Morchella esculenta fruit forest intercropping system nutrient circulation intelligent regulation and control method based on Internet of Things

Abstract

The invention discloses an intelligent nutrient circulation regulation and control method for a morchella esculenta fruit forest intercropping system based on the Internet of Things, particularly relates to the technical field of intelligent agriculture, and is used for solving the problem of symbiosis imbalance of hyphae and fruit trees caused by nutrient regulation and control lag and static threshold setting in the prior art. Hypha density, electric signal dominant frequency and root system absorption rate data are collected in real time through a multispectral impedance sensor and an ion selective electrode array, and a space-time aligned dynamic matrix is constructed; dynamically updating a nutrient balance threshold value based on a real-time ratio of a hypha metabolism rate to an absorption rate, and generating a nutrient transfer priority by combining collaborative analysis of a hypha transfer direction included angle and an electric signal main frequency offset gradient; a continuous transfer path is identified through a region growing algorithm, water and fertilizer amounts are distributed according to priorities, and a local area instruction is generated; and inverting the equipment response delay by using the high-frequency pulse signal, synchronously adjusting the execution frequency and updating the dynamic matrix and the threshold parameter, thereby realizing the dynamic adaptation and autonomous optimization of the nutrient supply and demand of the intercropping system.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent agriculture, and more specifically, to an intelligent regulation method for nutrient cycling in a Morchella and fruit tree intercropping system based on the Internet of Things. Background Art

[0002] In the intercropping cultivation system of Morchella and fruit trees, the optimization of nutrient cycling can improve yield and ecological benefits. Existing technologies usually use Internet of Things monitoring devices to collect soil environment parameters in real time, and adjust water and fertilizer supply through a closed-loop control system to achieve dynamic management of nutrient resources. It depends on the comparison of sensor data with preset thresholds to trigger the start and stop operations of irrigation or fertilization devices, so as to maintain the balance of soil nutrients.

[0003] However, due to the significant differences in nutrient absorption rates between Morchella and fruit trees, and the response delay of the closed-loop control system is difficult to match the rapid changes in mycelial metabolism, resulting in the execution of regulation instructions lagging behind the actual demand. The system cannot distinguish short-term fluctuations from real gaps, frequently triggers excessive irrigation or fertilization, causing local soil salt accumulation, inhibition of mycelial activity, and decline in the absorption efficiency of fruit tree roots. Eventually, it destroys the symbiotic balance and resource utilization efficiency of the intercropping system and cannot adapt to the dynamic nutrient demand contradictions in the multi-species co-growth scenario. Summary of the Invention

[0004] In order to overcome the above defects of the prior art, the present invention provides an intelligent regulation method for nutrient cycling in a Morchella and fruit tree intercropping system based on the Internet of Things to solve the problems raised in the above background art.

[0005] To achieve the above object, the present invention provides the following technical solutions: An intelligent regulation method for nutrient cycling in a Morchella and fruit tree intercropping system based on the Internet of Things, comprising: S1. Based on the Internet of Things, collect the Morchella mycelial density and the fruit tree root absorption rate in real time, and generate a dynamic matrix representing the nutrient competition relationship between the mycelium and the roots; S2. Dynamically update the nutrient balance threshold according to the ratio of the mycelial metabolism rate to the fruit tree root absorption rate, and generate the current nutrient competition intensity based on the dynamic matrix; S3. When the nutrient competition intensity exceeds the nutrient balance threshold, locate the region with the largest nutrient gap and the region with the smallest nutrient gap based on the dynamic matrix; S4. Based on the collaborative analysis of the main frequency shift gradient difference of the mycelial electrical signal and the included angle of the mycelial migration direction in the region with the largest nutrient gap and the region with the smallest nutrient gap, generate the nutrient transfer priority; S5. Identify the continuous region as the transfer path according to the spatial gradient distribution of the Morchella mycelial density, and allocate the water and fertilizer amounts according to the nutrient transfer priority to generate local instructions; S6. Adjust the execution frequency of the local instruction through a high-frequency pulse signal, and synchronously update the dynamic matrix and the nutrient balance threshold.

[0006] Furthermore, based on the real-time collection of Morchella hypha density and fruit tree root absorption rate by the Internet of Things, generate a dynamic matrix representing the nutrient competition relationship between hyphae and roots, including: Real-time collect the fruit tree root absorption rate through an ion-selective electrode array buried in the fruit tree root zone. The fruit tree root absorption rate includes the absorption amount of nitrogen, phosphorus, and potassium ions per unit time; Real-time collect the Morchella hypha density through a multispectral impedance sensor deployed in the hyphal layer. The Morchella hypha density includes the hyphal length and biomass per unit volume; Align the fruit tree root absorption rate and the Morchella hypha density according to the collection timestamp and spatial coordinates, and generate a dynamic matrix with timestamps as rows and spatial coordinates as columns. Each element of the dynamic matrix contains the coupling value of the fruit tree root absorption rate and the Morchella hypha density at the corresponding spatio-temporal position.

[0007] Furthermore, dynamically update the nutrient balance threshold according to the ratio of the hyphal metabolic rate to the fruit tree root absorption rate, and generate the current nutrient competition intensity based on the dynamic matrix, including: Calculate the hyphal metabolic rate based on the change in hyphal dry weight between adjacent timestamps in the dynamic matrix. The hyphal metabolic rate is the change value of hyphal dry weight per unit time; According to the real-time ratio of the hyphal metabolic rate to the fruit tree root absorption rate at each spatio-temporal position in the dynamic matrix, statistically calculate the mean and standard deviation of the real-time ratio within the historical time window; When the real-time ratio exceeds the sum of the mean and twice the standard deviation, calculate the nutrient balance threshold adjustment amount through the square root relationship between the real-time ratio and the mean, and superimpose the nutrient balance threshold adjustment amount on the initial threshold to generate the updated nutrient balance threshold; Generate the nutrient competition intensity based on the product of the updated nutrient balance threshold and the spatial gradients of the hyphal dry weight and the fruit tree root absorption rate in the dynamic matrix. The spatial gradient is the numerical difference between adjacent grid cells at the same timestamp.

[0008] Furthermore, when the nutrient competition intensity exceeds the nutrient balance threshold, locate the regions with the largest and smallest nutrient gaps based on the dynamic matrix, including: Based on the nutrient competition intensity in the dynamic matrix, screen all grid regions where the nutrient competition intensity exceeds the nutrient balance threshold; Perform spatial clustering analysis on the selected grid regions, and merge adjacent grids into candidate regions through an unsupervised learning algorithm. The unsupervised learning algorithm is K-means clustering based on Euclidean distance; Calculate the nutrient gap amount for each candidate area, where the nutrient gap amount is the difference between the total absorption rate of fruit tree roots and the total metabolic rate of hyphae within the candidate area; Sort according to the absolute value of the nutrient gap amount, mark the candidate area with the largest absolute value as the area with the largest nutrient gap, and mark the one with the smallest absolute value as the area with the smallest nutrient gap.

[0009] Furthermore, based on the collaborative analysis of the main frequency shift gradient difference of hyphal electrical signals and the included angle of hyphal migration directions between the area with the largest nutrient gap and the area with the smallest nutrient gap, generate the nutrient transfer priority, including: Extract the main frequency shift gradients of hyphal electrical signals in the area with the largest nutrient gap and the area with the smallest nutrient gap from the dynamic matrix; Calculate the main frequency shift gradient difference between the area with the largest nutrient gap and the area with the smallest nutrient gap, where the main frequency shift gradient difference is the absolute value difference between the main frequency shift gradients of the area with the largest nutrient gap and the area with the smallest nutrient gap; Calculate the included angle of hyphal migration directions between the area with the largest nutrient gap and the area with the smallest nutrient gap; When the main frequency shift gradient difference is greater than the preset difference threshold and the included angle of hyphal migration directions is less than the preset angle threshold, increase the priority of the area with the largest nutrient gap by one level; When the main frequency shift gradient difference is less than or equal to the preset difference threshold and the included angle of hyphal migration directions is greater than or equal to the preset angle threshold, decrease the priority of the area with the smallest nutrient gap by one level; Generate a nutrient transfer priority sequence according to the results of priority increase and decrease.

[0010] Furthermore, the main frequency shift gradient of hyphal electrical signals is the difference in the main frequencies of hyphal electrical signals between adjacent grid cells at the same time stamp; the included angle of hyphal migration directions is the spatial geometric angle between the hyphal migration path direction vectors of the area with the largest nutrient gap and the area with the smallest nutrient gap.

[0011] Furthermore, identify continuous areas as transfer paths according to the spatial gradient distribution of Morchella hypha density, and allocate water and fertilizer amounts according to the nutrient transfer priority to generate local instructions, including: Based on the spatial gradient distribution of Morchella hypha density in the dynamic matrix, identify continuous areas as transfer paths through the region growing algorithm, where the continuous area is a set of adjacent grid cells with a gradient difference less than the preset threshold; According to the nutrient transfer priority sequence, allocate water and fertilizer amounts to each continuous area from high to low according to the priority; Bind the allocated water and fertilizer amounts to the spatio-temporal coordinates of the continuous area to generate local instructions; Sort the local instructions by the execution time stamp through the edge computing node and transmit them to the irrigation and fertilization device.

[0012] Furthermore, the local instructions include the target grid coordinates, water and fertilizer types, and spraying amounts.

[0013] Furthermore, the execution frequency of the local instructions is adjusted by the high-frequency pulse signal, and the dynamic matrix and the nutrient balance threshold are updated synchronously, including: Generate a high-frequency pulse signal based on the execution feedback data of the local instruction, wherein the frequency of the high-frequency pulse signal is inversely proportional to the response delay of the irrigation device; The execution frequency of local instructions is adjusted synchronously according to the high-frequency pulse signal to ensure that the timestamps of irrigation and fertilization operations are aligned with the acquisition cycle of the dynamic matrix; The dynamic matrix is ​​updated based on the adjusted execution frequency. The method of updating the dynamic matrix is ​​as follows: aligning the row index of the dynamic matrix according to the execution timestamp, and removing the spatiotemporal data corresponding to the instructions that have timed out and not been executed; The nutrient balance threshold was recalculated based on the real-time ratio of mycelium metabolic rate to fruit tree root absorption rate in the updated dynamic matrix, which was recalculated by replacing the historical mean with the sliding mean of the current time window.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. Through dynamic matrix and multi-parameter collaborative analysis, a closed-loop nutrient regulation system for the mycelium-fruit tree symbiotic system was constructed, which significantly improved the accuracy of resource allocation and system stability in intercropping scenarios; based on the dynamic data fusion of mycelium density, electrical signal main frequency offset and root absorption rate, the regulation lag problem caused by single parameter threshold judgment in traditional methods was solved; the nutrient balance threshold was updated by the real-time ratio of mycelium metabolism rate to absorption rate, and the spatiotemporal coupling characteristics of high-competition areas were dynamically identified by combining the collaborative analysis of migration direction angle and electrical signal gradient to avoid local salt accumulation and mycelium activity inhibition; the generation mechanism of nutrient transfer priority transforms the biological law of mycelium chemotaxis migration into quantitative rules for water and fertilizer allocation, realizing the leap from static zoning to dynamic path planning, and ensuring that resource allocation is accurately matched with the dynamic interaction needs of mycelium-fruit trees; 2. Through the dynamic adaptation of high-frequency pulse signals and execution frequencies, the IoT control delay is converted into the input variable of the closed-loop feedback, and the dynamic matrix and nutrient threshold are updated synchronously to form an adaptive regulation and iteration mechanism. By eliminating timeout instruction data and correcting threshold parameters in real time, it can effectively respond to short-term nutrient fluctuations and equipment response errors, and ensure the robustness of the system under complex disturbances. The coupling of the electrophysiological characteristics of mycelium and the IoT control enables the nutrient cycle to have the endogenous regulatory ability to autonomously adapt to environmental changes, while improving the efficiency of water and fertilizer utilization, maintaining the ecological synergy and sustainable productivity of the intercropping system. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a flow chart of the intelligent control method of nutrient cycle in Morchella fruit forest intercropping system based on Internet of Things of the present invention; Figure 2 Flow chart for generating nutrient transfer priorities for the present invention. Detailed implementation manners

[0016] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0017] Embodiment: Figure 1 A method for intelligent regulation of nutrient cycling in the Morchella intercropping system between fruit trees based on the Internet of Things according to the present invention is provided, including: S1. Based on the Internet of Things, the Morchella hyphal density and the fruit tree root absorption rate are collected in real time to generate a dynamic matrix representing the nutrient competition relationship between hyphae and roots; S2. The nutrient balance threshold is dynamically updated according to the ratio of the hyphal metabolic rate to the fruit tree root absorption rate, and the current nutrient competition intensity is generated based on the dynamic matrix; S3. When the nutrient competition intensity exceeds the nutrient balance threshold, the regions with the largest and smallest nutrient gaps are located based on the dynamic matrix; S4. Based on the collaborative analysis of the main frequency shift gradient difference of the hyphal electrical signals and the included angle of the hyphal migration direction in the regions with the largest and smallest nutrient gaps, the nutrient transfer priorities are generated; S5. The continuous regions are identified as transfer paths according to the spatial gradient distribution of the Morchella hyphal density, and the water and fertilizer amounts are allocated according to the nutrient transfer priorities to generate local instructions; S6. The execution frequency of the local instructions is adjusted through high-frequency pulse signals, and the dynamic matrix and the nutrient balance threshold are updated synchronously.

[0018] S1. Based on the Internet of Things, the Morchella hyphal density and the fruit tree root absorption rate are collected in real time to generate a dynamic matrix representing the nutrient competition relationship between hyphae and roots. The specific implementation is as follows: The concentration data of nitrogen, phosphorus, and potassium ions are collected in real time through an ion-selective electrode array buried in the fruit tree root zone. The ion-selective electrode array consists of a nitrate ion-selective electrode, a phosphate ion-selective electrode, and a potassium ion-selective electrode. Three groups of electrodes are arranged radially in the main root distribution area of each fruit tree. The horizontal spacing of each group of electrodes is 10% to 20% of the fruit tree crown diameter, and the vertical insertion depth is 20 cm to 40 cm to cover the main root absorption layer.

[0019] The electrode array collects the ion concentrations in the soil solution every 10 minutes through the Internet of Things nodes. The collected signals are transmitted to the edge computing node after analog-to-digital conversion. The edge computing node calculates the absorption rates of nitrogen, phosphorus, and potassium ions by the roots per unit time based on the time difference between two adjacent collections (in hours) and the fresh weight of the roots (obtained by regularly sampling and weighing). For example, if the nitrate ion concentration drops from 100 micromoles per liter to 90 micromoles per liter in two adjacent collections, the time interval is 10 minutes, and the fresh weight of the roots is 10 grams, then the nitrogen absorption rate is (100 - 90) micromoles per liter × 1 cubic decimeter (soil volume) ÷ (0.167 hours × 10 grams of fresh root weight) = 6 micromoles per hour per gram of fresh weight.

[0020] The multi-spectral impedance sensors deployed in the mycelium layer are used to collect the mycelium density data of Morchella esculenta in real time. The sensor probes are arranged in equilateral triangles with a spacing of 5 cm to 10 cm and an insertion depth of 5 cm to 15 cm, which matches the depth of the mycelium active layer. The sensors excite the soil medium with a swept-frequency signal of 1 - 100 kHz, measure the impedance amplitude (unit: ohm) and phase angle (unit: degree) at different frequencies, and the data is transmitted to the edge computing node in real time. The mycelium biomass is inverted from the impedance amplitude in the low-frequency band of 1 - 10 kHz. The inversion method is: the impedance amplitude is negatively correlated with the dry weight of the mycelium. For example, when the impedance amplitude at 1 kHz is 500 ohms, the dry weight of the mycelium is 20 mg / cm³; the mycelium length is inverted from the phase angle in the high-frequency band of 50 - 100 kHz. The inversion method is: the phase angle is positively correlated with the total length of the mycelium. For example, when the phase angle at 100 kHz is 30 degrees, the mycelium length is 150 cm / cm³.

[0021] Align the root absorption rate and the mycelium density data according to the timestamp and spatial coordinates. The timestamp alignment method is: all sensor nodes are synchronized with the edge computing node clock through the Network Time Protocol (NTP), the timestamp accuracy is millisecond-level, and the data is merged in 5-minute time windows. The data within the same window is regarded as collected at the same moment. The spatial coordinate alignment method is: the orchard is divided into grid cells of 1 square meter, and the coordinates of the grid center points are measured by RTK-GPS and pre-stored in the system; the installation positions of the ion-selective electrodes and the multi-spectral impedance sensors are matched to the corresponding grids through GPS coordinates to ensure that the data within the same grid belongs to the same spatial unit. When the sensor position deviation exceeds 1 / 5 of the grid side length, the coordinate calibration process is triggered.

[0022] Generate a dynamic matrix with timestamps as rows and spatial coordinates as columns. The row index of the dynamic matrix is the acquisition timestamp accurate to the second, the column index is the longitude and latitude coordinates of the grid center point, and each matrix element contains the nitrogen, phosphorus, and potassium absorption rates (unit: micromoles per hour per gram fresh weight), hyphal dry weight (unit: milligrams per cubic centimeter), and hyphal length (unit: centimeters per cubic centimeter) at the corresponding spatio-temporal position. The generation rule of the coupling value is: normalize the absorption rate and hyphal density parameter at the same spatio-temporal position. The normalization method is to divide each parameter by its historical maximum value to compress the value range to 0-1, and then multiply the normalized values to generate a six-dimensional coupling parameter group. For example, if the nitrogen absorption rate at a certain position is 5 micromoles per hour per gram fresh weight (historical maximum is 10), and the hyphal dry weight is 20 milligrams per cubic centimeter (historical maximum is 50), then the normalized coupling value is (5 / 10)×(20 / 50)=0.2.

[0023] Filter and correct the abnormal data collected by the sensor. When the change rate of ion concentration exceeds 3 times the standard deviation of the historical average, it is determined as abnormal data, and the replacement rule is to take the median of the same type of data in the adjacent three time windows within the same grid. When the phase angle of the impedance sensor exceeds the range of 0-180 degrees, trigger the probe self-check program: retract the probe to the ground and then re-insert it into the soil, and collect three times repeatedly. If the range of the three measurement values exceeds 30 degrees, it is determined as a hardware failure and an alarm is issued.

[0024] The dynamic matrix is stored in the HDF5 format. Each timestamp corresponds to a file, and the six-dimensional coupling parameters are stored in groups according to the grid coordinates within the file. Data compression uses the Zlib algorithm.

[0025] S2. Dynamically update the nutrient balance threshold according to the ratio of the hyphal metabolic rate to the fruit tree root absorption rate, and generate the current nutrient competition intensity based on the dynamic matrix. The specific implementation is as follows: Calculate the hyphal metabolic rate based on the change in hyphal dry weight between adjacent timestamps in the dynamic matrix. The hyphal dry weight data comes from the dynamic matrix generated in step S1. Each element of the dynamic matrix contains the hyphal dry weight value in milligrams per cubic centimeter and the corresponding acquisition timestamp. The calculation method of the hyphal metabolic rate is: take the difference between the hyphal dry weight at the current timestamp and the hyphal dry weight at the previous timestamp, and divide it by the interval duration between adjacent timestamps (unit: hours) to obtain the change value of hyphal dry weight per unit time, with the unit of milligrams per cubic centimeter per hour. The interval duration between adjacent timestamps is determined by the data acquisition frequency in step S1. For example, if the data is collected once every 10 minutes, the time interval is fixed at 0.167 hours. If the hyphal dry weights of a certain grid cell at two consecutive timestamps are 20 milligrams per cubic centimeter and 22 milligrams per cubic centimeter respectively, then the hyphal metabolic rate is (22-20)÷0.167 = 12, with the unit of milligrams per cubic centimeter per hour.

[0026] According to the real-time ratio of the mycelial metabolic rate to the fruit tree root absorption rate at each spatio-temporal position in the dynamic matrix, the mean and standard deviation of the real-time ratio within the historical time window are statistically calculated. The calculation method of the real-time ratio is: divide the mycelial metabolic rate by the sum of the nitrogen, phosphorus, and potassium absorption rates at the corresponding spatio-temporal position to avoid the accidental interference of the single-ion absorption rate. The unit of the nitrogen, phosphorus, and potassium absorption rates is micromoles per hour per gram of fresh weight, and the values are from the acquisition data of the ion-selective electrode array in step S1. The duration of the historical time window is the data of the past 24 hours stored in the dynamic matrix, and the time window contains 144 time stamps at 10-minute intervals. The calculation methods of the mean and standard deviation are: perform an arithmetic average on all real-time ratios within the window to obtain the mean, and then calculate the unbiased estimated standard deviation of these ratios. For example, if the average value of the real-time ratio in the past 24 hours is 1.0 and the standard deviation is 0.2, the threshold adjustment is triggered when the current real-time ratio exceeds 1.4 (the mean plus twice the standard deviation).

[0027] When the real-time ratio exceeds the sum of the mean and twice the standard deviation, calculate the nutrient balance threshold adjustment amount through the square root relationship between the real-time ratio and the mean. The setting basis of the square root relationship is: verified by field experiments, when the square root proportional relationship between the real-time ratio and the historical mean is used to adjust the threshold, it can effectively balance the sudden and slow change scenarios of nutrient supply and demand. The calculation method of the adjustment amount is: multiply the initial threshold by the square root of the real-time ratio and the mean. The initial threshold is calibrated by the absorption relationship between the mycelium and the root under the long-term monitoring of the nutrient balance state. For example, when the initial threshold is calibrated to 1.0, when the real-time ratio is 2.0 and the historical mean is 1.0, the adjustment amount is , and the updated threshold is the initial threshold plus the adjustment amount, that is, the updated threshold is 1.0 + 1.414 = 2.414. If the real-time ratio does not exceed the threshold, the initial threshold remains unchanged.

[0028] Generate the nutrient competition intensity based on the product of the updated nutrient balance threshold and the spatial gradients of hyphal dry weight and fruit tree root absorption rate in the dynamic matrix. The calculation method of the spatial gradient is as follows: Take the numerical difference of the hyphal dry weight or absorption rate of adjacent grid cells at the same timestamp. The adjacent grid cells are defined as a 1-square-meter area sharing a boundary. For example, if the hyphal dry weight of a certain grid is 20 milligrams per cubic centimeter and the adjacent grid on its east side is 18 milligrams per cubic centimeter, then the spatial gradient of hyphal dry weight is 20 - 18 = 2, with the unit of milligrams per cubic centimeter; if the nitrogen absorption rate of this grid is 5 micromoles per hour per gram of fresh weight and the adjacent grid on its east side is 6 micromoles per hour per gram of fresh weight, then the spatial gradient of absorption rate is 5 - 6 = -1, with the unit of micromoles per hour per gram of fresh weight. The generation rule of the nutrient competition intensity is: Multiply the updated threshold by the spatial gradient of hyphal dry weight and then divide by the absolute value of the spatial gradient of absorption rate. For example, if the threshold is 2.414, the hyphal gradient is 2, and the absolute value of the absorption rate gradient is 1, then the competition intensity is 2.414×2÷1 = 4.828.

[0029] Set boundary handling rules for extreme input data. When the real-time ratio is zero or negative, it is determined that the sensor data is abnormal, skip the threshold update step and trigger an abnormal alarm; when the spatial gradient of hyphal dry weight or absorption rate is zero, it is regarded as a state of uniform nutrient distribution, and the competition intensity is directly assigned a value of zero. The hardware configuration depends on the data processing ability of the edge computing node. The competition intensity data of the dynamic matrix is stored in the HDF5 format. Each file is named according to the timestamp, and the data is grouped and stored according to the grid coordinates within the file. The compression algorithm uses Zlib to reduce the storage overhead.

[0030] S3. When the nutrient competition intensity exceeds the nutrient balance threshold, locate the regions with the largest and smallest nutrient gaps based on the dynamic matrix. The specific implementation is as follows: Based on the nutrient competition intensity data in the dynamic matrix, screen all grid regions where the nutrient competition intensity exceeds the nutrient balance threshold. Each grid cell in the dynamic matrix contains the nutrient competition intensity value generated in step S2 and the corresponding spatio-temporal position coordinates. The unit of the nutrient competition intensity value is a dimensionless parameter. The screening rule is: Traverse all grid cells in the dynamic matrix. If the nutrient competition intensity value of a certain grid is greater than the updated nutrient balance threshold in step S2, then mark this grid as a high-competition region. For example, if the nutrient balance threshold is 2.414 and the competition intensity value of a certain grid is 4.828, then this grid is selected; if the competition intensity value is 1.5, then it is not selected.

[0031] Perform spatial clustering analysis on the selected grid areas, and merge adjacent grids into candidate areas through unsupervised learning algorithms. The unsupervised learning algorithm uses K-means clustering based on Euclidean distance. The specific operation is as follows: Input the longitude and latitude coordinates of each highly competitive grid as two-dimensional data points into the algorithm. The number of clusters is set to the integer value obtained by taking the square root of the total number of highly competitive grids selected. When initializing the cluster centers, the K-means++ algorithm is used to avoid local optimal solutions. The iteration termination condition is that the sum of the Euclidean distances of the cluster center changes for two consecutive times is less than 0.01. For example, if 9 highly competitive grids are selected, the number of clusters is set to 3, and the algorithm merges adjacent grids into 3 candidate areas through iterative calculation. Adjacent grids are defined as grid cells that share boundaries or vertices to ensure spatial continuity.

[0032] Calculate the nutrient gap amount for each candidate area. The calculation method of the nutrient gap amount is as follows: Sum the fruit tree root absorption rates of all grid cells within each candidate area, and then subtract the total mycelium metabolic rate of all grid cells within the same area. The unit of the fruit tree root absorption rate is micromoles per hour per gram of fresh weight, which is derived from the data collected by the ion-selective electrode array in step S1; the unit of the mycelium metabolic rate is milligrams per cubic centimeter per hour, which is derived from the calculation result in step S2. For example, if a candidate area contains two grids, with absorption rates of 5 micromoles per hour per gram of fresh weight and 3 micromoles per hour per gram of fresh weight respectively, and mycelium metabolic rates of 12 milligrams per cubic centimeter per hour and 8 milligrams per cubic centimeter per hour respectively, then the nutrient gap amount is (5 + 3) - (12 + 8) = -12, and the unit is micromoles per hour per gram of fresh weight.

[0033] Sort according to the absolute value of the nutrient gap amount. Mark the candidate area with the largest absolute value as the area with the largest nutrient gap, and the one with the smallest absolute value as the area with the smallest nutrient gap. The sorting method is as follows: Arrange the absolute values of the nutrient gap amounts of all candidate areas in descending order, take the first one as the largest area, and the last one as the smallest area. If there are multiple candidate areas with the same absolute value, preferentially select the area with a larger spatial area. For example, if the absolute values of the nutrient gap amounts of three candidate areas are 12, 10, and 10 respectively, and one of the two areas with a value of 10 contains 3 grids and the other contains 2 grids, then the largest area is the candidate area with an absolute value of 12, and the smallest area is the candidate area with an absolute value of 10 that contains 2 grids.

[0034] Handle extreme cases where there are no candidate areas. When there are no grids in the dynamic matrix with a nutrient competition intensity exceeding the threshold, it is determined that the entire area is nutrient balanced, skip the subsequent steps and generate a global balance status report; when the number of candidate areas exceeds 20, limit the maximum number of clusters to 10 to improve calculation efficiency. The clustering results and nutrient gap amount data are stored in JSON format to support calls from the visualization platform. Data compression uses the Zlib algorithm to reduce storage space occupancy.

[0035] In step S3, through dynamic matrix and spatial clustering analysis, the regions with the largest and smallest nutrient gaps are accurately located, solving the problems of regulation lag and regional misjudgment caused by static thresholds and manual division in traditional methods. Based on the real-time nutrient competition intensity, high-conflict regions are screened, and unsupervised learning algorithms are used to merge adjacent grids into candidate regions, breaking through the subjectivity and inefficiency of manually defined boundaries; by calculating the nutrient supply-demand difference between hyphae and roots within the candidate regions, the gap contradiction is directly quantified, avoiding the one-sidedness of single-parameter analysis. Compared with the prior art, in this step, through dynamic data-driven and adaptive clustering, accurate spatial identification of nutrient gaps is achieved, combining IoT real-time data with unsupervised learning, effectively solving the detection problem of local nutrient imbalance in the multi-species symbiosis scenario, and significantly improving the targeting of resource regulation and the ecological stability of the system.

[0036] Figure 2 The flowchart for generating the nutrient transfer priority of the present invention is given. S4. Based on the collaborative analysis of the main frequency shift gradient difference of the hyphal electrical signals and the included angle of the hyphal migration direction between the region with the largest nutrient gap and the region with the smallest nutrient gap, the nutrient transfer priority is generated. The specific implementation is as follows: Extract the main frequency shift gradients of the hyphal electrical signals in the regions with the largest and smallest nutrient gaps in the dynamic matrix. The main frequency data of the hyphal electrical signals stored in each grid unit of the dynamic matrix is from the acquisition results of the multispectral impedance sensor in step S1, with the unit of Hertz. The calculation method of the main frequency shift gradient of the hyphal electrical signal is: take the difference in the main frequency values of adjacent grid units at the same time stamp. Adjacent grid units are defined as 1-square-meter regions sharing a boundary. For example, if the main frequency of a grid is 100 Hertz and its adjacent grid to the east is 95 Hertz, then the main frequency shift gradient is 100 - 95 = 5, with the unit of Hertz per meter. The main frequency shift gradients of the regions with the largest and smallest nutrient gaps are the average values of the main frequency shift gradients of all their grid units. The determination methods of the regions with the largest and smallest nutrient gaps are from the results of sorting according to the absolute value of the nutrient gap amount in step S3. The largest region is the candidate region with the largest absolute value, and the smallest region is the candidate region with the smallest absolute value.

[0037] Calculate the difference in the main frequency shift gradients between the region with the largest nutrient gap and the region with the smallest nutrient gap. The difference in the main frequency shift gradient is the absolute value difference between the average values of the main frequency shift gradients of the largest region and the smallest region. For example, if the average main frequency shift gradient of the largest region is 8 Hertz per meter and the average main frequency shift gradient of the smallest region is 3 Hertz per meter, then the difference is , with the unit of hertz per meter. The preset difference threshold is determined by statistical analysis of historical data, specifically, it is the median of the main frequency shift gradient differences in all regions within the past 30 days. For example, through the analysis of historical data, it is found that the median of the main frequency shift gradient differences is 10 hertz per meter, then the difference threshold is set to 10 hertz per meter. If the current difference exceeds this threshold, it is determined that the hyphal electrical signal is significantly abnormal.

[0038] Calculate the included angle between the hyphal migration directions of the region with the largest nutrient gap and the region with the smallest nutrient gap. The hyphal migration direction vector is derived from the spatial gradient direction of the hyphal density in step S1. The calculation method of the spatial gradient direction is as follows: take the change rates of the hyphal density in the east-west and north-south directions to form a two-dimensional vector. The calculation method of the included angle is: input the migration direction vectors of the largest and smallest regions into the vector included angle formula to calculate their spatial geometric included angle, with the unit of degree. For example, if the migration direction vector of the largest region is (1, 0) (due east direction) and the migration direction vector of the smallest region is (0, 1) (due north direction), then the dot product of the two vectors is zero, and the included angle obtained by the inverse cosine calculation is 90 degrees. The preset angle threshold is calibrated through the hyphal population migration experiment. Specifically, it is found that when the included angle is less than 45 degrees, the hyphal migration paths are highly convergent under laboratory conditions. Therefore, the angle threshold is set to 45 degrees.

[0039] When the main frequency shift gradient difference is greater than the preset difference threshold and the included angle of the hyphal migration direction is less than the preset angle threshold, increase the priority level of the region with the largest nutrient gap by one level. For example, if the difference is 12 hertz per meter (greater than the threshold of 10 hertz per meter) and the included angle is 30 degrees (less than the threshold of 45 degrees), then the priority level of the largest region is increased from level 1 to level 2. The priority increase rule is integer-level adjustment. Each time the condition is met, the level is increased by one level, and the maximum does not exceed level 3. When the main frequency shift gradient difference is less than or equal to the preset difference threshold and the included angle of the hyphal migration direction is greater than or equal to the preset angle threshold, decrease the priority level of the region with the smallest nutrient gap by one level. For example, if the included angle is 60 degrees (greater than the threshold of 45 degrees), then the priority level of the smallest region is decreased from level 3 to level 2. If multiple conditions are met simultaneously, execute according to the priority increase rule first. If there is a conflict at the same level after the priority adjustment, sort by the spatial area from largest to smallest, and the region with the larger area takes precedence.

[0040] Generate a nutrient transfer priority sequence according to the results of the priority increase and decrease. The sequence generation rule is: sort by the priority level from high to low, and for regions at the same level, arrange them from largest to smallest according to the spatial area. For example, if the priority level of the largest region is level 2, the priority level of the smallest region is level 2, and the area of the largest region is 5 square meters and the area of the smallest region is 3 square meters, then the largest region is ranked before the smallest region in the sequence. When the priorities of all regions are the same, arrange them in descending order according to the absolute value of the nutrient gap in step S3. The generated priority sequence is stored in JSON format for step S5 to call for allocating water and fertilizer resources.

[0041] Set processing rules for extreme cases where there are no qualified regions or the threshold is exceeded. When the main frequency offset gradient difference is zero or negative, it is determined that the sensor data is abnormal, the priority adjustment is skipped, and the calibration process is triggered. The calibration process includes re-acquiring the electrical signal data and verifying the contact state of the sensor probe. When the calculation of the included angle of the hyphal migration direction fails (such as when the vector modulus is zero or the direction vector cannot be resolved), the priority sequence of the previous timestamp is used, and the abnormal event is recorded in the log. In terms of hardware dependence, the calculation process is executed by the edge computing node. The priority sequence data is stored in JSON format for step S5 to call, and the compression algorithm uses Zlib to reduce the storage overhead.

[0042] Step S4 accurately identifies the nutrient transfer priority through the collaborative analysis of the main frequency offset gradient difference of the hyphal electrical signal and the included angle of the migration direction, solving the problems of regulation lag and regional adaptation deviation caused by traditional technologies relying on a single static parameter. The main frequency offset of the hyphal electrical signal reflects the intensity of local nutrient competition, and the included angle of the migration direction reveals the spatial coordination of the hyphal population behavior. The combination of the two can dynamically capture the essence of the nutrient supply and demand contradiction. Through multi-parameter dynamic coupling and non-linear logical judgment, the limitations of manual experience threshold setting are broken through, and the hyphal biological characteristics are combined with the real-time data of the Internet of Things to achieve a transformation from "passive response" to "active prediction", significantly improving the nutrient regulation accuracy and ecological stability of the intercropping system, and ensuring that the resource allocation meets the symbiotic needs of the hyphae and fruit trees.

[0043] S5. Identify continuous regions as transfer paths according to the spatial gradient distribution of the Morchella hypha density, and allocate the amount of water and fertilizer according to the nutrient transfer priority to generate local instructions. The specific implementation is as follows: Based on the spatial gradient distribution of the Morchella hypha density in the dynamic matrix, continuous regions are identified as transfer paths through the region growing algorithm. The hypha density data stored in each grid cell of the dynamic matrix comes from the acquisition results of the multispectral impedance sensor in step S1, with the unit of milligrams per cubic centimeter. The spatial gradient distribution of the hypha density is the difference in hypha density between adjacent grid cells at the same timestamp, and adjacent grid cells are defined as 1-square-meter regions sharing a boundary. The merging rule of the region growing algorithm is: if the difference in the hypha density gradient between adjacent grids is less than the preset gradient threshold, they are merged into the same continuous region. The preset gradient threshold is determined through historical data statistics, specifically the median of all regional gradient differences in the past 30 days. For example, if it is found through analyzing historical data that the median of the gradient difference is 5 Hertz per meter (Hz / m), the threshold is set to 5 Hertz per meter (Hz / m). The specific operation of the region growing algorithm is: starting from any unallocated grid, check the gradient difference of its adjacent grids. If the difference is less than the threshold, merge them into the same region until no adjacent grid meets the condition, forming a set of continuous regions.

[0044] According to the nutrient transfer priority sequence generated in step S4, water and fertilizer amounts are allocated to each continuous area from high to low priority. The nutrient transfer priority sequence is derived from the collaborative analysis results based on the difference in main frequency offset gradient and the included angle of migration direction in step S4, and the priority weight is an integer value (such as level 1, level 2, level 3). The calculation method of water and fertilizer amount is: multiply the priority weight of the continuous area by the area of the area, and the area of the area is the product of the number of grid cells contained in the continuous area and 1 square meter. For example, if the priority of a certain continuous area is level 2 and it contains 3 grid cells, then the water and fertilizer amount is 2×3 = 6, with the unit of unit water and fertilizer amount. The actual spraying amount of the unit water and fertilizer amount is calibrated through field experiments. For example, each unit corresponds to 1 liter of water and fertilizer solution.

[0045] Bind the allocated water and fertilizer amounts to the spatio-temporal coordinates of the continuous area to generate local instructions. The format of the local instructions includes the target grid coordinates, water and fertilizer type, and spraying amount. The target grid coordinates are the set of longitude and latitude coordinates of all grid cells in the continuous area, and the coordinates are derived from the grid center point data measured by RTK-GPS in step S1. The water and fertilizer type is dynamically selected according to the priority weight. Nitrogen fertilizer is allocated to the area with priority level 1, phosphorus-potassium compound fertilizer is allocated to the area with priority level 2, and trace element solution is allocated to the area with priority level 3. The spraying amount is the water and fertilizer amount multiplied by the fertilization amount per unit area, and the fertilization amount per unit area is calibrated through field experiments to be 1 liter per square meter. For example, if the water and fertilizer amount of a certain continuous area is 6 units, then the spraying amount is 6 liters, and the water and fertilizer type is phosphorus-potassium compound fertilizer.

[0046] Sort the local instructions by the execution timestamp through the edge computing node and transmit them to the irrigation and fertilization device. The local instructions are encapsulated in JSON format, including fields such as timestamp, target coordinates, water and fertilizer type, and spraying amount. For example, an instruction is: timestamp "2024-10-01 08:00:00", coordinates "116.40,39.90", type "phosphorus-potassium compound fertilizer", spraying amount "6 liters". After the instructions are sorted in ascending order of timestamp, they are transmitted to the intelligent irrigation device deployed in the field through the LoRa wireless communication protocol. The irrigation device is an intelligent sprinkler based on the ESP32 microcontroller, which supports receiving and parsing JSON instructions for execution, and the flow error range of the sprinkler is controlled within ±5%.

[0047] Handle extreme cases where there are no continuous regions or the amount of water and fertilizer is zero. When there are no continuous regions in the dynamic matrix that meet the gradient difference threshold, it is determined that the hyphal density is balanced across the entire region, and a global balance report is generated and the water and fertilizer allocation is skipped; when the calculated amount of water and fertilizer is zero or negative, an anomaly detection process is triggered to re-verify the hyphal density data and the priority sequence. The anomaly detection process includes: re-collecting the hyphal density data, verifying the contact status of the sensor probe. If the data anomaly persists for three acquisition cycles, an artificial intervention alarm is triggered. In terms of hardware dependencies, the communication protocol of the irrigation equipment is LoRaWAN, the operating frequency band is from 470 MHz to 510 MHz, and the communication distance covers 1 km to ensure full-region connectivity of the field equipment.

[0048] Step S5 precisely plans the water and fertilizer transfer path through the dynamic adaptation of the hyphal density gradient distribution and the nutrient priority, solving the problems of resource misallocation and ecological interference caused by the traditional technology relying on a fixed fertilization mode. The hyphal density gradient reflects the local nutrient competition intensity, and combined with the priority sequence, it can dynamically identify the regions that urgently need to be regulated; the regional growth algorithm ensures the spatial continuity of the transfer path and avoids fragmented operations. Compared with the existing technology, through the collaborative control of the gradient threshold and the priority weight, it breaks through the limitations of static partition fertilization, combines the spatial characteristics of the hyphal population behavior with the real-time decision-making of the Internet of Things, realizes the leap from "homogeneous spraying" to "targeted regulation", and significantly improves the nutrient utilization efficiency and ecological synergy of the intercropping system.

[0049] S6. Adjust the execution frequency of the local instruction through a high-frequency pulse signal, and synchronously update the dynamic matrix and the nutrient balance threshold. The specific implementation is as follows: Generate a high-frequency pulse signal based on the execution feedback data of the local instruction. The local instruction comes from the control instruction generated in step S5, which includes the target grid coordinates, the type of water and fertilizer, and the spraying amount. The execution feedback data includes the response delay time of the irrigation and fertilization devices, with the unit of seconds. The calculation method of the frequency of the high-frequency pulse signal is: take the reciprocal of the response delay time and limit the frequency range between 0.1 Hz and 10 Hz. For example, if the response delay time of a certain irrigation operation is 2 seconds, the frequency of the pulse signal is 0.5 Hz. The pulse signal is generated through the general-purpose input / output interface of the edge computing node, and the signal waveform is a square wave with a fixed duty cycle of 50% to ensure the compatibility of the drive circuit.

[0050] Synchronously adjust the execution frequency of local instructions according to the high-frequency pulse signal. The adjustment method is as follows: Align the pulse signal frequency with the acquisition period of the dynamic matrix. The acquisition period is derived from the data acquisition frequency of the sensor node in step S1, for example, once every 10 minutes. If the pulse signal frequency is higher than the reciprocal of the acquisition period (for example, the reciprocal of 10 minutes is 0.0017 Hz), then reduce the execution frequency to be consistent with the acquisition period; if it is lower than the reciprocal of the acquisition period, then increase the execution frequency to the pulse signal frequency. For example, if the acquisition period is 10 minutes and the pulse signal frequency is 0.5 Hz, the execution frequency is increased to 0.5 Hz. The adjusted execution frequency ensures that the time stamps of irrigation operations are precisely aligned with the acquisition time stamps of the dynamic matrix, with the error controlled within ±1 second.

[0051] Update the dynamic matrix based on the adjusted execution frequency. The update method is as follows: Align the row index of the dynamic matrix according to the execution time stamp. The row index of the dynamic matrix is the acquisition time stamp accurate to the second. For instructions that have not been executed within the timeout period, the timeout is defined as not responding within more than 3 acquisition periods (for example, 30 minutes), and the corresponding spatio-temporal data is removed. For example, if the time stamp of a certain instruction is "2023-10-01 08:00:00" and it has not been executed at 08:30:00, then the data of hyphal density, absorption rate, and competition intensity associated with this time stamp in the dynamic matrix is deleted. The removal operation is implemented through the data processing software of the edge computing node. The software completes data filtering and matrix reconstruction based on the Pandas library (version 1.2.3) of Python 3.8.

[0052] Recalculate the nutrient balance threshold according to the real-time ratio of the hyphal metabolic rate to the fruit tree root absorption rate in the updated dynamic matrix. The calculation method of the real-time ratio is the same as that in step S2, that is, the hyphal metabolic rate divided by the sum of absorption rates. When recalculating the nutrient balance threshold, replace the historical mean with the sliding average of the current time window. The length of the time window is the same as the 24 hours defined in step S2. For example, if the sliding average of the real-time ratio within the current time window is 1.2, then the updated threshold is replaced with 1.2. After the threshold is updated, it is synchronized to the operation processes of steps S2 to S5 to ensure the real-time nature of subsequent cyclic regulation.

[0053] Set handling rules for extreme cases where the pulse signal is abnormal or the dynamic matrix update fails. When the pulse signal frequency exceeds the range of 0.1 Hz to 10 Hz, it is determined as a hardware driver abnormality, stop adjusting the execution frequency and switch to the default acquisition period synchronization mode; when the alignment of the dynamic matrix row index fails, trigger the data rollback mechanism to restore to the previous valid version matrix.

[0054] The technical solution of this embodiment constructs a closed-loop adaptive nutrient cycling system through the deep coupling of the dynamic interaction data between hyphae and fruit trees and the real-time regulation of the Internet of Things. Traditional methods rely on static thresholds or manual experience to divide regions, making it difficult to accurately adapt to the dynamic changes of symbiotic systems. However, this solution realizes the non-linear correlation modeling between the collective behavior of hyphae and the absorption rate of fruit trees through the spatio-temporal alignment of the dynamic matrix and the multi-parameter coupling analysis (such as the main frequency shift of hyphal electrical signals, the included angle of migration direction, and the intensity of nutrient competition). In particular, through the dynamic adaptation of high-frequency pulse signals and execution frequencies, a feedback control mechanism is introduced into the real-time correction of nutrient transfer paths. This solution proposes a full-link closed loop of "data acquisition - competition analysis - path planning - feedback iteration" through the cross-domain integration of hyphal electrophysiological characteristics and Internet of Things control logic. Through the continuous update of the dynamic matrix and the threshold adaptive mechanism, the system has endogenous stability to cope with complex ecological disturbances. For example, the collaborative analysis of hyphal migration direction and electrical signal gradient, the inverse control of pulse signal and execution frequency, etc. are all based on the biological essence of hyphal chemotactic migration, transforming biological laws into quantifiable control parameters, thus realizing the transformation from "manual intervention" to "system self-consistency".

[0055] It should be noted that this method constructs a closed-loop nutrient cycling regulation system for the Morchella-fruit tree intercropping system through Internet of Things perception, dynamic data fusion, and adaptive feedback control, deeply coupling the biological characteristics of hyphae with real-time decision-making of the Internet of Things to form a full-link dynamic adaptation of "perception - analysis - decision - execution - iteration".

[0056] In the data perception layer (S1), through the heterogeneous deployment of multi-spectral impedance sensors and ion-selective electrode arrays, spatio-temporally heterogeneous data of hyphal density, main frequency of electrical signals, and root absorption rate of fruit trees are captured in real time. Hyphal density is inverted from the low-frequency impedance amplitude and phase angle, and the root absorption rate is dynamically calculated through ion concentration gradient and fresh weight, ensuring the interpretability of data sources at the physical dimension and biological metabolism levels. The row and column indexes of the dynamic matrix are strictly aligned with spatio-temporal coordinates, providing a standardized data container for subsequent analysis.

[0057] In the competition analysis layer (S2 - S3), based on the temporal alignment and spatial clustering of the dynamic matrix, the nutrient competition contradiction between hyphae and fruit trees is revealed. Through the real-time ratio statistics of hyphal metabolic rate and absorption rate, combined with the mean-standard deviation threshold update mechanism of the sliding time window, the critical state of nutrient supply-demand imbalance is dynamically captured; further, through the collaborative analysis of the included angle of hyphal migration direction and the main frequency gradient of electrical signals, the spatial coupling characteristics of high-competition regions are identified, breaking through the limitations of traditional single-parameter threshold determination.

[0058] At the decision-making and planning layer (S4 - S5), based on unsupervised learning and region growing algorithm, highly competitive regions are clustered into continuous transfer paths, and water and fertilizer resources are allocated in combination with the product relationship between priority weights and region areas. The dynamic adjustment mechanism of priority weights (such as being triggered by the product threshold of angle - gradient) ensures that resource allocation conforms to the biological law of hyphal chemotactic migration. The spatio - temporal coordinate binding and JSON - format encapsulation of local instructions achieve seamless docking between control parameters and Internet of Things devices.

[0059] At the feedback and iteration layer (S6), through the dynamic adaptation of high - frequency pulse signals and execution frequencies, the irrigation response delay is converted into the inverse - frequency input of pulse signals, driving the precise synchronization of the execution cycle and data acquisition cycle. The real - time update of the dynamic matrix eliminates timeout instruction data, and the historical threshold is replaced by the moving - window mean value, forming a closed - loop feedback chain of "data acquisition → regulation execution → matrix update → threshold correction".

[0060] The calculations involved in the embodiments are all dimensionless numerical calculations. The preset parameters and threshold selections in the calculations are set by those skilled in the art according to the actual situation.

[0061] It should be noted that the present invention can be deployed on the device itself to achieve embedded applications, or can also run on a PC or other terminals with a user interface, so as to meet various hardware environments and usage requirements.

[0062] The above - mentioned embodiments can be implemented in whole or in part by software, hardware, firmware, or any other arbitrary combination. When implemented using software, the above - mentioned embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general - purpose computer, a special - purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer - readable storage medium, or transmitted from one computer - readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wired (such as infrared, wireless, microwave, etc.) means. The computer - readable storage medium can be any available medium that the computer can access, or a data storage device such as a server or data center that contains one or more collections of available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid - state drive.

[0063] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the systems, devices and modules described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0064] In several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division. In actual implementation, there may be other division methods. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the couplings or direct couplings or communication connections shown or discussed with each other can be through some interfaces. The indirect couplings or communication connections of the devices or modules can be in electrical, mechanical or other forms.

[0065] The modules described as separate components may or may not be physically separated. The components shown as modules may or may not be physical modules. They can be located in one place or distributed to multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0066] In addition, in each embodiment of the present application, the functional modules can be integrated in a processing module, or each module can exist physically alone, or two or more modules can be integrated in one module.

[0067] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks or optical discs that can store program codes.

[0068] As described above, it is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

[0069] Finally: The above are only the preferred embodiments of the present invention and are not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. An intelligent regulation method for nutrient cycling in a Morchella intercropping system in fruit forests based on the Internet of Things, characterized in that, Including: S1. Based on the Internet of Things, real-time collect the Morchella hypha density and the fruit tree root system absorption rate, and generate a dynamic matrix representing the nutrient competition relationship between the hyphae and the roots. S2. Dynamically update the nutrient balance threshold according to the ratio of the hypha metabolic rate to the fruit tree root system absorption rate, and generate the current nutrient competition intensity based on the dynamic matrix. S3. When the nutrient competition intensity exceeds the nutrient balance threshold, locate the region with the largest nutrient gap and the region with the smallest nutrient gap based on the dynamic matrix. S4. Based on the collaborative analysis of the main frequency shift gradient difference of the hypha electrical signal and the included angle of the hypha migration direction in the region with the largest nutrient gap and the region with the smallest nutrient gap, generate the nutrient transfer priority. S5. Identify the continuous region as the transfer path according to the spatial gradient distribution of the Morchella hypha density, and allocate the water and fertilizer amount according to the nutrient transfer priority to generate a local instruction. S6. Adjust the execution frequency of the local instruction through a high-frequency pulse signal, and synchronously update the dynamic matrix and the nutrient balance threshold.

2. The intelligent regulation method for nutrient cycling of the morel and fruit forest intercropping system based on the Internet of Things according to claim 1, wherein Based on the Internet of Things, real-time collect the Morchella hypha density and the fruit tree root system absorption rate, and generate a dynamic matrix representing the nutrient competition relationship between the hyphae and the roots, including: Real-time collect the fruit tree root system absorption rate through an ion-selective electrode array buried in the fruit tree root zone. The fruit tree root system absorption rate includes the unit-time absorption amounts of nitrogen, phosphorus, and potassium ions. Real-time collect the Morchella hypha density through a multispectral impedance sensor deployed on the hypha layer. The Morchella hypha density includes the hypha length per unit volume and the biomass. Align the fruit tree root system absorption rate and the Morchella hypha density according to the collection timestamp and the spatial coordinates, and generate a dynamic matrix with timestamps as rows and spatial coordinates as columns. Each element of the dynamic matrix contains the coupling value of the fruit tree root system absorption rate and the Morchella hypha density at the corresponding spatio-temporal position.

3. The intelligent regulation method for nutrient cycling of the morel and fruit forest intercropping system based on the Internet of Things according to claim 1, wherein Dynamically update the nutrient balance threshold according to the ratio of the hypha metabolic rate to the fruit tree root system absorption rate, and generate the current nutrient competition intensity based on the dynamic matrix, including: Calculate the hypha metabolic rate based on the change amount of the hypha dry weight between adjacent timestamps in the dynamic matrix. The hypha metabolic rate is the change value of the hypha dry weight per unit time. According to the real-time ratio of the hypha metabolic rate to the fruit tree root system absorption rate at each spatio-temporal position in the dynamic matrix, statistically calculate the mean and standard deviation of the real-time ratio within the historical time window. When the real-time ratio exceeds the sum of the mean and twice the standard deviation, calculate the nutrient balance threshold adjustment amount through the square root relationship between the real-time ratio and the mean, and superimpose the nutrient balance threshold adjustment amount on the initial threshold to generate the updated nutrient balance threshold. Based on the product of the updated nutrient balance threshold and the spatial gradients of the hypha dry weight and the fruit tree root system absorption rate in the dynamic matrix, generate the nutrient competition intensity. The spatial gradient is the numerical difference between adjacent grid cells at the same timestamp.

4. The intelligent regulation method for nutrient cycling of the morel and fruit tree intercropping system based on the Internet of Things according to claim 1, characterized in that, When the nutrient competition intensity exceeds the nutrient balance threshold, locate the region with the largest nutrient gap and the region with the smallest nutrient gap based on the dynamic matrix, including: Based on the nutrient competition intensity in the dynamic matrix, screen all grid regions where the nutrient competition intensity exceeds the nutrient balance threshold. Perform spatial clustering analysis on the screened grid regions, and merge adjacent grids into candidate regions through an unsupervised learning algorithm. The unsupervised learning algorithm is K-means clustering based on the Euclidean distance. Calculate the nutrient gap of each candidate area, which is the difference between the sum of the root absorption rate of fruit trees and the sum of the mycelium metabolic rate in the candidate area; The candidate areas are sorted according to the absolute value of the nutrient gap, and the candidate area with the largest absolute value is marked as the area with the largest nutrient gap, and the one with the smallest absolute value is marked as the area with the smallest nutrient gap.

5. The intelligent regulation method for nutrient cycling of the morel and fruit forest intercropping system based on the Internet of Things according to claim 1, characterized in that, Based on the collaborative analysis of the difference in the main frequency offset gradient of the mycelial electrical signal in the area with the largest nutrient gap and the area with the smallest nutrient gap and the angle of mycelial migration direction, the nutrient transfer priority is generated, including: Extract the main frequency offset gradient of mycelial electrical signal in the area with the largest and smallest nutrient gap in the dynamic matrix; Calculate the difference in the main frequency offset gradient between the area with the largest nutrient gap and the area with the smallest nutrient gap. The main frequency offset gradient difference is the absolute value difference between the main frequency offset gradients of the area with the largest nutrient gap and the area with the smallest nutrient gap. Calculate the angle between the mycelial migration direction in the area with the largest nutrient gap and the area with the smallest nutrient gap; When the difference in the main frequency offset gradient is greater than the preset difference threshold and the angle of hyphae migration direction is less than the preset angle threshold, the priority of the area with the largest nutrient gap is increased by one level; When the difference in the main frequency offset gradient is less than or equal to the preset difference threshold and the angle of hyphae migration direction is greater than or equal to the preset angle threshold, the priority of the area with the smallest nutrient gap is reduced by one level; Generate a nutrient transfer priority sequence based on the results of increasing and decreasing priorities.

6. The intelligent regulation method for nutrient cycling of the morel-fruit forest intercropping system based on the Internet of Things according to claim 5, characterized in that, The main frequency offset gradient of mycelium electrical signal is the main frequency difference of mycelium electrical signal of adjacent grid cells at the same time stamp; the mycelium migration direction angle is the spatial geometric angle between the mycelium migration path direction vectors in the area with the largest nutrient gap and the area with the smallest nutrient gap.

7. The intelligent regulation method for nutrient cycling of the morel and fruit tree intercropping system based on the Internet of Things according to claim 1, characterized in that, According to the spatial gradient distribution of the mycelium density of Morchella, continuous areas are identified as transfer paths, and water and fertilizer are allocated according to the nutrient transfer priority to generate local instructions, including: Based on the spatial gradient distribution of Morchella mycelium density in the dynamic matrix, the continuous region is identified as the transfer path through the region growing algorithm. The continuous region is a set of adjacent grid cells whose gradient difference is less than the preset threshold. According to the nutrient transfer priority sequence, water and fertilizer are allocated to each consecutive area from high to low priority; Bind the amount of water and fertilizer allocated to the spatiotemporal coordinates of the continuous area to generate local instructions; Local instructions are sorted by execution timestamps through edge computing nodes and transmitted to irrigation and fertilization devices.

8. The intelligent regulation method for nutrient cycling of the morel and fruit forest intercropping system based on the Internet of Things according to claim 7, characterized in that, Local instructions include target grid coordinates, water and fertilizer types, and spraying amounts.

9. The intelligent regulation method for nutrient cycling of the Morchella intercropping system in fruit forests based on the Internet of Things according to claim 1, characterized in that, The execution frequency of local instructions is adjusted through high-frequency pulse signals, and the dynamic matrix and nutrient balance threshold are updated synchronously, including: Generate a high-frequency pulse signal based on the execution feedback data of the local instruction, wherein the frequency of the high-frequency pulse signal is inversely proportional to the response delay of the irrigation device; The execution frequency of local instructions is adjusted synchronously according to the high-frequency pulse signal to ensure that the timestamps of irrigation and fertilization operations are aligned with the acquisition cycle of the dynamic matrix; The dynamic matrix is ​​updated based on the adjusted execution frequency. The method of updating the dynamic matrix is ​​as follows: aligning the row index of the dynamic matrix according to the execution timestamp, and removing the spatiotemporal data corresponding to the instructions that have timed out and not been executed; The nutrient balance threshold was recalculated based on the real-time ratio of mycelium metabolic rate to fruit tree root absorption rate in the updated dynamic matrix, which was recalculated by replacing the historical mean with the sliding mean of the current time window.

Images