Improved rubber forest soil environment real-time monitoring method and system based on internet of things

Abstract

This invention proposes a method and system for real-time monitoring of improved rubber plantation soil environment based on the Internet of Things (IoT). The method includes: setting up a fixed monitoring profile in the field where nano-biochar-based fertilizer is applied, and collecting baseline data at each depth layer; during operation, comparing real-time soil data with baseline data at the same depth, calculating the improvement proximity and comprehensive index direction to generate the improvement status of each depth layer; reading the status of each layer from shallow to deep along the profile, identifying the effective layer, and registering anomalies such as downward displacement of water content accompanied by increased conductivity or heavy metals in the migration risk layer as downward displacement risk, thereby determining the fertilizer effect boundary position and judging the boundary movement status through adjacent periodic changes; identifying the risk type based on the fertilizer effect boundary status, generating a risk warning record containing warning level and management instructions, and pushing it to the management terminal. This invention transforms IoT monitoring data into boundary status and disposal instructions, and is suitable for continuous monitoring and precise management of improved rubber plantation soil environment.

Description

Technical Field

[0001] This invention belongs to the field of real-time monitoring of improved rubber plantation soil environment, and particularly relates to a method and system for real-time monitoring of improved rubber plantation soil environment based on the Internet of Things. Background Technology

[0002] Rubber plantations are mostly distributed in areas with high temperature and humidity, concentrated rainfall, and significant soil acidification. Long-term cultivation can easily lead to problems such as aggravated soil acidification, depletion of organic matter, unstable nitrogen and phosphorus nutrient supply, abnormal fluctuations in electrical conductivity, increased availability of heavy metals in certain areas, and leaching and migration during the rainy season. To improve these soil environments, nano-biochar-based fertilizers and other soil amendments are increasingly being used in production. These materials improve soil pH buffering capacity, water and fertilizer retention capacity, slow nutrient release capacity, and adsorption and passivation capacity for heavy metal ions through their pore structure, surface functional groups, and fertilizer-carrying properties. However, the soil-improving effect of nano-biochar-based fertilizers is not synchronous, uniform, and stable throughout the entire soil profile. Instead, it is affected by factors such as fertilization depth, rubber tree root distribution, rainfall infiltration, slope runoff, initial soil acidification level, nutrient release rate, and competitive adsorption of heavy metals, resulting in different response processes in the fertilized layer, the root-active layer, and the layer below the roots at risk of migration. Current soil environmental monitoring in rubber plantations mainly relies on manual sampling, laboratory testing, or monitoring by ordinary IoT sensors. Manual sampling has high accuracy, but the sampling cycle is long and the spatial coverage is limited, making it difficult to capture processes such as post-rain leaching, acidification rebound, rapid decay of fertilizer effectiveness, and short-term migration of heavy metals in a timely manner. Although ordinary IoT monitoring can continuously acquire data such as soil pH, conductivity, moisture, nutrients, organic matter, and concentration of some ions, most systems still use single-point values ​​or fixed thresholds as alarm criteria, focusing more on whether a certain indicator exceeds the limit, without establishing vertical correlations between the fertilization layer, root activity layer, and migration risk layer around the same fixed monitoring profile, and without continuously comparing real-time readings with the baseline state of the same node, the same depth, and the same indicator. Therefore, it is difficult to determine whether changes at a certain depth truly reflect the improvement effect of nano-biochar-based fertilizer. Especially in the context of improving rubber plantation soil, changes in the same type of index may correspond to different soil processes. For example, a decrease in pH may be due to acidification rebound, leaching after rain, or fertilizer inefficiency; an increase in heavy metal ion concentration may be due to insufficient adsorption and passivation, acidification-induced reactivation, or water infiltration and migration; and an increase in conductivity may be due to nutrient release, salt accumulation, or leaching-carried ion migration. It is difficult to distinguish the causes of anomalies by relying on a single threshold. Even if the existing system detects anomalies in data at a certain depth, it is difficult to further determine whether the effective effect of nano-biochar-based fertilizer has covered the active root layer of rubber trees, whether there is advancement, stagnation, regression, or abnormal migration in a fixed profile, and it is also difficult to locate the risk assessment to the boundary layer and its adjacent layers. This results in a lack of clear correspondence between the early warning results and actual management measures such as fertilization, acidification adjustment, drainage, mulching to reduce leaching, or heavy metal control.

[0003] Therefore, there is a need for a real-time monitoring method and system that, based on existing soil IoT monitoring, can establish fixed monitoring profile benchmarks, generate deep improvement status, identify fertilizer efficiency boundaries, and output targeted early warnings and management instructions for improving the vertical structure of rubber plantation soil and the action process of nano-biochar-based fertilizers. Summary of the Invention

[0004] The purpose of this invention is to propose a method and system for real-time monitoring of the soil environment in rubber plantations based on the Internet of Things (IoT), thereby solving the aforementioned problems.

[0005] To achieve the above objectives, a method for real-time monitoring of improved rubber plantation soil environment based on the Internet of Things is provided in a first aspect of the present invention, the method comprising the following steps: Within the area where the nano-biochar-based fertilizer is applied, fixed monitoring profiles are established according to the fertilizer application layer, the active root layer of the rubber tree, and the migration risk layer below the root system. Several benchmark profile data are constructed based on the benchmark data of each depth layer of the fixed monitoring profiles. Real-time soil data at various depths are collected and compared with the baseline profile data at the same point and depth. By comparing the baseline representative value and real-time reading of the same depth layer and the same indicator, the improvement proximity amount for the corresponding depth layer and the corresponding indicator is generated. Based on the improvement proximity amount, directional markers are generated according to the technical meaning of the indicator, and the depth improvement status is generated according to the hierarchical attributes. The depth improvement status includes the plot number, profile number, node number, depth number, hierarchical attributes, sampling time, directional markers for each indicator, and the improvement status of the corresponding depth layer for each depth layer within the fixed monitoring profile. The improvement status includes improved, weakly improved, and suspected abnormal improvement. The system acquires the state of all depth layers arranged from shallow to deep within the same fixed monitoring profile. The improved or weakly improved depth layers are identified as effective action layers. Abnormal states in the root migration risk layer where water shifts downwards accompanied by increased conductivity or heavy metal ion concentration are registered as downward migration risks. Based on the effective action layers and downward migration risks, the effective improvement boundary position of the nano-biochar-based fertilizer is determined, and the boundary movement state is judged by the boundary position changes between adjacent sampling periods. The fertilizer effect boundary state includes the boundary movement state, root active layer coverage markers, and abnormal downward migration markers. Based on the fertilizer efficiency boundary state, risk type identification is performed, and risk warning instruction records are generated for the corresponding time.

[0006] Preferably, the acquisition time of the reference profile data is based on the stable record before fertilization. If the sensor is deployed after fertilization, the record of the initial stable period after fertilization is used as the reference.

[0007] Preferably, the reference data includes pH, conductivity, water content, organic matter, nitrogen content, phosphorus content, and heavy metal ion concentration data; The benchmark profile data consists of benchmark records for each depth layer within a fixed monitoring profile; each benchmark record includes a plot number, profile number, node number, depth number, layer attribute, index type, benchmark representative value, benchmark acquisition time period marker, and detection source marker.

[0008] Preferably, the step of generating directional markers based on the improved proximity quantity according to the technical meaning of the index, and generating a depth improvement state according to the hierarchical attributes, specifically involves: Based on the improved proximity measurement shown, determine whether the same indicator at the same layer and the same node is close to the target interval: When the baseline data forms a positive directional marker and the conductivity remains stable, the current layer is marked as improved; when some baseline data has improved but nutrient release or heavy metal passivation has not yet been fully demonstrated, the current layer is marked as weakly improved; when the pH or organic matter changes little relative to the baseline profile data and does not show a clear direction of improvement, the current layer is marked as unimproved; when the pH is far from the suitable range, the conductivity increases continuously, or the concentration of heavy metal ions increases, the current layer is marked as suspected abnormally improved.

[0009] Preferably, determining the fertilizer effect boundary position reached by the effective improvement effect of the nano-biochar-based fertilizer based on the effective action layer and the risk of downward migration specifically involves: In the depth layers arranged from shallow to deep in the fixed profile, the deepest layer that meets the effective action layer and is not registered as having a risk of downward displacement is selected as the current fertilizer effect boundary position; when no effective action layer meets the conditions at the current sampling time, the fertilizer effect boundary position is registered as not having formed an effective boundary.

[0010] Preferably, the boundary motion state is determined by comparing the current fertilizer effect boundary position with the fertilizer effect boundary position of the previous sampling period; Specifically, when the boundary motion state is greater than 0 and the downward risk marker is not triggered, the current boundary motion state is registered as normal advancement; when the boundary motion state is equal to 0 and the current boundary motion state is continuously maintained at the same depth layer for a preset period, the current boundary motion state is registered as stagnation; when the boundary motion state is less than 0, the current boundary motion state is registered as regression; when the downward risk marker of the migration risk layer below the root system is triggered, the current boundary motion state is registered as abnormal downward movement. The root active layer coverage marker is obtained by comparing the fertilizer effect boundary position with the depth number of the root active layer in the depth improvement state.

[0011] Preferably, the step of identifying the risk type based on the fertilizer efficiency boundary state specifically involves: Based on the risk assessment path triggered by the described fertilizer efficiency boundary state, within the boundary layer and its adjacent layers, the sequence of the three event chains is determined according to the first occurrence time of each key event; the risk of leaching or heavy metal migration during the rainy season corresponds to the event chain of decreased water content, pH moving away from the suitable range, and increased heavy metal ion concentration; the risk of acidification rebound-type fertilizer efficiency decline corresponds to the event chain of boundary regression, pH of the root active layer moving away from the suitable range, and decreased improvement status; the risk of fertilizer efficiency not reaching the root layer corresponds to the event chain of boundary stagnation, the fertilized layer remaining improved, and the root active layer remaining unimproved; the risk of insufficient heavy metal passivation corresponds to the event chain of improved pH of the root active layer, no decrease in heavy metal ion concentration, and continuous existence of insufficient passivation status; the risk of insufficient nutrient release corresponds to the event chain of decreased nitrogen or phosphorus content, no improvement in organic matter, and boundary stagnation or regression.

[0012] Preferably, the risk warning instruction record includes the plot number, profile number, sampling time, risk type, warning level, target depth layer, depth range adjacent to the boundary, management instructions, and transmission record.

[0013] Preferably, the warning level is generated in the following way: The warning level determination quantity is obtained by adding the risk trigger quantity, the target level aggravation item, and the management urgency item. The risk trigger quantity is determined based on the logical determination relationship of the fertilizer effect boundary state, the sequence of the three event chains, and the continuous maintenance condition. When the target depth layer includes the root active layer or the risk layer of migration below the root system, the aggravation item takes the higher value. The management urgency item is determined by the boundary movement state in the fertilizer effect boundary state. The warning level determination quantity is used to register the level as attention level, warning level, or disposal level. The management instructions are generated jointly by the risk type, target depth layer, and early warning level.

[0014] A second aspect of the invention provides an Internet of Things-based real-time monitoring system for improved rubber plantation soil environment, the system comprising: The benchmark profile management module is used to fix the fertilizer application layer, the active root layer of rubber trees, and the migration risk layer below the root system as fixed monitoring profiles within the area where nano-biochar-based fertilizer is applied; and to construct several benchmark profile data based on the benchmark data of each depth layer of the fixed monitoring profile. The depth improvement status generation module is used to collect real-time soil data at various depth layers, compare it with the baseline profile data at the same point and depth, and generate the improvement proximity amount for the corresponding depth layer and the corresponding indicator by comparing the baseline representative value and real-time reading of the same depth layer and the same indicator. Based on the improvement proximity amount, directional markers are generated according to the technical meaning of the indicator, and the depth improvement status is generated according to the hierarchical attributes. The depth improvement status includes the plot number, profile number, node number, depth number, hierarchical attributes, sampling time, directional markers of each indicator, and the improvement status of the corresponding depth layer for each depth layer in the fixed monitoring profile. The improvement status includes improved, weakly improved, and suspected abnormal improvement. The fertilizer effect boundary identification module is used to acquire the state of all depth layers arranged from shallow to deep within the same fixed monitoring profile, identify the improved or weakly improved depth layers as effective action layers, and register abnormal states in the migration risk layer below the root system where water content shifts downwards and is accompanied by increased conductivity or heavy metal ion concentration as downward migration risk. Based on the effective action layers and downward migration risk, the module determines the fertilizer effect boundary position reached by the effective improvement effect of the nano-biochar-based fertilizer, and judges the boundary movement state by the boundary position change in adjacent sampling periods. The fertilizer effect boundary state includes the boundary movement state, root active layer coverage marker, and abnormal downward migration marker. The risk warning and management instruction output module is used to identify the risk type based on the fertilizer effect boundary state and generate a risk warning instruction record at the corresponding time.

[0015] The beneficial technical effects of the present invention are at least as follows: This invention addresses the need for real-time monitoring of the improved soil environment in rubber plantations after the application of nano-biochar-based fertilizers. It proposes a monitoring method and system based on fixed vertical monitoring profiles, with the fertilizer efficiency boundary status as the core judgment object and the sequence of events in adjacent layers as the basis for risk identification. The system sets up fixed monitoring profiles within the fertilization area according to the fertilization layer, the active root layer of rubber trees, and the migration risk layer below the roots. It collects and saves benchmark profile data for the same plot, profile, node, depth, and indicator. In subsequent operation, real-time soil data is compared with the corresponding benchmark records at the same point and depth. Based on whether each indicator is close to its improvement target range or shows a change opposite to the improvement target, the improvement status of each depth layer is determined. The system further reads the state of each depth layer in a fixed profile from shallow to deep, identifying the improved or weakly improved layers as effective layers. Abnormal states in the root migration risk layer, such as downward shift in water content accompanied by increased conductivity or heavy metal ion concentration, are separately registered as downward migration risks. This determines the fertilizer effect boundary reached by the normal improvement effect of nano-biochar-based fertilizer. Changes in the boundary position between adjacent sampling periods are used to determine boundary advancement, stagnation, regression, failure to cover the active root layer, or abnormal downward migration. Addressing the difficulty of explaining the causes of anomalies in existing technologies, after determining the fertilizer effect boundary state, the system only calls the indicator direction markers and real-time parameter records of the boundary layer and its adjacent layers. Based on the chronological relationship of a finite event sequence, it identifies risk types such as rainy season leaching or heavy metal migration, acidification rebound-type fertilizer effect attenuation, fertilizer effect not reaching the root layer, insufficient heavy metal passivation, and insufficient nutrient release. This transforms risk assessment from a single indicator exceeding limit to a process assessment constrained by both the boundary state and adjacent soil processes.

[0016] To address the difficulty of translating existing monitoring results into management measures, the system generates early warning levels based on risk type, target depth layer, boundary movement status, and risk persistence. It then records the target plot, target profile, target depth layer, and corresponding management actions into a risk warning instruction log, which is pushed to the management terminal via an IoT platform. This allows measures such as applying nano-biochar-based fertilizer, adjusting fertilization depth, acidification treatment, drainage treatment, mulching to reduce leaching, and heavy metal control to be mapped to specific risk types and soil layers. Through this design, the invention transforms discrete parameter readings from existing IoT soil monitoring into baseline profiles, depth status, fertilizer effectiveness boundaries, and treatment instructions suitable for rubber plantation improvement scenarios. This enables more accurate determination of whether nano-biochar-based fertilizer is truly effective in the active root layer of rubber trees, more timely identification of fertilizer effectiveness decay and heavy metal migration risks, and provides clear execution guidelines for continuous monitoring and on-site management of improved rubber plantation soil. Attached Figure Description

[0017] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0018] Figure 1 This is a flowchart of the method for real-time monitoring of improved rubber plantation soil environment based on the Internet of Things, according to the present invention. Detailed Implementation

[0019] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0020] like Figure 1 As shown in the embodiment of the present invention, a method for real-time monitoring of improved rubber plantation soil environment based on the Internet of Things is provided. The method includes: S1. Within the area where the nano-biochar-based fertilizer is applied, a fixed monitoring profile is established, consisting of the fertilization layer, the active root layer of the rubber tree, and the migration risk layer below the root system. Several reference profile data are constructed based on the reference data of each depth layer of the fixed monitoring profile.

[0021] Specifically, within the area where the nano-biochar-based fertilizer is applied, the system fixes the fertilization layer, the active root layer of rubber trees, and the migration risk layer below the root system into the same vertical monitoring profile, and preserves the corresponding initial soil conditions for each depth layer.

[0022] The actual inputs for this step include deployment data for fixed monitoring profiles, fertilization operation data, and initial soil testing data. Deployment data for fixed monitoring profiles is entered by on-site installers into the IoT gateway or management terminal, including plot number, profile number, node number, depth number, and hierarchical attributes. Hierarchical attributes include the fertilization layer, root activity layer, and root migration risk layer. After sensor installation, the above numbering relationships are permanently saved for subsequent real-time data attribution and baseline data matching. Fertilization operation data is entered from the fertilization record terminal, operation registration interface, or on-site management terminal, including the plot where nano-biochar-based fertilizer was applied, fertilization time, fertilization location, and fertilization depth, used to determine the baseline data collection period. The system prioritizes using stable records before fertilization as the baseline; if the sensor deployment is completed after fertilization, the initial stable period record after fertilization is used as the baseline. Initial soil testing data comes from sensor nodes or corresponding detection units on fixed monitoring profiles. pH is collected using buried soil pH electrodes, electrical conductivity using soil conductivity probes, and moisture content using capacitive or time-domain reflectometry soil moisture sensors. Organic matter, nitrogen content, phosphorus content, and heavy metal ion concentration are uploaded via in-situ detection modules, plug-in rapid detection modules, or periodic detection interfaces bound to node numbers. When using periodic detection interfaces, field personnel take samples from a fixed depth and use portable detectors for testing. The test results are then uploaded via a management terminal, bound to the corresponding plot number, profile number, node number, and depth number.

[0023] Furthermore, to ensure the verifiability of subsequent comparisons at the same point and depth, the system generates benchmark profile data. The system performs consistency checks on all types of input data. For continuous sensor data, the system saves the sensor number, installation depth, calibration time, measurement range, and unit diameter. For portable detectors or laboratory testing interface data, the system saves the testing batch, sampling depth, testing item, and unit diameter. pH is recorded as dimensionless acidity / alkalinity readings. Conductivity, water content, organic matter, nitrogen content, phosphorus content, and heavy metal ion concentration are each saved using fixed units registered on the same land management platform. The baseline record and real-time record for the same indicator must use the same or equivalent units before proceeding to step two. If the detection source for the same indicator changes, the system does not directly use the old baseline for numerical comparison. Instead, it marks the change in detection source in the record and requires confirmation through a transitional comparison record at the same depth that the reading diameters of the two testing interfaces are consistent. If consistency cannot be confirmed, the indicator is only retained as an observation record and does not participate in the depth improvement status determination.

[0024] In practical implementation, the location of fixed monitoring profiles should be determined first. Fixed monitoring profiles should be set within the actual application area of ​​the nano-biochar-based fertilizer, avoiding obviously non-representative locations such as the edges of drainage ditches, mechanically compacted areas, temporary composting points, or long-term waterlogging points. For rubber plantations on slopes, fixed monitoring profiles can be set up at the top, middle, and bottom of the slope. Each profile should be fixedly numbered according to the fertilization layer, root active layer, and root migration risk layer to preserve slope differences and ensure clear vertical relationships within the same profile. After the sensor nodes are installed, the system establishes a fixed hierarchical relationship for each profile. For example, if a plot is numbered A01, a monitoring profile within that plot is numbered P01, the fertilization layer nodes are recorded as P01-D1, the root active layer nodes as P01-D2, and the root migration risk layer nodes as P01-D3; where D1, D2, and D3 represent fixed depth layer numbers used to distinguish different soil layers within the profile. When subsequent real-time data enters the system, the data records carry the same parcel number, profile number, node number, and depth number, enabling the system to accurately match the real-time data with the corresponding benchmark profile data. .

[0025] Furthermore, when collecting baseline data, the system determines the baseline time period based on fertilization operation data. When continuous records exist prior to fertilization, the system selects the period during which no fertilization, acidification, drainage, heavy rainfall, or significant irrigation occurred as the baseline time period. When using initial stabilization period records after fertilization, the system selects the period after the fertilization operation where sensor readings are no longer disturbed by construction and no significant external operations such as rainfall, re-fertilization, acidification, or drainage occurred on-site as the baseline time period. For example, if significant short-term changes in soil moisture content and conductivity occur on the day of fertilization due to ditching, backfilling, or watering, the system marks this period as a construction disturbance period, and the baseline is established only after continuous records stabilize. For pH, conductivity, and moisture content continuously collected by sensors, the IoT gateway receives data records according to a fixed sampling period. Each record includes plot number, profile number, node number, depth number, index type, reading, collection time, and sensor status marker. If a record shows that the sensor is offline, the probe is malfunctioning, or the reading is empty, the system marks the record as invalid. If an isolated jump occurs at a sampling point, such as a sudden abnormal increase in conductivity at a certain moment, without corresponding changes at adjacent moments and depths, the system marks the record as abnormal. For organic matter, nitrogen content, phosphorus content, and heavy metal ion concentration, when using an in-situ detection module, the data recording method is consistent with that of continuous sensors. When using a portable detector or laboratory detection interface, the detection results are entered or imported through a management terminal and bound to the corresponding profile number and depth number.

[0026] For example, after field personnel obtain soil samples at depths corresponding to P01-D2, they use a portable nitrogen and phosphorus analyzer to obtain the test results. When uploading these results, they are labeled as A01, P01, D2, and the corresponding index type, enabling the test results to be written into the baseline profile data. The baseline record corresponding to P01-D2 in the middle.

[0027] The determination of the benchmark representative value is completed within the same indicator. For continuously acquired indicators, the system selects a representative record from stable records within the benchmark period as the benchmark representative value for that indicator; for periodically detected indicators, the system selects a valid detection record within the benchmark period as the benchmark representative value for that indicator. When there are multiple valid records for an indicator, the system can use the median record as a representative to reduce the impact of a single random fluctuation on the benchmark. Each benchmark record also stores a detection source marker, indicating that the indicator was obtained from an embedded sensor, in-situ detection module, portable detector, or laboratory detection interface. In the subsequent step two, when performing real-time comparison, the system confirms that the real-time data and the benchmark data have the same detection source or belong to an equivalent detection interface based on the detection source marker. For example, if the benchmark profile data... The heavy metal ion concentrations at P01-D3 are obtained from the portable detector interface. Therefore, subsequent heavy metal ion concentrations at the same depth are also uploaded through the same or equivalent detection interface. If the detection source is changed, the system marks the change in the record as a quality indicator for subsequent comparisons. Before applying nano-biochar-based fertilizer to plot A01 of a rubber plantation, a P01 monitoring profile was set up near a representative rubber tree. P01-D1 corresponds to the fertilization layer. D2 corresponds to the root active layer, P01 D3 corresponds to the migration risk layer below the root system. Before fertilization, the system continuously collects pH, conductivity, and water content. Simultaneously, a portable detector obtains the organic matter, nitrogen, phosphorus, and heavy metal ion concentrations for D1, D2, and D3. If an isolated jump in water content reading occurs in layer D2 at a certain moment, and D1 and D3 do not show corresponding changes, this record is marked as an anomaly. Finally, the system saves P01... D1, P01 D2, P01 D3 is the baseline record for three depth layers, and subsequent step two receives P01. After determining the real-time pH and heavy metal ion concentration of D2, it was compared with PO1 in D0. Compare the corresponding indicators for D2.

[0028] This step outputs reference profile data. The baseline profile data consists of baseline records for each depth layer within a fixed monitoring profile. Each baseline record includes a plot number, profile number, node number, depth number, hierarchical attribute, indicator type, baseline representative value, baseline data collection period marker, and detection source marker. The plot number and profile number are used to identify the monitoring object; the node number and depth number are used for subsequent comparisons at the same point and depth; the hierarchical attribute is used to distinguish between the fertilization layer, the root active layer, and the root migration risk layer; the indicator type and baseline representative value are used to generate the depth improvement status in step two; and the detection source marker is used to identify how the detection data was obtained.

[0029] S2. Collect real-time soil data at each depth layer and compare it with the baseline profile data at the same point and depth. By comparing the baseline representative value and real-time reading of the same depth layer and the same indicator, generate the improvement proximity amount for the corresponding depth layer and the corresponding indicator. Based on the improvement proximity amount, generate directional markers according to the technical meaning of the indicator, and generate the depth improvement status according to the hierarchical attributes. The depth improvement status includes the plot number, profile number, node number, depth number, hierarchical attributes, sampling time, directional markers of each indicator, and improvement status of the corresponding depth layer for each depth layer in the fixed monitoring profile. The improvement status includes improved, weakly improved, and suspected abnormally improved.

[0030] Specifically, the system at the sampling time After receiving real-time soil data from each depth layer within a fixed monitoring profile, the data is first processed according to the baseline profile data. The system matches the plot number, profile number, node number, depth number, and indicator type, then reads the corresponding hierarchical attributes, benchmark representative value, benchmark collection period marker, and detection source marker. The plot number and profile number are used to confirm the rubber plantation improvement area to which the real-time data belongs; the node number and depth number are used to confirm whether the real-time data belongs to the fertilization layer, root active layer, or root migration risk layer; and the indicator type is used to determine whether the real-time reading should be compared with the benchmark profile data. The system compares the values ​​against which baselines are represented. The baseline acquisition period marker identifies whether the baseline represents the pre-fertilization state or the initial stable period after fertilization. The detection source marker confirms the acquisition method of the real-time detection record and the baseline record. After the above matching, the system determines the meaning of changes in pH, conductivity, water content, organic matter, nitrogen content, phosphorus content, and heavy metal ion concentration for each depth layer, and converts these changes into a state record for that depth layer. This state record is... The components will be read according to the depth number in the subsequent step three. The fertilizer efficiency boundary state is determined by the state changes between adjacent depth layers.

[0031] Understandably, the system performs availability processing before real-time data is included in this calculation. If the real-time record lacks a parcel number, profile number, depth number, or indicator type, or cannot be processed... If a baseline record with the same depth and the same metric is found, the real-time record is marked as an unmatchable record and will not be included in the improved proximity measurement. The calculation process is as follows: If the real-time reading exceeds the sensor's range, the sensor is offline or faulty, the sampling time exceeds the allowable range of the current sampling period, or there is an isolated jump without adjacent depth response support compared to adjacent time points, the record is marked as a quality anomaly record. For a certain depth layer, the system only generates an improved, weakly improved, unimproved, or suspected abnormally improved state when the key indicators required for entering the state determination meet the valid recording requirements; if the key indicators are insufficient, the system records the depth layer as a data insufficiency state and temporarily does not use it as an effective layer in step three to avoid generating false fertilizer effect boundaries due to missing data.

[0032] Understandably, the core calculations used in this step are derived from the mathematical concept of "distance from a point to a target interval." For pH, conductivity, organic matter, nitrogen content, phosphorus content, and heavy metal ion concentration, the system reads the target or control intervals for each indicator in the rubber plantation soil improvement. For example, pH corresponds to the rhizosphere suitable interval, conductivity to the stable control interval, heavy metal ion concentration to the safe control interval, and nitrogen and phosphorus content to the nutrient supply intervals. For water content, the system registers water retention control intervals in the fertilization layer and the root active layer to determine the water content maintenance status of these layers. In the migration risk layer below the roots, the system simultaneously reads the water content direction markers at adjacent sampling times to identify post-rainwater content downward shift events. The system first calculates the deviation of the baseline representative value from the target interval of the indicator, and then calculates the deviation of the real-time reading from the target interval. The difference between the two is used to determine whether the real-time reading is closer to the improvement target than the baseline state. Different indicators are calculated separately within their respective categories, and the results are used for direction determination and status marking. The calculation relationship is as follows: The aforementioned target intervals, control intervals, and engineering variation scales are not arbitrarily generated by the system, but are jointly registered based on the site soil background survey, rubber tree cultivation and management requirements, local agronomic control standards, amendment material application plans, sensor resolution capabilities, and on-site commissioning records. For the same site, the system can register target intervals separately for the fertilization layer, the root active layer, and the root migration risk layer; for example, the pH target interval for the root active layer is used to express the rhizosphere suitability state, and the heavy metal ion concentration control interval for the migration risk layer is used to express the downward migration safety state. It only serves as a normalization scale within the same indicator, and is not used to directly add different indicators, nor does it change the physical or chemical meaning of the indicator itself; when an indicator is not registered with a target range, control range, or engineering variation scale, the system does not calculate the improvement approximation of that indicator, but instead lists it as an indicator not involved in the judgment: ; in, Indicates the sampling time lower depth layer Indicators The improvement is close to the amount; To fix the depth layer numbering in the monitoring profile, by The depth number is obtained from the depth, such as D1, D2, or D3; As an indicator type, it consists of real-time data records and baseline profile data. The index type fields are jointly determined, including pH, conductivity, water content, organic matter, nitrogen content, phosphorus content, and heavy metal ion concentration; Sampling time lower depth layer Indicators Real-time readings are uploaded from fixed sensor nodes, in-situ detection modules, portable detector interfaces, or laboratory detection interfaces. for same depth layer and the same indicator The benchmark representative value; As an indicator The deviation calculation process for its target range is performed by the system after reading the target lower limit and target upper limit of the indicator. When the reading is lower than the target lower limit, the difference between the reading and the target lower limit is taken. When the reading is within the target range, the value is zero. When the reading is higher than the target upper limit, the difference between the reading and the target upper limit is taken. As an indicator The scale of engineering changes is pre-registered based on the sensor resolution, land management threshold, or agronomic control range of this indicator. , and All results were obtained for the same index m, and the numerator and denominator used the same measurement scale for the same index. This formula is used to represent the degree to which the indicator approaches the target range relative to a baseline. It is derived from the question of whether the real-time deviation from the baseline has decreased: if the real-time reading is closer to the target range than the baseline value, then... Less than , If the real-time reading is close to the reference value, then... Approaching zero; if the real-time reading is further away from the target range than the reference value, then The value is negative. Moisture migration events are formed by moisture direction markers from adjacent sampling times and adjacent depth layers, serving as event markers for subsequent migration risk assessment.

[0033] Understandably, taking the pH determination of the root active layer as an example, in plot A01, P01... D2 is the root active layer. The pH baseline value for this layer is recorded as 5.0 in D0. The system registers the target pH range for the root active layer of this plot as 4.5 to 6.5, with an engineering variation scale of 0.5. At sampling time t, P01 If the real-time pH reading uploaded by the D2 electrode is 5.6, then the reference value is lower than the target range with a deviation of 0.5; if the real-time reading is within the target range with a deviation of 0, then substituting these values ​​into the above relationship yields... The system labels this pH indicator as an acid-base buffer improvement. Taking the determination of heavy metal ion concentration as another example, P01... The baseline value for D2 is 0.42, the upper limit for safety control of this indicator is 0.30, the engineering variation scale is taken as 0.10, and the sampling time is... If the real-time reading is 0.30, the deviation of the baseline value from the upper limit of safety control is 0.12, the deviation of the real-time reading is 0, and the corresponding improvement approach value is 1.2. The system marks this heavy metal index as passivated and stable. If the real-time heavy metal ion concentration at the same depth layer rises to 0.48, the real-time deviation is greater than the baseline deviation, the improvement approach value is negative, and the system marks this index as evidence of reactivation or migration risk. The above values ​​are all from the baseline and real-time records of the same node, the same depth, and the same index, and can be directly written into the field commissioning record for verifying the system status judgment results.

[0034] Furthermore, after obtaining the approximate improvement values ​​for each indicator, the system generates directional markers according to the technical meaning of the indicators. For pH, a positive change indicates improved acid-base buffering, while a negative change indicates enhanced acidification. For conductivity, a reading remaining within the stable control range without continuous increases is recorded as a stable ionic environment, while a continuous increase and departure from the stable control range is recorded as salinity or ion accumulation. For moisture, the system combines the current sampling time with continuous records from adjacent sampling times to determine its change process. In the fertilized layer and root active layer, enhanced water retention capacity corresponds to improved water retention. In the migration risk layer below the roots, a short-term increase in moisture after rain, accompanied by a subsequent increase in conductivity or heavy metal ion concentration in the lower layer, is recorded as evidence of water migration risk. For organic matter, nitrogen content, and phosphorus content, entering or maintaining the supply range is recorded as organic matter retention and improved nutrient supply, respectively, while a continuous drop below the supply range is recorded as insufficient nutrient supply. For heavy metal ion concentration, a decrease or entry into the safe control range is recorded as passivation stability, while an increase and departure from the safe control range is recorded as evidence of reactivation or migration risk. The aforementioned directional markers, together with the hierarchical attributes, explain that the same reading changes have different engineering implications at different depth layers. For example, an increase in moisture content in the fertilization layer may support the judgment of improved water retention, while in the migration risk layer, when combined with an increase in conductivity and an increase in heavy metal ion concentration, it will support the judgment of downward migration risk.

[0035] Furthermore, after generating the direction markers, the system generates a deeply modified state according to the hierarchical attributes. The determination of the fertilization layer's condition focuses on whether the nano-biochar-based fertilizer begins to exert its effects of acid-base buffering, water and fertilizer retention, nutrient release, and heavy metal passivation in the applied area. When most indicators in the fertilization layer—pH, organic matter, nitrogen content, phosphorus content, and heavy metal ion concentration—show positive directional markers, and the conductivity remains stable, the layer is considered improved. When pH or organic matter has improved, but nutrient release or heavy metal passivation is not yet fully manifested, the layer is considered weakly improved. When the main indicators are compared to the baseline profile data... When changes are minor and do not show a clear direction of improvement, the layer is recorded as unimproved; when pH deviates from the suitable range, conductivity continuously increases, or heavy metal ion concentration rises, the layer is recorded as suspected abnormal improvement. The determination of the root activity layer focuses on whether the improvement effect has been transmitted to the main absorption area of ​​the rubber tree; when pH moves towards the suitable range, heavy metal ion concentration decreases or maintains a safe direction of change, and water, nitrogen, and phosphorus content support root absorption, the layer is recorded as improved; when the fertilized layer has shown improvement while the root activity layer only shows improvement in some indicators, the layer is recorded as weakly improved; when the root activity layer continuously approaches the baseline profile data... When the recorded state is as described, the layer is marked as unimproved. When the pH of the active root layer has improved but the concentration of heavy metal ions has not decreased, or the conductivity has continuously increased, the layer is marked as suspected abnormally improved. The state determination of the migration risk layer below the root system focuses on whether the fertilizer effect or pollution risk has shifted downward. When the layer remains close to the baseline state and there is no combination of sudden increase in moisture, increase in conductivity, and increase in heavy metal ion concentration, the state remains unimproved or weakly improved. When the layer experiences a sudden increase in moisture after rain, followed by an increase in conductivity or increase in heavy metal ion concentration, the layer is marked as suspected abnormally improved, and is used in the subsequent step three to identify whether the fertilizer effect boundary has shifted abnormally downward.

[0036] The aforementioned "most indicators," "continuous increase," "continuous approach," and "partial indicator improvement" are all fixedly registered by the system in the plot configuration table according to the sampling cycle and the number of indicators. Typically, pH, organic matter, nitrogen content, phosphorus content, and heavy metal ion concentration are used as core indicators of improvement effect, while moisture and electrical conductivity are used as auxiliary process indicators. Only when more than half of the core indicators reach a positive directional marker and no adverse risk is observed in the auxiliary process indicators can an improvement be registered. The determination cycle for continuous increase or continuous approach is jointly determined by the sampling frequency, the rainy season management cycle, and sensor stability; for example, it can be registered as at least two adjacent effective sampling cycles. If positive improvement evidence and migration risk evidence appear simultaneously at a certain depth layer, the system prioritizes registering it as a suspected abnormal improvement and writes the conflicting indicators into the status record for further judgment in steps three and four, rather than directly treating it as a normal fertilizer effect advancement result.

[0037] Based on the on-site operation process, three fixed depth layers were set in the P01 profile of a rubber plantation A01 plot, among which P01 D1 is the fertilization layer, P01 D2 is the root active layer, P01 D3 is the migration risk layer below the root system. Sampling time The system obtains data from the reference profile. Read P01 D1, P01 D2 and P01 The D3 indexes' benchmark values ​​and detection source markers are displayed, and real-time data at the corresponding depth is received. If P01 As the pH of D1 shifts towards the optimal range, organic matter increases, nitrogen and phosphorus contents are within the supply range, heavy metal ion concentrations decrease, and conductivity remains stable. middle Record as improved; if P01 The pH of D2 has shifted towards the optimal range, and the moisture content has improved, but the concentration of heavy metal ions has not yet decreased, and the nitrogen and phosphorus contents have changed little. Therefore, P01 D2 is denoted as a weak improvement; if P01 D3 experiences a sudden increase in moisture after rain, followed by an increase in conductivity and a departure of heavy metal ion concentration from the safe control range. Therefore, P01... D3 is designated as a suspected anomalous improvement. This leads to... It is not a set of alarms for a single indicator, but a depth state sequence arranged along a fixed profile, which can directly express the on-site process of "fertilization layer has taken effect, root activity layer is showing improvement response, and the lower migration risk layer shows signs of abnormal downward movement".

[0038] The output of this step is a deeply improved state. . By a fixed monitoring profile at the sampling time The data consists of state records for each depth layer. Each state record includes the plot number, profile number, node number, depth number, layer attribute, sampling time, directional markers for each indicator, and the improved state of that depth layer. The plot number, profile number, node number, depth number, and layer attribute are inherited from the baseline profile data. Sampling time is from real-time soil data; the orientation markers for each indicator are determined by the amount of soil amendment. Together with the dedicated judgment process for indicators, the improvement status of this depth layer is generated by the status judgment rules corresponding to the fertilization layer, root activity layer and the migration risk layer below the root system. The values ​​include improved, weakly improved, not improved, suspected abnormal improvement and insufficient data status.

[0039] This step uses reference profile data. By matching the same point and depth with real-time soil data, various real-time readings are converted into depth improvement status. The system utilizes the mathematical calculation relationship of the distance from a point to the target interval to determine whether each indicator is closer to the improvement target relative to the baseline state. Then, it combines the engineering meaning of different depth layers in the improvement process of nano-biochar-based fertilizer to form the state records of the fertilization layer, root activity layer, and migration risk layer.

[0040] S3. Obtain the state of all depth layers arranged from shallow to deep in the same fixed monitoring profile, identify the improved or weakly improved depth layers as effective action layers, and register the abnormal state of water displacement in the migration risk layer below the root system accompanied by increased conductivity or increased heavy metal ion concentration as downward migration risk; determine the fertilizer effect boundary position reached by the effective action layer and downward migration risk based on the effective action layer and downward migration risk, and judge the boundary movement state by the boundary position change in adjacent sampling cycles; wherein, the fertilizer effect boundary state includes the boundary movement state, root active layer coverage marker and abnormal downward migration marker.

[0041] Specifically, this step directly organizes these depth layer states into a profile state sequence according to the vertical order of the fixed profile. Plot number and profile number are used to determine a fixed monitoring profile within the same rubber plantation improvement area; node number and depth number are used to determine the arrangement of each soil layer in that profile; hierarchical attributes are used to identify the fertilization layer, the active rubber tree root layer, and the layer below the root system with migration risk; sampling time is used to establish a boundary change relationship with the previous sampling period; directional markers for each indicator are used to identify whether there are abnormal processes near the boundary, such as water displacement, increased conductivity, or increased heavy metal ion concentration; and the improvement state of this depth layer is used to determine the deepest layer reached by the normal and effective action of the nano-biochar-based fertilizer. After this step, the layers are distributed sequentially... It is converted into fertilizer effect boundary location, boundary movement state, root active layer coverage marker, abnormal downward movement marker, and boundary adjacent depth range, which together constitute the fertilizer effect boundary state. .

[0042] Among them, the system first follows The plot number and profile number are used to extract the state of all depth layers of the same fixed monitoring profile, and they are arranged from shallow to deep according to the depth number.

[0043] For example, the P01 section of plot A01 includes P01. D1, P01 D2 and P01 D3 consists of three depth layers, with hierarchical attributes of fertilization layer, root activity layer, and root-below migration risk layer, arranged in an ordered sequence from the fertilization layer to the deep risk layer. This ordered sequence inherits the fixed monitoring profile relationship determined in step one and the layer-by-layer improvement status formed in step two, enabling the system to determine the arrival location of the improvement effect in the vertical profile. For each depth layer, the system reads its improvement status and generates an effective effect marker: depth layers with a status of "improved" or "weakly improved" are registered as effective effect layers; depth layers with a status of "not improved" or "insufficient data" are registered as unreached layers; depth layers with a status of "suspected abnormal improvement" combine their indicator direction markers to generate a downward migration risk marker or abnormal status record. If the root-below migration risk layer has a water downward migration direction marker and is accompanied by at least one direction marker of increased conductivity or increased heavy metal ion concentration, the system registers it as a downward migration risk layer; other suspected abnormal improvement states are entered into... The abnormal downward shift marker or the risk type judgment in step four. Thus, The textual status and indicator direction markers are converted into vertical status sequences that can be used for boundary extraction and risk registration.

[0044] The determination of the fertilizer effect boundary location is derived from the boundary element selection method of finite ordered sets in mathematics. This type of method typically selects the last element in a set of ordered elements that satisfies a certain condition as the boundary. This application uses this method to identify the fertilizer effect range in a fixed soil profile: in a set of depth layers arranged from shallow to deep, the deepest layer that satisfies the effective effect condition and is not registered as a layer at risk of downward displacement is selected as the current fertilizer effect boundary location; when no effective effect layer satisfying the condition appears at the current sampling time, the system registers the fertilizer effect boundary location as not having formed an effective boundary. The specific calculation relationship is as follows: ; in, Indicates the sampling time The location of the fertilizer effect boundary is extracted from the depth state sequence of the current fixed monitoring profile; This represents a depth layer number within a fixed monitoring profile, derived from the output of step two. ; Indicates the sampling time The set of depth layer numbers within the same plot and the same profile, by The depth numbering in the data is composed of; Indicates satisfaction and The effective action layer set; Representing depth layer The depth sequence coding is determined by the depth number of the fixed monitoring profile; Indicates the sampling time lower depth layer The effective function marker is determined by The improved state transition of the middle depth layer is obtained. When the state is "improved" or "weakly improved", it is 1, and when the state is "not improved", "suspected abnormal improvement" or "insufficient data state", it is 0. Indicates the sampling time lower depth layer The downward shift risk marker, by The hierarchical attributes of the depth layer and the directional markers of each indicator are used to determine the value of 1 when water moves downward in the migration risk layer below the root system and is accompanied by an increase in conductivity or heavy metal ion concentration; otherwise, the value is 0. This indicates a vacancy state where no effective fertilizer effect boundary has been formed; the system sequentially encodes and registers this state as... The left and right sides of this equation represent the registration results of the fertilizer effect boundary position. The calculation process is used to select boundary elements or register states where no effective boundary has been formed in a finite ordered set. This equation is derived from the engineering criterion of "deepest effective layer": first use... Limit the normal effective layer, then use Excluding abnormal layers caused by downward shift risk, finally using Select the depth layer with the largest sequential code; when the set of effective layers is empty, the system registers... It is used for subsequent boundary displacement and determination of uncovered root layers.

[0045] In actual calculations, if the P01 section of plot A01 includes P01... D1, P01 D2 and P01 D3, the system can register its sequential codes as 1, 2, and 3 respectively. At sampling time t, if... Display P01 D1 is the improved version, P01 D2 is a weak improvement, P01 D3 is unmodified, then P01 D1 and P01 The effective action markers for D2 are all 1, P01 The effective action flag for D3 is 0, and none of the three depth layers have triggered the downshift risk flag. The system selects P01, which has the largest sequential code, from the depth layers that meet the conditions. D2, obtained .like The system shows that all three depth layers are unmodified, and the system obtains... Register Lt as This indicates that an effective fertilizer effect boundary has not yet been formed in the current profile. If at another sampling time... Display P01 D1 is the improved version, P01 D2 is the improved version, P01 D3 is a suspected anomalous improvement, and P01 The directional markers at D3 indicate a downward shift of water and an increase in heavy metal ion concentration, therefore P01 The risk of D3 shifting downwards is marked as 1, and the system will set P01... D3 is used as the record for the downshifted risk layer, and P01 is selected in the normal effective layer. D2 as .

[0046] Furthermore, determine the location of the fertilizer efficiency boundary. Then, the system reads the fertilizer effect boundary state saved in the previous sampling period. To obtain the fertilizer efficiency boundary position of the previous cycle Here This is the boundary record formed and stored in the previous sampling period in this step. This represents the sampling interval between two adjacent boundary determinations. The system employs the adjacent sequence difference concept from finite state transitions, comparing the sequential codes of the current boundary position with those of the previous cycle to form the boundary displacement. : ; in, Indicates the sampling time The displacement of the fertilizer effect boundary; The location of the fertilizer effect boundary at the current sampling moment is obtained from the aforementioned boundary extraction relationship; This indicates the location of the fertilizer effect boundary in the previous sampling period, based on the fertilizer effect boundary state saved in the previous period. Read; This represents the sequential code corresponding to the current fertilizer effect boundary position, when Time to take ; This represents the sequential code corresponding to the fertilizer effect boundary position of the previous cycle, when Time to take Both sides of this formula are depth layer sequence differences, and the calculation results are used to describe the relative movement of the boundary from shallow to deep along a fixed profile. This formula is derived from the boundary sequence difference between adjacent sampling periods: when the current boundary sequence code is greater than that of the previous period, the boundary advances into deeper layers; when the current boundary sequence code is equal to that of the previous period, the boundary remains at the same depth layer or remains in a state where no effective boundary has been formed; when the current boundary sequence code is less than that of the previous period, the boundary retreats into shallower layers.

[0047] Furthermore, when the fertilizer-efficiency boundary state is first formed, if the previous cycle's... The system registers the fertilizer efficiency boundary position of the previous cycle as This initial record is marked as the initial boundary determination in the record; it is only used to establish a subsequent displacement comparison benchmark and does not directly generate boundary regression or stagnation conclusions. The interval between adjacent sampling periods... The sampling period of the IoT platform and the soil improvement response observation period are jointly determined. The stagnation judgment period of continuous records in the same layer is registered by the plot configuration table, which is usually not less than two valid sampling periods. Within this period, there must be no records of sensor offline, missing key indicators, or unresolved external operation disturbances.

[0048] For example, in the previous sampling period Its sequential code is 1; the current sampling time Its sequential encoding is ,but The system registration boundary is progressing normally. If the boundaries of both the previous cycle and the current cycle are... ,but The system continues to monitor the number of consecutive cycles where the boundary remains at the same layer; if multiple consecutive sampling cycles remain at the same layer, and the root active layer has not changed from weakly improved to improved, the system registers it as stagnant. If the previous cycle... Current cycle ,but The system registers boundary rollback. If both the previous period and the current period are... ,but The system registers the situation as not having formed a valid boundary and simultaneously registers the uncovered active root layer. If the risk marker for the downward shift of the migration risk layer in the current cycle is triggered, the system is in the fertilizer effectiveness boundary state. The system synchronously registers abnormal downward movement markers, enabling S4 to identify risk types by calling upon the sequence of changes in moisture, conductivity, and heavy metal ion concentrations within the depth range adjacent to the boundary.

[0049] Furthermore, the system then determines the fertilizer efficiency boundary position. fertilizer effect boundary displacement And shifting risk markers to generate fertilizer efficiency boundary states .when And when the downward risk marker is not triggered, the boundary motion state is registered as normal advancement; when And the boundary remains continuously at the same depth layer for a preset period, and the boundary motion state is registered as stagnant; when When the boundary motion state is initially registered as regressive, and when the downward migration risk marker of the migration risk layer below the root system is triggered, the boundary motion state is registered as abnormal downward migration. The root active layer coverage marker is determined by... and The middle-level attribute is obtained by comparing the depth number of the root active layer: If If an effective action layer is located in the root active layer or deeper than the root active layer, and the downward risk layer is not interpreted as a normal effective action boundary, it is registered as a covered root active layer; if If the root active layer is located in the fertilization layer and is not improved, it is registered as an uncovered root active layer; if the root active layer is weakly improved but has not been continuously converted to improved, it is registered as covered but insufficiently improved, used for stagnation observation or fertilization depth verification in step four; if the root active layer is in a state of insufficient data, it is registered as coverage cannot be confirmed, and only a supplementary sampling or verification prompt is generated in step four, without directly triggering the risk of fertilizer effect not reaching the root layer. The depth range adjacent to the boundary is determined by... The depth layer and its adjacent layers constitute, for example At that time, the depth range near the boundary includes P01. D1, P01 D2 and P01 D3; At that time, the depth range near the boundary includes P01. D1 and P01 D2. The depth range adjacent to this boundary is determined by... The depth number and hierarchical attributes are directly generated and used as the target area for analyzing the order of parameter changes in step four.

[0050] In field applications, the fertilizer-efficiency boundary state It can simultaneously express engineering information on both fertilizer effectiveness coverage and risk mitigation. land parcel Taking the cross-section as an example, the fertilizer efficiency boundary position in the previous cycle Current cycle Display P01 D1 is the improved version, P01 D2 is a weak improvement, P01 If D3 is unmodified, then the system obtains , The boundary movement is in a normal progression state, the root active layer is marked as covered, and the depth range of the boundary's vicinity is P01. D1, P01 D2 and P01 D3. If P01 in the next cycle D2 changed from weak improvement to improved, while P01 If D3 is still not improved, then Still , The system registration boundary maintains and continuously covers the active root layer. If a certain cycle occurs after rain... If the system is registered as a suspected abnormal improvement in step two, and the directional markers show a downward shift in moisture content, an increase in conductivity, and an increase in heavy metal ion concentration, then the system is in... Register anomaly downshift markers in the middle and keep the depth range near the boundary at P01. D1, P01 D2 and P01 D3 enables step four to focus on reading the sequence of changes in moisture, pH, conductivity, and heavy metal ion concentrations within this range. If P01 is in the drought or acidification rebound phase... D2 changed from improved to unimproved, while P01 If D1 is still a weak improvement, then the system obtains , The boundary motion state is registered as a regression, and a risk analysis of fertilizer effect decay or acidification rebound is triggered in step four.

[0051] The output of this step is the fertilizer effect boundary state. This includes plot number, profile number, sampling time, fertility boundary location, boundary movement status, root active layer cover marker, anomaly downslope marker, and boundary proximity depth range. Plot number, profile number, and sampling time are inherited from... The location of the fertilizer effect boundary is determined by... Indicates; the boundary motion state is determined by The root active layer coverage markers are determined jointly by continuous records in the same layer and downshift risk markers; the root active layer coverage markers are determined by... The relative position with the active root layer is determined; the abnormal downward migration marker is determined by the downward migration risk marker of the migration risk layer below the root system; the depth range near the boundary is determined by... The depth of the layers and their adjacent depths is determined.

[0052] Understandably, this is a deeply improved state. Derived as fertilizer efficiency boundary state This allows for further assessment of whether the fertilizer effect reaches the root-active layer, whether it stagnates, regresses, or abnormally migrates downwards, based on the results of each layer's improvement. This treatment directly serves the core scenario of real-time monitoring of the rubber plantation soil environment: after the application of nano-biochar-based fertilizer, its effective effect is manifested layer by layer in the soil profile. The root protection effect of rubber trees depends on whether the fertilizer effect boundary covers the root-active layer, while the risk of heavy metal migration during the rainy season depends on whether abnormal downward migration of water and ions occurs near the boundary and in the risk layer below.

[0053] S4. Based on the fertilizer effect boundary state, identify the risk type and generate a risk warning instruction record for the corresponding time.

[0054] Specifically, the plot number and profile number in the fertilizer efficiency boundary state are used to determine the rubber plantation improvement area corresponding to this warning; the sampling time is used to form the time index of the warning record; the fertilizer efficiency boundary location is used to determine the core soil layer for risk assessment; the boundary movement status is used to determine which boundary process (advancement, stagnation, retreat, or abnormal downward shift) the current risk assessment begins from; the root activity layer coverage marker is used to determine whether the improvement effect of nano-biochar-based fertilizer has acted on the main absorption area of ​​the rubber trees; the abnormal downward shift marker is used to determine whether the risk assessment of leaching and heavy metal migration during the rainy season has begun; and the boundary proximity depth range is used to limit the range of soil parameters called in this step. The system reads the indicator direction markers and corresponding real-time parameter records formed in step two within the boundary proximity depth range given by the fertilizer efficiency boundary state, including whether pH is moving towards the suitable range, whether electrical conductivity is continuously increasing, whether moisture is shifting downward, whether organic matter is improving, whether nitrogen and phosphorus content are within the supply range, and whether heavy metal ion concentration is decreasing or increasing. Thus, the object of risk assessment is limited to the layer where the fertilizer efficiency boundary is located and its adjacent layers, rather than a generalized assessment of the entire plot; for example, the fertilizer efficiency boundary state... The location of the fertilizer efficiency boundary of the P01 section of plot A01 is shown as P01. D2, the boundary adjacent depth range is P01 D1, P01 D2 and P01 D3, the system calls the direction markers and real-time parameter records of these three depth layers at the current sampling time and adjacent historical sampling times, among which P01 D1 is used to determine whether the fertilizer source in the fertilization layer is stable, P01 D2 is used to determine whether the active root layer is protected, P01 D3 is used to determine whether water, ions, and heavy metals are undergoing downward migration.

[0055] The risk type identification adopts the concept of finite event sequence discrimination in engineering monitoring. The original form of this concept involves recording the first occurrence of several key events within a limited time window and determining the process type based on the order of these events. In this application, this concept has been modified into a risk identification method based on "fertilizer-effect boundary state triggering + boundary adjacent soil layer event chain matching + continuous confirmation." Each risk type... This corresponds to a three-event chain determined by the soil improvement process in rubber plantations. The risks of leaching or heavy metal migration during the rainy season correspond to water displacement, pH deviating from the suitable range, and increased heavy metal ion concentration; the risk of acidification rebound-type fertilizer effectiveness decline corresponds to fertilizer effectiveness boundary regression, root active layer pH deviating from the suitable range, and decreased root active layer improvement status; the risk of fertilizer effectiveness not reaching the root layer corresponds to boundary stagnation, the fertilized layer remaining improved, the root active layer remaining unimproved, or the root active layer remaining weakly improved but not turning into an improved state; the risk of insufficient heavy metal passivation corresponds to root active layer pH improvement, no decrease in heavy metal ion concentration, and continuous existence of insufficient passivation; the risk of insufficient nutrient release corresponds to decreased nitrogen or phosphorus content, no improvement in organic matter, and boundary stagnation or regression. The system reads the first occurrence time of these events within the depth range near the boundary specified in the fertilizer effectiveness boundary state, and combines this with the boundary movement state, root active layer coverage marker, and abnormal downward displacement marker in the fertilizer effectiveness boundary state to obtain the risk trigger value: Furthermore, the identification window for the event chain is jointly determined by the sampling period, the time of soil moisture displacement after local rainfall, the sensor upload period, and the update period of the periodic detection index, and is written into the risk type configuration table. For continuous sensor indices, the first occurrence time is the first valid sampling time that meets the direction marking condition; for periodic detection indices, the first occurrence time is the time when the detection result is uploaded and passes the depth binding verification. If an event in the three event chains cannot be confirmed due to missing data, unverified changes in the detection source, sampling time exceeding the window, or the depth layer not being within the depth range adjacent to the boundary, then the sequential relationship of that event chain is invalid; the system can retain early warnings formed by boundary triggers, but does not treat unconfirmed event chains as complete risk processes.

[0056] ; in, Indicates the sampling time Risk type The risk trigger level is used as input for subsequent warning level determination; The system indicates the preset risk types, including the risk of leaching or heavy metal migration during the rainy season, the risk of fertilizer effectiveness decline due to acidification rebound, the risk of fertilizer effectiveness not reaching the root zone, the risk of insufficient passivation of heavy metals, and the risk of insufficient nutrient release. Indicates risk type The boundary trigger marker is determined by the boundary motion state, root active layer coverage marker, and abnormal downward movement marker in the fertilizer effect boundary state. The trigger is determined by... If not triggered, take For example, when an abnormal downward shift marker is triggered, there is a risk of leaching during the rainy season or migration of heavy metals. Pick When the boundary recedes, the risk of acidification rebound-type fertilizer effect decline increases. Pick When the fertilizer application rate is low and the root zone is not covered, the risk of fertilizer not reaching the root zone is high. Pick ; , and These represent the risk types respectively. The first occurrence time of the first, second and third key events in the three-event chain within the depth range near the boundary is determined by the indicator direction marker and real-time parameter record within that range. If a key event does not occur, the sequential relationship of the event chain is determined to be invalid. This indicates a logical decision process; if the condition within the parentheses is true, then... Take when the condition is not met ; Indicates risk type The number of consecutive hold times in the current sampling time and previous consecutive sampling periods is determined by the boundary trigger marker of the current sampling time, the event chain determination result, and historical risk warning instructions. Jointly determine: Risk type at the current sampling time When a boundary trigger or event chain is established, if the previous cycle If the same type of risk remains unresolved, then If this risk type was not recorded in the previous period, then Risk type at the current sampling time When not triggered, This formula is derived from the discrimination of finite event sequences: first from... Once it is confirmed that the fertilizer effect boundary has reached a boundary state corresponding to the risk type, then... It was confirmed that soil incidents occurred sequentially according to the formation process of this risk type, ultimately led by... This confirms that the risk process persists in adjacent sampling periods. Each term in the formula represents a status marker, event sequence, or consecutive count determination result; the calculation result is used for risk process triggering records.

[0057] It is important to note that this risk triggering factor does not perform a weighted summation of raw readings of different dimensions such as pH, conductivity, moisture, nutrient concentration, or heavy metal ion concentration. Instead, based on the directional markers established in step two and the boundary states established in step three, it counts the three logical results: boundary triggering, event sequence, and continuous maintenance. Therefore, The numerical value represents the completeness of risk evidence, not the physical quantity of soil parameters. Subsequent warning levels will be graded using only this logical count.

[0058] In the on-site calculation of the risk of leaching or heavy metal migration during the rainy season, it is assumed that the fertilizer effect boundary status of the P01 profile of plot A01 shows that the fertilizer effect boundary is located at P01. D2, the boundary adjacent depth range is P01 D1, P01 D2 and P01 D3, the abnormal downward shift flag has been triggered, so Br,t for this risk type is set to 1. The system reads the direction flags of the three most recent sampling periods within this range and finds the first sampling time P01. D3 shows a marker indicating the direction of moisture migration, and the second sampling time is P01. D2 shows a pH direction marker indicating a departure from the suitable range, and the third sampling time is P01. D3 shows a marker indicating an increasing direction of heavy metal ion concentration, and all three conditions are met. The corresponding logical judgment is taken If the previous cycle If the same risk type remains unresolved, then Logical judgment Substituting into the equation yields... If this is the first time this risk type has been triggered, the previous cycle... If this risk type is not recorded, then The logical decision corresponding to the continuous maintenance condition is taken. ,at this time The above results are used to distinguish between warning records where the event chain has been formed but is only in its initial triggering stage, and handling records where the event chain has been continuously maintained. If, under the same circumstances, only the abnormal downward movement marker is triggered, and the pH decrease and heavy metal ion concentration increase in the event chain have not yet occurred in sequence, then the event chain logic judgment takes... The system generates early warning records, primarily based on drainage and leaching reduction measures.

[0059] Furthermore, for the risk of fertilizer effectiveness decline due to acidification rebound, the system reads the boundary movement state and location from the fertilizer effectiveness boundary status. When the boundary movement state is regression, the system reads the pH direction marker, improvement state change, and fertilization layer status within the depth range adjacent to the boundary. If the event chain shows that the fertilizer effectiveness boundary first retreats from the root active layer to the fertilization layer, then the pH of the root active layer moves away from the suitable range, and then the root active layer changes from improved or weakly improved to unimproved, then the event chain for this risk type is established. The system sets the target depth layer as the root active layer, and the management actions are directed towards acidification treatment and supplementation of buffer-type amendment materials. If nitrogen and phosphorus content also decrease within the depth range adjacent to the same boundary, the system records the risk warning instruction. The system adds fertilization prompts to cover both acidification rebound and fertilizer effectiveness decay in on-site treatment. For risks where fertilizer effectiveness does not reach the root zone, the system reads the root activity layer coverage marker and boundary movement status from the fertilizer effectiveness boundary status. When the root activity layer coverage marker shows no coverage, the boundary movement status is continuously registered as stagnant, and the depth improvement status generated in step two shows that the fertilization layer remains improved while the root activity layer remains unimproved, the system registers this risk type. When the root activity layer has been weakly improved but has not turned improved for multiple consecutive effective sampling cycles, the system registers it as evidence of insufficient improvement as a lower intensity case of the same risk, and positions the management action as checking the depth of the fertilization trench, adjusting the fertilization depth, or applying nano-biochar-based fertilizer near the root activity layer.

[0060] For the risk of insufficient heavy metal passivation, the system uses the root active layer coverage marker in the fertilizer effect boundary state as the starting point. When the fertilizer effect boundary has covered the root active layer, and the pH of the root active layer has moved towards the suitable range, but the heavy metal ion concentration has not decreased, remains high, or shows an increasing direction marker, the system registers the risk of insufficient heavy metal passivation. This risk indicates that the acid-base buffering effect has entered the root layer, but the heavy metal adsorption and passivation effect has not yet stabilized. The system sets the target depth layer as the root active layer and the adjacent layer below it, and the management actions include supplementing with biochar-based amendments or heavy metal blocking materials with adsorption and passivation effects. For the risk of insufficient nutrient release, the system reads the nitrogen content, phosphorus content, and organic matter direction markers within the depth range near the boundary, and makes a judgment based on the boundary stagnation or regression status in the fertilizer effect boundary state. When nitrogen or phosphorus content is continuously below the supply range, organic matter does not show signs of improvement, and the boundary movement is stagnant or regressive, the system registers the risk of insufficient nutrient release. If the risk mainly occurs in the fertilization layer, the management action is to supplement with nano-biochar-based fertilizer or adjust the fertilizer ratio. If the risk mainly occurs in the root active layer, the management action is to supplement fertilizer near the root active layer and adopt a slow-release nitrogen and phosphorus supply method.

[0061] Furthermore, after determining the risk type, the system generates an early warning level using the hierarchical response concept from engineering early warning systems. The original form of this concept is to classify response levels based on trigger intensity, affected objects, and urgency of response. In this application, trigger intensity is represented by the risk trigger quantity, affected objects by the target depth layer, and urgency of response by the movement state of the fertilizer-efficiency boundary. In rubber plantation soil improvement, the active root layer directly relates to rubber tree absorption and rhizosphere safety, while the migration risk layer below the roots directly relates to the downward movement of heavy metals and nutrients. Therefore, when the target depth layer enters the aforementioned layers, the system increases the early warning level determination quantity. The early warning level determination quantity is formed by the following formula: ; in, Indicates the sampling time Risk type The amount of information used to determine the warning level; The aforementioned risk trigger quantity is formed by the boundary trigger, the sequence of events, and the continuity of events. This indicates an increased level of the target, determined by the risk type. The target depth layer is determined by taking a value of 0 when the target depth layer is only located in the fertilization layer, and taking a value of 1 when the target depth layer includes the root active layer or the migration risk layer below the root system. The term "urgent management item" is determined by the boundary motion state in the fertilizer effect boundary state. It is set to 0 when the boundary state is normal advancement, 1 when the boundary state is stagnant or the active root layer is not covered, and 2 when the boundary state is retreating or abnormally shifting downwards. This formula is derived from the engineering graded response rule: first use... Indicate whether the risk process has been fully formed, then use This indicates whether the risk affects the root active layer or the migration risk layer, and finally uses... This indicates the urgency of the current boundary state requiring on-site handling. The system, based on... The value is used to register the early warning level: Registered as a follower. It was registered as a warning level at that time. It was registered as a disposal level at that time.

[0062] Furthermore, in the calculation and implementation of the warning level, if the aforementioned risks of rain leaching or heavy metal migration have been mitigated... The target depth layer includes the root active layer P01. D2 and the migration risk layer below the root system P01 D3, then Meanwhile, in the boundary motion state of the fertilizer effect boundary state, if the boundary motion state is abnormally downward, then... Substituting, we get The system registered this risk as a "handling" level risk. The corresponding generated... In the diagram, the target plot is A01, the target profile is P01, and the target depth layer is P01. D2 and P01 D3, the risk type is rainy season leaching or heavy metal migration risk, the warning level is the response level, and the management instructions include drainage treatment, leaching reduction treatment by covering, and supplementing heavy metal blocking materials in the layer near the boundary. If, at another sampling time, the fertilizer effect boundary is stagnant and the root active layer is not covered, the event chain is only partially established, the risk trigger value is 1, and the target depth layer includes the root active layer, then... If the boundary state is stagnation, then ,get The system registers the risk as an early warning level, and the management instruction is to check the fertilization depth and apply nano-biochar-based fertilizer near the active root layer.

[0063] The management instructions are generated by combining risk type, target depth layer, and warning level. Risks of leaching or heavy metal migration during the rainy season correspond to drainage treatment, mulching to reduce leaching, and supplementation of heavy metal blocking materials; the target depth layers are the root active layer and the layer below the root system at risk of migration. Risks of acidification rebound-type fertilizer effectiveness decline correspond to acidification treatment of the root active layer and supplementation of buffer-type amendment materials. Risks of fertilizer effectiveness not reaching the root layer correspond to adjusting the fertilization depth, verifying the depth of the fertilization trench, and supplementing with nano-biochar-based fertilizer near the root active layer. Risks of insufficient heavy metal passivation correspond to supplementing adsorption and passivation materials and rechecking heavy metal ion detection. Risks of insufficient nutrient release correspond to supplementing with nano-biochar-based fertilizer, adjusting the nitrogen and phosphorus supply method, and checking the stability of electrical conductivity. Each management instruction includes the target plot, target profile, target depth layer, risk type, warning level, and execution action. The IoT platform pushes these instructions to the management terminal and generates a transmission record. The transmission record includes the transmission time, receiving terminal number, risk type, warning level, and target layer, used by on-site personnel to locate specific rubber plantation areas and specific soil layers.

[0064] Finally, this step outputs a risk warning instruction record. This record includes the plot number, profile number, sampling time, risk type, warning level, target depth layer, boundary proximity depth range, management instructions, and transmission record. The plot number, profile number, sampling time, and boundary proximity depth range are inherited from the fertility boundary state; the risk type is determined by the boundary movement state of the fertility boundary state, the root active layer coverage marker, the abnormal downward movement marker, and the event chain sequence within the boundary proximity depth range. The warning level is determined by the risk trigger level, the target level aggravation factor, and the management urgency factor; the target depth layer is determined by the fertilizer effect boundary position, the root activity layer coverage marker, and the abnormal downward movement marker; management instructions are generated by corresponding risk types and target depth layers; and the sending record is generated by the IoT platform when pushing to the management terminal. Through risk warning instruction recording, the system transforms the fertilizer effect boundary state into a clear risk type, warning level, and on-site management action.

[0065] This invention also provides an IoT-based real-time monitoring system for improved rubber plantation soil environment, the system comprising: The benchmark profile management module is used to fix the fertilizer application layer, the active root layer of rubber trees, and the migration risk layer below the root system as fixed monitoring profiles within the area where nano-biochar-based fertilizer is applied; and to construct several benchmark profile data based on the benchmark data of each depth layer of the fixed monitoring profile. The depth improvement status generation module is used to collect real-time soil data at various depth layers, compare it with the baseline profile data at the same point and depth, and generate the improvement proximity amount for the corresponding depth layer and the corresponding indicator by comparing the baseline representative value and real-time reading of the same depth layer and the same indicator. Based on the improvement proximity amount, directional markers are generated according to the technical meaning of the indicator, and the depth improvement status is generated according to the hierarchical attributes. The depth improvement status includes the plot number, profile number, node number, depth number, hierarchical attributes, sampling time, directional markers of each indicator, and the improvement status of the corresponding depth layer for each depth layer in the fixed monitoring profile. The improvement status includes improved, weakly improved, and suspected abnormal improvement. The fertilizer effect boundary identification module is used to acquire the state of all depth layers arranged from shallow to deep within the same fixed monitoring profile, identify the improved or weakly improved depth layers as effective action layers, and register abnormal states in the migration risk layer below the root system where water shifts downwards accompanied by increased conductivity or increased heavy metal ion concentration as downward migration risk. Based on the effective action layers and downward migration risk, the module determines the fertilizer effect boundary position reached by the effective improvement effect of the nano-biochar-based fertilizer, and judges the boundary movement state by the boundary position change in adjacent sampling periods. The fertilizer effect boundary state includes the boundary movement state, root active layer coverage marker, and abnormal downward migration marker. The risk warning and management instruction output module is used to identify the risk type based on the fertilizer effect boundary state and generate a risk warning instruction record at the corresponding time.

[0066] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0067] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for real-time monitoring of improved rubber plantation soil environment based on the Internet of Things, characterized in that, Includes the following steps: Within the area where the nano-biochar-based fertilizer is applied, fixed monitoring profiles are established according to the fertilizer application layer, the active root layer of the rubber tree, and the migration risk layer below the root system. Several benchmark profile data are constructed based on the benchmark data of each depth layer of the fixed monitoring profiles. Real-time soil data at each depth layer is collected and compared with the baseline profile data at the same point and depth. By comparing the baseline representative value and real-time reading of the same depth layer and the same index, the improvement approximation amount of the corresponding depth layer and the corresponding index is generated. Based on the improved proximity quantity, directional markers are generated according to the technical meaning of the indicators, and depth improvement status is generated according to the hierarchical attributes. The depth improvement status includes the plot number, profile number, node number, depth number, hierarchical attributes, sampling time, directional markers of each indicator, and improvement status of the corresponding depth layer in each depth layer of the fixed monitoring profile. The improvement status includes improved, weakly improved, and suspected abnormal improvement. The system acquires and extracts the states of all depth layers arranged from shallow to deep within the same fixed monitoring profile. The improved or weakly improved depth layers are identified as effective action layers. Abnormal states in the migration risk layer below the root system, where water content shifts downwards accompanied by increased conductivity or heavy metal ion concentration, are registered as downward migration risks. Based on the effective action layers and downward migration risks, the effective improvement boundary position of the nano-biochar-based fertilizer is determined, and the boundary movement state is judged by the boundary position changes in adjacent sampling periods. The fertilizer effect boundary state includes the boundary movement state, root active layer coverage markers, and abnormal downward migration markers. Based on the fertilizer efficiency boundary state, risk type identification is performed, and risk warning instruction records are generated for the corresponding time.

2. The method for real-time monitoring of improved rubber plantation soil environment based on the Internet of Things according to claim 1, characterized in that, The acquisition time of the reference profile data is based on the stable record before fertilization. If the sensor is deployed after fertilization, the record of the initial stable period after fertilization is used as the reference.

3. The method for real-time monitoring of improved rubber plantation soil environment based on the Internet of Things according to claim 1, characterized in that, The baseline data includes pH, conductivity, water content, organic matter, nitrogen content, phosphorus content, and heavy metal ion concentration data; The benchmark profile data consists of benchmark records for each depth layer within a fixed monitoring profile; each benchmark record includes a plot number, profile number, node number, depth number, layer attribute, index type, benchmark representative value, benchmark acquisition time period marker, and detection source marker.

4. The method for real-time monitoring of improved rubber plantation soil environment based on the Internet of Things according to claim 3, characterized in that, The process of generating directional markers based on the improved proximity quantity according to the technical meaning of the index, and generating a depth-improved state according to the hierarchical attributes, specifically involves: Based on the improved proximity measurement shown, determine whether the same indicator at the same layer and the same node is close to the target interval: When the baseline data forms a positive directional marker and the conductivity remains stable, the current layer is marked as improved; when some baseline data has improved but nutrient release or heavy metal passivation has not yet been fully demonstrated, the current layer is marked as weakly improved; when the pH or organic matter changes little relative to the baseline profile data and does not show a clear direction of improvement, the current layer is marked as unimproved; when the pH is far from the suitable range, the conductivity increases continuously, or the concentration of heavy metal ions increases, the current layer is marked as suspected abnormally improved.

5. The method for real-time monitoring of improved rubber plantation soil environment based on the Internet of Things according to claim 1, characterized in that, The determination of the fertilizer effect boundary reached by the effective improvement effect of nano-biochar-based fertilizer based on the effective action layer and the risk of downward migration is specifically as follows: In the depth layers arranged from shallow to deep in the fixed profile, the deepest layer that meets the effective action layer and is not registered as having a risk of downward displacement is selected as the current fertilizer effect boundary position; when no effective action layer meets the conditions at the current sampling time, the fertilizer effect boundary position is registered as not having formed an effective boundary.

6. The method for real-time monitoring of improved rubber plantation soil environment based on the Internet of Things according to claim 5, characterized in that, The boundary motion state is determined by comparing the current fertilizer effect boundary position with the fertilizer effect boundary position of the previous sampling period. Specifically, when the boundary motion state is greater than 0 and the downward risk marker is not triggered, the current boundary motion state is registered as normal advancement; when the boundary motion state is equal to 0 and the current boundary motion state is continuously maintained at the same depth layer for a preset period, the current boundary motion state is registered as stagnation; when the boundary motion state is less than 0, the current boundary motion state is registered as regression; when the downward risk marker of the migration risk layer below the root system is triggered, the current boundary motion state is registered as abnormal downward movement. The root active layer coverage marker is obtained by comparing the fertilizer effect boundary position with the depth number of the root active layer in the depth improvement state.

7. The method for real-time monitoring of improved rubber plantation soil environment based on the Internet of Things according to claim 1, characterized in that, The risk type identification based on the fertilizer effect boundary state specifically includes: Based on the risk assessment path triggered by the fertilizer effect boundary state, within the boundary layer and its adjacent layers, the sequence of the three event chains is determined according to the first occurrence time of each key event; the risk of leaching or heavy metal migration during the rainy season corresponds to the event chain of decreased water content, pH moving away from the suitable range, and increased heavy metal ion concentration; the risk of acidification rebound-type fertilizer effect attenuation corresponds to the event chain of boundary regression, pH moving away from the suitable range in the root active layer, and decreased improvement status. The risk of fertilizer efficiency not reaching the root zone corresponds to an event chain of stagnation at the boundary, the fertilization layer remaining improved, and the root active layer remaining unimproved; the risk of insufficient heavy metal passivation corresponds to an event chain of unimproved pH in the root active layer, no decrease in heavy metal ion concentration, and continuous existence of insufficient passivation; the risk of insufficient nutrient release corresponds to an event chain of decreased nitrogen or phosphorus content, no improvement in organic matter, and stagnation or regression at the boundary.

8. The method for real-time monitoring of improved rubber plantation soil environment based on the Internet of Things according to claim 1, characterized in that, The risk warning instruction record includes the plot number, profile number, sampling time, risk type, warning level, target depth layer, depth range near the boundary, management instructions, and sending record.

9. The method for real-time monitoring of improved rubber plantation soil environment based on the Internet of Things according to any one of claims 7 or 8, characterized in that, The warning level is generated in the following way: The warning level determination quantity is obtained by adding the risk trigger quantity, the target level aggravation item, and the management urgency item. The risk trigger quantity is determined based on the logical determination relationship of the fertilizer effect boundary state, the sequence of the three event chains, and the continuous maintenance condition. When the target depth layer includes the root active layer or the risk layer of migration below the root system, the aggravation item takes the higher value. The management urgency item is determined by the boundary movement state in the fertilizer effect boundary state. The warning level determination quantity is used to register the level as attention level, warning level, or disposal level. The management instructions are generated jointly by the risk type, target depth layer, and early warning level.

10. A real-time monitoring system for improved rubber plantation soil environment based on the Internet of Things, characterized in that, include: The benchmark profile management module is used to fix the fertilizer application layer, the active root layer of rubber trees, and the migration risk layer below the root system as fixed monitoring profiles within the area where nano-biochar-based fertilizer is applied; and to construct several benchmark profile data based on the benchmark data of each depth layer of the fixed monitoring profile. The depth improvement status generation module is used to collect real-time soil data at each depth layer, compare it with the benchmark profile data at the same point and depth, and generate the improvement approximation amount of the corresponding depth layer and the corresponding index by comparing the benchmark representative value and real-time reading of the same depth layer and the same index. Based on the improved proximity quantity, directional markers are generated according to the technical meaning of the indicators, and depth improvement status is generated according to the hierarchical attributes. The depth improvement status includes the plot number, profile number, node number, depth number, hierarchical attributes, sampling time, directional markers of each indicator, and improvement status of the corresponding depth layer in each depth layer of the fixed monitoring profile. The improvement status includes improved, weakly improved, and suspected abnormal improvement. The fertilizer effect boundary identification module is used to acquire the state of all depth layers arranged from shallow to deep within the same fixed monitoring profile, identify the improved or weakly improved depth layers as effective action layers, and register abnormal states in the migration risk layer below the root system where water content shifts downwards and is accompanied by increased conductivity or heavy metal ion concentration as downward migration risk. Based on the effective action layers and downward migration risk, the module determines the fertilizer effect boundary position reached by the effective improvement effect of the nano-biochar-based fertilizer, and judges the boundary movement state by the boundary position change in adjacent sampling periods. The fertilizer effect boundary state includes the boundary movement state, root active layer coverage marker, and abnormal downward migration marker. The risk warning and management instruction output module is used to identify the risk type based on the fertilizer effect boundary state and generate a risk warning instruction record at the corresponding time.

Images