Intelligent monitoring system and method for rice storage environment based on multi-sensor fusion

Abstract

The present application relates to the technical field of warehouse environment monitoring, and particularly relates to a rice warehouse environment intelligent monitoring system and method based on multi-sensor fusion, wherein the monitoring method comprises the following steps: S1, collecting the conventional measurement values of the embedded sensor in the grain pile, and constructing a multi-frequency dielectric response spectrum based on the complex impedance data; S2, converting into a characteristic response time spectrum; S3, extracting characteristic parameters; S4, constructing a contact coupling factor; S5, obtaining the real measurement values after compensation; through the technical chain of excitation collection, decoupling analysis, parameter extraction, factor construction and dynamic correction, the perception of the probe-grain interface contact state and compensation are creatively introduced into the grain condition monitoring field, and the measurement deviation problem caused by the settlement of the grain pile in the background technology is fundamentally solved, and the intelligent and accurate management of the rice warehouse is realized.

Description

Technical Field

[0001] This invention relates to the technical field of warehouse environment monitoring, and in particular to an intelligent monitoring system and method for rice warehouse environment based on multi-sensor fusion. Background Technology

[0002] Current rice storage monitoring systems generally employ embedded probes to collect grain temperature and moisture data. Temperature and moisture sensors deployed at different depths and locations within the grain pile acquire real-time grain condition data, providing a basis for decision-making regarding ventilation, temperature control, and other storage operations. However, as a typical granular porous medium, grain undergoes irreversible viscoelastic creep and gravitational settling during long-term storage due to its own weight and the coupling effect of damp-heat stress. This microstructural evolution leads to stress redistribution within the grain pile, changes in porosity, and consequently alters the physical contact interface between the sensor probe and the surrounding grain grains. This is mainly manifested in the formation of micropores, relaxation of contact stress, and localized delamination.

[0003] In existing technologies, the evolution of the aforementioned contact state has two significant impacts on sensor measurements. Thermally, deteriorating contact leads to a non-linear increase in the contact thermal resistance at the probe-grain interface, obstructing the heat transfer path and causing a severe lag in the sensor's response to changes in ambient temperature—a thermal inertia error. This lag effect is particularly critical in the early stages of grain pile heating. When the grain has begun to slowly heat up, the sensor fails to respond promptly due to thermal inertia, resulting in lower or no temperature readings, thus delaying the optimal time for early mold warning. Electrically, traditional capacitive moisture sensors invert moisture content based on the dielectric constant of the grain pile. However, the dielectric constant is affected by both the moisture content of the grain itself and the porosity of the probe-grain interface. When micropores appear at the interface, the low dielectric constant of air lowers the overall measurement value, leading to a false decrease in moisture readings; conversely, when the interface is excessively compressed, the measurement value may be falsely high. Therefore, interface impedance fluctuations caused by settling are often misinterpreted as actual moisture changes, triggering false drift. Summary of the Invention

[0004] To address the technical problems existing in the background art, this invention proposes an intelligent monitoring system and method for rice storage environment based on multi-sensor fusion, the specific solution of which is as follows: A smart monitoring method for rice storage environment based on multi-sensor fusion includes the following steps: S1. Collect the conventional measurement values ​​of the embedded sensor in the grain pile, apply a multi-frequency AC excitation signal to the embedded sensor, collect the complex impedance data of the embedded sensor in the grain pile, and construct a multi-frequency dielectric response spectrum based on the complex impedance data. S2. Decouple the multi-frequency dielectric response spectrum and convert it into a characteristic response time spectrum that characterizes the relaxation time of different physical processes inside the grain pile. S3. Based on the characteristic response time spectrum, extract the characteristic parameters corresponding to the contact state between the embedded sensor and the grain interface; S4. Construct a contact coupling factor based on the comparison results between the characteristic parameters and the preset benchmark value; S5. Based on the contact coupling factor, the conventional measurement values ​​collected in S1 are dynamically corrected to obtain the compensated true measurement values.

[0005] Furthermore, in S1, the embedded sensor includes a temperature sensor and a moisture sensor, the frequency range of the multi-frequency AC excitation signal is 100Hz to 1MHz, the complex impedance data includes the real and imaginary parts of the impedance at different frequencies, and the multi-frequency dielectric response spectrum is composed of the real and imaginary parts of the impedance.

[0006] Furthermore, in S2, the multi-frequency dielectric response spectrum is decoupled and analyzed to transform it into a characteristic response time spectrum characterizing the relaxation time of different physical processes inside the grain pile, as follows: The multi-frequency dielectric response spectrum is numerically deconvolved to transform it into a characteristic response time spectrum with the relaxation time constant as the abscissa and the polarization intensity as the ordinate. The characteristic response time spectrum contains multiple characteristic peaks, which correspond to the relaxation responses of different physical processes inside the grain pile.

[0007] Furthermore, in S2, the multi-frequency dielectric response spectrum is subjected to numerical deconvolution processing, as follows: The Tikhonov regularization algorithm is used to perform numerical deconvolution processing on the multi-frequency dielectric response spectrum. The mathematical model is as follows: ; By minimizing the objective function: ; Solving for the results ,in This is the regularization parameter.

[0008] Furthermore, the characteristic response time spectrum includes: The first characteristic peak located in the high-frequency band has a relaxation time constant. Less than The second corresponds to the dipole orientation polarization of the grain body and the equivalent resistance of the grain pile body; The second characteristic peak located in the low-frequency band has a relaxation time constant. Greater than The seconds correspond to the polarization capacitance and contact gap resistance at the interface between the embedded sensor and the grain.

[0009] Furthermore, in S3, based on the aforementioned characteristic response time spectrum, characteristic parameters corresponding to the contact state between the embedded sensor and the grain interface are extracted, as follows: In the characteristic response time spectrum, a first characteristic peak corresponding to the polarization process of the grain body and a second characteristic peak corresponding to the contact polarization process of the embedded sensor and the grain interface are identified, and the current contact gap resistance value is extracted from the second characteristic peak as the characteristic parameter.

[0010] Furthermore, in S4, the formula for calculating the contact coupling factor is: ; in, This is the contact gap resistance value extracted from the characteristic response time spectrum at the current moment. This is the initial contact gap resistance value preset when the grain pile has not settled in the early stage of storage.

[0011] Furthermore, in S5, based on the contact coupling factor, the conventional measurement values ​​collected in S1 are dynamically corrected to obtain the compensated true measurement values, as follows: When the embedded sensor is a moisture sensor, the compensation amount is calculated based on the contact coupling factor using a pre-built compensation model. The compensation amount is superimposed on the currently collected measurement value to obtain the corrected true measurement value.

[0012] Furthermore, in S5, based on the contact coupling factor, the conventional measurement values ​​collected in S1 are dynamically corrected to obtain the compensated true measurement values, as follows: When the embedded sensor includes a temperature sensor, the thermal time constant of the current probe is estimated based on the contact coupling factor; an inverse thermal conduction filtering algorithm is used to dynamically compensate the temperature measurement value collected in S1 based on the thermal time constant to eliminate thermal hysteresis error.

[0013] A smart monitoring system for rice storage environment based on multi-sensor fusion includes: The acquisition spectrum module is used to acquire conventional measurement values ​​of the embedded sensor in the grain pile, apply multi-frequency AC excitation signals to the embedded sensor, acquire the complex impedance data of the embedded sensor in the grain pile, and construct a multi-frequency dielectric response spectrum based on the complex impedance data. The decoupling transformation module is used to perform decoupling analysis on the multi-frequency dielectric response spectrum and transform it into a characteristic response time spectrum that characterizes the relaxation time of different physical processes inside the grain pile. The parameter extraction module is used to extract feature parameters corresponding to the contact state between the embedded sensor and the grain interface based on the feature response time spectrum. A factor construction module is used to construct a contact coupling factor based on the comparison results between the feature parameters and preset benchmark values; The numerical correction module is used to dynamically correct the conventional measurement values ​​acquired in the acquisition spectrum module based on the contact coupling factor, so as to obtain the compensated true measurement values.

[0014] Compared with the prior art, the present invention can achieve at least the following beneficial effects: 1. This invention applies multi-frequency AC excitation signals to an embedded sensor and constructs a multi-frequency dielectric response spectrum, which is then decoupled and analyzed to transform into a characteristic response time spectrum, thereby extracting characteristic parameters that can quantitatively describe the interface contact state. Based on the comparison between the extracted interface contact characteristic parameters and a preset benchmark value, a contact coupling factor is constructed, reflecting the degree of deviation of the current contact state from the initial storage stage. This solves the technical problem of indistinguishable interface impedance and bulk moisture content, laying a physical foundation for eliminating false moisture drift. The constructed contact coupling factor not only reflects the electrical contact state, but its physical essence is consistent with the thermal contact state. By introducing the contact coupling factor, real-time input parameters are provided for subsequent dynamic adjustment of the thermal time constant and compensation for thermal hysteresis errors. This allows temperature compensation to adaptively adjust with changes in the contact state, ensuring that even if the sensor's response is slow due to thermal inertia in the early stages of grain pile heating, the compensated true temperature value can still reflect the heating trend in a timely manner, thus restoring early warning capabilities. This enables the embedded sensor to output data consistent with the actual grain condition, avoiding misjudgments and misoperations caused by sensor reading distortion, significantly improving the long-term stability and data reliability of the grain condition monitoring system.

[0015] 2. This invention applies multi-frequency excitation to an embedded sensor and constructs a dielectric response spectrum. Through decoupling analysis, characteristic parameters characterizing the contact state of the probe-grain interface are extracted, and a contact coupling factor is constructed. This enables online sensing of interface contact changes caused by grain pile settling. Based on this factor, conventional temperature and humidity measurements are dynamically corrected. On the one hand, this eliminates the interference of interface impedance fluctuations on moisture measurement and solves the problem of false drift caused by the inability of traditional methods to distinguish between grain moisture and interface voids. On the other hand, it provides an adaptive adjustment basis for thermal hysteresis compensation, restoring the early heating warning capability lost due to increased contact thermal resistance. Thus, the accuracy and reliability of monitoring data are significantly improved over a storage period of several years. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a flowchart of the method of the present invention.

[0017] Figure 2 This is a system principle block diagram of the present invention. Detailed Implementation

[0018] Embodiments of the present invention are described in detail below. Examples of these embodiments are illustrated in the accompanying drawings, wherein the same or similar symbols 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.

[0019] Please refer to Figure 1 This invention provides an intelligent monitoring method for rice storage environment based on multi-sensor fusion, comprising the following steps: S1. Collect the conventional measurement values ​​of the embedded sensor in the grain pile, apply a multi-frequency AC excitation signal to the embedded sensor, collect the complex impedance data of the embedded sensor in the grain pile, and construct a multi-frequency dielectric response spectrum based on the complex impedance data.

[0020] It should be noted that the conventional measurement values ​​refer to the basic physical quantities of the grain pile collected by the embedded sensors using traditional monitoring methods. These mainly include grain temperature data measured by temperature sensors and moisture data measured by moisture sensors. These data are the basic basis for the daily management of grain depots. However, during long-term storage, the contact state between the sensors and grain grains may change due to grain pile settling, which may cause deviations in these conventional measurement values. Therefore, subsequent steps are required for correction.

[0021] In an optional embodiment, in S1, the embedded sensor includes a temperature sensor and a moisture sensor, the frequency range of the multi-frequency AC excitation signal is 100Hz to 1MHz, the complex impedance data includes the real and imaginary parts of the impedance at different frequencies, and the multi-frequency dielectric response spectrum is composed of the real and imaginary parts of the impedance.

[0022] It should be noted that the multi-frequency AC excitation signal is a detection signal specifically applied in this invention to obtain the dielectric properties of the grain pile. By applying a series of weak AC voltages covering a wide frequency range to the sensor probe, polarization responses at different scales within the grain pile can be excited. The preferred frequency range is 100Hz to 1MHz. This range is selected based on the following physical mechanism: low-frequency signals (<100Hz) are sensitive to space charge polarization and contact impedance changes at the probe-grain interface, reflecting the contact state between the sensor and the grain; high-frequency signals (>1kHz) mainly reflect the dipole reorientation polarization within the grain itself, which is closely related to the grain moisture content. Covering such a wide frequency band allows for subsequent spectral decoupling to separate the influence of different physical processes.

[0023] It should be noted that complex impedance data refers to the complex expression of impedance measured at different frequencies, including the real part (resistive component) and the imaginary part (reactant component). The real part reflects the current loss effect of the grain pile, while the imaginary part reflects the energy storage characteristics of the grain pile. By sweeping frequency measurements, a set of complex impedance data varying with frequency can be obtained, which contains superposition information of multiple polarization mechanisms within the grain pile.

[0024] It should be noted that the multi-frequency dielectric response spectrum is a spectral curve composed of complex impedance data at different frequencies, typically represented by frequency on the x-axis and impedance magnitude and phase angle (or real and imaginary parts) on the y-axis. This spectrum serves as the fundamental input for decoupling analysis in subsequent steps, implicitly reflecting the time-scale differences between the bulk polarization process of the grain (fast response) and the interfacial polarization process (slow response). In practice, the multi-frequency dielectric response spectrum can be collected daily or weekly to track long-term changes in the grain pile condition.

[0025] It should be noted that the embedded sensor typically employs a composite probe structure, integrating a temperature sensor, a moisture sensor, and an excitation electrode into one unit. This allows for real-time acquisition of conventional grain temperature and moisture content, as well as the application of multi-frequency excitation and measurement of complex impedance response. This integrated design ensures the consistency of conventional measurement values ​​with complex impedance data in both time and space, providing a foundation for subsequent corrections based on the same probe.

[0026] S2. Decouple the multi-frequency dielectric response spectrum and convert it into a characteristic response time spectrum that characterizes the relaxation time of different physical processes inside the grain pile.

[0027] In an optional embodiment, in S2, the multi-frequency dielectric response spectrum is decoupled and analyzed to transform it into a characteristic response time spectrum characterizing the relaxation time of different physical processes inside the grain pile, as follows: The multi-frequency dielectric response spectrum is numerically deconvolved to transform it into a spectrum with a relaxation time constant. The x-axis represents the polarization intensity. The characteristic response time spectrum is represented by the ordinate. The characteristic response time spectrum contains multiple characteristic peaks, which correspond to the relaxation responses of different physical processes inside the grain pile.

[0028] It should be noted that decoupling analysis refers to a mathematical processing method that separates the responses of multiple superimposed physical processes. The multi-frequency dielectric response spectrum obtained from S1 is the result of the combined effects of multiple polarization mechanisms within the grain pile—processes such as grain bulk dipole orientation, space charge accumulation at interfaces, and ion migration are superimposed in the frequency domain and are difficult to distinguish directly. Through decoupling analysis, the aliased signals of these physical processes at different time scales in the frequency domain can be separated into independent response components in the time domain, thereby enabling independent observation and analysis of different physical processes within the grain pile.

[0029] It should be noted that relaxation time is a physical quantity describing the time required for a physical system to recover from a non-equilibrium state to an equilibrium state. In dielectric response theory, different polarization mechanisms have different relaxation time constants. Dipole reorientation polarization within the grain is a rapid process with an extremely short relaxation time (typically on the order of microseconds to milliseconds); while space charge polarization at the probe-grain interface is a slow process with a longer relaxation time (typically on the order of milliseconds to seconds). This difference in the order of magnitude of relaxation time is the physical basis for the present invention's ability to separate different physical processes in the frequency domain—high-frequency excitation mainly induces rapid relaxation processes, while low-frequency excitation mainly induces slow relaxation processes.

[0030] It should be noted that the characteristic response time spectrum is a spectrum obtained by transforming the impedance data in the frequency domain to the time domain. Its horizontal axis represents the relaxation time constant τ, and the vertical axis represents the polarization intensity at the corresponding relaxation time. This spectrum visually illustrates the relaxation processes present within the grain pile and the relative strength of each process. Mathematically, the characteristic response time spectrum is a representation of the relaxation time distribution of the dielectric response function. It transforms frequency domain information into time domain information, separating multiple physical processes that originally overlapped in the frequency domain into independent characteristic peaks in the time domain, greatly improving the ability to identify different physical mechanisms.

[0031] In an optional embodiment, in S2, the multi-frequency dielectric response spectrum is subjected to numerical deconvolution processing as follows: The Tikhonov regularization algorithm is used to perform numerical deconvolution processing on the multi-frequency dielectric response spectrum. The mathematical model is as follows: ; By minimizing the objective function: ; Solving for the results ,in This is the regularization parameter.

[0032] It should be noted that numerical deconvolution is a key mathematical tool for converting from the frequency domain to the time domain. Since the frequency domain response Z(ω) and the time domain response G(τ) have a mathematical relationship described by the aforementioned integral equation, directly solving G(τ) is mathematically an ill-posed inverse problem, meaning that even small fluctuations in the measured data can lead to violent oscillations in the solution. Therefore, regularization methods are needed to obtain stable and physically meaningful solutions.

[0033] In an optional embodiment, the characteristic response time spectrum includes: The first characteristic peak located in the high-frequency band has a relaxation time constant. Less than The second corresponds to the dipole orientation polarization of the grain body and the equivalent resistance of the grain pile body; The second characteristic peak located in the low-frequency band has a relaxation time constant. Greater than The seconds correspond to the polarization capacitance and contact gap resistance at the interface between the embedded sensor and the grain.

[0034] It should be noted that the first characteristic peak in the high-frequency band ( Less than The characteristic peak (seconds) corresponds to a rapid polarization process, mainly including the reorientation polarization of water molecule dipoles within the grain and the contribution of bulk ionic conductivity. The intensity of this characteristic peak is primarily affected by the grain's moisture content—higher moisture content results in more polarizable dipoles and a stronger peak. Simultaneously, the position of this characteristic peak (relaxation time) is temperature-dependent; as temperature increases, the relaxation time decreases slightly (the peak shifts to the left). This characteristic peak forms the physical basis for traditional moisture measurement.

[0035] It should be noted that the second characteristic peak in the low-frequency band ( Greater than The second characteristic peak corresponds to the slow polarization process, and its physical origin is the space charge polarization at the interface between the probe's metal surface and the grain. When the probe is in close contact with the grain, the charge accumulated at the interface can relax rapidly, and the peak position is shifted to the left (τ is small). When an air gap appears around the probe, charge relaxation is hindered, the relaxation time is prolonged, the peak position shifts to the right, and the peak area (i.e., the contact gap resistance value) increases. Therefore, the position and intensity of the second characteristic peak directly reflect the contact state between the probe and the grain—the rightward shift and enhancement of the peak are typical signs of contact relaxation and the appearance of micro-gaps. This physical relationship is the core basis for extracting contact state characteristic parameters in this invention.

[0036] Through the decoupling analysis described above, the moisture response and contact response, which were previously confused in conventional monitoring, were separated, laying the physical foundation for independently correcting moisture and temperature measurements in subsequent steps. This technical approach breaks through the limitation of traditional grain condition monitoring that treats sensor readings as absolutely true, and for the first time achieves self-sensing and self-compensation of the sensor's own coupling state.

[0037] S3. Based on the characteristic response time spectrum, extract the characteristic parameters corresponding to the contact state between the embedded sensor and the grain interface.

[0038] In an optional embodiment, in S3, based on the characteristic response time spectrum, characteristic parameters corresponding to the contact state between the embedded sensor and the grain interface are extracted as follows: In the characteristic response time spectrum, a first characteristic peak corresponding to the polarization process of the grain body and a second characteristic peak corresponding to the contact polarization process of the embedded sensor and the grain interface are identified, and the current contact gap resistance value is extracted from the second characteristic peak as the characteristic parameter.

[0039] It should be noted that the characteristic parameter refers to the physical quantity extracted from the characteristic response time spectrum that can quantitatively describe the contact state of the probe-grain interface. According to the characteristic response time spectrum obtained from the aforementioned decoupling analysis, the grain bulk polarization process and the interface contact polarization process exhibit two independent characteristic peaks in the time domain. These two peaks have different relaxation time constants and different physical origins. Among them, the second characteristic peak directly originates from the charge accumulation and relaxation behavior at the probe-grain interface. Therefore, the characteristic quantities of this peak (such as peak position, peak area, and peak height) are closely related to the interface contact state. This invention selects the contact gap resistance value extracted from the second characteristic peak as the characteristic parameter because this physical quantity has a clear electrical meaning. It corresponds to the real part of the impedance encountered when current passes through the tiny gap between the probe and the grain, and can most directly reflect the existence and size of the gap.

[0040] S4. Construct a contact coupling factor based on the comparison results between the characteristic parameters and the preset benchmark value.

[0041] In an optional embodiment, in S4, the formula for calculating the contact coupling factor is: ; in, This is the contact gap resistance value extracted from the characteristic response time spectrum at the current moment. This is the initial contact gap resistance value preset when the grain pile has not settled in the early stage of storage.

[0042] It should be noted that the contact gap resistance value The extraction method is as follows: After obtaining the feature response time spectrum... Then, first determine the relaxation time interval where the second characteristic peak is located (usually 1000-10000 pm). Instant (seconds). Within this interval Integrating the integral yields the total intensity of the interfacial polarization process, which is inversely proportional to the equivalent resistance at the interface. By appropriately calibrating the integral value (e.g., establishing a mapping relationship between the integral value and the actual gap resistance under laboratory conditions), the contact gap resistance value at the current moment can be obtained. When the probe is in close contact with the grain, charge relaxation at the interface is unimpeded. The current is relatively small; when an air gap exists around the probe, the current needs to pass through the low-conductivity air layer. Significantly increased; when the probe is excessively compacted or the grains squeeze together, leading to an increased contact area. It may decrease slightly. Therefore, The numerical change directly reflects the evolution of the contact state.

[0043] It should be noted that when identifying two characteristic peaks in the characteristic response time spectrum, a peak detection algorithm combined with prior knowledge of relaxation time can be used. This is because the relaxation time of the grain bulk polarization process (the first characteristic peak) is typically within... Instant The time is on the order of seconds, while the relaxation time of the interface polarization process (second characteristic peak) is usually on the order of seconds. Instant The two peaks are on the order of seconds, differing by at least an order of magnitude, and are usually clearly separated in actual spectra. However, in certain special cases (such as extremely high or low moisture content in grain piles), the boundaries of the two peaks may overlap. In this case, Gaussian peak fitting or partial least squares methods can be used to mathematically decouple the overlapping peaks and extract the contribution of each peak separately.

[0044] It should be noted that the preset baseline value This refers to the initial contact gap resistance value measured at the beginning of the grain pile's entry into the warehouse, before significant settlement has occurred. At this time, the contact state between the probe and the grain is considered ideal—the grain has just finished loading, is in close contact with the probe, and has not yet developed micro-gaps due to gravity creep. The timing for obtaining the data is typically within the stabilization period after the grain has been stored (e.g., within 1-2 weeks after storage). Multiple measurements are taken at different time points, and the average value is calculated to eliminate the impact of short-term fluctuations. This baseline value serves as a reference point for subsequent comparisons, and its physical meaning is the interfacial resistance under optimal contact conditions between the probe and the grain.

[0045] By constructing this dimensionless factor, this scheme quantifies the complex interface contact state into an index that is easy to apply in engineering, providing a unified input parameter for dynamically correcting measurement values ​​in subsequent steps. The advantage of this factor lies in its normalization properties, which eliminates the absolute numerical differences under different probes, different grain piles, and different environmental conditions, making it applicable to various deployment scenarios based on the same correction model.

[0046] It should be noted that traditional monitoring systems treat sensors as constant measuring tools, ignoring the fact that the interface between the sensor and the measured medium evolves over time. Self-sensing capabilities provide the prerequisite for subsequent intelligent correction and also provide grain depot managers with a quantitative basis for assessing the reliability of sensor data. For example, when the CCF value of a sensor deviates significantly from 1 for a long period, managers can determine that the data at that location may be distorted, requiring priority for on-site verification or sensor maintenance.

[0047] S5. Based on the contact coupling factor, the conventional measurement values ​​collected in S1 are dynamically corrected to obtain the compensated true measurement values.

[0048] It should be noted that dynamic correction refers to the process of adjusting conventional measurement values ​​online based on the contact coupling factor (CCF) monitored in real time. Traditional grain condition monitoring systems treat sensor readings as true values, ignoring the impact of changes in the contact state between the sensor and the measured medium on the measurement. This invention achieves dynamic correction of measurement values ​​by introducing CCF—when the CCF deviates from 1, the system automatically determines that the current measurement value has a deviation and initiates the corresponding correction algorithm based on the magnitude and direction of the CCF (greater than 1 indicates contact relaxation, less than 1 indicates contact compaction). This dynamic correction mechanism enables the monitoring system to adaptively compensate for measurement errors caused by grain pile settlement, thereby maintaining the accuracy of monitoring data over a storage period of several years.

[0049] In an optional embodiment, in S5, based on the contact coupling factor, the conventional measurement values ​​collected in S1 are dynamically corrected to obtain the compensated true measurement values, as follows: When the embedded sensor is a moisture sensor, the compensation amount is calculated based on the contact coupling factor using a pre-built compensation model. The compensation amount is superimposed on the currently collected measurement value to obtain the corrected true measurement value.

[0050] It's important to note that the physical basis of moisture correction lies in the fact that the measurement principle of capacitive moisture sensors is based on the dielectric constant of the grain pile. The dielectric constant of water (approximately 80) is much higher than that of air (approximately 1) and the dry matter content of grain (approximately 3-5). When the probe is in close contact with the grain, the sensor's sensitive area is primarily filled with grain and water, and the measured value reflects the true moisture content. When air gaps appear around the probe, the low dielectric constant of air lowers the overall dielectric constant, leading to an underestimation of the moisture content. This deviation is directly related to the size of the air gap, which in turn is reflected in the contact coupling factor (CCF) (the larger the CCF, the more severe the air gap). Therefore, a quantitative relationship between CCF and moisture measurement deviation can be established to correct the moisture reading.

[0051] It should be noted that the pre-built compensation model establishes a mapping relationship between CCF and moisture compensation ΔM through laboratory calibration experiments. The calibration method is as follows: Under laboratory conditions, moisture sensors are buried in grain piles with different moisture contents (e.g., 12%, 14%, 16%). A mechanical device controls the gap between the sensor and the grain grains (simulating different CCF values), and the moisture reading deviation under different gaps is measured. A large amount of calibration data is then fitted into a functional relationship. Alternatively, a lookup table can be created to obtain the compensation model. In practical applications, the system uses the current CCF and the original moisture measurement value... The compensation amount can be obtained by looking up a table or interpolating. Then follow the formula Calculate the corrected true moisture value. When CCF=1 (good contact), ΔM≈0, the correction is minimal; when CCF>1 (loose contact), ΔM is positive, compensating for the lower reading caused by air gaps; when CCF<1 (too tight contact), ΔM may be negative, compensating for the higher reading caused by over-compaction.

[0052] In an optional embodiment, in S5, based on the contact coupling factor, the conventional measurement values ​​collected in S1 are dynamically corrected to obtain the compensated true measurement values, as follows: When the embedded sensor includes a temperature sensor, the thermal time constant of the current probe is estimated based on the contact coupling factor; an inverse thermal conduction filtering algorithm is used to dynamically compensate the temperature measurement value collected in S1 based on the thermal time constant to eliminate thermal hysteresis error.

[0053] It's important to note that the physical basis of temperature correction lies in heat conduction: temperature sensors are typically encapsulated within a metal probe. Heat must travel from the grain through the probe-grain interface and then through the probe material to reach the sensing element. When the probe is in close contact with the grain, the interfacial thermal resistance is low, and the sensor responds quickly. However, when an air gap appears around the probe, the low thermal conductivity of air (approximately 0.026 W / m·K, far lower than the grain's approximately 0.1 W / m·K) significantly increases the interfacial thermal resistance, slowing heat transfer. Consequently, the temperature change measured by the sensor lags behind the actual grain temperature change. This thermal hysteresis effect causes the sensor reading to be lower in the initial stage of grain pile heating and higher in the initial stage of cooling, thus masking early temperature anomalies.

[0054] In an optional embodiment, the thermal time constant is estimated using the following formula: ; Wherein, (CCF) is the contact coupling factor, and (k) is the pre-calibrated thermoelectric conversion coefficient, which is obtained by laboratory calibration under different moisture and temperature conditions; The compensation formula for the inverse heat conduction filtering algorithm is: ; in, This is the currently collected temperature measurement value. The temperature change rate after smoothing pretreatment. This is the compensated actual temperature value.

[0055] It should be noted that the thermal time constant It is a characteristic parameter describing the response speed of a temperature sensor. Its physical meaning is the time required for the sensor to reach 63.2% response to a step temperature change. The larger the value, the slower the sensor response and the more severe the thermal hysteresis. This invention experimentally reveals an approximately linear relationship between the thermal time constant and the contact coupling factor (CCF). Where k is the thermoelectric conversion coefficient. The physical interpretation of this relationship is: CCF reflects the tightness of the interfacial contact; the worse the interfacial contact (the larger the CCF), the greater the interfacial thermal resistance and the larger the thermal time constant. The thermoelectric conversion coefficient k is obtained through laboratory calibration—under different moisture and temperature conditions, the actual thermal response time corresponding to different CCFs is measured, and the k value is obtained by fitting the data. Since the thermal conductivity of grains varies with moisture and temperature, the k value may not be constant. Therefore, a two-dimensional lookup table of k with respect to moisture and temperature is usually established, and a suitable k value is dynamically selected based on the current moisture and temperature measurements in practical applications.

[0056] It should be noted that a typical application scenario for temperature correction is early warning of mold growth. In the early stages of mold growth, the temperature inside the grain pile may slowly rise at a rate of 0.1℃ / h. With good probe contact, the sensor can detect this minute change; however, when an air gap appears in the probe, the thermal hysteresis effect may cause the temperature rise rate measured by the sensor to decrease to 0.02℃ / h, far below the alarm threshold. Using the temperature correction algorithm proposed in this invention, even if the sensor only measures an apparent temperature rise rate of 0.02℃ / h, the system can still detect it based on the current CCF (e.g., CCF=1.5). By magnifying the data, the true temperature rise rate of 0.1℃ / h can be deduced, thus issuing an early warning several hours or even days in advance. This ability to see beyond the surface is a significant advancement of this invention compared to traditional monitoring methods.

[0057] It should be noted that moisture correction and temperature correction can be implemented independently or in combination. For example, when grain pile settling causes gaps in the probe, moisture correction can prevent false readings of moisture loss due to air gaps, avoiding incorrect ventilation and drying by operators. Simultaneously, temperature correction ensures timely capture of the true temperature rise in the early stages of heating. Together, these two mechanisms enable the sensor system to accurately monitor chemical changes (moisture and temperature) even when physical changes (settling) occur in the grain pile, achieving a more fundamental understanding of the grain's condition.

[0058] Please refer to Figure 2 This invention provides an intelligent monitoring system for rice storage environment based on multi-sensor fusion, comprising: The acquisition spectrum module is used to acquire conventional measurement values ​​of the embedded sensor in the grain pile, apply multi-frequency AC excitation signals to the embedded sensor, acquire the complex impedance data of the embedded sensor in the grain pile, and construct a multi-frequency dielectric response spectrum based on the complex impedance data. The decoupling transformation module is used to perform decoupling analysis on the multi-frequency dielectric response spectrum and transform it into a characteristic response time spectrum that characterizes the relaxation time of different physical processes inside the grain pile. The parameter extraction module is used to extract feature parameters corresponding to the contact state between the embedded sensor and the grain interface based on the feature response time spectrum. A factor construction module is used to construct a contact coupling factor based on the comparison results between the feature parameters and preset benchmark values; The numerical correction module is used to dynamically correct the conventional measurement values ​​acquired in the acquisition spectrum module based on the contact coupling factor, so as to obtain the compensated true measurement values.

[0059] In summary, this patent application creatively introduces the perception and compensation of the contact state of the grain interface of the probe into the field of grain condition monitoring through a technical chain of excitation acquisition, decoupling analysis, parameter extraction, factor construction, and dynamic correction. It fundamentally solves the measurement deviation problem caused by grain pile settlement in the background technology, and provides core technical support for the intelligent and precise management of rice storage.

[0060] Furthermore, any content not described in detail in this specification is existing technology known to those skilled in the art.

[0061] In the embodiments provided by this invention, it should be understood that the disclosed system or method can be implemented in other ways. For example, the embodiments of the invention described above are merely illustrative; for instance, the division of modules is only a logical functional division, and there may be other division methods in actual implementation.

[0062] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0063] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated module can be implemented in hardware or in the form of hardware plus software functional modules.

[0064] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the basic characteristics of the present invention.

[0065] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for intelligent monitoring of rice storage environment based on multi-sensor fusion, characterized in that, Includes the following steps: S1. Collect the conventional measurement values ​​of the embedded sensor in the grain pile, apply a multi-frequency AC excitation signal to the embedded sensor, collect the complex impedance data of the embedded sensor in the grain pile, and construct a multi-frequency dielectric response spectrum based on the complex impedance data. S2. Decouple the multi-frequency dielectric response spectrum and convert it into a characteristic response time spectrum that characterizes the relaxation time of different physical processes inside the grain pile. S3. Based on the characteristic response time spectrum, extract the characteristic parameters corresponding to the contact state between the embedded sensor and the grain interface; S4. Construct a contact coupling factor based on the comparison results between the characteristic parameters and the preset benchmark value; S5. Based on the contact coupling factor, the conventional measurement values ​​collected in S1 are dynamically corrected to obtain the compensated true measurement values.

2. The intelligent monitoring method for rice storage environment based on multi-sensor fusion as described in claim 1, characterized in that: In S1, the embedded sensor includes a temperature sensor and a moisture sensor, the frequency range of the multi-frequency AC excitation signal is 100Hz to 1MHz, the complex impedance data includes the real and imaginary parts of the impedance at different frequencies, and the multi-frequency dielectric response spectrum is composed of the real and imaginary parts of the impedance.

3. The intelligent monitoring method for rice storage environment based on multi-sensor fusion as described in claim 1, characterized in that: In S2, the multi-frequency dielectric response spectrum is decoupled and analyzed to transform it into a characteristic response time spectrum characterizing the relaxation time of different physical processes inside the grain pile, as follows: The multi-frequency dielectric response spectrum is numerically deconvolved to transform it into a spectrum with a relaxation time constant. The x-axis represents the polarization intensity. The characteristic response time spectrum is represented by the ordinate. The characteristic response time spectrum contains multiple characteristic peaks, which correspond to the relaxation responses of different physical processes inside the grain pile.

4. The intelligent monitoring method for rice storage environment based on multi-sensor fusion as described in claim 3, characterized in that: In S2, the multi-frequency dielectric response spectrum is subjected to numerical deconvolution processing as follows: The Tikhonov regularization algorithm is used to perform numerical deconvolution processing on the multi-frequency dielectric response spectrum. The mathematical model is as follows: ； By minimizing the objective function: ； Solving for the results ,in This is the regularization parameter.

5. The intelligent monitoring method for rice storage environment based on multi-sensor fusion as described in claim 4, characterized in that: The characteristic response time spectrum includes: The first characteristic peak located in the high-frequency band has a relaxation time constant. Less than The second corresponds to the dipole orientation polarization of the grain body and the equivalent resistance of the grain pile body; The second characteristic peak located in the low-frequency band has a relaxation time constant. Greater than The seconds correspond to the polarization capacitance and contact gap resistance at the interface between the embedded sensor and the grain.

6. The intelligent monitoring method for rice storage environment based on multi-sensor fusion as described in claim 1, characterized in that: In S3, based on the characteristic response time spectrum, characteristic parameters corresponding to the contact state between the embedded sensor and the grain interface are extracted as follows: In the characteristic response time spectrum, a first characteristic peak corresponding to the polarization process of the grain body and a second characteristic peak corresponding to the contact polarization process of the embedded sensor and the grain interface are identified, and the current contact gap resistance value is extracted from the second characteristic peak as the characteristic parameter.

7. The intelligent monitoring method for rice storage environment based on multi-sensor fusion as described in claim 1, characterized in that: In S4, the formula for calculating the contact coupling factor is: ； in, This is the contact gap resistance value extracted from the characteristic response time spectrum at the current moment. This is the initial contact gap resistance value preset when the grain pile has not settled in the early stage of storage.

8. The intelligent monitoring method for rice storage environment based on multi-sensor fusion as described in claim 1, characterized in that: In S5, based on the contact coupling factor, the conventional measurement values ​​collected in S1 are dynamically corrected to obtain the compensated true measurement values, as follows: When the embedded sensor is a moisture sensor, the compensation amount is calculated based on the contact coupling factor using a pre-built compensation model. The compensation amount is superimposed on the currently collected measurement value to obtain the corrected true measurement value.

9. The intelligent monitoring method for rice storage environment based on multi-sensor fusion as described in claim 1, characterized in that: In S5, based on the contact coupling factor, the conventional measurement values ​​collected in S1 are dynamically corrected to obtain the compensated true measurement values, as follows: When the embedded sensor includes a temperature sensor, the thermal time constant of the current probe is estimated based on the contact coupling factor; an inverse thermal conduction filtering algorithm is used to dynamically compensate the temperature measurement value collected in S1 based on the thermal time constant to eliminate thermal hysteresis error.

10. A smart monitoring system for rice storage environment based on multi-sensor fusion, employing the smart monitoring method for rice storage environment based on multi-sensor fusion as described in any one of claims 1-9, characterized in that, include: The acquisition spectrum module is used to acquire conventional measurement values ​​of the embedded sensor in the grain pile, apply multi-frequency AC excitation signals to the embedded sensor, acquire the complex impedance data of the embedded sensor in the grain pile, and construct a multi-frequency dielectric response spectrum based on the complex impedance data. The decoupling transformation module is used to perform decoupling analysis on the multi-frequency dielectric response spectrum and transform it into a characteristic response time spectrum that characterizes the relaxation time of different physical processes inside the grain pile. The parameter extraction module is used to extract feature parameters corresponding to the contact state between the embedded sensor and the grain interface based on the feature response time spectrum. A factor construction module is used to construct a contact coupling factor based on the comparison results between the feature parameters and preset benchmark values; The numerical correction module is used to dynamically correct the conventional measurement values ​​acquired in the acquisition spectrum module based on the contact coupling factor, so as to obtain the compensated true measurement values.

Images