementing soil replacement or dilution remediation, topsoil burial, and adding absorbent materials (such as bentonite); using composite materials such as heavy metal fixatives, biochar, and humic acid to seal or passivate pollutants; and promoting ecological restoration methods, such as planting aquatic purification plants (such as water hyacinth) for phytoremediation.

[0151] Assume the comprehensive health index calculated for a certain region is Approximately 76 points, indicating the soil quality is deemed healthy for growth.

[0152] According to the grading standards, this score is between 60 and 80, classifying it as Level 2: Light Pollution. Therefore, the area will be automatically treated as follows:

[0153] The pollution level is marked as "lightly polluted";

[0154] Assign a risk label: "Yellow Card Warning";

[0155] Recommended management suggestions: Apply conditioners appropriately, optimize water layer management, and enhance microbial activity.

[0156] The implementation and simulation prediction module implements soil regulation strategies, simulates the changing trends of paddy field soil health under different climatic conditions, predicts the impact of future climate change on paddy field soil, and builds a visualization dashboard.

[0157] Preferably, the specific steps for implementing soil conditioning strategies, simulating the changing trends of paddy field soil health under different climatic conditions, and predicting the impact of future climate change on paddy field soil are as follows:

[0158] The established soil conditioning strategies are matched with historical climate factors in different regions, the soil health indicators to be predicted are selected, and the domain of the target variable is determined based on the historical value range.

[0159] It should be explained that the following data were collected from the target area over the past 5–10 years: average annual temperature, maximum / minimum temperature, annual and monthly precipitation distribution, sunshine duration and photosynthetically active radiation, humidity and evapotranspiration, and records of meteorological disasters (such as floods and droughts); the effectiveness of the same regulation strategy (such as the application of humic acid) was compared in specific climatic years.

[0160] If the regulating effect decreases in high-temperature years, it indicates that the regulator is sensitive to temperature; if the heavy metal concentration decreases more significantly in rainy years, it indicates that moisture facilitates metal migration / fixation. A sensitivity correspondence table between different strategies and climate factors is constructed to assist in strategy adjustments under different climate conditions in the future.

[0161] The following key soil health indicators are preferred as prediction targets: soil organic matter content, soil pH, electrical conductivity, soluble heavy metal content, and soil microbial diversity index (optional). Based on existing historical monitoring data of soil samples, the prediction range (universe of discourse) for each target variable is reasonably set, as shown in the table below:

[0162] index Target variable type Theoretical domain range illustrate Organic matter content Continuous variables 0.5–5.0 Normal arable land should have a soil moisture content of ≥1.0%, and high-quality arable land should have a soil moisture content of ≥2.5%. pH value Continuous variables 4.5–8.5 The suitable depth for paddy fields is 5.5–6.8 meters. electrical conductivity Continuous variables 0.1–5.0 A value above 2.0 is considered moderate to severe salinization. Soluble heavy metal content Continuous variables 0–2.0 The safety threshold for arable land is ≤0.3 (general). Soil microbial diversity index Continuous variables 1.0–3.5 Approaching 3 indicates a healthy state of biodiversity.

[0163] Suppose a regulation strategy of heavy application of organic fertilizer + green manure rotation is implemented in a certain area, and we want to predict the organic matter content level of the area in the coming year;

[0164] Historical climate factors show that last winter and spring were characterized by low temperatures and little sunshine, resulting in a slow increase in organic matter content.

[0165] If this year is predicted to be a warm and humid year, the expected regulation effect will be even better;

[0166] Set the target domain of prediction as If the output organic matter content is 2.75%, it meets the health standards.

[0167] If the predicted value is below 1.0%, an adjustment to increase the intensity of the investment will be automatically triggered.

[0168] The universe of discourse of the target variable is divided into several fuzzy intervals, and the parameters of the wolf pack optimization algorithm are set. By randomly initializing the positions of artificial wolves in the universe of discourse, the corresponding fuzzy partitioning scheme is generated.

[0169] The artificial wolf constructs fuzzy logical relation groups based on a fuzzy partitioning scheme, evaluates fitness indicators using multi-source agricultural time-series data, and updates the fuzzy logical relation groups.

[0170] Preferably, the specific implementation steps for constructing fuzzy logical relation groups based on a fuzzy partitioning scheme and evaluating fitness indicators using multi-source agricultural time-series data are as follows:

[0171] Each artificial wolf, according to the corresponding fuzzy partitioning scheme, divides the target variable into a corresponding fuzzy set;

[0172] A fuzzy logic relation group containing fuzzy logic relations is constructed based on fuzzy sets, and historical agricultural multi-source time series data is input into the model for prediction to obtain a sequence of predicted values.

[0173] Calculate the root mean square error between the predicted value and the true value, and use the root mean square error as the fitness index of the fuzzy partitioning scheme.

[0174] The optimal artificial wolf is selected as the alpha wolf based on the fitness index, and the fuzzy logic relation group is updated based on the alpha wolf.

[0175] It needs to be explained that each artificial wolf represents a fuzzy partitioning scheme. For example, the target variable is divided into three fuzzy subsets: low [0.5, 1.5], medium [1.0, 3.0], and high [2.5, 5.0]. Based on the partitioning scheme of each artificial wolf, a corresponding fuzzy rule group is constructed to map the relationship between the input (time series features) and the target variable. Historical climate and field management data (such as once a week) are input into the fuzzy logic group corresponding to each artificial wolf to generate a set of predicted output sequences. These sequences are compared with the actual organic matter content sequences. For the predicted output of each artificial wolf, the root mean square error between its prediction and the actual value is calculated.

[0176]

[0177] In the formula, denoted as the root mean square error of the prediction model corresponding to the i-th artificial wolf; c represents the length of the time series. This represents the predicted value of the i-th artificial wolf at time t. denoted as the true value at time t; where, the smaller the root mean square error, the better the fuzzy partitioning scheme represented by the artificial wolf, the closer the constructed fuzzy logic group is to the true target trend, and the higher the fitness.

[0178] By inputting different future climate paths and combining them with soil regulation strategies for each region, the system uses updated fuzzy logic relation sets to predict the future trend of soil health changes in each region and outputs soil health change curves under each climate scenario.

[0179] Preferably, the specific implementation steps for inputting different future climate paths, combining soil regulation strategies for each region, using updated fuzzy logic relation sets to predict the future soil health change trend of each region, and outputting soil health change curves under each climate scenario are as follows:

[0180] Input multiple future climate path scenarios to generate corresponding future climate time series data for each region;

[0181] By combining established soil regulation strategies in each region with future climate time series data, a joint input sequence for future environment and management can be constructed.

[0182] The optimal fuzzy partitioning scheme determined by the alpha wolf and the updated fuzzy logic relation group model are used to predict the joint input sequence of future environment and management, and generate the predicted value sequence of soil health indicators.

[0183] Based on the predicted value sequence, the soil health evolution trend of each region under different climate scenarios is output, generating a change curve in the time dimension.

[0184] It should be explained that future climate path scenarios are usually generated based on different greenhouse gas emission scenarios and climate model predictions; for example, the following climate scenarios can be selected: low emission scenario: greenhouse gas emissions tend to stabilize and the degree of climate warming is low; medium emission scenario: greenhouse gas emissions increase moderately and the degree of climate warming is moderate; high emission scenario (RCP8.5): greenhouse gas emissions continue to increase and the degree of climate warming is high.

[0185] Each scenario will generate climate data for the next few years, including: temperature (average annual temperature, extreme temperature), precipitation (annual precipitation, precipitation distribution), sunshine duration (photosynthetically active radiation), and humidity, evapotranspiration, etc. (which affect soil evaporation and water management).

[0186] Different regions employ different soil conditioning strategies. For example, in Region A, increasing the application of organic fertilizers and microbial agents helps regulate soil pH; in Region B, green fertilizers and optimized irrigation strategies help regulate soil salinity; and in Region C, deep tillage and cover crops are used to restore soil structure.

[0187] Using predicted soil health indicators (such as organic matter content, pH, and electrical conductivity), time-varying curves are generated, and the soil health evolution trend of each region under different climatic scenarios is output. Specifically, this is manifested as follows:

[0188] X-axis: time (years); Y-axis: soil health indicators (such as organic matter content, pH value, etc.); multiple curves represent the soil health evolution trend under different climate scenarios (low emissions, medium emissions, high emissions).

[0189] For example, under low-emission scenarios, soil health remains stable with moderate climate change, and may even increase slightly;

[0190] Under high-emission scenarios, rising temperatures and reduced precipitation may lead to the loss of soil organic matter and a decline in soil health index;

[0191] Under the medium-emission scenario, soil health may fluctuate and be significantly affected by climate change.

[0192] Plot the predicted results under all climate scenarios as soil health change curves. For example, in region A, soil health steadily increases under the low emission scenario; soil health slightly decreases under the medium emission scenario; and soil health significantly decreases under the high emission scenario.

[0193] By introducing multiple future climate path scenarios and soil regulation strategies for different regions, and combining fuzzy logic relation groups and artificial wolf pack optimization algorithms, the evolution trend of soil health under different future climate scenarios is effectively predicted, and clear soil health change curves are output, which enhances the reliability of the prediction and provides technical support for agricultural managers.

[0194] This application also discloses a method for detecting and regulating paddy soil based on multi-source sensor data analysis, characterized in that the method is implemented based on the aforementioned paddy soil detection and regulation system based on multi-source sensor data analysis.

[0195] It should be noted that the calculation formulas and all parameters involved in the calculations in this application have been dimensionless beforehand. The process of dimensionless processing is well known in the industry and will not be described here.

[0196] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A paddy soil detection and regulation system based on multi-source sensor data analysis, characterized in that, The paddy field soil testing and conditioning system includes: The regional division module divides the paddy field soil into multiple regions and deploys sensing and monitoring equipment in each region; The data acquisition module uses the sensor monitoring device to collect soil parameters of paddy fields at different depths in each area, and acquires image data including rice leaves, rice roots and stems, and rice plants; the different depths are divided into topsoil, subsoil and subsoil. The soil health assessment module analyzes the soil health status scores of the topsoil, subsoil, and subsoil layers based on the screened paddy field soil parameters, and obtains the soil health score for each region based on the different depths of the paddy field gley layer distribution in the current region. The rice health assessment module analyzes the health of rice leaves, the development of rice roots, and the height growth of the plant based on the image data, and obtains a rice health status score for each region. The comprehensive health index calculation module comprehensively evaluates the comprehensive health index of paddy field soil in each region based on the health status score and the rice health status score. The soil quality grading module classifies the quality of paddy field soil according to the comprehensive health index, classifying the quality into healthy growth, light pollution, moderate pollution and heavy pollution, and formulates soil conditioning strategies for paddy field soil in each region. The implementation and simulation prediction module implements the soil regulation strategy, simulates the changing trend of paddy field soil health under different climatic conditions, predicts the impact of future climate change on paddy field soil, and builds a visualization dashboard. The expression for the soil health score is: In the formula, k Indicates soil parameters; K This represents the total number of soil parameters; Representing soil parameters k The weights; Indicates the first L Soil parameters in the soil layer k Health score; Indicates the first L Soil parameters in the soil layer k ; Indicates the first L Soil parameters in the soil layer k The mean; Indicates the first L Soil parameters in the soil layer k Standard deviation; The rice health assessment module includes: The leaf health scoring module analyzes rice leaf image data, examining leaf color, shape, and lesions to assess leaf health and generate a leaf health score. N ; The root health scoring module analyzes the length, distribution, and density of the roots using image data of rice roots and stems to assess the development of the rice root system and obtain a root health score. M ; The plant growth scoring module measures the height of rice plants based on image data, analyzes the growth status of the rice plants, and obtains a plant growth score. Q ; The comprehensive health status scoring module is used to integrate the leaf health score, the root health score and the plant growth score, assign weights according to the importance of preset indicators, and obtain the rice health status score for each region. By combining the leaf health score, the root health score, and the plant growth score, and assigning weights according to the importance of preset indicators, the rice health status score for each region is obtained, including: Obtain leaf health scores respectively N Root health score M Score for plant height Q And all scores range from 0 to 100; Score of leaf health N Root health score M Score for plant height Q The range is scaled down to 0-1 to obtain the leaf health score. N Root health score M ′ and plant height score Q ′; Leaf health scores were assigned based on the importance of the preset indicators. N Root health score M ′ and plant height score Q The weight F' N′ F M′ and F Q′ And substitute it into the following formula: The rice health status score was calculated. ; In the formula, N ′ indicates leaf health score; M ′ indicates the root health score; Q ′ indicates plant height score; F N′ Indicates the weight of the leaf health score; F M′ Indicates the weight of the root health score; The weighting of the plant height score; This represents the exponential penalty function; Based on the health status score and the rice health status score, the comprehensive health index of paddy field soil in each region is comprehensively evaluated, including: Based on the different impacts of paddy soil and rice health on agricultural production, weights are assigned to soil health scores and rice health status scores, and then substituted into the following formula: The comprehensive health index of paddy field soil in each region was calculated. ; In the formula, Indicates the location of the gley layer in the paddy field L Health score of the soil layer; Indicates the first L The weighting of the health score in the soil layer; The score represents the health status of the rice. The weights representing the health status scores of rice. This represents the exponential penalty function; Implementing the aforementioned soil conditioning strategy, simulating the changing trends of paddy field soil health under different climatic conditions, and predicting the impacts of future climate change on paddy field soils include: The established soil conditioning strategies are matched with historical climate factors in different regions, the soil health indicators to be predicted are selected, and the domain of the target variables is determined based on the historical value range. The universe of discourse of the target variable is divided into several fuzzy intervals, and the parameters of the wolf pack optimization algorithm are set. By randomly initializing the positions of artificial wolves in the universe of discourse, the corresponding fuzzy partitioning scheme is generated. The artificial wolf constructs fuzzy logical relation groups based on a fuzzy partitioning scheme, evaluates fitness indicators using multi-source agricultural time-series data, and updates the fuzzy logical relation groups. By inputting different future climate paths and combining them with soil regulation strategies for each region, the system uses updated fuzzy logic relation sets to predict the future trend of soil health changes in each region and outputs soil health change curves under each climate scenario.

2. The paddy field soil detection and regulation system based on multi-source sensor data analysis according to claim 1, characterized in that, The quality of paddy field soil is graded according to the comprehensive health index, and the quality is divided into healthy growth, slightly polluted, moderately polluted, and heavily polluted categories: Paddy field soil quality is divided into four levels based on a comprehensive health index; When the comprehensive health index When the soil quality is within the first level range, it is considered to be healthy for growth, indicating that the soil conditions and the rice growing environment are suitable. When the comprehensive health index Within the second level, the soil quality is judged as slightly polluted, indicating that the soil has a slight nutrient imbalance or minor pollution. When the comprehensive health index When the soil quality falls within the third level, it is classified as moderately polluted, indicating that the soil has begun to suffer serious damage. When the comprehensive health index When the soil quality falls within the fourth level, it is classified as severely polluted, indicating that the soil health has been seriously degraded.

3. The paddy field soil detection and regulation system based on multi-source sensor data analysis according to claim 1, characterized in that, The artificial wolf is divided into fuzzy logical relation groups according to a fuzzy partitioning scheme, and fitness indicators are evaluated using multi-source agricultural time-series data, including: Each artificial wolf, according to the corresponding fuzzy partitioning scheme, divides the target variable into a corresponding fuzzy set; A fuzzy logic relation group containing fuzzy logic relations is constructed based on fuzzy sets, and historical agricultural multi-source time series data is input into the fuzzy logic relation group model for prediction to obtain a predicted value sequence; Calculate the root mean square error between the predicted value and the true value, and use the root mean square error as the fitness index of the fuzzy partitioning scheme. The optimal artificial wolf is selected as the alpha wolf based on the fitness index, and the fuzzy logic relation group is updated based on the alpha wolf.

4. The paddy field soil detection and regulation system based on multi-source sensor data analysis according to claim 3, characterized in that, Given different future climate pathways, combined with soil regulation strategies for each region, and utilizing updated fuzzy logical relation sets, the soil health change trends for each region over future periods are predicted. The output includes soil health change curves for each climate scenario: Input multiple future climate path scenarios to generate corresponding future climate time series data for each region; By combining established soil regulation strategies in each region with future climate time series data, a joint input sequence for future environment and management can be constructed. The optimal fuzzy partitioning scheme determined by the alpha wolf and the updated fuzzy logic relation group model are used to predict the joint input sequence of future environment and management, and generate the predicted value sequence of soil health indicators. Based on the predicted value sequence, the soil health evolution trend of each region under different climate scenarios is output, generating a change curve in the time dimension.

5. A method for detecting and regulating paddy field soil based on multi-source sensor data analysis, characterized in that, This method is based on the paddy field soil detection and regulation system based on multi-source sensor data analysis as described in any one of claims 1-4.

Images