Intelligent agriculture cloud platform data acquisition and management system based on Internet of Things

Abstract

The invention belongs to the technical field of agricultural Internet of Things, and discloses an Internet of Things-based intelligent agricultural cloud platform data acquisition and management system, which comprises the following steps of: acquiring plant rhizosphere microenvironment information, generating rhizosphere original data and performing correlation analysis to obtain rhizosphere electrochemical interaction data; collecting canopy microclimate information, and performing temperature and humidity gradient quantification processing and pore dynamic response characteristic extraction to obtain pore response characteristic data; collecting field illumination information, and carrying out multispectral distribution analysis and photoperiod sensitivity differentiation analysis to obtain a crop photoperiod sensitivity spectrum; performing multi-dimensional association fusion to generate crop physiological status comprehensive evaluation data; agricultural measure analysis and resource allocation optimization are carried out, and finally a resource allocation instruction is generated and agricultural operation scheduling is carried out; according to the invention, intelligent acquisition, analysis and management of data in each link of agricultural production are realized, and the utilization efficiency of agricultural resources and the production accuracy are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of agricultural Internet of Things, and more particularly, to a smart agricultural cloud platform data acquisition and management system based on Internet of Things. BACKGROUND

[0002] With the deepening of the concept of precision agriculture and the rapid development of Internet of Things technology, smart agricultural systems have become an indispensable part of modern agricultural production. Current smart agricultural data acquisition and management systems aim to improve agricultural production efficiency and resource utilization through environmental parameter monitoring and crop growth analysis. However, these systems still have obvious deficiencies in the fine monitoring and correlation analysis of key parameters of plant physiology and ecology.

[0003] In the field of smart agricultural Internet of Things technology, current data acquisition and management systems often ignore the correlation between plant rhizosphere microenvironment material exchange dynamics and soil conductivity changes. The material exchange process of the plant rhizosphere microenvironment has a direct impact on crop growth and yield formation, and soil conductivity changes are one of the important indicators of plant rhizosphere activity, as root exudates and microbial activity can affect soil ion concentration distribution. Due to the lack of real-time correlation monitoring of plant rhizosphere microenvironment material exchange and soil conductivity changes, it is difficult to accurately grasp the dynamic changes of the rhizosphere environment and its impact on crop growth in actual agricultural production, resulting in inaccurate irrigation and fertilization decisions, and thus unable to effectively improve agricultural resource utilization efficiency. Crop photosynthesis is also affected by the combined effects of leaf temperature and environmental humidity gradients. This gradient change can change photosynthetic efficiency by affecting stomatal opening and closing, transpiration rate, etc. The existing smart agricultural systems fail to consider the dynamic monitoring of leaf microscale temperature and humidity gradients, ignoring the key role of this physical factor in crop physiological activities, which further reduces the accuracy of crop growth prediction models and underestimates the potential fluctuations in photosynthesis under certain microclimate conditions, especially in complex terrain and multi-layer planting structure farmland, increasing the error and uncertainty of yield prediction. Individual differences in crop photoperiod sensitivity affect the accuracy of agricultural production decisions, especially in mixed planting or rotation systems, where different varieties have significantly different demands for light duration and quality. The current smart agricultural platform lacks fine identification and analysis of crop individual photoperiod sensitivity, resulting in a "one-size-fits-all" approach to light supplementation strategies that cannot accurately regulate light environments for individual differences in field crops, affecting the relevance and effectiveness of agronomic measures.

[0004] In view of the above, the present application proposes a smart agricultural cloud platform data acquisition and management system based on Internet of Things to solve the above problems. SUMMARY

[0005] In order to overcome the above-mentioned defects of the prior art, in order to achieve the above-mentioned purpose, the present application provides the following technical scheme: a smart agricultural cloud platform data acquisition and management system based on Internet of Things, comprising:

[0006] Root domain perception module: collect rhizosphere microenvironment information, generate rhizosphere original data; according to the rhizosphere original data, carry out soil electrochemical property and rhizosphere material exchange correlation analysis, and generate rhizosphere electrochemical interaction data;

[0007] Canopy microenvironment analysis module: collect canopy microclimate information, generate canopy microclimate original data, carry out temperature and humidity gradient quantification processing according to the canopy microclimate original data, generate leaf temperature and humidity gradient data; based on the leaf temperature and humidity gradient data, carry out stomatal dynamic response characteristic extraction, and generate stomatal response characteristic data;

[0008] Illumination response analysis module: collect field illumination information, generate multi-point spectrum original data, carry out multi-spectrum distribution analysis based on the multi-point spectrum original data, and generate field spectrum distribution data; carry out crop individual photoperiod sensitivity differentiation analysis on the spectrum distribution data, and generate crop photoperiod sensitive spectrum;

[0009] Agricultural decision-making cooperation module: the rhizosphere electrochemical interaction data, the stomatal response characteristic data and the crop photoperiod sensitive spectrum are associated and fused in multiple dimensions, and crop physiological state comprehensive evaluation data is generated; based on the crop physiological state comprehensive evaluation data, agricultural measure analysis is carried out, and agricultural measure execution scheme is generated; according to the agricultural measure execution scheme, resource allocation optimization analysis is carried out, and resource allocation instruction is generated; based on the resource allocation instruction, agricultural operation scheduling is carried out.

[0010] The technical effect and advantages of the smart agricultural cloud platform data acquisition and management system based on Internet of Things of the present application are as follows:

[0011] The application realizes precise monitoring and management of the whole life cycle of crops by collecting crop growth environment data in multiple dimensions. The root domain perception module realizes accurate characterization of the interaction between root system and soil microorganisms by implanting a micro electrode array to monitor the rhizosphere microenvironment in real time, providing a scientific basis for crop nutrient absorption. The crown microenvironment analysis module accurately captures stomatal response changes by collecting and quantifying leaf temperature and humidity gradient data through a multi-point temperature and humidity sensor, realizing fine monitoring of plant water use efficiency. The light response analysis module identifies the spatial and temporal distribution characteristics of field light through a distributed spectral sensing network, realizing differentiated analysis of the light period sensitivity of individual crops and providing a basis for precise light management. The agricultural decision-making coordination module correlates and fuses multiple dimensional data for analysis, generating comprehensive evaluation data and precise execution plans, significantly improving the relevance of agricultural measures and resource utilization efficiency. The system realizes the transition from single environmental factor monitoring to comprehensive evaluation of all-round physiological state, from passive response to active prediction and early warning, providing technical support for modern precision agriculture and effectively improving the scientific nature, accuracy and sustainability of agricultural production. BRIEF DESCRIPTION OF DRAWINGS

[0012] Figure 1 FIG. 1 is a schematic diagram of a data acquisition and management system of a smart agricultural cloud platform based on the Internet of Things according to the present application;

[0013] Figure 2 FIG. 2 is a schematic diagram of a data acquisition and management method of a smart agricultural cloud platform based on the Internet of Things according to the present application. DETAILED DESCRIPTION

[0014] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of the present application.

[0015] Embodiment 1

[0016] Please refer to Figure 1 The data acquisition and management system of the smart agricultural cloud platform based on the Internet of Things according to the present embodiment includes:

[0017] Root domain perception module: collect rhizosphere microenvironment information through Internet of Things perception nodes to generate rhizosphere raw data; perform correlation analysis of soil electrochemical properties and rhizosphere material exchange based on the rhizosphere raw data to generate rhizosphere electrochemical interaction data;

[0018] The canopy microenvironment analysis module: collects canopy microclimate information, generates canopy microclimate raw data, performs temperature and humidity gradient quantification processing according to the canopy microclimate raw data, and generates leaf temperature and humidity gradient data; based on the leaf temperature and humidity gradient data, the stomatal dynamic response characteristics are extracted, and the stomatal response characteristic data are generated;

[0019] The light response analysis module: collects field light information, generates multi-point spectral raw data, performs multi-spectral distribution analysis based on the multi-point spectral raw data, and generates field spectral distribution data; the crop individual photoperiod sensitivity differentiation analysis is performed on the spectral distribution data, and the crop photoperiod sensitive spectrum is generated;

[0020] The agricultural decision-making cooperation module: multi-dimensional correlation fusion is performed on the rhizosphere electrochemical interaction data, the stomatal response characteristic data and the crop photoperiod sensitive spectrum, and the crop physiological state comprehensive evaluation data are generated; based on the crop physiological state comprehensive evaluation data, the agricultural measure analysis is performed, and the agricultural measure execution scheme is generated; according to the agricultural measure execution scheme, the resource allocation optimization analysis is performed, and the resource allocation instruction is generated; based on the resource allocation instruction, the agricultural operation scheduling is performed.

[0021] Preferably, plant rhizosphere microenvironment information is collected to generate rhizosphere raw data; soil electrochemical characteristics and rhizosphere material exchange correlation analysis is performed according to the rhizosphere raw data to generate rhizosphere electrochemical interaction data, including:

[0022] The plant rhizosphere microenvironment information is collected by implanting a microelectrode array to generate rhizosphere raw data;

[0023] Root exudates and soil ion concentration changes are associated and identified according to the rhizosphere raw data to generate root-soil ion exchange characteristic data;

[0024] Based on the root-soil ion exchange characteristic data, a rhizosphere microorganism activity and conductivity correlation map is constructed to generate a rhizosphere microecological conductivity characteristic spectrum;

[0025] According to the rhizosphere microecological conductivity characteristic spectrum, the soil electrochemical characteristics and rhizosphere material exchange correlation analysis is performed to generate the rhizosphere electrochemical interaction data.

[0026] Specifically, the implantable microelectrode array technology is used to monitor the plant rhizosphere region in real time. Each electrode array contains multiple ion selective electrodes (ISE), which can simultaneously detect K + , Ca 2+ , NO3 - , NH4 +various key ion concentrations. The electrode array is buried around the plant root system in a grid distribution manner within the depth range of 0 to 20 cm, with a horizontal spacing of 5 to 10 cm and a vertical spacing of 3 to 5 cm between adjacent probes, forming a three-dimensional monitoring network. The electrode array transmits data in real time to the on-site data collector through a low-power wireless transmission module. The sampling frequency can be adjusted according to the growth stage of the crops, with a sampling frequency of once every 5 minutes during the key growth period and once every 30 minutes during the non-key period. Each microelectrode probe is equipped with an independent temperature compensation unit to ensure that the measurement accuracy error is controlled within ±2% under temperature fluctuation conditions (within the range of 5 to 45°C). The collected rhizosphere raw data includes soil pH, conductivity, oxidation-reduction potential, temperature, and various key ion concentrations. The correlation analysis between root exudates and soil ion concentration changes is performed on the collected rhizosphere raw data. First, the time series analysis method is used to calculate the daily variation pattern of each ion concentration and the correlation of root activity. The time series correlation analysis is performed by the sliding window method (window size of 4 hours, step size of 1 hour) to identify the time delay relationship between the peak period of root activity and the change of each ion concentration within the day-night cycle. The characteristics of root exudates are evaluated by indirect indicators, mainly including short-term fluctuations (changes within ±0.3 units) of soil pH and small changes (within the range of ±0.5 mS / cm) of conductivity. When a decrease in pH value of more than 0.2 units within a short period of time (30 minutes) is detected, accompanied by an increase in the concentration of a specific ion (such as K + 、H + ), it is determined that the root is releasing organic acid exudates; when an increase in conductivity of more than 0.3 mS / cm within a short period of time is detected, accompanied by a significant decrease in NH4 + concentration, it is determined that the root is absorbing nitrogen. Through this pattern recognition, a corresponding relationship table between root exudates and changes in various ion concentrations is established, forming root-soil ion exchange characteristic data. These data are stored in matrix form, with rows representing different ion types, columns representing time series, and matrix element values representing concentration change rates, providing basic data support for the analysis of microbial activity in the next step.

[0027] Preferably, based on the root-soil ion exchange characteristic data, a rhizosphere microbial activity-conductivity correlation map is constructed to generate a rhizosphere micro-ecological conductivity characteristic spectrum, including:

[0028] According to the root-soil ion exchange characteristic data, the conductivity fluctuation pattern is identified in different time periods to generate a conductivity dynamic change sequence;

[0029] Based on the conductivity dynamic change sequence, the microbial community activity is deduced to generate a microbial activity index matrix;

[0030] The microbial activity index matrix and the conductivity dynamic change sequence are spatiotemporally fused to generate a transient conductivity-microbial activity correspondence table;

[0031] According to the transient electrical conductivity-microbial activity correspondence table, the correlation diagram of rhizosphere microbial activity and conductivity is constructed, and the rhizosphere microecological conductivity characteristic spectrum is generated.

[0032] Specifically, the conductivity measurements from the rhizosphere-soil ion exchange characteristic data were extracted, and the whole day of 24 hours was divided into 6 time periods (4 hours each), namely, early morning (00:00-04:00), morning (04:00-08:00), noon (08:00-12:00), afternoon (12:00-16:00), evening (16:00-20:00), and night (20:00-24:00). Wavelet transform (db4 wavelet basis was selected) was applied to the conductivity data in each time period for time-frequency analysis to extract the fluctuation characteristics of different frequency components. Three fluctuation frequency bands were set: high frequency band (period less than 30 minutes), medium frequency band (period 30-120 minutes), and low frequency band (period greater than 120 minutes). For the fluctuation amplitudes of different frequency bands, the average fluctuation intensity and fluctuation variance were calculated to form the fluctuation characteristic vector. According to the cluster analysis of the fluctuation characteristic vector, the conductivity fluctuation patterns were divided into four categories: stable type, periodic type, mutation type, and mixed type. The first and second derivatives of the conductivity were calculated for the data in each time period to represent the rate and acceleration of conductivity change, and combined with the fluctuation pattern classification results, the conductivity dynamic change sequence containing time information, absolute value, rate of change, acceleration of change, and fluctuation pattern type was generated. For example, when the morning period (04:00-08:00) conductivity was detected to present a periodic fluctuation and the first derivative reached a peak around 06:00, this feature was recorded as the "morning active type" fluctuation pattern; when the afternoon period (12:00-16:00) conductivity presented a mutation type fluctuation and the duration was more than 40 minutes, it was recorded as the "afternoon mutation type" fluctuation pattern. Based on the conductivity dynamic change sequence, combined with the theoretical model of rhizosphere microbial activity, the deduction analysis of microbial community activity was carried out. First, the mapping relationship model between conductivity change and microbial metabolic activity was established. This model was based on the following assumptions: microbial metabolic activity would affect the ion exchange balance in the rhizosphere region, which would then be reflected in the dynamic change of conductivity. The random forest algorithm was selected to construct the prediction model, and the input features included various indicators in the conductivity dynamic change sequence (such as fluctuation pattern type, rate of change, acceleration of change, etc.), and the output was the microbial activity index. The model training used a laboratory validation data set, which was obtained by simultaneously measuring the conductivity change and microbial activity (through indicators such as enzyme activity or respiratory rate) under controlled conditions. The model parameters were set as follows: 100 decision trees, maximum depth of each tree 8, minimum sample partition number 10, and Gini coefficient as the impurity measurement standard. The model verification used 10-fold cross-validation to ensure that the prediction accuracy was above 95%. The trained model was applied to the field-collected conductivity dynamic change sequence to generate the microbial activity prediction results. According to the prediction results, the activity indicators of six key microbial functional groups were calculated: nitrogen cycle bacteria (nitrifying bacteria, denitrifying bacteria), phosphorus-dissolving bacteria, potassium-releasing bacteria, organic matter-decomposing bacteria, plant growth-promoting bacteria, and pathogen-inhibiting bacteria.For each functional group, detect its activity level (range 0 to 10, 0 represents complete inhibition, 10 represents maximum activity, evaluated by oxygen consumption), activity trend (up, stable or down) and activity duration. Organize these indicators into a microbial activity indicator matrix, with rows representing different microbial functional groups, columns representing time series, and element values representing activity levels at corresponding time points. Temporally and spatially fuse the microbial activity indicator matrix with the conductivity dynamic change sequence to generate a transient conductivity-microbial activity correspondence table. First, time-align the two sets of data to ensure consistency of timestamps. For each time point in the microbial activity indicator matrix, extract the corresponding conductivity dynamic change data to establish the correspondence between conductivity features and microbial activity. Apply the dynamic time warping algorithm to handle the time delay of the two sets of data, which can identify the time lag relationship between conductivity changes and microbial activity changes, with a window size of 2 hours and a step size of 10 minutes. The identified time delay relationship is recorded as the "response delay time" and becomes an important attribute of the correspondence table. For each microbial functional group, calculate its activity correlation coefficient with different conductivity features (such as absolute value, change rate, fluctuation pattern, etc.), using both Pearson correlation coefficient and Spearman rank correlation coefficient, and taking the maximum of the two as the final correlation strength. Based on the correlation strength, identify the set of conductivity features that have the greatest impact on each microbial functional group activity. At the same time, for each conductivity fluctuation pattern, identify its corresponding typical microbial activity pattern. For example, when "periodic" conductivity fluctuations occur in the morning period with an amplitude in the range of 0.3 to 0.5 mS / cm, it usually corresponds to a significant increase in the activity of nitrogen cycle bacteria (activity level from 3 to 7) and an extension of the duration (from 1 hour to 3 hours). This correspondence is recorded in the transient conductivity-microbial activity correspondence table, which contains attributes such as time information, conductivity features, microbial activity level, correlation strength and response delay time. Based on the transient conductivity-microbial activity correspondence table, construct a rhizosphere microbial activity and conductivity correlation graph. The graph uses a directed weighted graph structure, with nodes representing conductivity features and edges representing their influence relationships. The weight of the edge represents the correlation strength, and the direction of the edge represents the causal relationship. For the construction of the graph, first extract all significant correlation relationships (correlation strength greater than 0.65) from the transient conductivity-microbial activity correspondence table as the initial edge set of the graph. Then, apply Granger causality test to identify the causal relationship in the correlation relationship to determine the direction of the edge. For each edge, calculate its time correlation characteristics, including whether there is a specific time pattern (such as morning specificity, all-day persistence, etc.) and conditional trigger characteristics (such as whether specific temperature, pH value, etc. conditions are required). These characteristics are used as edge attributes to enrich the information content of the graph.Based on the graph structure, the centrality indicators of each microbial functional group (including degree centrality, closeness centrality and betweenness centrality) are calculated to identify the functional groups that play a key role in the rhizosphere micro-ecological network. At the same time, the influence indicators of the electrical conductivity characteristics are evaluated to identify the types of electrical conductivity characteristics that have the greatest impact on microbial activity. In this way, a complete rhizosphere microbial activity and electrical conductivity correlation map is constructed, and a rhizosphere micro-ecological electrical conductivity feature spectrum is generated. This feature spectrum not only contains static correlation relationships, but also includes time dynamic characteristics and condition triggering mechanisms, and can comprehensively describe the electrical conductivity characteristics of the rhizosphere micro-ecosystem. According to the rhizosphere micro-ecological electrical conductivity feature spectrum, the correlation analysis between soil electrochemical characteristics and rhizosphere material exchange is carried out. First, the correlation pattern between electrical conductivity change and microbial activity is extracted from the feature spectrum to identify key electrochemical indicators (such as electrical conductivity threshold, pH value change range, etc.). Combined with these indicators, the correlation between the release pattern of root exudates (such as organic acids, sugars, amino acids, etc.) and the change of soil electrochemical environment is analyzed. Using multiple regression analysis method, a mathematical model is established between the release rate of root exudates and the changes of electrical conductivity, pH value and oxidation-reduction potential. Model parameters: significant variables are selected by stepwise regression method, with a significance level of 0.05, and the variance inflation factor is used for multiple collinearity test. Based on the established model, the release characteristics of root exudates under different electrochemical conditions are predicted, and the regulation of microbial activity on this process is evaluated. Through this analysis, rhizosphere electrochemical interaction data is generated, which includes electrochemical characteristic parameters (such as electrical conductivity, pH value, oxidation-reduction potential, etc.), root exudate parameters (such as type, concentration, release rate, etc.), microbial activity parameters (such as functional group activity level, metabolic products, etc.), and their interaction parameters (such as correlation coefficient, influence degree, response time, etc.).

[0033] Preferably, the crown microclimate information is collected to generate crown microclimate original data, and the temperature and humidity gradient quantization processing is performed according to the crown microclimate original data to generate leaf temperature and humidity gradient data; the stomatal dynamic response characteristics are extracted based on the leaf temperature and humidity gradient data to generate stomatal response characteristic data, including:

[0034] The crown microclimate information is collected by a multi-point micro wireless temperature and humidity sensor to generate crown microclimate original data;

[0035] The leaf surface and surrounding air temperature and humidity difference analysis is performed according to the crown microclimate original data to generate leaf-air interface temperature and humidity gradient data;

[0036] The leaf-air interface temperature and humidity gradient data are analyzed for spatio-temporal change trend to generate leaf temperature and humidity gradient dynamic change data;

[0037] The leaf temperature and humidity gradient dynamic change data are corrected according to the terrain and planting structure information to generate leaf temperature and humidity gradient data;

[0038] Based on the leaf temperature and humidity gradient data, the stomatal dynamic response characteristics are extracted, and the stomatal response characteristic data are generated.

[0039] Specifically, a multi-point micro wireless temperature and humidity sensor network is deployed to finely monitor the plant canopy microclimate. Each sensor module is composed of a high-precision digital temperature and humidity sensor, a micro computing unit, and a low-power wireless communication module. The sensors adopt a hierarchical arrangement strategy, and are installed at different heights (ground, lower canopy, middle canopy, and upper canopy) of each target plant, forming a vertical temperature and humidity monitoring profile. In the horizontal direction, the sensors are distributed in a grid pattern with a spacing of 3 to 5 meters, covering the entire monitoring area. The installation position of each sensor is accurately recorded to facilitate spatial interpolation analysis. The sampling frequency of the sensors is once every 5 minutes, which can be automatically adjusted to once every 1 minute during critical weather change periods (such as sunrise, sunset, or weather changes). The temperature and humidity sensors are equipped with a micro wind speed sensor to record micro air flow changes, which is crucial for understanding the formation and changes of canopy microclimate. The original data of canopy microclimate include: air temperature (℃), relative humidity (%), dew point temperature (℃), absolute humidity (g / m 3), wind speed (m / s) and direction, etc., as well as the three-dimensional spatial coordinates and timestamps of each sensor. Based on the original data of the canopy microclimate, the temperature and humidity differences between the leaf surface and the surrounding air were calculated. To achieve this goal, the leaf surface temperature was first measured directly by the leaf surface temperature sensor. The micro-dual temperature difference sensor technology was used to measure the air temperature at a distance of 1 cm from the leaf surface, directly obtaining the temperature gradient data. For relative humidity, gradient measurements were made by a micro-humidity sensor array at different distances (0 mm, 5 mm, 10 mm, 20 mm) around the leaf. Based on the temperature and humidity measurements, the water vapor pressure difference between the leaf surface and the surrounding air was calculated, which is a key physical parameter driving plant transpiration. For each measurement point, leaf-air interface temperature and humidity gradient data were generated, including time, location, leaf temperature, air temperature, leaf-air temperature difference, leaf surface relative humidity, air relative humidity, humidity gradient, and water vapor pressure difference. The leaf-air interface temperature and humidity gradient data were analyzed for temporal and spatial trends. In the time dimension, time series analysis methods were applied, including moving average (window size 30 minutes), trend decomposition (STL decomposition for extracting seasonal, trend, and residual components), and change point detection (PELT algorithm for identifying abrupt points in the data). The identified temporal features include: diurnal variation patterns (such as morning warming, midday cooling, etc.), fluctuation periods (hourly, daily), and abnormal change points (such as sudden cooling, humidity sudden change, etc.). In the spatial dimension, the data of discrete measurement points were extended to continuous spatial distribution through geostatistical methods (such as Kriging interpolation). Through this interpolation method, a three-dimensional spatial distribution map of temperature and humidity gradients was generated, identifying hotspots and areas of rapid change in temperature and humidity gradients within the canopy. By combining temporal and spatial features, a spatiotemporal joint analysis was conducted to identify the coordinated variation patterns of temperature and humidity gradients in time and space. For example, the temperature gradient variation characteristics of the canopy top and interior during the midday high temperature period, or the spatial distribution characteristics of leaf humidity gradient during the morning dew formation period. Through this analysis, leaf temperature and humidity gradient dynamic change data were generated, which describe the variation rules and characteristic patterns of temperature and humidity gradients in time and space. The leaf temperature and humidity gradient dynamic change data were corrected for topography and planting structure. First, topographic data (such as elevation, slope, aspect, etc.) and planting structure data (such as row spacing, plant spacing, canopy width, etc.) were introduced to establish a topography-planting structure influence model. Topographic data were obtained through high-precision RTK-GPS and unmanned aerial surveying technology, with an accuracy of centimeters. Planting structure data were obtained through field measurement and image recognition technology, recording the location, height, canopy width, and branch structure of each plant.Based on these data, the influence factors of micro-topography and planting structure on temperature and humidity gradient are established, including slope direction influence factor (temperature gradient change caused by the difference of sunlight radiation on different slope directions), slope gradient influence factor (the influence of slope on air movement and humidity distribution) and planting density influence factor (the influence of plant spacing on ventilation in canopy and temperature and humidity distribution). The relationship model between temperature and humidity gradient and these influence factors is established by using multiple regression method. The correction process adopts iterative least square method, and the convergence threshold of each iteration is set to 0.01, and the maximum iteration number is 100. Through this correction process, the interference of terrain and planting structure factors on temperature and humidity gradient is eliminated, and the change of temperature and humidity gradient caused by the physiological state of plants is extracted. The corrected data form leaf temperature and humidity gradient data, which more accurately reflect the temperature and humidity exchange characteristics between plant leaves and the surrounding environment.

[0040] Preferably, based on the leaf temperature and humidity gradient data, the stomatal dynamic response characteristics are extracted to generate stomatal response characteristic data, including:

[0041] According to the leaf temperature and humidity gradient data, the transpiration rate is estimated to generate transpiration potential prediction data;

[0042] The actual temperature distribution of the leaf is obtained by infrared thermal imaging technology to generate a leaf temperature distribution map;

[0043] The transpiration potential prediction data and the leaf temperature distribution map are compared and analyzed to generate stomatal opening and closing state inference data;

[0044] According to the stomatal opening and closing state inference data, the photosynthetic efficiency change trend is predicted to generate stomatal response characteristic data.

[0045] Specifically, the transpiration potential of the plant is calculated based on the leaf temperature and humidity gradient data. The transpiration potential is the maximum transpiration rate that the plant can reach under the current environmental conditions, reflecting the driving force of the environment on transpiration. The calculation uses an improved Penman-Monteith equation, which considers the influence of water vapor pressure difference, air temperature, light, wind speed and other factors on the transpiration process. The key improvement is to introduce the leaf temperature and humidity gradient data to replace the atmospheric layer temperature and humidity data used in the traditional equation, so that the calculation result more accurately reflects the leaf microenvironment conditions. The calculation formula is as follows: Where ET represents the reference evapotranspiration, Δ represents the slope of the saturation vapor pressure curve, Rn represents the net radiation, G represents the soil heat flux, γ represents the psychrometric constant, T represents the leaf surface temperature, u2 represents the wind speed at 2 cm around the leaf, es represents the leaf surface saturation vapor pressure, and ea represents the actual vapor pressure of the air around the leaf. These parameters are extracted from the leaf temperature and humidity gradient data and obtained through other sensors such as radiation sensors and wind speed sensors. The calculation result is expressed in terms of transpiration potential index, ranging from 0 to 10, with 0 indicating almost no transpiration potential and 10 indicating extremely high transpiration potential. For each leaf monitoring point, transpiration potential prediction data containing parameters such as time, location, transpiration potential index, and influence factor weight are generated. These data reflect the possible transpiration state of plant leaves under current environmental conditions, providing a theoretical basis for subsequent analysis of stomatal behavior. A high-precision infrared thermal imager is used to monitor the temperature distribution of plant leaves. The imaging frequency is once every 10 minutes, which can be increased to once every 1 minute during periods of rapid light or temperature changes. The thermal imager is installed on an adjustable height bracket, covering the entire monitoring area to ensure that each target leaf is within the field of view. The acquired thermal images are geometrically and radiometrically corrected to eliminate lens distortion and environmental radiation interference. Image processing uses a four-step process: (1) use adaptive thresholding to segment the leaf area; (2) apply median filtering (window size 5x5 pixels) to reduce noise; (3) calculate the temperature statistical features (mean temperature, maximum temperature, minimum temperature, temperature standard deviation, temperature gradient, etc.) of each leaf; (4) generate a leaf temperature distribution heat map using color coding to display temperature distribution. The processed data form a leaf temperature distribution map, which contains detailed temperature distribution information for each leaf, with a spatial resolution of millimeters and a temperature resolution of 0.1°C, providing intuitive data for analyzing the relationship between leaf temperature and stomatal activity. The transpiration potential prediction data are compared and analyzed with the leaf temperature distribution map to infer the stomatal opening and closing state. The analysis is based on the principle of stomatal opening and closing: when the stomata are open, the leaf dissipates heat through transpiration, and the leaf temperature is usually lower than the surrounding environment; when the stomata are closed, transpiration is limited, and the leaf temperature may rise. The comparison and analysis are divided into three steps: (1) calculate the difference between the actual leaf temperature and the theoretical temperature. The theoretical temperature is the leaf temperature that should be obtained according to the energy balance equation and the transpiration potential prediction. The larger the difference, the greater the deviation between the actual transpiration rate and the potential transpiration rate; (2) calculate the leaf temperature uniformity index, which is the ratio of the average temperature to the pre-set temperature threshold. When the stomata are uniformly open, the leaf temperature distribution is relatively uniform; when the stomata are partially open, the temperature distribution is uneven. The index ranges from 0 to 1, with 1 indicating complete uniformity; (3) calculate the leaf temperature dynamic response index. By analyzing the response speed and amplitude of leaf temperature to environmental changes, the sensitivity of stomatal regulation is inferred. The response index ranges from 0 to 10, with 0 indicating almost no response and 10 indicating extremely fast response. Based on these three indicators, the stomatal state is judged using fuzzy logic reasoning.The fuzzy rule set contains multiple rules, with the three aforementioned indicators as input and the output being the degree of stomatal opening (ranging from 0 to 1, with 0 indicating fully closed and 1 indicating fully open). The fuzzy inference uses the Mamdani model, with a trapezoidal membership function and the centroid method for defuzzification. The inference results form inferred data on stomatal opening and closing status, which includes information such as the degree of stomatal opening, stomatal reaction rate, and stomatal state change trends for each leaf at different time points. Based on this inferred data, the trend of plant photosynthetic efficiency is predicted. This prediction is based on the stomatal limitation theory, which states that stomatal opening and closing status directly affects the rate of CO2 entering the leaf, thereby affecting photosynthetic efficiency. The prediction model uses a multi-layer perceptron (MLP) neural network architecture. Input features include stomatal opening, leaf temperature, light intensity, air CO2 concentration, and leaf water status. The neural network architecture consists of five input layers (corresponding to the five input features), two hidden layers with eight and four nodes respectively, and two output layers (corresponding to photosynthetic efficiency and change trends). Activation function: The hidden layer uses the ReLU function, and the output layer uses the Sigmoid function (photosynthetic efficiency) and the Tanh function (change trend). Training parameters: learning rate is 0.01, batch size is 32, number of training rounds is 1000 rounds, and the early stopping condition is no significant improvement after 20 consecutive rounds on the validation set. The model training data comes from the gas exchange measurement results of specimens under laboratory conditions, covering photosynthetic parameters under different stomatal opening and closing states. The trained model is applied to real-time field data to predict the photosynthetic efficiency level (relative value, range 0 to 1) and change trend (range -1 to +1, negative values ​​indicate a downward trend, positive values ​​indicate an upward trend) of each monitored leaf. The prediction results also include the weights of the impact of different environmental factors on photosynthetic efficiency, which are used to identify the current limiting factors of photosynthesis. All these prediction results constitute the stomatal response characteristic data, which comprehensively describes the dynamic response characteristics of plant stomata under different environmental conditions and their impact on the photosynthetic process.

[0046] Preferably, field illumination information is collected to generate multi-point spectral raw data, multi-spectral distribution analysis is performed based on the multi-point spectral raw data to generate field spectral distribution data; and individual crop photoperiod sensitivity differential analysis is performed on the spectral distribution data to generate a crop photoperiod sensitivity spectrum, including:

[0047] Collect field light information through a distributed spectral sensing network to generate multi-point spectral raw data;

[0048] Perform spatiotemporal interpolation processing on multi-point spectral raw data to generate a continuous spectral distribution scene map;

[0049] Based on the continuous spectral distribution scene map, the spatial heterogeneity of light intensity and light quality is analyzed to generate field spectral distribution data;

[0050] The spectral distribution data is subjected to crop individual photoperiod sensitivity differentiation analysis to generate a crop photoperiod sensitivity spectrum.

[0051] Specifically, a distributed spectral sensing network is deployed to comprehensively monitor the field light environment. The spectral sensing network is composed of multiple types of sensors: (1) a full-spectrum spectrometer (wavelength range 350-1000 nm, spectral resolution 1.5 nm) for high-precision spectral acquisition; (2) a multi-channel spectral sensor (typical wavebands 450 nm, 550 nm, 650 nm, 730 nm, 810 nm, and 950 nm) for specific waveband light intensity monitoring; (3) a PAR (photosynthetically active radiation) sensor (wavelength range 400-700 nm) for monitoring the amount of photosynthetically active radiation available to crops; and (4) a red / far-red ratio sensor, which specifically monitors the ratio of red light (around 660 nm) and far-red light (around 730 nm), a key parameter affecting plant photomorphogenesis. The sensors are deployed in a hierarchical grid layout, with 3-5 monitoring layers set up in the vertical direction, from the ground to the top of the canopy, at intervals of 0.5-1 meter; and in the horizontal direction, covering the entire monitoring area. Each monitoring point is equipped with a GPS positioning and attitude sensor to record the precise position and orientation of the sensor for subsequent data processing. The sampling frequency is once every 10 minutes, automatically adjusted to once every 1 minute during periods of intense light changes such as sunrise and sunset. The collected multi-point spectral raw data includes parameters such as the spatial position, timestamp, waveband light intensity, total PAR value, red / far-red ratio, and spectral curve of each monitoring point. Data resolution: light intensity accuracy ±5 μmol / m 2 / s, with a spectral resolution of 1.5 nm, a temporal resolution of 1 to 10 minutes, and a spatial resolution depending on the sensor arrangement density. The collected multi-point spectral raw data were processed with spatial and temporal interpolation to generate a continuous spectral distribution scene map covering the entire monitoring area. In the temporal dimension, spline interpolation was used to handle the intervals between data sampling, ensuring continuous time series data with a uniform temporal resolution of 1 minute. In the spatial dimension, a combination of geographically weighted regression (GWR) and kriging interpolation was used. GWR takes into account the influence of spatial location on light distribution, calculating specific weight coefficients for each location, while kriging interpolation considers spatial autocorrelation, enabling the prediction of unknown point values based on known point values. For different waveband light intensities, separate interpolation processing was performed to preserve waveband-specific information. Important parameters for interpolation processing included search radius (set to twice the monitoring point spacing), kernel function type (Gaussian kernel function), interpolation weight (calculated inversely proportional to distance), and outlier handling threshold (3 times standard deviation). For spectral curve data, principal component analysis (PCA) was used to extract main features, which were then spatially interpolated and reconstructed into complete spectral curves. Through this processing, discrete multi-point spectral data were converted into continuous four-dimensional data (three-dimensional space + time dimension), forming a continuous spectral distribution scene map. This scene map is presented in a three-dimensional grid format, with each grid point containing complete spectral information and a time variation sequence. Based on the continuous spectral distribution scene map, spatial heterogeneity analysis of light intensity and light quality was conducted. Light intensity analysis focused on the spatial distribution and temporal variation of PAR values. The PAR daily integral value at each spatial location was calculated to reflect the cumulative light exposure at that location. The spatial coefficient of variation (CV, ratio of standard deviation to mean) of PAR was calculated to quantify the unevenness of light intensity distribution. Light hotspots (areas with PAR values 1.5 times higher than the average) and light weak areas (areas with PAR values 50% lower than the average) were identified, and a light intensity zoning map was drawn. Light quality analysis focused on the spatial differences in spectral composition. Important spectral indicators were calculated for each location, including red / far-red ratio, blue light proportion (percentage of blue light in PAR), red light proportion, and green light proportion. These indicators have important effects on plant photomorphogenesis and photoperiodic response. Through principal component analysis (PCA) and cluster analysis, the monitoring area was divided into different light quality type zones, such as "high blue light zone". Time variation analysis focused on the diurnal variation patterns and seasonal trends of light parameters. For each spatial location, time series features such as sunrise peak time, sunset change rate, and midday stable period length were calculated for light parameters (such as PAR). Fourier transform was used to analyze the periodic variation of light parameters, identifying different scale variation patterns such as daily, weather, and seasonal cycles. The results of spatial and temporal analysis were integrated to generate field spectral distribution data.The data comprehensively describes the illumination environment of the monitoring area, including light intensity distribution, light quality distribution, and time variation law, and other multi-dimensional information.

[0052] Preferably, the spectral distribution data is subjected to crop individual photoperiod sensitivity differentiation analysis to generate a crop photoperiod sensitivity spectrum, including:

[0053] Based on the field spectral distribution data combined with plant growth conditions, individual photosynthetic efficiency is evaluated to generate individual photosynthetic efficiency data;

[0054] According to the individual photosynthetic efficiency data, illumination conditions and growth performance correlation analysis is performed to generate an illumination-growth correlation matrix;

[0055] The illumination-growth correlation matrix is subjected to intra-species individual difference clustering to generate photoperiod response grouping data;

[0056] According to the photoperiod response grouping data, a crop photoperiod sensitivity individualized identification standard is constructed to generate a crop photoperiod sensitivity spectrum.

[0057] Specifically, in combination with the field spectral distribution data and plant growth monitoring data, the photosynthetic efficiency of each crop individual is evaluated. Plant growth monitoring uses multiple methods: (1) automatic measurement, including high-precision laser scanners (point cloud density greater than 1000 points / m 2 ) and multi-angle RGB cameras (resolution greater than 12 million pixels), which collect plant three-dimensional structure and appearance features daily at regular times; (2) portable chlorophyll fluorometers, which measure representative leaf fluorescence parameters, including maximum photochemical efficiency, ΦPSII (actual photochemical efficiency), NPQ (non-photochemical quenching), etc., reflecting plant photosystem activity, on a regular basis (every 3 to 5 days); (3) gas exchange measurement, which measures typical leaf parameters such as net photosynthetic rate, stomatal conductance, and intercellular CO2 concentration on a regular basis (every 7 to 10 days). These measurement data are matched with the spectral distribution data of the location where the plants are located to calculate the photosynthetic efficiency index of each plant. The photosynthetic efficiency index includes: net photosynthetic rate under unit light intensity, biomass accumulation rate per unit leaf area (g / m 2 / day), light use efficiency (LUE, biomass energy / absorbed light energy), and light response curve parameters (e.g., maximum photosynthetic rate Pmax, light compensation point LCP, light saturation point LSP, etc.). These indices collectively reflect the photosynthetic capacity and light energy use efficiency of the plants under the current light conditions. For each plant, individual photosynthetic efficiency data containing spatial location, time period, photosynthetic efficiency indices, and growth performance parameters are generated. Based on the individual photosynthetic efficiency data, the correlation between light conditions and plant growth performance is analyzed. First, the key light parameters are determined, including: daily average PAR value, daily light duration (hour), red / far-red ratio, blue light proportion, and light fluctuation index (standard deviation of light intensity to mean value). Then, the key growth performance indices are determined, including: plant height growth rate (cm / day), leaf area growth rate (cm 2 / day), dry matter accumulation rate (g / day), branch number, flower bud differentiation rate, and yield-related indicators. For each plant, the correlation coefficient (Pearson correlation coefficient) between the light parameters and the growth performance indicators during the growth period was calculated, forming a correlation coefficient matrix. There was a nonlinear relationship between the correlation, and the mutual information method was used to quantify the nonlinear association between the light parameters and the growth performance. To identify the importance differences of different light parameters, the random forest algorithm was applied for feature importance analysis, and the influence weight of each light parameter on the growth performance was calculated. The correlation coefficient, mutual information value, and feature importance weight were integrated to generate a comprehensive correlation strength score (range 0 to 1). Finally, a light-growth correlation matrix was formed, with the rows representing different plant individuals, the columns representing the combination of different light parameters and growth performance indicators, and the matrix element values representing the corresponding correlation strength score. Cluster analysis was performed on the light-growth correlation matrix to identify the individual differences within the variety. Hierarchical clustering method was used for clustering, Euclidean distance was used for distance measurement, and Ward minimum variance method was used for connection. Dynamic tree cutting method was used for clustering tree pruning to ensure that each cluster had sufficient discrimination and internal consistency. The number of clusters was determined by multiple evaluation indicators, including silhouette coefficient and Davies-Bouldin index. Generally, the final number of clusters was set to 3 to 5, representing different photoperiod response types. Feature analysis was performed on each cluster to identify the key light parameters and growth performance features that distinguish the cluster. The photoperiod sensitivity index of each cluster was calculated, which reflects the response sensitivity of plants within the cluster to light changes. The sensitivity index was calculated based on the growth performance changes caused by small changes in light parameters. Typical photoperiod response types include "light-sensitive type" (rapid and significant response to light changes), "light-stable type" (slow and small response to light changes), and "intermediate type" (between the two). The clustering results were associated with the individual information of the plants to generate photoperiod response grouping data, which indicated the photoperiod response type of each plant and its key feature parameters. Based on the photoperiod response grouping data, an individualized identification standard for crop photoperiod sensitivity was established. The identification standard was based on multiple feature indicators: (1) morphological feature indicators, such as plant type characteristics (plant height to crown width ratio), leaf arrangement (sparse or compact), leaf angle (upright or horizontal), etc.; (2) physiological feature indicators, such as photosynthetic characteristic parameters (such as light compensation point, light saturation point), chlorophyll content, carotenoid to chlorophyll ratio, etc.; (3) light response indicators, such as light response time (delay time of physiological response after light change), light response amplitude (change amplitude of physiological parameters), and light adaptation rate, etc. Based on these indicators, an individualized identification model was established, which could quickly judge the photoperiod sensitivity type of any plant. The model used the random forest algorithm, with the input features being the above morphological, physiological, and light response indicators, and the output being the photoperiod sensitivity type and its probability score. The model training used 10-fold cross-validation.Based on the trained model, the light period sensitivity of all monitored plants is evaluated, and a dataset containing the sensitivity type, sensitivity score and key characteristic parameters of each plant is generated, i.e. the crop light period sensitivity spectrum. This sensitivity spectrum not only reflects the overall light period characteristics of the variety, but also reveals the individual differences within the variety, providing an important basis for precision agriculture management.

[0058] Preferably, the rhizosphere electrochemical interaction data, stomatal response characteristic data and crop light period sensitivity spectrum are multi-dimensionally associated and fused to generate comprehensive evaluation data of the physiological state of the crop, including:

[0059] The rhizosphere electrochemical interaction data, stomatal response characteristic data and crop light period sensitivity spectrum are time-series aligned to generate multi-dimensional time-series collaborative data;

[0060] The multi-dimensional time-series collaborative data is analyzed for physiological process interaction influence to generate a physiological process coupling relationship map;

[0061] Based on the physiological process coupling relationship map, the current growth state and potential stress risk of the crop are evaluated to generate comprehensive evaluation data of the physiological state of the crop.

[0062] Specifically, the data from different modules are time-aligned to ensure consistency and comparability in the time dimension. First, the time labels of the data are unified, converting all data into timestamps in the same time zone and time format. Then, time resampling is performed for data with different collection frequencies. High-frequency data (e.g., minute-level data) are converted to low-frequency data (e.g., hour-level or day-level data) through moving average or downsampling methods; low-frequency data generate high-frequency data points through interpolation methods (e.g., linear interpolation, spline interpolation, or ARIMA model prediction). The aligned time resolution is usually set to 1 hour, and in special cases (e.g., rapid change periods), it can be increased to 10 minutes. During the data alignment process, missing values and outliers are handled. Missing values are filled by multiple imputation methods, which consider the correlation between variables; outliers are identified by the IQR method (interquartile range method), and values identified as potential outliers are decided to be retained, corrected, or deleted after confirmation by a person skilled in the art. After completing the time alignment, the spatial reference of different data sources is ensured to be consistent. All data are mapped to a unified spatial coordinate system, facilitating spatial correlation analysis. Finally, multi-dimensional time-series collaborative data are generated, which integrate rhizosphere electrochemical interaction data, stomatal response characteristic data, and crop photoperiod sensitivity spectrum, and ensure the consistency of time and space dimensions, laying the foundation for subsequent interaction analysis. Physiological process interaction analysis is performed on the multi-dimensional time-series collaborative data to reveal the correlation and mutual influence mechanism between different physiological processes. Various statistical and machine learning methods are used in the analysis, including: (1) partial correlation analysis, which controls the influence of other variables and calculates the direct correlation between two physiological process parameters; (2) Granger causality test, which identifies the time series causal relationship between physiological processes and reveals which processes are leading indicators of other processes; (3) structural equation model, which constructs the path relationship between physiological processes and quantifies the strength of direct and indirect effects; (4) dynamic Bayesian network, which establishes the probabilistic dependence relationship between physiological variables and reflects the dynamic influence between variables at different time points. Through these analyses, key physiological process interaction patterns are identified, such as the coordination relationship between root absorption and stomatal opening and closing, the feedback mechanism between photosynthesis and root exudate release, etc. For each interaction pattern, its strength (correlation coefficient or influence coefficient), direction (positive or negative correlation), time lag (hours or days), and conditional dependence (under what conditions is it significant) are calculated. These interaction relationships are presented in the form of a network graph, with nodes representing physiological processes or parameters and edges representing their correlation relationships. The thickness of the edge represents the correlation strength, and the color of the edge represents the correlation direction (positive or negative correlation). Network analysis uses centrality measures (such as degree centrality and betweenness centrality) to identify key nodes in the network, which represent physiological processes that have the greatest impact on the overall physiological state.Based on the above analysis, a physiological process coupling relationship graph is generated, which comprehensively describes the mutual influence relationship between different physiological processes of crops, providing a systematic perspective for understanding the overall physiological state of crops. Based on the physiological process coupling relationship graph, the current growth state and potential stress risk of crops are comprehensively evaluated. The evaluation adopts a multi-index comprehensive evaluation method, including the following steps: (1) determining the evaluation index system, including growth state indicators (such as photosynthetic efficiency, transpiration efficiency, root activity, etc.) and stress risk indicators (such as water stress risk, nutrient stress risk, pest and disease risk, etc.); (2) establishing an index weight system, using a combination of the analytic hierarchy process (AHP) and entropy weight method to determine the weight of each index; (3) building an evaluation model, using a fuzzy comprehensive evaluation method to combine qualitative judgment and quantitative analysis, generating a comprehensive score; (4) setting threshold values and risk levels, dividing the score results into different levels (such as "excellent", "good", "general", "poor" and "risk"). During the evaluation process, special attention is paid to the coordination and balance between physiological processes. Through the physiological process coupling relationship graph, uncoordinated physiological states are identified, such as imbalance between photosynthesis and root absorption, imbalance between stomatal regulation and water supply, etc. These uncoordinated states are usually early signals of potential problems, even if individual indicators have not reached the warning threshold. At the same time, the dynamic trend of the physiological state is analyzed to identify potential signs of deterioration or improvement. Trend analysis uses a time series decomposition method to decompose the original data into trend, seasonal and residual terms, focusing on the change direction and rate of the trend term. Combined with the current value, historical trend and prediction model, a 7-day physiological state prediction is generated, providing a basis for early intervention. The evaluation results form the crop physiological state comprehensive evaluation data, which includes current physiological state score, specific indicator values, potential stress risk rating, unbalanced physiological process identification and future state prediction, etc., providing comprehensive support for subsequent agronomic decisions.

[0063] Preferably, based on the crop physiological state comprehensive evaluation data, agronomic measure matching is performed to generate an agronomic measure execution scheme; resource allocation optimization analysis is performed according to the agronomic measure execution scheme to generate resource deployment instructions; agricultural operation scheduling is performed based on the resource deployment instructions, including:

[0064] According to the crop physiological state comprehensive evaluation data, decision logic reasoning is performed in combination with the agronomic knowledge base to generate preliminary intervention strategy suggestions;

[0065] The feasibility and resource consumption of the preliminary intervention strategy suggestions are evaluated to generate an agronomic measure execution scheme;

[0066] Resource allocation optimization analysis is performed according to the agronomic measure execution scheme to generate resource deployment instructions;

[0067] The terminal receives resource allocation instructions and performs precise irrigation, fertilization and light supplement operations through the Internet of Things.

[0068] Preferably, the preliminary intervention strategy suggestion is generated by combining the comprehensive evaluation data of the crop physiological state with the agronomic knowledge base for decision logic reasoning.

[0069] Based on the comprehensive evaluation data of the crop physiological state, the current crop growth limiting factor is identified, and a limiting factor ranking list is generated.

[0070] According to the limiting factor ranking list, the related intervention scheme in the agronomic knowledge base is retrieved, and a set of optional intervention schemes is generated.

[0071] The resource input and expected output ratio of the set of optional intervention schemes is evaluated, and a scheme benefit score table is generated.

[0072] Based on the scheme benefit score table, the preliminary intervention strategy suggestion is generated by decision logic reasoning.

[0073] Specifically, the physiological state of the crop is comprehensively evaluated by multi-dimensional analysis to identify the limiting factors of current crop growth. The analysis methods include: (1) comparative analysis, comparing the current physiological indicators with the standard values or historical optimal values to identify the indicators with the largest gap; (2) sensitivity analysis, simulating the influence of small changes in different parameters on the growth state to identify the parameters with the most significant influence; (3) relative deficiency degree calculation, calculating the deficiency proportion of each factor relative to the optimal value according to Liebig's minimum factor law. Combining these methods, the main limiting factors of current growth are identified, such as water stress, nitrogen deficiency, insufficient light, and unsuitable temperature, etc. The identified limiting factors are ranked in terms of importance, including: influence degree (restriction intensity on growth), urgency (how quickly intervention is needed), and intervenability (the ease and effectiveness of intervention). A comprehensive score (range 0 to 100) is calculated for each limiting factor, and the list is sorted according to the score to generate a limiting factor ranking list, which includes information such as limiting factor name, current value, target value, gap ratio, influence degree score, urgency score, intervenability score, and comprehensive score. According to the limiting factor ranking list, relevant intervention schemes are retrieved from the agronomic knowledge base. The agronomic knowledge base is a pre-set structured database containing rich agronomic management knowledge and intervention schemes. The main components of the knowledge base include: (1) crop physiological model library, containing growth response models of different crops under various conditions; (2) agronomic measures library, containing specific methods and parameters of various agronomic measures such as irrigation, fertilization, light regulation, temperature management, etc.; (3) case library, recording intervention cases and effects similar to the current situation in history; (4) expert rule library, containing experience rules and decision rules summarized by agricultural experts. The retrieval process uses fuzzy matching and similarity calculation methods to find the most relevant intervention scheme set for each limiting factor. For complex situations (multiple limiting factors exist simultaneously), a combination strategy is used to generate a comprehensive intervention scheme. The retrieval results form a set of optional intervention schemes, each containing intervention targets, specific measures, operation parameters (such as irrigation amount, fertilizer formula, light supplement intensity, etc.), applicable conditions, and expected effects, etc. The set of optional intervention schemes is evaluated in terms of resource input and expected output ratio. The evaluation includes two main aspects: cost evaluation and benefit evaluation. Cost evaluation considers various resource inputs, including: material resources (such as water, fertilizer, energy, etc.), equipment resources (such as irrigation equipment, light supplement equipment, etc.), human resources (such as operators, technicians, etc.), and time cost. Each resource is assigned a unit cost coefficient, and the total cost estimate is calculated. Benefit evaluation considers the expected output increase, including: yield improvement, quality improvement, production cycle shortening, and risk reduction, etc. The benefit estimate is based on the prediction results of the crop physiological model and the statistical analysis of historical cases. The input-output ratio (ROI, expected benefit divided by total cost) of each scheme is calculated as the key indicator of scheme benefit.Meanwhile, the difficulty, uncertainty, and environmental impact of each scheme are considered as non-quantitative factors, and weights are assigned by expert scoring. The overall benefit score of each scheme (range 0-100) is calculated by integrating all evaluation indicators. The final benefit score table of each intervention scheme is generated, which provides a quantitative basis for decision-making. Based on the benefit score table, the decision-making logic reasoning is performed to generate the preliminary intervention strategy recommendations. The decision-making logic reasoning adopts a combination of multi-criteria decision-making methods and expert systems. The multi-criteria decision-making methods include the analytic hierarchy process and the ideal solution method, which are used to rank schemes under multiple evaluation indicators. The decision-making process considers various factors: (1) benefit score, which tends to select the scheme with the highest overall benefit score; (2) risk balance, which seeks a balance between high benefits and low risks; (3) resource constraints, which ensure that the selected scheme can be implemented under current resource conditions; (4) time window, which considers the appropriate implementation time of intervention measures; (5) synergy between measures, which ensures that there is no conflict between multiple measures. The decision logic is expressed in a set of "if-then" rules, such as "if the water stress severity is greater than 80 and the available water resources are sufficient, then the precision irrigation scheme is preferred." The rule set is constructed by expert knowledge and machine learning methods, containing hundreds of decision rules for different situations. The current state and rule set are processed by rule-based reasoning methods (such as the Rete algorithm) to generate preliminary intervention strategy recommendations. The strategy recommendations include recommended combinations of intervention measures, specific parameters for each measure, implementation schedules, priority orders, and expected effects, providing a framework for subsequent scheme refinement. The feasibility and resource consumption of the preliminary intervention strategy recommendations are evaluated to further refine and optimize the intervention scheme. The feasibility evaluation is conducted from three aspects: technical feasibility, operational feasibility, and economic feasibility. Technical feasibility examines whether the recommended intervention measures are supported by current equipment and technology; operational feasibility assesses the difficulty of implementing measures and their impact on existing workflows; economic feasibility analyzes the balance between intervention costs and expected benefits. The resource consumption evaluation accurately calculates the demand for various resources, including water resources (cubic meters), fertilizer resources (kilograms, classified by components), energy resources (kilowatt-hours), and labor resources (man-hours). Based on the evaluation results, the preliminary strategy recommendations are adjusted and optimized, which may include narrowing the intervention range, reducing the intervention intensity, implementing in stages, or replacing resource-intensive measures. The optimized scheme is more practical and has a clear quantitative estimate of resource demand. The final agronomic measure execution scheme is formed, which includes specific operation guidelines, parameter settings, implementation schedules, and resource demand lists, and is a direct basis for subsequent resource allocation and operation execution. Based on the agronomic measure execution scheme, resource allocation optimization analysis is performed. The optimization goal is to maximize resource utilization efficiency and minimize resource waste and cost while meeting the demand for agronomic measures. The optimization method uses a combination of linear programming and heuristic algorithms.First, a mathematical model of resource allocation is established, the decision variable is the resource allocation amount of each execution unit, the constraint conditions include resource total amount constraint, time window constraint and technical parameter constraint, and the objective function is to minimize the total cost or maximize the resource utilization efficiency. For complex nonlinear problems, genetic algorithm or particle swarm optimization algorithm is used to solve the approximate optimal solution. The optimization process considers the space-time distribution characteristics of resources, such as the balance of water resources in different regions, the load balancing of energy resources in different time periods, etc. For critical resources (such as water resources), emergency reserve strategies are also considered to ensure sufficient resources to respond in emergency situations. The optimization results form resource allocation instructions, which include resource allocation amount, execution time arrangement and operation parameter setting of each execution terminal, and are transmitted to the execution terminal in the form of electronic instructions to ensure accurate execution and dynamic adjustment of agronomic measures, and realize fine management of smart agriculture.

[0074] The embodiment realizes precise monitoring and management of the whole life cycle of crops by collecting crop growth environment data in multiple dimensions. The root domain perception module realizes real-time monitoring of rhizosphere microenvironment through implantable microelectrode array, realizes accurate characterization of root-microbe interaction, and provides scientific basis for crop nutrient absorption. The crown microenvironment analysis module collects and quantifies leaf temperature and humidity gradient data through multi-point temperature and humidity sensors, accurately captures stomatal response changes, and realizes fine monitoring of plant water use efficiency. The light response analysis module identifies the spatio-temporal distribution characteristics of field light through a distributed spectral sensor network, realizes differential analysis of crop individual photoperiod sensitivity, and provides a basis for precise light management. The agronomic decision-making collaboration module correlates and analyzes multi-dimensional data to generate comprehensive evaluation data and precise execution plan, significantly improving the pertinence of agronomic measures and resource utilization efficiency. The system realizes the transition from single environmental factor monitoring to comprehensive physiological state evaluation, from passive response to active prediction and early warning, providing technical support for modern precision agriculture, effectively improving the scientific nature, accuracy and sustainability of agricultural production.

[0075] Embodiment 2;

[0076] Please refer to Figure 2 The embodiment does not describe part of the content in detail, see the description of embodiment 1, and provides a data acquisition and management method of a smart agriculture cloud platform based on Internet of Things, comprising:

[0077] S1, collecting plant rhizosphere microenvironment information to generate rhizosphere original data; performing soil electrochemical property and rhizosphere material exchange correlation analysis according to the rhizosphere original data to generate rhizosphere electrochemical interaction data;

[0078] S2, collect canopy microclimate information, generate canopy microclimate original data, perform temperature and humidity gradient quantification processing according to the canopy microclimate original data, and generate leaf temperature and humidity gradient data; perform stomatal dynamic response characteristic extraction based on the leaf temperature and humidity gradient data, and generate stomatal response characteristic data;

[0079] S3, collect field illumination information, generate multi-point spectral original data, perform multi-spectral distribution analysis based on the multi-point spectral original data, and generate field spectral distribution data; perform crop individual photoperiod sensitivity differentiation analysis on the spectral distribution data, and generate crop photoperiod sensitive spectrum;

[0080] S4, perform multi-dimensional correlation fusion on the rhizosphere electrochemical interaction data, the stomatal response characteristic data, and the crop photoperiod sensitive spectrum, and generate crop physiological state comprehensive evaluation data; perform agronomic measure analysis based on the crop physiological state comprehensive evaluation data, and generate an agronomic measure execution scheme; perform resource allocation optimization analysis according to the agronomic measure execution scheme, and generate a resource allocation instruction; and perform agricultural operation scheduling based on the resource allocation instruction.

[0081] The above merely describes preferred embodiments of the present application and is not intended to limit the present application. Although the foregoing embodiments of the present application have been described in detail, those skilled in the art can make modifications to the technical solutions described in the foregoing embodiments, or make equivalent replacements to some technical features, without departing from the spirit and principle of the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall fall within the scope of protection of the present application.

[0082] It should be noted that in this document, the terms "comprising", "containing", or any other variant thereof are intended to cover non-exclusive inclusion, so that processes, methods, articles, or devices that include a series of elements not only include those elements, but also include other elements not explicitly listed, or inherent to such processes, methods, articles, or devices. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article, or device that includes the element.

[0083] In the description of the present application, it should be understood that the terms "first", "second", etc. are only used for differentiation and description, and cannot be understood as indicating or implying relative importance.

[0084] In the description of the present application, unless otherwise specified, the meaning of "a plurality of" is two or more.

[0085] In the description of the present application, the meaning of "several" is one or more, and the meaning of "a large number" is two or more.

[0086] In the description of the specification, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example" or "some examples" etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are contained in at least one embodiment or example of the present application. In the specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in one or more embodiments or examples.

[0087] The formula in the specification is a dimensionless value calculation, the formula is obtained by collecting a large amount of data to simulate the formula of the nearest real situation, and the preset parameters and threshold values in the formula are set by the person skilled in the art according to the actual situation.

[0088] Although the embodiments of the present application have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, replacements and variations can be made to the embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the claims and their equivalents.

Claims

1. A data collection and management system for a smart agricultural cloud platform based on the Internet of Things, characterized in that: include: Root zone perception module: collects plant rhizosphere microenvironment information and generates rhizosphere raw data; Based on the original rhizosphere data, the correlation analysis between soil electrochemical characteristics and rhizosphere material exchange was performed to generate rhizosphere electrochemical interaction data; Canopy microenvironment analysis module: collects canopy microclimate information, generates canopy microclimate raw data, performs temperature and humidity gradient quantification based on the canopy microclimate raw data, and generates leaf temperature and humidity gradient data; Extract stomatal dynamic response characteristics based on leaf temperature and humidity gradient data to generate stomatal response characteristic data; Light response analysis module: collects field light information, generates multi-point spectral raw data, performs multi-spectral distribution analysis based on the multi-point spectral raw data, and generates field spectral distribution data; Perform differential analysis of crop individual photoperiod sensitivity on spectral distribution data to generate crop photoperiod sensitivity spectrum; Agronomic decision-making collaborative module: This module integrates rhizosphere electrochemical interaction data, stomatal response characteristic data, and crop photoperiod sensitivity spectrum into a multi-dimensional correlation to generate comprehensive assessment data on crop physiological status. Analyze agronomic measures based on comprehensive assessment data of crop physiological status and generate implementation plans for agronomic measures; Conduct resource allocation optimization analysis based on the agronomic measures implementation plan and generate resource allocation instructions; then schedule agricultural operations based on the resource allocation instructions.

2. The data acquisition and management system of the smart agricultural cloud platform based on the Internet of Things according to claim 1 is characterized in that: The plant rhizosphere microenvironment information is collected to generate rhizosphere raw data; Based on the original rhizosphere data, the soil electrochemical characteristics and rhizosphere material exchange correlation analysis are performed to generate rhizosphere electrochemical interaction data, including: The plant rhizosphere microenvironment information is collected through an implantable microelectrode array to generate rhizosphere raw data; Based on the original rhizosphere data, the correlation between root exudates and soil ion concentration changes was identified to generate root-soil ion exchange characteristic data; Based on the root-soil ion exchange characteristic data, a correlation map between rhizosphere microbial activity and conductivity was constructed to generate the rhizosphere microecological conductivity characteristic spectrum; Based on the rhizosphere microecological conductivity characteristic spectrum, the correlation analysis between soil electrochemical characteristics and rhizosphere material exchange was performed to generate rhizosphere electrochemical interaction data.

3. The data acquisition and management system of the smart agricultural cloud platform based on the Internet of Things according to claim 2 is characterized in that: The method of constructing a correlation map between rhizosphere microbial activity and electrical conductivity based on root-soil ion exchange characteristic data and generating a rhizosphere microecological electrical conductivity characteristic spectrum includes: Conductivity fluctuation patterns are identified in different time periods based on root-soil ion exchange characteristic data to generate a conductivity dynamic change sequence. The microbial community activity was deduced based on the dynamic change sequence of conductivity to generate a microbial activity index matrix; The microbial activity index matrix and the conductivity dynamic change sequence are temporally and spatially fused to generate a transient conductivity-microbial activity correspondence table; According to the transient conductivity-microbial activity correspondence table, a correlation map between rhizosphere microbial activity and conductivity was constructed to generate the rhizosphere microecological conductivity characteristic spectrum.

4. The data acquisition and management system of the smart agricultural cloud platform based on the Internet of Things according to claim 3 is characterized in that: The canopy microclimate information is collected to generate canopy microclimate raw data, and temperature and humidity gradient quantification processing is performed based on the canopy microclimate raw data to generate leaf temperature and humidity gradient data; Extract stomatal dynamic response characteristics based on leaf temperature and humidity gradient data to generate stomatal response characteristic data, including: Collect canopy microclimate information through multi-point miniature wireless temperature and humidity sensors to generate canopy microclimate raw data; Analyze the temperature and humidity differences between the leaf surface and the surrounding air based on the original canopy microclimate data to generate the leaf-air interface temperature and humidity gradient data; Analyze the spatiotemporal trend of leaf-air interface temperature and humidity gradient data to generate dynamic change data of leaf temperature and humidity gradient; Correct the dynamic change data of leaf temperature and humidity gradient according to the terrain and planting structure information to generate leaf temperature and humidity gradient data; Stomatal dynamic response characteristics are extracted based on leaf temperature and humidity gradient data to generate stomatal response characteristic data.

5. The data acquisition and management system of the smart agricultural cloud platform based on the Internet of Things according to claim 4 is characterized in that: The extraction of stomatal dynamic response characteristics based on leaf temperature and humidity gradient data to generate stomatal response characteristic data includes: Estimate transpiration rate based on leaf temperature and humidity gradient data to generate transpiration potential prediction data; The actual temperature distribution of the leaves is obtained through infrared thermal imaging technology, and a leaf temperature distribution map is generated; Compare and analyze the transpiration potential prediction data with the leaf temperature distribution map to generate stomatal opening and closing status inference data; The photosynthetic efficiency change trend is predicted based on the data inferred from the stomatal opening and closing status, and the stomatal response characteristic data is generated.

6. The data acquisition and management system of the smart agricultural cloud platform based on the Internet of Things according to claim 5 is characterized in that: The field illumination information is collected to generate multi-point spectral raw data, and multi-spectral distribution analysis is performed based on the multi-point spectral raw data to generate field spectral distribution data; Spectral distribution data is used to analyze the individual photoperiod sensitivity of crops and generate crop photoperiod sensitivity spectra, including: Collect field light information through a distributed spectral sensing network to generate multi-point spectral raw data; Perform spatiotemporal interpolation processing on multi-point spectral raw data to generate a continuous spectral distribution scene map; Based on the continuous spectral distribution scene map, the spatial heterogeneity of light intensity and light quality is analyzed to generate field spectral distribution data; The spectral distribution data are used to analyze the differences in the photoperiod sensitivity of individual crops and generate the crop photoperiod sensitivity spectrum.

7. The data acquisition and management system of the smart agricultural cloud platform based on the Internet of Things according to claim 6 is characterized in that: The step of performing differential analysis of individual crop photoperiod sensitivity on the spectral distribution data to generate a crop photoperiod sensitivity spectrum includes: Based on the field spectral distribution data and plant growth conditions, individual photosynthetic efficiency is evaluated to generate individual photosynthetic efficiency data; The correlation analysis between light conditions and growth performance was conducted based on individual photosynthetic efficiency data to generate a light-growth correlation matrix; The light-growth correlation matrix was clustered based on individual differences within varieties to generate photoperiod response grouping data; Based on the photoperiod response grouping data, an individualized identification standard for crop photoperiod sensitivity is constructed to generate a crop photoperiod sensitivity spectrum.

8. The data acquisition and management system of the smart agricultural cloud platform based on the Internet of Things according to claim 7 is characterized in that: The multi-dimensional correlation fusion of rhizosphere electrochemical interaction data, stomatal response characteristic data, and crop photoperiod sensitivity spectrum is performed to generate comprehensive crop physiological status assessment data, including: Perform time-series alignment processing on rhizosphere electrochemical interaction data, stomatal response characteristic data, and crop photoperiod sensitivity spectrum to generate multi-dimensional time-series synergistic data; Conduct physiological process interaction analysis on multi-dimensional time series collaborative data to generate a physiological process coupling relationship map; Based on the coupled relationship map of physiological processes, the current growth status and potential stress risks of crops are evaluated, and comprehensive assessment data of crop physiological status is generated.

9. The data acquisition and management system of the smart agricultural cloud platform based on the Internet of Things according to claim 8 is characterized in that: The agronomic measures analysis is performed based on the comprehensive evaluation data of the crop physiological status to generate an agronomic measures implementation plan; Conduct resource allocation optimization analysis based on agronomic measures implementation plan and generate resource allocation instructions; Agricultural operation scheduling based on resource allocation instructions, including: Based on the comprehensive assessment data of crop physiological status and the agronomic knowledge base, decision-making logic reasoning is carried out to generate preliminary intervention strategy recommendations; Conduct feasibility and resource consumption assessments on preliminary intervention strategy recommendations and generate agronomic implementation plans; Conduct resource allocation optimization analysis based on agronomic measures implementation plan and generate resource allocation instructions; The IoT execution terminal receives resource allocation instructions and performs precise irrigation, fertilization and light supplement operations.

10. The data acquisition and management system of the smart agricultural cloud platform based on the Internet of Things according to claim 9 is characterized in that: The above mentioned comprehensive assessment data of crop physiological status is combined with the agronomic knowledge base to perform decision-making logic reasoning and generate preliminary intervention strategy recommendations, including: Identify current crop growth limiting factors based on comprehensive crop physiological status assessment data and generate a ranked list of limiting factors; Relevant intervention plans in the agronomic knowledge base are retrieved according to the sorted list of restriction factors to generate a set of optional intervention plans; Evaluate the resource input and expected output ratio of the optional intervention plan set and generate a plan benefit score sheet; Based on the program benefit score sheet, decision-making logic reasoning is performed to generate preliminary intervention strategy recommendations.

Images