Organic fertilizer field soil nutrient real-time monitoring system

Abstract

The application discloses a kind of organic fertilizer soil nutrient real-time monitoring systems, the present application relates to agricultural sensing technology field, including: three-dimensional gradient sensitive field electrode array, neural morphological coding module, dual-channel space-time inversion verification module and closed-loop control module;The electrode array includes the main electrode and auxiliary electrode of the non-uniformly arranged micro-needle cluster, main electrode works in conventional concentration detection mode, auxiliary electrode is configured as high dynamic threshold trigger mode, both cooperates and monitors soil solution ion concentration.The organic fertilizer soil nutrient real-time monitoring system, overcome the sensor response lag problem caused by the dramatic change of nutrient concentration in the initial stage of organic fertilizer mineralization, combined with dual-channel space-time inversion verification and chaos early warning driven hardware dynamic reconstruction technology, eliminate high concentration area nonlinear distortion, realize the accurate capture of real dynamic in the process of dramatic change, improve the reliability of monitoring data in key window period.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of agricultural sensing technology, in particular to a real-time monitoring system for soil nutrients after organic fertilizer is applied. BACKGROUND

[0002] In the prior art, real-time monitoring systems for soil nutrients based on physical or chemical sensing principles have been widely used, which usually deploy in-situ sensors in the field to continuously measure the target nutrient concentration. However, when applied to the scenario of organic fertilizer application, these systems face a prominent technical bottleneck. After organic fertilizer is applied to the soil, it will undergo a significant initial mineralization stage under the action of microorganisms, resulting in a sharp and large change in the concentration of key nutrients in the soil solution within a short period of time. However, the core response mechanism of the mainstream in-situ sensors has inherent limitations in such high-dynamic and rapidly changing environments. The response of the sensitive element to the step change in ion concentration has a non-negligible lag, making it difficult to keep up with the minute-to-hour level concentration surge; at the same time, in the high concentration range, the response characteristics of the sensor often deviate significantly from the ideal linear relationship, showing a tendency of saturation or sensitivity decline. This makes the real-time data obtained by the monitoring system in the most critical initial nutrient release window period after the application of organic fertilizer severely distorted, and unable to accurately reflect the true dynamic trajectory of the nutrient concentration in the soil solution. This distortion directly affects the timeliness and accuracy of the farming decisions based on the monitoring data, and hinders the accurate assessment of the initial mineralization process and potential environmental risks of organic fertilizer. Therefore, the technical problem to be solved at present is: how to overcome the response lag and non-linear distortion of the existing sensors in the condition of rapid and large dynamic change of soil nutrient concentration in the initial stage of organic fertilizer application, and realize high-precision real-time capture of the true nutrient dynamics in this critical window period. SUMMARY

[0003] To achieve the above purpose, the present application is implemented by the following technical scheme: a real-time monitoring system for soil nutrients after organic fertilizer is applied, comprising:

[0004] a three-dimensional gradient-sensitive field electrode array, a neuromorphic encoding module, a dual-channel spatio-temporal inversion verification module, and a closed-loop control module;

[0005] The electrode array includes a main electrode and an auxiliary electrode composed of non-uniformly arranged micro-needle clusters, the main electrode works in a conventional concentration detection mode, and the auxiliary electrode is configured in a high-dynamic threshold trigger mode, and the two work together to monitor the ion concentration in the soil solution;

[0006] The neuromorphic encoding module is connected to the output end of the electrode array, converts the analog signal of the main electrode into a pulse frequency sequence, and converts the analog signal of the auxiliary electrode into a pulse phase shift sequence;

[0007] The double-channel space-time inversion verification module is connected with the neuromorphic encoding module at the input end and connected with the closed-loop control module at the output end.

[0008] The signal output end of the closed-loop control module is connected with the working mode switching interface of the electrode array.

[0009] Preferably, the sensitive surface of the auxiliary electrode is covered with a hydrophobic nano coating, and the activation threshold is set as the critical value of the target nutrient concentration change at the initial stage of organic fertilizer application; and the signal output end of the auxiliary electrode is integrated with a logarithmic compression circuit.

[0010] Preferably, the neuromorphic encoding module comprises a main channel encoding unit and an auxiliary channel encoding unit.

[0011] The main channel encoding unit converts the main electrode signal into a pulse sequence according to the concentration-frequency mapping relationship;

[0012] The auxiliary channel encoding unit generates a phase shift pulse with time delay characteristics according to the auxiliary electrode signal intensity, taking the main channel pulse sequence as the reference clock.

[0013] Preferably, the double-channel space-time inversion verification module comprises:

[0014] a) a forward reconstruction unit for analyzing the pulse frequency sequence and the phase shift pulse to generate a real-time concentration waveform;

[0015] b) a field inversion unit for inversely deducing the concentration field distribution according to the space-time distribution characteristics of the pulse sequence based on a preset soil pore network and organic matter mineralization dynamics coupled model;

[0016] c) a dynamic error analyzer for comparing the deviation amount of the real-time concentration waveform and the concentration field distribution in a moving time window.

[0017] Preferably, the coupled model of the field inversion unit is established by the following steps:

[0018] i) constructing a nutrient diffusion equation containing soil pore structure parameters;

[0019] ii) embedding an organic matter mineralization rate equation as a source term;

[0020] iii) setting the spatial coordinates of the electrode array as the boundary monitoring points;

[0021] iv) dynamically correcting the model diffusion coefficient using the space-time correlation of the pulse sequence.

[0022] Preferably, the closed-loop control module comprises:

[0023] The mode switching controller sends a high-speed sampling instruction to the main electrode and a differential amplification instruction to the auxiliary electrode when the dynamic error amount exceeds the set threshold.

[0024] The chaotic feature extractor calculates the Lyapunov exponent of the pulse interval as a switching trigger signal in real time.

[0025] Preferably, the high-speed sampling instruction shortens the main electrode sampling period; and the differential amplification instruction causes the auxiliary electrode to output a linear amplification signal in the critical concentration interval.

[0026] Preferably, the system further comprises a soil temperature and humidity sensor integrated with the electrode array, and an output end of the soil temperature and humidity sensor is connected to a compensation interface of the neuromorphic encoding module.

[0027] Preferably, the target nutrient includes ammonium nitrogen and available potassium.

[0028] Preferably, the needle tip spacing of the microneedle cluster is 50-200 microns, and the needle length difference is 0.5-2 mm.

[0029] The application provides an organic fertilizer soil nutrient real-time monitoring system.

[0030] The organic fertilizer soil nutrient real-time monitoring system overcomes the sensor response lag problem caused by the dramatic change in nutrient concentration in the initial mineralization stage of organic fertilizer through the critical trigger mechanism of the three-dimensional gradient electrode array and the neuromorphic pulse topology coding; in combination with the double-channel space-time inversion verification and the hardware dynamic reconstruction technology driven by the chaos early warning, the nonlinear distortion in the high concentration area is eliminated, the real dynamic of the nutrient in the dramatic change process is accurately captured, and the reliability of the monitoring data in the key window period is improved.

[0031] The organic fertilizer soil nutrient real-time monitoring system is based on the three-dimensional arrangement structure of the microneedle cluster and the environmental self-adaptive compensation mechanism, and the system still maintains stable operation under complex field conditions, and the autonomous closed-loop control characteristic reduces the need for manual intervention; by deeply integrating the soil physical model and real-time sensing data, not only dynamic basis is provided for precision fertilization, but also reliable technical support is created for evaluating the mineralization process of organic fertilizer and environmental risk early warning, and the efficient use of agricultural resources and the coordinated development of ecological protection are promoted. BRIEF DESCRIPTION OF DRAWINGS

[0032] Figure 1 Fig. 1 is a module interaction schematic diagram of the organic fertilizer soil nutrient real-time monitoring system of the application;

[0033] Figure 2 Fig. 4 is a coupling model establishment process schematic diagram of the field inversion unit of the application. DETAILED DESCRIPTION

[0034] The technical solutions in the embodiments of the present application will be clearly and completely described with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative effort belong to the scope of protection of the present application.

[0035] Referring to Figure 1 and Figure 2 The present application provides a technical solution: an organic fertilizer field soil nutrient real-time monitoring system, comprising:

[0036] a three-dimensional gradient sensitive field electrode array, a neuromorphic encoding module, a double-channel space-time inversion verification module and a closed-loop control module;

[0037] The electrode array comprises a main electrode and an auxiliary electrode formed by non-uniformly arranged micro-needle clusters, the main electrode works in a conventional concentration detection mode, and the auxiliary electrode is configured in a high dynamic threshold trigger mode, and the two electrodes cooperatively monitor the ion concentration of the soil solution;

[0038] The neuromorphic encoding module is connected to the output end of the electrode array, converts the analog signal of the main electrode into a pulse frequency sequence, and converts the analog signal of the auxiliary electrode into a pulse phase shift sequence;

[0039] The double-channel space-time inversion verification module is connected to the neuromorphic encoding module at the input end and connected to the closed-loop control module at the output end;

[0040] The closed-loop control module is connected to the working mode switching interface of the electrode array at the signal output end.

[0041] It should be further explained that in the specific implementation process, after the system is started, the main electrode of the three-dimensional gradient sensitive field electrode array continuously monitors the baseline concentration change of the target nutrient in the soil solution, and the auxiliary electrode is in a standby trigger state; when the organic fertilizer is applied to the soil to trigger the initial mineralization reaction, if the concentration of the target nutrient sharply rises to a preset critical value, the critical value is experimentally calibrated according to the type of organic fertilizer, and the auxiliary electrode is automatically activated and outputs a signal processed by logarithmic compression.

[0042] The neuromorphic encoding module synchronously receives the main and auxiliary electrode signals: the analog voltage signal of the main electrode is converted into a pulse frequency sequence in proportion, wherein the higher the concentration, the faster the frequency; at the same time, the main electrode pulse sequence is used as a reference clock, and a phase shift pulse with a specific time delay is dynamically generated according to the signal strength of the auxiliary electrode, wherein the higher the concentration, the longer the time delay; the two together form a pulse topology network representing the space-time evolution characteristics of the concentration.

[0043] The dual-channel space-time inversion verification module analyzes the pulse topology in real time: the forward reconstruction unit generates the first concentration waveform by decoding the pulse frequency and phase shift; the field inversion unit calls the preset soil pore network-mineralization dynamics coupled model, and the model takes the electrode spatial coordinates as the monitoring points, dynamically corrects the diffusion coefficient combined with the real-time pulse space-time correlation, and reversely deduces the second concentration field distribution. The dynamic error analyzer compares the deviation between the first concentration waveform and the second concentration field distribution within a moving time window of 5 minutes, and sends a correction instruction to the closed-loop control module when the continuous three comparison errors exceed the set threshold.

[0044] The closed-loop control module responds immediately: the Lyapunov exponent of the main electrode pulse interval is calculated by the chaos feature extractor, and if the exponent mutation exceeds the stable domain, a high-speed sampling instruction is sent to the main electrode, and the auxiliary electrode is commanded to switch to the differential amplification mode to enhance the linear response in the critical concentration interval. This process continues until the dynamic error falls within the threshold, and the system returns to the normal monitoring mode. The soil temperature and humidity sensor collects environmental parameters in real time and inputs them into the neuromorphic encoding module for signal compensation, ensuring that the concentration inversion result is not affected by temperature and humidity disturbances.

[0045] The sensitive surface of the auxiliary electrode is covered with a hydrophobic nano coating, and its activation threshold is set as the critical value of the target nutrient concentration change at the initial stage of organic fertilizer application; the auxiliary electrode signal output end integrates a logarithmic compression circuit. It needs to be further explained that in the specific implementation process, the sensitive surface of the auxiliary electrode is treated by plasma-enhanced chemical vapor deposition process before the micro-needle cluster packaging, forming a fluorosilane nano coating, the contact angle of which is greater than 150 degrees, effectively blocking the adsorption of humic acid colloid in organic fertilizer; its activation threshold is set to the lower limit value of the peak interval of the target nutrient at the initial stage of mineralization of a specific type of organic fertilizer by a programmable potentiometer, and when the soil solution ion concentration breaks through this threshold, the auxiliary electrode switches from standby state to active state; the microampere-level current signal output after activation is first converted to a voltage signal by a transimpedance amplifier, and then input into a two-stage signal conditioning circuit: the first stage uses an instrumentation amplifier to suppress common-mode interference, and the second stage connects a logarithmic compression chip to compress the 0.5-5V input dynamic range to a 0-1.5V linear output, ensuring that the high-concentration area signal is not saturated. This processing process reduces the nonlinear distortion of the original signal captured by the auxiliary electrode in the concentration surge stage to less than 1 / 5 of that of traditional electrodes, while the response delay is controlled within 200 milliseconds.

[0046] The neuromorphic encoding module includes a main channel encoding unit and an auxiliary channel encoding unit:

[0047] The main channel encoding unit converts the main electrode signal into a pulse sequence according to the concentration-frequency mapping relationship;

[0048] The auxiliary channel encoding unit generates a phase shift pulse with time delay characteristics based on the main channel pulse sequence as the reference clock according to the auxiliary electrode signal strength.

[0049] It needs to be further explained that, in the specific implementation process, the main channel encoding unit receives the 0-3V analog signal output by the main electrode, and converts it into a 1-100Hz pulse frequency sequence through a voltage-controlled oscillation circuit, wherein the concentration increases by 50mg / L per liter, and the frequency increases by 10Hz; the auxiliary channel encoding unit synchronously monitors the rising edge of the main channel pulse sequence, and inputs it as a reference clock signal into a delay-locked loop circuit. When the auxiliary electrode is activated, the 0-1.5V compressed signal output by it generates a time delay control voltage through a voltage-time converter, which drives a programmable delay line to generate a phase offset pulse relative to the reference clock. The offset pulse and the main channel pulse form a spatiotemporal topological relationship: in the stable concentration stage, the auxiliary pulse maintains a fixed time difference with the main pulse; when the mineralization of organic fertilizer causes a sharp change in concentration, the time delay of the auxiliary pulse increases superlinearly with the concentration, forming the characteristics of spatiotemporal diffusion of pulse clusters. The built-in time reference resynchronization mechanism in the encoding unit automatically switches to the internal crystal oscillator clock when the main pulse is lost for more than 3 cycles to prevent system lockout.

[0050] The double-channel spatiotemporal inversion verification module comprises:

[0051] a) a forward reconstruction unit for analyzing the pulse frequency sequence and the phase offset pulse to generate a real-time concentration waveform;

[0052] b) a field inversion unit for inversely deducing the concentration field distribution based on a preset soil pore network and organic matter mineralization dynamics coupled model according to the spatiotemporal distribution characteristics of the pulse sequence;

[0053] c) a dynamic error analyzer for comparing the deviation between the real-time concentration waveform and the concentration field distribution in a moving time window.

[0054] It needs to be further explained that, in the specific implementation process, the forward reconstruction unit receives the pulse frequency sequence and the phase offset pulse output by the neuromorphic encoding module, first analyzes the main channel pulse density distribution through a pulse counter and a time stamp recorder to generate a basic concentration change curve; at the same time, the auxiliary pulse time delay value is mapped into a high concentration correction coefficient by using a phase offset amount solver, and the two are superimposed to generate a first concentration waveform.

[0055] The field inversion unit is synchronously started, and a three-dimensional concentration field is constructed with the electrode array spatial coordinates as the boundary monitoring points by calling the preset soil physical model and biochemical model; wherein the soil physical model includes the pore water conductivity parameter described by Darcy's law, and the biochemical model includes the organic matter mineralization rate based on the Michaelis-Menten equation; when receiving the real-time pulse sequence, the model automatically extracts the time difference of adjacent electrode pulses to calculate the nutrient diffusion rate, and if the standard deviation of the XY plane pulse spatial distribution in the same time window exceeds the threshold value, the model clay layer resistance parameter is dynamically adjusted.

[0056] The dynamic error analyzer uses a 5-minute moving window to calculate the root mean square deviation between the first concentration waveform and the model-inverted concentration field. When the deviation of three consecutive windows shows a monotonically increasing trend and the final value exceeds the set tolerance, the system is determined to be in a dynamic distortion state and a correction command is output.

[0057] The coupled model of the field inversion unit is established through the following steps:

[0058] i) Construct nutrient diffusion equations that include soil pore structure parameters;

[0059] ii) Embed the organic matter mineralization rate equation as a source term;

[0060] iii) Set the spatial coordinates of the electrode array as the boundary monitoring points;

[0061] iv) Dynamically correct the model diffusion coefficient using the spatiotemporal correlation of pulse sequences.

[0062] It should be further explained that, in the specific implementation process, the construction of the coupling model begins with the digitization of the soil physical structure. Micro-CT scanning is used to obtain the three-dimensional distribution of pores in the target field, and the median of the equivalent pore size and the tortuosity factor are extracted and substituted into the Darcy law framework to establish the initial diffusion equation. The biochemical module is embedded with the Michaelis-Menten kinetic equation, and its maximum reaction rate parameter is preset according to the type of organic fertilizer. The three-dimensional coordinates of the electrode array are used as the boundary condition monitoring node.

[0063] When the system is deployed and receives pulse sequences in real time, the model performs four dynamic correction steps: First, it detects the arrival time difference of the same nutrient pulses from adjacent electrodes. If the time difference exceeds 1.5 times the theoretical propagation time, the drag coefficient of the diffusion equation is increased. Second, it analyzes the uniformity of the pulse distribution in the XY plane. When the standard deviation is greater than the 15% threshold for two consecutive cycles, the clay adsorption compensation sub-model is activated. Then, it dynamically updates the reaction rate parameters of the Michaelis-Menten equation based on the rate of change of the pulse frequency of the main electrode. Finally, it uses the time delay growth rate of the phase shift pulse of the auxiliary electrode to inversely determine the accelerated trend of organic matter mineralization. When the time delay growth rate exceeds 10 μs / min, a transient source term is added to the diffusion equation. The entire correction process iterates every 10 minutes to ensure that the model output concentration field matches the actual dynamic range with an error of less than 3%.

[0064] The closed-loop control module includes:

[0065] The mode switching controller sends a high-speed sampling command to the main electrode and a differential amplification command to the auxiliary electrode when the dynamic error exceeds the set threshold.

[0066] The chaotic feature extractor calculates the Lyapunov exponent of the pulse interval in real time as the switching trigger signal.

[0067] It should be further explained that during the specific implementation process, when the system is running, the chaotic feature extractor continuously monitors the time interval of the main electrode pulse sequence and generates the Lyapunov exponent spectrum through the phase space reconstruction algorithm; when the dominant exponent value continuously crosses the zero line from the negative region within 3 minutes and stabilizes in the 0.3-0.5 range, it is determined that the concentration field has entered a drastic unstable state.

[0068] The mode switching controller synchronously receives the out-of-range signal from the dynamic error analyzer and immediately sends a hardware reconfiguration command to the electrode array: the main electrode switches to sampling mode; the auxiliary electrode simultaneously activates the differential amplifier circuit, subtracting the output signal of the original logarithmic compression channel from the newly added linear amplification channel to generate an amplified signal with flat gain characteristics within the critical concentration range. During this switching process, pulse width modulation technology is used to smoothly transition the signal output, preventing data jumps from exceeding 1 / 10 of the range. When the Lyapunov index falls below 0.1 and remains below it for 10 minutes, the system automatically resumes the normal monitoring mode.

[0069] High-speed sampling commands shorten the sampling period of the main electrode; differential amplification commands enable the auxiliary electrode to output a linearly amplified signal within the critical concentration range. It should be further explained that, in the specific implementation, when the closed-loop control module issues a high-speed sampling command, the main electrode internally switches its integrating capacitor array to compress the signal integration time and simultaneously starts a charge-balanced analog-to-digital converter to acquire transient signals. This process sacrifices some signal-to-noise ratio by shortening the integration period, but reduces the concentration transition capture delay. After receiving the differential amplification command, the auxiliary electrode connects the logarithmic compression channel output to the inverting input of the subtractor, and the newly added broadband linear amplification channel output to the non-inverting input. The difference between the two generates a differential signal. Within the critical concentration range, this differential processing effectively cancels the nonlinear component of the logarithmic channel in the high-concentration region, and the gain fluctuation of the output signal is controlled within 0.5 dB across the entire concentration range.

[0070] A hardware interlock mechanism is used during the switching process: before the sampling rate of the main electrode is increased, the previous integration result is frozen and stored in the buffer register; when the differential channel of the auxiliary electrode is enabled, the old and new channels run in parallel for 0.5 seconds, and the original channel is automatically cut off when the output deviation is less than 1% of the full scale.

[0071] It also includes a soil temperature and humidity sensor integrated with the electrode array, whose output is connected to the compensation interface of the neuromorphic coding module. Further explanation is needed: in the specific implementation process, the chaotic feature extractor continuously captures the interval time of the main electrode pulse sequence. Every 100 accumulated interval values ​​construct a time series vector, which is then mapped to a seven-dimensional embedding space using phase space reconstruction technology. An improved Wolf algorithm is used to calculate the maximum Lyapunov exponent. When this exponent value continuously rises from the negative region and stably exceeds 0.3 within a 180-second monitoring window, it is determined that the concentration field has entered a stage of unstable and drastic change. At this time, a three-level warning signal is sent to the mode switching controller: if the exponent value is in the 0.3-0.4 range, a preparatory command is initiated, preloading high-speed sampling parameters; when it reaches 0.4-0.5, a switching command is sent, triggering electrode mode reconstruction; if it exceeds 0.5, emergency calibration is activated, injecting test pulses to verify system integrity. Temperature drift compensation is introduced in the exponent calculation process, with a correction of 0.02 exponent units for every degree Celsius change, ensuring that fluctuations in the field environment do not affect the reliability of the criterion. When the index value falls below 0.1 and remains below that value for 600 seconds, a recovery command is generated to cancel the high-speed sampling mode.

[0072] The target nutrients include ammonium nitrogen and available potassium. It should be further explained that, during implementation, temperature and humidity sensors integrated into the base of the electrode array collect soil environmental parameters in real time. The temperature monitoring range covers -10°C to 50°C, and the humidity detection range reaches 100% of the saturated moisture content. When the temperature change exceeds 5°C per hour or the humidity fluctuation is greater than 15%, the neuromorphic coding module initiates dynamic compensation: temperature compensation uses a piecewise linear interpolation algorithm, correcting the main electrode gain by 0.05 volts per degree Celsius in the -10 to 10°C range, and by 0.02 volts in the 10 to 50°C range; humidity compensation calls a preset conductivity-moisture content curve based on the soil type, using an exponential model for sandy soil and a logarithmic model for clay, converting humidity changes into equivalent ion concentration shifts.

[0073] The compensation data is input to the dedicated calibration port of the encoding module via a serial interface and processed synchronously with the main and auxiliary electrode signals to ensure that the pulse timing topology is not affected by environmental disturbances. Especially during the peak period of organic fertilizer mineralization, when the temperature and humidity sensor detects a sudden rise in temperature and humidity saturation, the compensation weight is automatically increased to twice that of the normal mode to prevent signal drift caused by drastic changes in microbial activity.

[0074] The microneedle clusters have a tip spacing of 50–200 micrometers and a needle length difference of 0.5–2 millimeters. Further explanation is needed: in practical implementation, the microneedle clusters employ a gradient three-dimensional arrangement design. The tip spacing of the main electrode is set at 80 micrometers, while the tip spacing of the auxiliary electrode is expanded to 180 micrometers, forming a non-uniform sensitive field distribution. The needle length difference is strictly controlled at 1.2 millimeters, with short needles of 5 millimeters used to capture rapid ion exchange in shallow soil solutions, and long needles extending to a depth of 6.2 millimeters to monitor the nutrient diffusion delay effect. When organic fertilizer is applied, the short needle array responds first to the sudden change in nutrient concentration released from surface mineralization, and its dense arrangement increases the signal acquisition density to three times that of conventional electrodes. The long needles, by increasing the needle spacing, avoid clogging by deep soil particles. Simultaneously, the three-dimensional potential gradient generated by the needle length difference allows for the quantification and inversion of the deep diffusion delay time; each 0.1 millimeter needle length difference corresponds to a theoretical diffusion delay of 3 seconds. This structure is particularly crucial in clayey soils: when the short needle detects a sudden change in concentration and the long needle signal lags behind by more than 1.8 times the theoretical delay, the soil compaction compensation algorithm is automatically triggered to correct the distortion of the diffusion model caused by pore closure.

[0075] It should be further explained that, during the specific implementation process, when the system starts up, the three-dimensional gradient sensitive field electrode array penetrates deep into the soil monitoring layer. The main electrode continuously collects the electrochemical signals generated by the target nutrient ions in the soil solution at a constant sampling frequency, while the auxiliary electrode remains in standby mode. Only when organic fertilizer is applied and the mineralization reaction causes the target nutrient concentration to jump to a preset critical value, is the specially processed compressed signal automatically activated to avoid the delay problem caused by the full-process response of conventional sensors.

[0076] The neuromorphic coding module simultaneously receives two signals and converts the analog voltage of the main electrode into a pulse frequency sequence through a dedicated integrated circuit. The pulse density increases accordingly when the concentration increases. At the same time, the module uses the rising edge of the main electrode pulse as the reference clock and dynamically generates a phase-shifted pulse with time-shifting characteristics based on the signal strength of the auxiliary electrode.

[0077] The two pulses together constitute a topological network characterizing the spatiotemporal evolution of concentration, laying the foundation for subsequent processing. A dual-channel spatiotemporal inversion verification module analyzes this network in real time: the forward reconstruction unit generates the first concentration waveform field by decoding the pulse frequency distribution and phase shift; the inversion unit calls a pre-set multiphysics coupling model, which integrates soil pore structure, water conductivity, and organic matter mineralization kinetic equations, using electrode spatial coordinates as boundary constraints. When abnormalities or uneven planar distribution of the pulse sequence are detected during propagation between adjacent electrodes, the model's diffusion resistance parameters and adsorption compensation factors are automatically corrected.

[0078] The dynamic error analyzer continuously compares the deviation between the forward waveform and the inverted concentration field within the moving-time window. When the deviation exceeds the tolerance multiple times and shows an upward trend, the system is determined to have entered a dynamic distortion state. The closed-loop control module responds immediately: first, it analyzes the time series of the main electrode pulse interval through the chaotic feature extractor, reconstructs the phase space, and calculates the Lyapunov exponent; if the exponent remains high, it sends a hardware reconstruction command to the electrode array, the main electrode switches to ultra-high-speed sampling mode, significantly compressing the signal integration time, and the auxiliary electrode simultaneously activates the differential amplifier circuit, subtracting the output of the original nonlinear channel from the newly added linear channel to generate a flat response signal in the high-concentration region; this switching process uses gradual control technology to ensure continuous and stable output.

[0079] Once the index falls back and the error returns to normal, the system automatically switches back to normal mode. Soil temperature and humidity sensors collect environmental parameters in real time, and the neuromorphic coding module performs signal compensation according to preset rules to eliminate the influence of temperature drift. The unique geometric arrangement of the microneedle clusters—densely distributed short needles capture rapid surface changes, while sparsely distributed long needles penetrate deeper to monitor nutrient migration delays—creates a three-dimensional baseline that provides a basis for quantifying diffusion delay.

[0080] When the time delay of the long and short needle signals exceeds the theoretical threshold, the compaction compensation algorithm is automatically triggered to correct the model distortion caused by variations in deep soil structure. The entire system achieves accurate capture of the true trajectory of nutrients under the highly dynamic changes in the early stage of organic fertilizer mineralization through a four-order collaborative mechanism of sensing, encoding, verification, and control.

[0081] Electrode dynamic response mechanism: The activation threshold of the auxiliary electrode is preset with differentiated calibration according to the type of organic fertilizer. For example, a higher threshold is used for poultry and livestock manure, and a lower threshold is used for straw compost, to ensure that the mineralization initiation stage of various organic fertilizers is reliably captured; the sensitive surface is treated with superhydrophobicity to effectively block the adhesion of humus; the signal conditioning adopts a two-stage architecture of logarithmic compression and differential amplification, which not only prevents high concentration saturation, but also improves the linearity in the nonlinear region; the high-speed sampling mode of the main electrode is realized through a switchable integral capacitor array to improve transient response capability.

[0082] Neuromorphic coding principle: Pulse frequency coding strictly follows the monotonic relationship that pulse density increases with concentration; phase offset coding uses delay-locked loop technology to make the auxiliary pulse delay accurately reflect the high concentration change trend; built-in clock redundancy mechanism, when the main pulse is abnormal, seamlessly switches to the backup oscillation source to ensure continuous system operation.

[0083] Multi-model cross-validation logic: Forward reconstruction relies on pulse topology analysis to directly map the concentration change field; inversion is based on the physical mechanism model and dynamically adjusts parameters through pulse spatiotemporal characteristics; if the deviation between the two continues to exceed the limit, it indicates that the dynamic performance of the sensor is insufficient. At this time, the error analyzer outputs a correction command to trigger hardware reconstruction rather than simple software compensation.

[0084] Chaos early warning decision-making process: The Lyapunov exponent calculation adopts an improved algorithm to reduce computational complexity; a sustained high exponent indicates that the system is entering a stage of instability and drastic change; the mode switching controller executes a three-level response accordingly: pre-loading parameters, immediate hardware reconfiguration, or emergency self-check to ensure reliable decision-making.

[0085] Environmental adaptation strategies: Temperature and humidity compensation are matched with specific models according to soil type, using exponential relationships for sandy soil and logarithmic relationships for clay soil; the compensation weight is automatically increased during peak mineralization periods to suppress interference from microbial activity; the compensation signal is injected before encoding to ensure the purity of the pulse topology.

[0086] The gradient arrangement of microneedles enables high-resolution monitoring in shallow layers and anti-blocking capability in deep layers; the three-dimensional baseline constructed by the difference in needle length allows for the quantification and inversion of diffusion delay; when the measured time delay exceeds the standard, the model resistance parameters are automatically enhanced to suppress the inversion distortion caused by compacted soil.

[0087] A method for real-time monitoring of soil nutrients when organic fertilizer is returned to the field includes the following steps:

[0088] Step S1: The main electrode continuously collects basic soil nutrient signals, while the auxiliary electrode is on standby. When the target concentration jumps to a preset critical value due to organic fertilizer mineralization, the auxiliary electrode is activated and outputs an anti-saturation compression signal, while the main electrode signal maintains a linear response.

[0089] Step S2: Convert the analog signal of the main electrode into a pulse frequency sequence. Using the rising edge of the main pulse as the reference clock, generate a phase offset pulse according to the signal strength of the auxiliary electrode to form a spatiotemporal pulse topology network.

[0090] Step S3: The first concentration waveform is reconstructed by analyzing the pulse topology of the forward channel, and the pore-mineralization coupling model is called in the inversion channel. The parameters are dynamically corrected according to the pulse spatiotemporal distribution to generate the second concentration field; the output deviation of the two channels is continuously compared within the moving time window.

[0091] Step S4: Extract the main pulse interval sequence and calculate the Lyapunov exponent. When the exponent remains high and the deviation exceeds the limit, it is determined that a sudden unstable state has been entered. Perform a three-level response based on the exponent value range: pre-load / hardware reconfiguration / emergency self-test.

[0092] Step S5: Send a high-speed sampling command to the main electrode and switch the integrating capacitor array to achieve a fast response; send a differential amplification command to the auxiliary electrode and fuse the logarithmic channel and linear channel outputs to generate a flat signal in the high-concentration region; the switching process uses gradual control to ensure data continuity;

[0093] Step S6: Temperature and humidity sensors collect environmental parameters in real time and match compensation models according to soil type: exponential compensation is used for sandy soil and logarithmic compensation is used for clay soil; the compensation weight is automatically increased during peak mineralization periods.

[0094] Step S7: Use the difference in needle length of microneedle cluster to quantize and invert the deep diffusion delay; when the signal delay of short needle and long needle exceeds the theoretical threshold, trigger the compactness compensation algorithm to correct the model parameters;

[0095] Step S8: Once the chaos index falls back to the stable range and the deviation returns to normal, gradually increase the sampling time of the main electrode and gradually restore the normal monitoring mode.

[0096] By using the critical triggering mechanism of a three-dimensional gradient electrode array and neuromorphic pulse topology coding, the sensor response lag problem caused by drastic changes in nutrient concentration during the early stage of organic fertilizer mineralization is overcome. Combined with dual-channel spatiotemporal inversion verification and hardware dynamic reconstruction technology driven by chaotic early warning, nonlinear distortion in high-concentration areas is eliminated, enabling accurate capture of the true dynamics of nutrients during drastic changes and improving the reliability of monitoring data during critical window periods.

[0097] Based on the three-dimensional arrangement structure of microneedle clusters and the environmental adaptive compensation mechanism, the system maintains stable operation under complex field conditions, and its autonomous closed-loop control characteristics reduce the need for manual intervention. By deeply integrating soil physical models and real-time sensor data, it not only provides dynamic basis for precision fertilization, but also creates reliable technical support for assessing the mineralization process of organic fertilizer and early warning of environmental risks, thus promoting the coordinated development of efficient utilization of agricultural resources and ecological protection.

[0098] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0099] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A real-time monitoring system for soil nutrients when organic fertilizer is returned to the field, characterized in that, include: Three-dimensional gradient sensitive field electrode array, neuromorphic coding module, dual-channel spatiotemporal inversion verification module and closed-loop control module; The electrode array comprises a main electrode and an auxiliary electrode composed of non-uniformly arranged microneedle clusters. The main electrode operates in a conventional concentration detection mode, while the auxiliary electrode is configured in a high dynamic threshold triggering mode. The two work together to monitor the ion concentration in the soil solution. The neuromorphic coding module is connected to the output of the electrode array and converts the analog signal of the main electrode into a pulse frequency sequence and the analog signal of the auxiliary electrode into a pulse phase offset sequence. The input end of the dual-channel spatiotemporal inversion verification module is connected to the neuromorphic coding module, and the output end is connected to the closed-loop control module. The signal output terminal of the closed-loop control module is connected to the working mode switching interface of the electrode array. The sensitive surface of the auxiliary electrode is covered with a hydrophobic nano-coating, and its activation threshold is set as the critical value of the change in the target nutrient concentration at the initial stage of organic fertilizer application. The auxiliary electrode signal output terminal integrates a logarithmic compression circuit. The dual-channel spatiotemporal inversion verification module includes: a) Forward reconstruction unit, analyzes pulse frequency sequence and phase offset pulse to generate real-time concentration waveform; b) Field inversion unit: Based on a pre-set coupled model of soil pore network and organic matter mineralization kinetics, the concentration field distribution is inversely deduced according to the spatiotemporal distribution characteristics of the pulse sequence. c) Dynamic error analyzer, which compares the deviation between the real-time concentration waveform and the concentration field distribution within the moving time window; The coupling model of the field inversion unit is established through the following steps: i) Construct nutrient diffusion equations that include soil pore structure parameters; ii) Embed the organic matter mineralization rate equation as a source term; iii) Set the spatial coordinates of the electrode array as the boundary monitoring points; iv) Dynamically correct the model diffusion coefficient using the spatiotemporal correlation of pulse sequences; The closed-loop control module includes: The mode switching controller sends a high-speed sampling command to the main electrode and a differential amplification command to the auxiliary electrode when the dynamic error exceeds the set threshold. The chaotic feature extractor calculates the Lyapunov exponent of the pulse interval in real time as the switching trigger signal.

2. The real-time monitoring system for soil nutrients when organic fertilizer is returned to the field according to claim 1, characterized in that: The neuromorphic coding module includes a main channel coding unit and an auxiliary channel coding unit: The main channel encoding unit converts the main electrode signal into a pulse sequence according to the concentration-frequency mapping relationship; The auxiliary channel encoding unit uses the main channel pulse sequence as a reference clock and generates a phase offset pulse with time delay characteristics based on the auxiliary electrode signal strength.

3. The real-time monitoring system for soil nutrients when organic fertilizer is returned to the field according to claim 2, characterized in that: The high-speed sampling command shortens the sampling period of the main electrode; the differential amplification command enables the auxiliary electrode to output a linearly amplified signal in the critical concentration range.

4. The real-time monitoring system for soil nutrients when organic fertilizer is returned to the field according to claim 3, characterized in that: It also includes a soil temperature and humidity sensor integrated with the electrode array, whose output is connected to the compensation interface of the neuromorphic coding module.

5. The real-time monitoring system for soil nutrients when organic fertilizer is returned to the field according to claim 4, characterized in that: The target nutrients include ammonium nitrogen and available potassium.

6. The real-time monitoring system for soil nutrients when organic fertilizer is returned to the field according to claim 5, characterized in that: The microneedle cluster has a needle tip spacing of 50–200 micrometers and a needle length difference of 0.5–2 millimeters.

Images