Intelligent agricultural environment data real-time analysis method based on Internet of Things

Abstract

The invention belongs to the technical field of smart agriculture, and relates to a smart agriculture environment data real-time analysis method based on Internet of Things, which comprises the following steps: acquiring soil parameters of a first detection layer and a second detection layer of a farmland soil layer, and calculating a soil porosity compensation coefficient by combining a pressure difference value between the two layers, a soil convection measurement value is generated by combining the temperature difference value between the two layers; matching a preset growth period characteristic database to generate a growth period fingerprint code; constructing a dynamic data matrix containing the temperature gradient and the humidity change rate, giving a weighting coefficient to the time in combination with the growth period fingerprint code, and generating a smooth trend fluctuation curve; generating a three-dimensional imbalance index with economic weight adjustment; and selecting an irrigation mode according to the three-dimensional imbalance index hierarchy, and dynamically adjusting the communication priority of the equipment. According to the method, a hard coding decision-making mechanism which easily ignores market price factors and economic weights is solved, so that a resource allocation strategy deviates from actual benefit requirements.

Description

Technical Field

[0001] The present invention belongs to the technical field of smart agriculture and relates to a real-time analysis method for smart agricultural environmental data based on the Internet of Things. Background Art

[0002] Agricultural environmental monitoring systems based on the Internet of Things (IoT) face significant bottlenecks in real-time data processing and analysis. They typically use discrete sensors to collect single-dimensional environmental parameters, but the ability to concurrently process heterogeneous data from multiple sources is insufficient, resulting in delayed integration of key indicators. In complex farmland scenarios, cross-domain correlation analysis of soil thermal convection dynamics, crop growth stage identification, and market price fluctuations is difficult to achieve simultaneously. This not only hinders the timely generation of environmental control instructions but also easily leads to deviations in irrigation timing and resource mismatches.

[0003] Traditional solutions rely on offline batch processing architectures, using fixed thresholds to trigger basic operations like irrigation and ventilation. Common approaches include building static environmental models based on historical averages or employing single-factor decision trees for device control. Some systems attempt to incorporate sliding window mechanisms for time series data, but these fail to dynamically adjust analysis weights based on crop growth characteristics. This results in insufficiently smooth prediction curves and an inability to accurately capture the risk of imbalances caused by sudden changes in temperature and humidity during the day.

[0004] Based on the above problems, existing technologies have multi-dimensional defects due to algorithm limitations, and are prone to ignoring the hard-coded decision-making mechanism of market price factors and economic weights, causing resource allocation strategies to deviate from actual benefit needs and making it difficult to support the real-time operation needs of large-scale smart agricultural scenarios. Summary of the Invention

[0005] In order to solve the above problems, the present invention provides a real-time analysis method for smart agricultural environmental data based on the Internet of Things.

[0006] In a first aspect, the present invention provides a method for real-time analysis of smart agricultural environmental data based on the Internet of Things, which adopts the following technical solutions:

[0007] A real-time analysis method for smart agricultural environmental data based on the Internet of Things includes the following steps:

[0008] S1. Obtain soil parameters of the first detection layer and the second detection layer of the farmland soil layer, calculate the soil porosity compensation coefficient based on the pressure difference between the two layers, and generate soil convection measurement values ​​based on the temperature difference between the two layers;

[0009] S2. Capture crop canopy images using a multi-channel image sensor, extract vein clarity and chlorophyll reflectance intensity to form a feature vector, and match it with a preset growth period feature database to generate a growth period fingerprint code;

[0010] S3, constructing a dynamic data matrix including temperature gradient and humidity change rate, assigning a time weighting coefficient in combination with the growth period fingerprint code, and generating a smooth trend fluctuation curve;

[0011] S4. Compare soil convection measurements with the current dynamic environmental baseline, introduce temperature difference rate parameters and market price fluctuation factors, and generate a three-dimensional imbalance index with economic weight adjustment;

[0012] S5. Call the action plan library to match the quick response or gradual adjustment strategy, select the irrigation mode according to the three-dimensional imbalance index level, and dynamically adjust the device communication priority.

[0013] A further embodiment of the present invention, step S1, comprises the following steps:

[0014] Obtain soil parameters of the first detection layer and the second detection layer of the farmland soil layer;

[0015] The initial contact pressure value of the first detection layer and the stable contact pressure value of the second detection layer are measured respectively, and the soil porosity compensation coefficient is calculated according to the difference between the initial contact pressure value and the stable contact pressure value;

[0016] The temperature difference between the first detection layer and the second detection layer is obtained, and the soil convection measurement value is generated by combining the soil porosity compensation coefficient.

[0017] A further embodiment of the present invention combines a soil porosity compensation coefficient to generate a soil convection measurement value, comprising the following steps:

[0018] The temperature difference between the first detection layer and the second detection layer is multiplied by the absolute value of the soil porosity compensation coefficient to obtain the current thermal gradient;

[0019] The current thermal gradient is corrected based on the empirical conductivity of the soil medium to generate soil convection measurements.

[0020] A further embodiment of the present invention, step S2, comprises the following steps:

[0021] Collect visible and near-infrared images of crop canopies under constant illumination conditions;

[0022] Edge enhancement processing was performed on the visible light band image, and the ratio of the total length of leaf vein pixels to the leaf outline area was calculated as the leaf vein clarity;

[0023] Analyze the chlorophyll reflectance intensity in a specific wavelength range of near-infrared band images;

[0024] The normalized vein clarity and chlorophyll reflectance intensity form a two-dimensional feature vector.

[0025] A further solution of the present invention is to match a preset growth period feature database to generate a growth period fingerprint code, comprising the following steps:

[0026] The Mahalanobis distance is calculated based on the cluster center of the feature vector and the growth period feature database;

[0027] The growth phase discrimination results were obtained based on the minimum Mahalanobis distance matching;

[0028] If the change rate of chlorophyll reflectance intensity in multiple consecutive collection cycles exceeds the maturity threshold, it triggers the dynamic update of the maturity fingerprint code.

[0029] A further embodiment of the present invention, step S3, comprises the following steps:

[0030] Collect temperature gradient, humidity change rate, and light intensity within a predetermined time window to form a dynamic data matrix;

[0031] An exponential decay time weight is applied to each time node data, and the time weight decreases as the interval between the acquisition time and the current moment increases;

[0032] According to the growth period fingerprint code query time sensitivity coefficient, adjust the curvature parameter of dynamic data;

[0033] The trend fluctuation curve with confidence mark is generated by cubic spline interpolation and Gaussian filtering.

[0034] A further embodiment of the present invention, step S4, comprises the following steps:

[0035] Based on the current dynamic environmental baseline, calculate the minute-by-minute rate of change of the current temperature difference relative to the historical mean;

[0036] Access market price data streams and calculate price fluctuations to generate economic adjustment coefficients;

[0037] The soil porosity compensation coefficient, temperature difference rate parameter, and economic adjustment coefficient are given independent weights and then multiplied together to generate a three-dimensional imbalance index.

[0038] A further embodiment of the present invention generates an economic adjustment coefficient, comprising the following steps:

[0039] Obtain market price data of target agricultural products in real time and calculate daily price fluctuations;

[0040] If the price volatility exceeds the preset threshold, the sensitivity weight of the temperature parameter is logarithmized and normalized to generate an economic weight correction factor.

[0041] A further embodiment of the present invention, step S5, comprises the following steps:

[0042] Matching quick response plans or gradual adjustment plans according to the numerical range of the three-dimensional imbalance index;

[0043] If the index exceeds the limit threshold, the impact spraying strategy is called and the irrigation equipment is operated at full power;

[0044] If the index is in the middle range, start pulse spraying and adjust the water release rate in stages.

[0045] A further solution of the present invention dynamically adjusts the device communication priority, including the following steps:

[0046] Monitor the temperature ratio in environmental parameters. If it exceeds the preset threshold, increase the control priority of the ventilation equipment; generate a sequence of equipment collaborative work instructions, and terminate the operation of non-core equipment according to the urgency of the three-dimensional imbalance index.

[0047] In summary, the present invention has the following beneficial technical effects:

[0048] 1. Through multi-dimensional sensor linkage analysis of the soil pressure and temperature difference between the first and second detection layers, combined with the dynamic insertion mechanism of the double-helix probe, the soil porosity compensation coefficient and convection measurement value are accurately calculated, effectively identifying internal heat flow and structural changes in the soil. Multispectral imagery is used to extract the vein clarity and chlorophyll reflectance intensity of the crop canopy and generate a growth period fingerprint, enabling a deep integration of environmental monitoring data with the crop physiological status.

[0049] 2. Innovatively introduce market price volatility factors and economic weight correction parameters to dynamically link agricultural product trading data with the three-dimensional imbalance index. Real-time calculation of price volatility adjusts the sensitivity threshold of irrigation inputs, forming a flexible regulatory mechanism based on market benefits. When the economic value of crops plummets, the priority of monitoring non-core parameters is automatically reduced, establishing a quantitative correlation between environmental control costs and agricultural product returns, avoiding the waste of resources caused by traditional methods that rely solely on environmental thresholds.

[0050] 3. The dynamic communication priority algorithm analyzes parameters such as temperature ratio to build a coordinated execution sequence for equipment such as ventilation, irrigation, and lighting. This mechanism prioritizes the full operation of core equipment in scenarios with sudden high temperature increases or soil drying, reducing redundant energy consumption by up to 35% and significantly improving the overall operational efficiency of agricultural facilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. The drawings are used to provide a further understanding of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0052] Figure 1 A flowchart of a real-time analysis method for smart agricultural environmental data based on the Internet of Things is disclosed.

[0053] Figure 2 A structural schematic diagram of a real-time analysis system for smart agricultural environmental data based on the Internet of Things is disclosed. DETAILED DESCRIPTION

[0054] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0055] The following is combined with Figure 1-Figure 2 The preferred embodiments of the present invention are described in detail.

[0056] Refer to the attached Figure 1 The present invention proposes a real-time analysis method for smart agricultural environmental data based on the Internet of Things, comprising the following steps:

[0057] S1. Obtain soil parameters of the first detection layer and the second detection layer of the farmland soil layer, calculate the soil porosity compensation coefficient based on the pressure difference between the two layers, and generate soil convection measurement values ​​based on the temperature difference between the two layers;

[0058] S2. Capture crop canopy images using a multi-channel image sensor, extract vein clarity and chlorophyll reflectance intensity to form a feature vector, and match it with a preset growth period feature database to generate a growth period fingerprint code;

[0059] S3, constructing a dynamic data matrix including temperature gradient and humidity change rate, assigning a time weighting coefficient in combination with the growth period fingerprint code, and generating a smooth trend fluctuation curve;

[0060] S4. Compare soil convection measurements with the current dynamic environmental baseline, introduce temperature difference rate parameters and market price fluctuation factors, and generate a three-dimensional imbalance index with economic weight adjustment;

[0061] S5. Call the action plan library to match the quick response or gradual adjustment strategy, select the irrigation mode according to the three-dimensional imbalance index level, and dynamically adjust the device communication priority.

[0062] In one embodiment of the present invention, step S1 includes the following steps:

[0063] A double-helix probe is set up and inserted into the soil layer of the farmland in an alternating rotation manner. The first detection layer is formed at a depth of 0-5 cm in the soil layer, and the second detection layer is formed at a depth of 15-20 cm in the soil layer;

[0064] Specifically, the double helix probe has two sets of independent spiral probes. The motor of the double helix probe drives the independent spiral probes, which are inserted into the soil layer of the farmland in an alternating rotation mode of first rotating 180 degrees clockwise and then rotating 90 degrees counterclockwise, so as to avoid soil particles blocking the gaps of the spiral probes.

[0065] The soil layer at a depth of 0-5 cm refers to the cultivated layer, the active soil contact surface. The first detection layer, formed primarily by the circumferential cutting force of the spiral blades on the upper portion of the double-helix probe, captures temperature and humidity changes caused by surface evaporation. The soil layer at a depth of 15-20 cm represents the primary root distribution area for most crops. The second detection layer, formed primarily by the penetration of the tapered threads on the lower portion of the double-helix probe, monitors the absorption characteristics of crop roots. A soil heat conduction barrier is located between the first and second detection layers.

[0066] For the detection layers of the first detection layer and the second detection layer, the initial contact pressure value and the stable contact pressure value of each detection layer are measured respectively, and the difference between the initial contact pressure value and the stable contact pressure value of each detection layer is converted into a soil porosity compensation coefficient;

[0067] Specifically, the initial contact pressure is the instantaneous peak resistance value when the tip of the double-helix probe first touches the soil layer; the stable contact pressure is the average energy consumption of the double-helix probe tip during the continuous downward pressure to the target depth. Both the initial contact pressure and the stable contact pressure are recorded by the embedded piezoelectric film sensor built into the double-helix probe. The difference between the two pressure values ​​reflects the internal pore structure characteristics of the soil, and after normalization, the soil porosity compensation coefficient is obtained.

[0068] Among them, the soil porosity compensation coefficient is used to characterize the gas and water migration capacity of the farmland soil layer. The calculation of the soil porosity compensation coefficient satisfies the following formula:

[0069]

[0070] Among them, K p represents the soil porosity compensation coefficient; P i Indicates the initial contact pressure value; P s Indicates the stable contact pressure value; P cal Indicates the probe calibration reference value of the double helix probe, the result of laboratory measurement. i >P s , indicating K p A positive value indicates that the first detection layer of the farmland soil is compact but the second detection layer is loose.

[0071] For example, in actual measurement of sandy loam, the embedded piezoelectric film sensor built into the double helical probe is set to record the P i =42kPa, P s =28kPa, laboratory measurement probe calibration reference value P cal =50kPa;

[0072] Calculation shows that K p =(42-28) / 50=0.28, indicating that K p A positive value indicates that the first test layer of the farmland soil layer has hardened and formed a crust due to sunlight, but the second test layer of the farmland soil layer still maintains good air permeability.

[0073] The soil convection measurement value is obtained by combining the temperature difference between the first detection layer and the second detection layer and the obtained porosity compensation coefficient; the limit threshold of the soil convection measurement value is pre-set to determine whether the heat flow inside the soil poses a hazard;

[0074] Specifically, the temperature difference, the thermal gradient between the monitoring points of the first detection layer and the second detection layer of the double helix probe, the product of the temperature difference between the first detection layer and the second detection layer and the porosity compensation coefficient are used to obtain the soil convection measurement value;

[0075] Among them, the soil convection measurement value is used to quantify the comprehensive parameter of the energy flow rate within the farmland soil layer. The calculation of the soil convection measurement value satisfies the following formula:

[0076] Q c =|ΔT|×K p ×δ

[0077] Among them, Q c is the soil convection measurement; K p is the soil porosity compensation coefficient, which takes the absolute value when it is negative; δ represents the empirical conductivity coefficient of sand-clay mixed medium, which is the result of experimental determination;

[0078] ΔT represents the temperature difference between the first detection layer and the second detection layer, which is converted into the soil thermal convection intensity index after being corrected by the porosity compensation coefficient, satisfying the following formula:

[0079] ΔT=T1-T2

[0080] Among them, T1 represents the temperature of the first detection layer, reflecting the heat exchange state between the surface and the atmosphere;

[0081] T2 represents the temperature of the second detection layer, which characterizes the energy balance in the root activity area.

[0082] For example, in a corn field at noon, the double helix probe measured T1 = 32.5°C, T2 = 27.8°C, and the calculated ΔT = 4.7°C. The calculation formula for the soil porosity compensation coefficient is K p =(42-28) / 50=0.28; δ is 0.17; Q c =4.7×0.31×0.17≈0.25W / m 2 , set the soil convection measurement limit threshold to 0.3W / m 2 , the soil convection measurement value did not exceed the limit threshold, indicating that although there was high temperature on the surface, the heat flow inside the soil did not reach the critical point that would harm crop growth.

[0083] In one embodiment of the present invention, step S2 includes the following steps:

[0084] S21. Take crop canopy images at a fixed time every morning, segment leaf contours, and extract texture density distribution features.

[0085] Specifically, under constant light conditions between 1 and 2 hours after sunrise, crop canopy image data is continuously collected using a multi-channel image sensor equipped with a dual-spectral imaging module. Natural light is stable in the 450-600 lumens range. The multi-channel image sensor simultaneously captures visible and near-infrared wavelengths. The visible light band is used to identify leaf vein clarity, while the near-infrared band is used to analyze the cellular structure of the mesophyll.

[0086] After Gaussian filtering and noise reduction of the original image, an improved watershed algorithm was used to delineate leaf boundaries. The key was to set a dynamic gradient threshold to prevent over-segmentation. Texture density distribution features were calculated based on a 64×64 pixel sampling window. The consistency of the grayscale gradient direction of each cell was calculated to form a two-dimensional matrix reflecting the degree of wrinkling on the leaf surface.

[0087] Among them, the improved watershed algorithm is an image segmentation method based on the combination of region growing and edge detection. When setting the initial seed point, a color similarity constraint is introduced to avoid misidentifying the soil background as the leaf area.

[0088] The texture density distribution characteristics represent the morphological quantitative indicators of the leaf surface microstructure within a unit area. The high-density area corresponds to the densely distributed areas of leaf veins, and the low-density area represents the flat area of ​​mesophyll tissue.

[0089] For example, during the podding period of soybean plants, a multi-channel image sensor automatically photographs the canopy 30 cm below the top of the plant between 1 and 2 hours after sunrise. The processed images separate the outlines of 63 effective leaves, and the average texture density of the leaves in the southeast quadrant of the soybean plant is calculated to be 42% higher than that of the leaves in the northwest quadrant.

[0090] S22, the leaf vein clarity in the visible light band and the chlorophyll reflectance intensity in the near-infrared band, constitute a feature vector;

[0091] Specifically, the visible light band of the segmented leaf area uses the green light channel of 530nm to 570nm, the Laplace operator is used to enhance the sharpness of the leaf vein edge, and the ratio of the leaf vein length to the leaf area is calculated as the clarity index;

[0092] The near-infrared wavelength range of 720nm to 750nm was selected, and the reflectance intensity was calculated based on the integral value of the albedo of the chlorophyll absorption peak. The vein clarity in the visible light band and the chlorophyll reflectance intensity in the near-infrared band were normalized to form a two-dimensional feature vector, which was mapped to a preset variety feature space for coordinate positioning.

[0093] Among them, vein clarity is the quantitative value of the visual recognition of primary and secondary veins in visible light images. Chlorophyll reflectance intensity is the total reflected energy of mesophyll tissue to photons of a specific wavelength in the near-infrared band.

[0094] Multi-dimensional data fusion is achieved through coordinate system projection, and the eigenvector is calculated to satisfy the following formula:

[0095]

[0096] Among them, F v represents the eigenvector, L v represents the total length of leaf vein pixels, A l represents the blade outline area, and R(λ) represents the reflectivity value at wavelength λ.

[0097] For example, during the growth period of new shoots of apple trees, the visible light band analysis L v =86.3mm, A l =12.5cm 2 , calculated that L v / A l =6.9, and the reflection intensity in the near-infrared band analysis from 720nm to 750nm is 224.7, forming a characteristic vector [6.9, 224.7], which successfully matches the standard vector of the apple tree's extension period [5.8±0.5, 210±15].

[0098] S23, matching the maturity threshold preset in the growth model database, triggering the fingerprint code update when the chlorophyll reflection intensity exceeds the maturity threshold;

[0099] Specifically, historical data is collected to construct a growth model database for multiple crop varieties. For each variety, reference value intervals are set for six key stages, including the tillering, heading, and filling stages. The growth model database pre-stores the boundary conditions of the feature vectors for each crop variety at different growth stages. The Mahalanobis distance algorithm is used to calculate the similarity between the current feature vector and the cluster center of each growth stage. The growth stage corresponding to the minimum distance is selected as the discrimination result, and the growth stage fingerprint code is generated based on the discrimination results.

[0100] The historical data collection represents a benchmark dataset of variety development stages calibrated through field trials. The Mahalanobis distance algorithm is a multivariate statistical distance calculation method that considers inter-feature covariance, effectively eliminating the impact of parameter dimensionality differences. The growth stage fingerprint is a 16-bit code sequence used to uniquely identify a crop's growth stage, with each bit corresponding to a specific combination of physiological characteristics.

[0101] The chlorophyll activity coefficient change rate between two consecutive collection cycles is continuously monitored. If the chlorophyll activity coefficient change rate exceeds the maturity threshold of 15% for three consecutive cycles, a binary bit flip update of the growth period fingerprint code is triggered. The updated growth period fingerprint code adopts a multi-layer encoding structure: the first byte marks the main growth period code, the second byte records the transition state flag, and the third byte stores the feature vector offset;

[0102] The eigenvector offset represents the percentage deviation between the current parameters and the standard growth period parameters. The transition state flag identifies the specific physiological state during the growth period transition, such as the hormone level changes during the transition from flowering to fruiting. The maturity threshold is the critical point of chlorophyll activity decay derived from historical data.

[0103] For example, during the color change period monitoring of grapes, the chlorophyll reflectance intensity value gradually decreases from 185 to 168. When the third detection value falls below the preset maturity threshold of 170, the growth period fingerprint code is updated to a combination code of the maturity flag F1 and the offset +12%.

[0104] The beneficial effects of this step: leaf vein analysis in the visible light channel reveals the level of organ development, and chlorophyll monitoring in the near-infrared channel reflects the degree of metabolic activity. The spatiotemporal coupling of the two effectively eliminates the risk of single-factor misjudgment.

[0105] In one embodiment of the present invention, step S3 includes the following steps:

[0106] Collect environmental data from the most recent time period and create a dynamic matrix containing temperature gradient and humidity change rate every hour;

[0107] Specifically, we collected 240 hours of farmland environmental monitoring data over 10 consecutive days, using each natural hour as the smallest unit to construct a time series. For each time node, we extracted two parameters: temperature gradient and humidity change rate. The temperature gradient represents the difference between the highest and lowest temperatures at that time node, indicating the severity of heat distribution; the humidity change rate represents the percentage change in soil moisture content at the current time node compared to the data from the previous hour.

[0108] The temperature gradient, humidity change rate, original temperature, and light intensity are combined into a four-dimensional data vector, which is then arranged in chronological order to form a dynamic data matrix. The dynamic data matrix represents a multidimensional array structure arranged in an ordered manner according to a fixed time window. Each row of the dynamic data matrix corresponds to a collection moment, and each column of the dynamic data matrix records the characteristic values ​​of different environmental parameters.

[0109] Matrix M∈R N×4 Represents the dynamic data matrix of the last 240 hours, where N = 240 is the length of the time series (one data point per hour), and each row of the dynamic data matrix corresponds to the four-dimensional environmental feature vector at a specific moment.

[0110]

[0111] Where M is the dynamic data matrix; ΔT N is the temperature gradient at hour N; Indicates the humidity change rate in the Nth hour, reflecting the increase or decrease trend of soil moisture content; Indicates the average temperature in hour N; L N Indicates the light intensity at hour N.

[0112] For example, take the monitoring data of a strawberry greenhouse for 3 consecutive hours as an example.

[0113] At 08:00, the temperature gradient was 2.8°C (minimum 26.3°C / maximum 29.1°C), the humidity change rate was 4.17% (soil moisture content 60.0% in the previous hour / soil moisture content 62.5% in the current hour), the average temperature was 27.7°C, and the light intensity was 13,500 lux.

[0114] At 09:00, the temperature gradient was 3.5°C (minimum 27.9°C / maximum 31.4°C), the humidity change rate was -6.88% (soil moisture content 62.5% in the previous hour / 58.2% in the current hour), the average temperature was 29.6°C, and the light intensity was 21,200 lux.

[0115] At 10:00, the temperature gradient was 4.2°C (minimum 29.5°C / maximum 33.7°C), the humidity change rate was -5.33% (soil moisture content 58.2% in the previous hour / soil moisture content 55.1% in the current hour), the average temperature was 31.6°C, and the light intensity was 28,800 lux.

[0116] The corresponding expression of the dynamic data matrix M is,

[0117]

[0118] The beneficial effects of this verification example are: it accurately reflects the coordinated evolution characteristics of temperature, humidity, and light in the form of a matrix, with the temperature gradient increasing from 2.8°C to 4.2°C and the humidity change rate changing from a positive value to a continuously negative value, identifying the trend of the environment transitioning from a humid and stable environment to a high-temperature and dry environment; combined with the longitudinal comparison of each column of data, it quickly locates the associated nodes of temperature mutation and humidity drop.

[0119] Assign a time decay factor to each data point, and the value closer to the current moment has a higher weight;

[0120] Specifically, the sliding window algorithm dynamically screens the data set. The window length is set to 24 hours, and the sliding step is 1 hour each time. Each sliding window introduces an exponential decay weight function. For the data vector at each time point of the dynamic data matrix M, a time-varying weight coefficient is applied. The calculation of the weight coefficient satisfies the following formula:

[0121]

[0122] Among them, w t represents the time decay factor; λ represents the decay rate adjustment factor, which is used to control the steepness of the weight change. The larger the value, the faster the historical data decays; T curret Represents the timestamp of the current moment; t represents the timestamp corresponding to the data point; time decay factor, a time correlation weight distribution mechanism based on an exponential function.

[0123] For example, when processing the detection at the current time 12:00, the time difference with the data point at 11:00 the previous day is T curret -t=25 hours; the default value of the decay rate adjustment factor λ is 0.05.

[0124] w t =e -0.05×25 ≈0.286

[0125] When processing the detection at the current time 12:00, the time difference with the data point at 23:00 the previous day is T curret -t=13 hours; the default value of the decay rate adjustment factor λ is 0.05.

[0126] w t =e -0.05×13 ≈0.522

[0127] Combining the two sets of experimental data, each data point is assigned a time decay factor, and the value closer to the current moment has a higher weight.

[0128] The calculation of the environmental reference value for each sliding window satisfies the following formula:

[0129]

[0130] Among them, x t represents the actual measured value of the environmental parameter corresponding to the tth hour; w t represents the time attenuation factor at hour t; η is the dynamic environmental baseline value, which represents the dynamic average level of environmental parameters in the current time window.

[0131] For example, set the temperature monitoring value of a greenhouse for 3 consecutive hours.

[0132] Hour 1: T curret -t=3,λdefault 0.12,x1=25.3℃,calculated w1=0.697;

[0133] Hour 2: T curret -t=2,λdefault 0.12,x2=26.1℃,calculated w2=0.787;

[0134] Hour 3: T curret -t=1,λdefault 0.12,x3=27.4℃,calculated w3=0.887;

[0135] Calculate the current dynamic environmental reference value, η = 26.5°C, which represents the dynamic average level of the environmental parameters in the current time window.

[0136] Combined with the time sensitivity coefficient corresponding to the growth period fingerprint code, the curvature of the trend fluctuation curve is corrected;

[0137] Specifically, the growth period fingerprint of the historical data is extracted by combining the historical record data. The growth period intensity coefficient and the time sensitivity coefficient are manually matched for the growth period fingerprint. The matched data set is composed of a sensitivity parameter comparison set, and a sensitivity parameter comparison table is constructed based on the sensitivity parameter comparison set. After completing the weighted processing, the generated growth period fingerprint is called and the corresponding time sensitive parameter is queried in the predefined sensitivity parameter comparison table.

[0138] Among them, the time sensitivity coefficient represents the adjustment coefficient that quantifies the adaptability of crops to environmental fluctuations at different growth stages, such as the sensitivity weight of increasing temperature fluctuations during the flowering period.

[0139] Correct the curvature of the trend fluctuation curve to satisfy the following formula:

[0140] K adj =K ori ×(1+α×f gc )

[0141] Among them, K adj represents the corrected curvature; K ori represents the curvature of the original trend line; α represents the growth period intensity coefficient, which is used to control the upper limit of the correction amplitude; f gc Indicates the stage sensitivity value corresponding to the fingerprint code.

[0142] For example, if the crop is detected to be in the mature stage, the sensitivity parameter comparison table shows that α=0.3, f gc =0.8; if the original curvature K ori =1.276, adjusted K adj =1.276×(1+0.3×f gc 0.8)=1.581, the curvature of the trend line increases by 15%, which is closer to the sensitivity of mature crops to sudden temperature rise.

[0143] The corrected data is input into a Gaussian filter processor, and after eliminating occasional noise interference, a smooth and continuous trend fluctuation curve is generated using the cubic spline interpolation method. The peaks and troughs of the curve correspond to the periodic change characteristics of the environmental parameters, and the curvature radius parameter of the sampling point reflects the severity of the data change. The final output trend curve is accompanied by a time label and a credibility mark. For example, the credibility interval of the baseline value at 5:00 a.m. every day is marked as ±0.5%.

[0144] In one embodiment of the present invention, step S4 includes the following steps:

[0145] S41. Calculate the per-minute rate of change between the current temperature difference and the historical average for the same period. If the rate of change exceeds 2°C / h, a rate compensation correction mechanism is triggered.

[0146] Specifically, every minute, the temperature of the current first detection layer and the historical reference value of the average temperature of the same period in the past five years are obtained, and the instantaneous temperature difference between the two is calculated; the instantaneous temperature difference between the two is substituted into the moving time window, and the calculation of the temperature change rate satisfies the following formula:

[0147] V=ΔT / Δt

[0148] Here, ΔT represents the difference between the current temperature and the historical baseline, quantifying the rate at which the current temperature deviates from the historical baseline and reflecting the urgency of the environmental deterioration. Δt represents the time span of the 10-minute sliding window. When V ≥ 2°C / h, the rate compensation correction mechanism is automatically activated, injecting a dynamic tolerance factor to slow down the control response and avoid misjudgments caused by short-term, sharp fluctuations.

[0149] For example, a greenhouse was suddenly attacked by hot air between 11:00 and 11:10, and the temperature change rate was measured to be 2.3°C / h. After the rate compensation correction mechanism was triggered, the original control strategy was maintained. The temperature naturally dropped within the next 15 minutes, avoiding the invalid start-up of the sprinkler system and preventing energy waste and crop damage caused by frequent malfunctions of the equipment.

[0150] S42. Receive market price data streams. When the price of a specific agricultural product drops by more than a threshold value, automatically reduce the sensitivity weight of the temperature parameter to save energy and activate the economic optimization mode.

[0151] Specifically, by accessing the real-time data interface of the agricultural futures exchange, the latest transaction price of the target crop category is extracted; the sensitivity weight represents the influence coefficient of the temperature parameter in the three-dimensional imbalance index, and by reducing the weight, a higher temperature deviation is tolerated. The calculation of price changes satisfies the following formula:

[0152]

[0153] Among them, P change Indicates price change, P curret Indicates the current price, P base Indicates the benchmark price for the same period of the previous day.

[0154] For example, by accessing the real-time data interface of the agricultural futures exchange, the price of tomatoes fell by P in a single day. change =7.2%. When the unit price of specific agricultural products drops by more than 5%, the temperature sensitivity weight is reduced from 0.8 to 0.65. The 35.5℃ scenario that originally required a cooling trigger is now allowed to continue for 3 hours, saving 38 kWh of energy.

[0155] S43. The weight product of soil, air, and economy is used as the three-dimensional imbalance index;

[0156] Specifically, the porosity compensation coefficient, a microstructural indicator calculated by comparing the pressure difference before and after soil probe insertion, reflects the actual water and air permeability of the root zone, and derives the soil activity index from the porosity compensation coefficient. Temperature differential rate compensation is a dynamic adjustment mechanism activated when the temperature difference exceeds a specific threshold. It is used to prevent misjudgments caused by sensor noise interference and extract the temperature differential rate compensation value of air parameters. The economic adjustment coefficient is a parameter correction factor automatically adjusted based on the real-time price increase of agricultural products on the market trading platform. It is calculated by converting the price fluctuation range into a logarithmic function and then normalizing it. The economic adjustment coefficient is generated based on market price fluctuations.

[0157] The soil activity index, temperature difference rate compensation value, and economic adjustment coefficient are assigned independent weights and then multiplied to meet the following formula:

[0158] K=S w ×A w ×M w

[0159] Among them, K is the three-dimensional imbalance index, S w is the weight value corresponding to the soil activity index, A w is the weight value corresponding to the temperature difference rate compensation value, M w It represents the weight value corresponding to the economic adjustment coefficient. The range of each weight value is limited to the interval of 0.5 to 1.5. The interference caused by oversensitivity or slowness of a single parameter is eliminated through multiplication operation.

[0160] For example, the soil activity index dropped to 0.8 due to continued drought, the air temperature difference rate compensation value was detected to rise to 1.2 due to a sudden drop in temperature, and the tomato market price drop triggered the economic adjustment coefficient to be adjusted to 1.1. It was calculated that K = 0.8 × 1.2 × 1.1 = 1.056.

[0161] In one embodiment of the present invention, step S5 includes the following steps:

[0162] A pre-set action plan library includes quick response plans and gradual adjustment plans. When the three-dimensional imbalance index exceeds the set limit threshold, the quick response plan will be called first;

[0163] Specifically, a set of control strategies with two different response modes, including a fast response plan and a gradual adjustment plan;

[0164] The rapid response plan is a set of control solutions for sudden environmental crises, including full-power operation instructions for equipment and cross-module forced synchronization protocols; the gradual adjustment plan is a set of control solutions for periodic parameter fluctuations, using a phased progressive adjustment strategy; two sets of control strategies with different response modes are used to build an action plan library. The set limit threshold is the critical value obtained through historical data analysis. If the three-dimensional imbalance index is greater than or equal to the set limit threshold, the rapid response plan will be called first. Satisfying the following formula,

[0165] K≥θ cri

[0166] Where K is the three-dimensional imbalance index, θ cri Indicates the preset threshold parameter, the default value is 1.0.

[0167] For example, a modern agricultural park measures a three-dimensional imbalance index of 1.25 during the dry season. If the three-dimensional imbalance index is greater than or equal to the set limit threshold of 1.0, the rapid response plan will be called first, and all irrigation nodes will be immediately started and run at maximum flow. At the same time, non-essential lighting equipment will be turned off within 30 seconds to centrally supply power.

[0168] For irrigation equipment, set three levels of water release mode and select corresponding impact, pulse or penetration spraying strategies according to the imbalance value range;

[0169] Specifically, the irrigation equipment is set with three levels of water release modes, including impact spraying, pulse spraying, and penetration spraying;

[0170] Impact spraying refers to a short-duration, high-pressure, high-flow irrigation method that delivers 300% of the normal water flow per unit time. Pulse spraying is an intermittent, medium-pressure irrigation method that sprays for 60 seconds followed by a 20-second pause to form a pulse cycle. Penetration spraying is a long, slow, low-pressure irrigation method with a water flow rate 15% of that of impact spraying.

[0171] The selection of the three-level water release mode is based on the dynamic matching of the three-dimensional imbalance index interval and satisfies the following formula: if 0.8≤K<1.0, start the penetration spraying; if 1.0≤K<1.5, start the pulse spraying; if K≥1.5, force switch to impact spraying.

[0172] For example, during the rainy season, the K level of a rice field dropped to 0.75 due to continuous rainfall. When the infiltration spraying was started, only two waterings were required per day. A week later, a sudden high temperature caused the K level to rise to 1.3, and the field was automatically adjusted to pulse mode, operating three times per hour.

[0173] Establish a communication priority list between devices and give ventilation equipment priority control in high-temperature emergency situations;

[0174] Specifically, a collaborative working sequence between devices is dynamically generated. The communication priority list is automatically sorted by environmental parameter type, preventing resource conflicts caused by simultaneous startup of multiple devices. When temperature parameters account for more than 40% of the environmental imbalance, ventilation equipment is elevated to the top of the priority list.

[0175] For example, the greenhouse temperature parameters account for 52% due to the sudden increase in solar radiation at noon, and the temperature parameters account for more than 40% of the environmental imbalance. The running fertilization equipment is forced to stop, and the fan array is started first and maintained at the maximum speed until the temperature drops.

[0176] See attached Figure 2 The present invention also proposes a real-time analysis system for smart agricultural environmental data based on the Internet of Things, which includes the following modules:

[0177] Soil parameter acquisition module, used to obtain soil parameters of the first detection layer and the second detection layer of the farmland soil layer, calculate the soil porosity compensation coefficient based on the pressure difference between the two layers, and generate soil convection measurement values ​​based on the temperature difference between the two layers;

[0178] The canopy image analysis module is used to collect crop canopy images, extract vein clarity and chlorophyll reflectance intensity to form feature vectors, match them with the preset growth period feature database, and generate growth period fingerprint codes;

[0179] Dynamic matrix construction module, used to construct a dynamic data matrix containing temperature gradient and humidity change rate, combined with the growth period fingerprint code to assign a time weighting coefficient, thereby generating a smooth trend fluctuation curve;

[0180] Imbalance index generation module, used to generate a three-dimensional imbalance index with economic weight adjustment;

[0181] The strategy execution module is used to call the preset action plan library to match the irrigation strategy and dynamically adjust the device communication priority.

[0182] The modules can be implemented in whole or in part through software, hardware, or a combination thereof, supporting hardware embedded in or independent of a processor in a computer device, and also supporting software stored in a memory in a computer device, so that the processor can call and execute operations corresponding to the modules.

[0183] It should be noted that the human body information (including but not limited to human device information and personal information, etc.) and data (including but not limited to data used for analysis, stored data and displayed data, etc.) involved in the present invention are all information and data authorized by the human body or fully authorized by all parties. The collection, use and processing of relevant data require relevant legal standards.

[0184] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention, and should all be included in the scope of protection of the present invention.

Claims

1. A real-time analysis method for smart agricultural environmental data based on the Internet of Things, characterized in that: The following steps are involved: S1. Obtain soil parameters of the first detection layer and the second detection layer of the farmland soil layer, calculate the soil porosity compensation coefficient based on the pressure difference between the two layers, and generate soil convection measurement values ​​based on the temperature difference between the two layers; S2. Capture crop canopy images using a multi-channel image sensor, extract vein clarity and chlorophyll reflectance intensity to form a feature vector, and match it with a preset growth period feature database to generate a growth period fingerprint code; S3, constructing a dynamic data matrix including temperature gradient and humidity change rate, assigning a time weighting coefficient in combination with the growth period fingerprint code, and generating a smooth trend fluctuation curve; S4. Compare soil convection measurements with the current dynamic environmental baseline, introduce temperature difference rate parameters and market price fluctuation factors, and generate a three-dimensional imbalance index with economic weight adjustment; S5. Call the action plan library to match the quick response or gradual adjustment strategy, select the irrigation mode according to the three-dimensional imbalance index level, and dynamically adjust the device communication priority.

2. A real-time analysis method for smart agricultural environmental data based on the Internet of Things according to claim 1, characterized in that: Step S1 includes the following steps: Obtain soil parameters of the first detection layer and the second detection layer of the farmland soil layer; The initial contact pressure value of the first detection layer and the stable contact pressure value of the second detection layer are measured respectively, and the soil porosity compensation coefficient is calculated according to the difference between the initial contact pressure value and the stable contact pressure value; The temperature difference between the first detection layer and the second detection layer is obtained, and the soil convection measurement value is generated by combining the soil porosity compensation coefficient.

3. A real-time analysis method for smart agricultural environmental data based on the Internet of Things according to claim 2, characterized in that: The soil convection measurement is generated by combining the soil porosity compensation factor, which includes the following steps: The temperature difference between the first detection layer and the second detection layer is multiplied by the absolute value of the soil porosity compensation coefficient to obtain the current thermal gradient; The current thermal gradient is corrected based on the empirical conductivity of the soil medium to generate soil convection measurements.

4. A real-time analysis method for smart agricultural environmental data based on the Internet of Things according to claim 1, characterized in that: Step S2 includes the following steps: Collect visible and near-infrared images of crop canopies under constant illumination conditions; Edge enhancement processing was performed on the visible light band image, and the ratio of the total length of leaf vein pixels to the leaf outline area was calculated as the leaf vein clarity; Analyze the chlorophyll reflectance intensity in a specific wavelength range of near-infrared band images; The normalized vein clarity and chlorophyll reflectance intensity form a two-dimensional feature vector.

5. A real-time analysis method of smart agricultural environmental data based on the Internet of Things according to claim 4, characterized in that: Generating a growth period fingerprint code by matching a preset growth period feature database includes the following steps: The Mahalanobis distance is calculated based on the feature vector and the cluster center of the growth period feature database; The growth phase discrimination results were obtained based on the minimum Mahalanobis distance matching; If the change rate of chlorophyll reflectance intensity in multiple consecutive collection cycles exceeds the maturity threshold, it triggers the dynamic update of the maturity fingerprint code.

6. A real-time analysis method for smart agricultural environmental data based on the Internet of Things according to claim 5, characterized in that: Step S3 includes the following steps: Collect temperature gradient, humidity change rate, and light intensity within a predetermined time window to form a dynamic data matrix; An exponential decay time weight is applied to each time node data, and the time weight decreases as the interval between the acquisition time and the current moment increases; According to the growth period fingerprint code query time sensitivity coefficient, adjust the curvature parameter of dynamic data; The trend fluctuation curve with confidence mark is generated by cubic spline interpolation and Gaussian filtering.

7. A real-time analysis method for smart agricultural environmental data based on the Internet of Things according to claim 6, characterized in that: Step S4 includes the following steps: Based on the current dynamic environmental baseline, calculate the minute-by-minute rate of change of the current temperature difference relative to the historical mean; Access market price data streams and calculate price fluctuations to generate economic adjustment coefficients; The soil porosity compensation coefficient, temperature difference rate parameter, and economic adjustment coefficient are given independent weights and then multiplied together to generate a three-dimensional imbalance index.

8. A real-time analysis method of smart agricultural environmental data based on the Internet of Things according to claim 7, characterized in that: Generating the economic adjustment coefficient includes the following steps: Obtain market price data of target agricultural products in real time and calculate daily price volatility; If the price volatility exceeds the preset threshold, the sensitivity weight of the temperature parameter is logarithmized and normalized to generate an economic weight correction factor.

9. A real-time analysis method for smart agricultural environmental data based on the Internet of Things according to claim 8, characterized in that: Step S5 includes the following steps: Matching quick response plans or gradual adjustment plans according to the numerical range of the three-dimensional imbalance index; If the index exceeds the limit threshold, the impact spraying strategy is called and the irrigation equipment is operated at full power; If the index is in the middle range, start pulse spraying and adjust the water release rate in stages.

10. A real-time analysis method of smart agricultural environmental data based on the Internet of Things according to claim 9, characterized in that: Dynamically adjust device communication priorities, including the following steps: Monitor the temperature ratio in environmental parameters. If it exceeds the preset threshold, increase the control priority of the ventilation equipment; generate a sequence of equipment collaborative work instructions and terminate the operation of non-core equipment according to the urgency of the three-dimensional imbalance index.

Images