A grain moisture content acquisition and detection system based on microwave detection

Abstract

The present application belongs to the technical field of grain nondestructive detection and automatic measurement and control, and discloses a kind of grain moisture content acquisition and detection system based on microwave detection, comprising: vector network analyzer is used to emit microwave signal to measured sample in preset frequency band and receive reflected or transmitted signal, to measure the S parameter of measured sample;Microwave probe is used for electromagnetic coupling with measured sample;Two-dimensional motor drive mechanism is used to drive microwave probe to move relative to measured sample in two-dimensional plane;Motor drive unit is used to receive motion control instruction and drive stepper motor to operate;LabVIEW measurement and control module is used to execute bottom motor control instruction according to received parameter;MATLAB master control module is used to collect and save S parameter and spatial coordinate information;Data processing and automatic detection module generates moisture content distribution map and moisture content abnormal area marking result based on S parameter.

Description

Technical Field

[0001] This invention belongs to the field of non-destructive testing and automated measurement and control technology for grains, specifically relating to a grain moisture content acquisition and detection system based on microwave detection. Background Technology

[0002] Grain moisture content is one of the key parameters characterizing grain quality and storage safety. For example, the safe storage moisture content of wheat usually needs to be controlled within a certain range. When the moisture content is too high, the grain respiration rate increases significantly, which can easily lead to grain pile heating, mold, and pests, resulting in post-harvest losses and quality decline. Faced with large-scale wheat storage and long-term storage management, rapid and accurate detection of grain moisture content and its spatial distribution is a crucial foundation for ensuring food security and building smart grain depots.

[0003] Currently, methods for detecting grain moisture content can be broadly categorized into two types: direct measurement methods and indirect measurement methods. Direct measurement methods, such as drying and distillation, offer high accuracy and the results can be used for calibration. However, they are destructive, require small sample sizes per test, and have long testing cycles, making them unsuitable for large-scale, online testing scenarios. Indirect measurement methods, such as resistance methods, capacitance methods, and near-infrared spectroscopy, utilize the influence of moisture on electrical or optical properties to achieve rapid detection. They are relatively simple in structure and have fast response times, but they generally suffer from limited penetration depth, susceptibility to interference from stacking morphology and environmental factors, and difficulty in obtaining moisture information from large-volume grain piles.

[0004] Microwave detection technology infers moisture content by analyzing the attenuation and phase shift characteristics of electromagnetic waves propagating in a water-containing medium. It boasts strong penetration capabilities, is suitable for large sample volumes, and enables non-contact, rapid detection, making it particularly suitable for characterizing the internal moisture state of grains. In recent years, it has gradually become one of the important directions in grain moisture content research. Existing studies have constructed laboratory detection platforms based on microwave transmission or reflection methods. Using vector network analyzers or self-developed microwave detectors, they measure the S-parameters (such as the amplitude and phase of the transmission coefficient S21) after penetrating the grain sample. Based on this, they analyze the influence of multiple factors such as temperature, density, and thickness on the detection results, and establish data-driven moisture content prediction models.

[0005] In terms of hardware platforms, existing microwave detection systems typically include waveguides or antennas, a guide rail motor platform, a host computer, a data acquisition card, and a motor driver. Some systems use virtual instrument environments such as LabVIEW to develop measurement and control software to achieve motor drive control of the guide rail and microwave signal data acquisition. These platforms can simulate storage environments and multi-factor operating conditions, providing a hardware foundation for moisture content prediction model construction and frequency characteristic screening. However, from the perspective of "automated detection systems," existing microwave-based grain moisture content detection technologies and devices still have the following shortcomings: (1) The degree of automation in the measurement process is limited. Existing microwave testing platforms primarily focus on scanning control for single tests or limited combinations of operating conditions. Parameters such as scanning path, step size, and frequency points need to be manually set by the operator in advance through the interface, making it difficult to flexibly manage a large number of scanning tasks. For scenarios requiring high-density two-dimensional scanning or long-term continuous testing over a large area, the testing process still requires significant manual intervention.

[0006] (2) The motion control and data analysis processes are relatively separate. Most platforms use LabVIEW to complete motor motion control and signal acquisition, while data-driven methods such as moisture content prediction models and frequency feature subset screening are often implemented offline in environments such as MATLAB. The two interact with each other through file import and export. The measurement process and the automatic analysis process lack unified scheduling at the software level, and it is impossible to call the algorithm in real time during the scanning process to perform feature extraction, moisture content estimation or abnormal area identification.

[0007] (3) Lack of high-level unified control over motion and measurement LabVIEW programs often handle both low-level motor control and simple measurement process organization, but they lack support for complex scanning tasks, cross-platform algorithm calls, and unified management of multiple detection modes, making it difficult to flexibly expand the system into a fully automated, multi-mode detection platform.

[0008] (4) Insufficient automatic detection and result visualization capabilities While existing research has proposed various data-driven modeling and frequency feature screening methods, most of them focus on modeling and validating offline datasets. A complete software chain has not yet been formed that can automatically generate the S-parameter spatial distribution matrix, automatically call the detection algorithm, and output the moisture content distribution map or abnormal area marker in real time during the two-dimensional scanning process. The detection process still relies to a considerable extent on the manual observation and experience interpretation of the experimenters.

[0009] Therefore, there is an urgent need for a fully automated microwave detection data acquisition and automatic detection system and method that can introduce high-level main control and automatic detection algorithms on the basis of existing microwave detection hardware platforms, realize automatic scanning path planning, synchronous control of motion and measurement, and automatic storage, processing and discrimination of measurement data, so as to improve the automation, accuracy and intelligence of grain moisture content and related dielectric properties detection. Summary of the Invention

[0010] This invention aims to address the shortcomings of existing technologies and provides the following solutions: A grain moisture content acquisition and detection system based on microwave detection includes: a vector network analyzer, a microwave probe, a two-dimensional motor drive mechanism, a motor drive unit, a LabVIEW measurement and control module, a MATLAB main control module, and a data processing and automatic detection module; The vector network analyzer is used to transmit microwave signals to the sample under test in a preset frequency band and receive reflected or transmitted signals in order to measure the S-parameters of the sample under test. The microwave probe is connected to the vector network analyzer and is used for electromagnetic coupling with the sample under test; The two-dimensional motor drive mechanism is connected to the microwave probe and includes two linear motion units arranged along a first and a second direction that are perpendicular to each other, for driving the microwave probe to move relative to the sample under test in a two-dimensional plane. The motor drive unit is connected to the stepper motor in the two-dimensional motor drive mechanism, and is used to receive motion control commands and drive the stepper motor to run; The LabVIEW measurement and control module is connected to the motor drive unit and is used to execute low-level motor control commands based on the received parameters. The MATLAB main control module is connected to the LabVIEW measurement and control module and the vector network analyzer, and is used to collect and save the S-parameters and spatial coordinate information; The data processing and automatic detection module is integrated into the MATLAB main control module, and generates a moisture content distribution map and a mark result of abnormal moisture content areas based on the S parameters.

[0011] Preferably, each linear motion unit of the two-dimensional motor drive mechanism includes a stepper motor, a lead screw transmission mechanism, and a linear guide rail. The stepper motor drives the slider mounted on the linear guide rail to move through the lead screw transmission mechanism, thereby driving the microwave probe to move in the corresponding direction.

[0012] Preferably, the workflow of the MATLAB main control module includes: It receives user-defined parameters such as scan area size, scan step size, scan path type, measurement frequency band, and detection mode. Generate a 2D scanning path and a list of measurement tasks; The LabVIEW measurement and control module controls the two-dimensional motor drive mechanism, and triggers the vector network analyzer to collect the S-parameters when the microwave probe reaches each measurement point. Acquire and save the S-parameters and spatial coordinate information of each measurement point.

[0013] Preferably, the data processing and automatic detection module includes: a data preprocessing unit, a feature extraction unit, a discrimination and classification unit, and a result visualization unit; The data preprocessing unit is used to preprocess the S-parameters to obtain preprocessed parameters; The feature extraction unit is used to extract a feature vector from the preprocessed parameters to characterize the moisture content of the sample. The discrimination and classification unit constructs a moisture content prediction model based on the feature vector, and obtains the moisture content estimate of the tested sample based on the moisture content prediction model; The result visualization unit generates the moisture content distribution map and the marking results of the moisture content anomaly areas based on the moisture content estimate.

[0014] Preferably, the preprocessing method includes: Noise removal is performed on the collected S-parameters: ; in, This represents the parameter data after denoising. This indicates repeated data collection at the same measurement point. n The amplitude of the transmission coefficient. m =1,2,..., n ; The denoised parameter data from different measurement batches were then standardized: ; in, This represents the standardized parameter data. μk Indicates the first in the training data k The average amplitude of each frequency point σk Indicates the first in the training data k Standard deviation of amplitude at each frequency point; A smoothing filter is applied to the standardized parameter data to obtain the preprocessed parameters.

[0015] Preferably, the feature vector extracted from the preprocessed parameters is: ; in, fd This represents the representative frequency point sensitive to moisture content selected for each sample, and T indicates the transpose symbol.

[0016] Preferably, the moisture content prediction model includes: a linear regression model, an SVR model, or an AdaBoost-SVR model.

[0017] Preferably, the method for obtaining the moisture content estimate includes: ; in, This represents the estimated moisture content. F (·) indicates a moisture content prediction model.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) Realize fully automatic data acquisition for microwave detection: The LabVIEW measurement and control program, motor control and vector network analyzer are uniformly scheduled through the MATLAB main control module, and the scanning path planning, motion control and measurement triggering are automatically executed, reducing manual intervention and realizing unattended two-dimensional scanning and data acquisition.

[0019] (2) Realize the integration of measurement and control with automatic detection: Data preprocessing, feature extraction and automatic discrimination algorithms are integrated under the same software framework. Data analysis and result visualization can be completed during the measurement process, avoiding offline data transmission between LabVIEW and MATLAB in traditional methods, and improving detection efficiency and system integration.

[0020] (3) Improve the consistency of measurement data and spatial registration accuracy: Through unified path planning, motion positioning judgment and measurement triggering mechanism, ensure that the spatial position of each measurement point corresponds one-to-one with the S parameter data, reduce missed measurements and repeated measurements, improve the structure of two-dimensional data, and provide a reliable data foundation for moisture content modeling and spatial distribution analysis.

[0021] (4) Reduce reliance on human experience and improve the level of intelligent detection: Based on the automatic detection algorithm, feature extraction and automatic discrimination of two-dimensional S-parameter data can be performed, which can automatically identify areas with abnormal moisture content or other specific state areas, significantly reducing reliance on operator experience judgment and improving the consistency and objectivity of detection results.

[0022] (5) The system has strong scalability and wide applicability: The present invention adopts a hierarchical software architecture with MATLAB as the high-level master control and LabVIEW as the low-level measurement and control, which is convenient for expanding different types of motor platforms, microwave probes and detection algorithms. It can be widely used in the detection of moisture content of wheat and other grains and the microwave characteristic test of other media materials. Attached Figure Description

[0023] To more clearly illustrate the technical solution of the present invention, the drawings used in the embodiments are 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.

[0024] Figure 1 This is a schematic diagram of the system structure according to an embodiment of the present invention; Figure 2 This is a block diagram illustrating the data interaction and control logic between the MATLAB main control module, the LabVIEW measurement and control module, and the hardware in an embodiment of the present invention. Figure 3 This is a schematic diagram of a two-dimensional serpentine scanning path according to an embodiment of the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0027] Example: In this embodiment, as Figure 1 As shown, a grain moisture content acquisition and detection system based on microwave detection includes: a vector network analyzer, a microwave probe, a two-dimensional motor drive mechanism, a motor drive unit, a LabVIEW measurement and control module, a MATLAB main control module, and a data processing and automatic detection module.

[0028] A vector network analyzer is used to transmit microwave signals to the sample under test in a preset frequency band and receive reflected or transmitted signals in order to measure the S-parameters of the sample under test.

[0029] In this embodiment, the vector network analyzer is directly controlled by the MATLAB main control module via LAN, USB, or GPIB interfaces to execute low-level motor control and measurement commands based on received parameters. In another implementation, the vector network analyzer can also communicate indirectly with the MATLAB main control module via the LabVIEW measurement and control module to achieve scan task management, path planning, measurement scheduling, data storage, and automatic detection.

[0030] The microwave probe is connected to the vector network analyzer for electromagnetic coupling with the sample under test.

[0031] The two-dimensional motor drive mechanism is connected to the microwave probe and includes two linear motion units arranged along a first and a second direction that are perpendicular to each other, which are used to drive the microwave probe to move relative to the sample under test in a two-dimensional plane.

[0032] In this embodiment, each linear motion unit of the two-dimensional motor drive mechanism includes a stepper motor, a lead screw transmission mechanism, and a linear guide rail. The stepper motor drives a slider mounted on the linear guide rail to move via the lead screw transmission mechanism, thereby driving the microwave probe to move in the corresponding direction. The motor drive unit is connected to the stepper motor in the two-dimensional motor drive mechanism and is used to receive motion control commands and drive the stepper motor to run. The sample to be tested, such as a sample box or sample column filled with wheat, is placed below or at an appropriate position on the two-dimensional motor drive mechanism.

[0033] The LabVIEW Measurement and Control Module connects to the motor drive unit and is used to execute low-level motor control commands based on the received parameters.

[0034] The MATLAB main control module is connected to the LabVIEW measurement and control module and the vector network analyzer to collect and save S-parameters and spatial coordinate information.

[0035] In this embodiment, the MATLAB main control module and the LabVIEW measurement and control module can interact through ActiveX, .NET interface, or TCP / IP communication, such as... Figure 2 The diagram shown illustrates the data interaction and control logic between the MATLAB main control module and the LabVIEW measurement and control module and hardware.

[0036] The workflow of the MATLAB main control module includes: receiving the scan area size, scan step size, scan path type, measurement frequency band, and detection mode set by the user; generating a two-dimensional scan path and measurement task list; controlling the two-dimensional motor drive mechanism through the LabVIEW measurement and control module, and triggering the vector network analyzer to collect S-parameters when the microwave probe reaches each measurement point; and acquiring and saving the S-parameters and spatial coordinate information of each measurement point.

[0037] In this embodiment, parameters such as the length and width of the scanning area, the scanning step size, and the measurement frequency band are set by the operator in the interface of the MATLAB main control module. The MATLAB main control module generates a two-dimensional serpentine scanning path based on these parameters, such as... Figure 3 As shown, the path information is converted into a series of motion control commands and measurement tasks. For each measurement point, the MATLAB main control module calls the LabVIEW measurement and control module to control the movement of the two-dimensional motor drive mechanism. After the motor reaches its position and stabilizes, the vector network analyzer is triggered to collect S-parameter data, forming a three-dimensional dataset. ,in( i , j () represents the spatial grid coordinates. fk For the first kEach measurement frequency point. The obtained S-parameter data and corresponding spatial coordinates are structured and stored by the data processing and automatic detection module, providing a data foundation for subsequent automatic detection.

[0038] The data processing and automatic detection module is integrated into the MATLAB main control module, generating moisture content distribution maps and marking results of moisture content anomaly areas based on S-parameters.

[0039] The data processing and automatic detection module includes: a data preprocessing unit, a feature extraction unit, a discrimination and classification unit, and a result visualization unit.

[0040] The data preprocessing unit is used to preprocess the S-parameters to obtain the preprocessed parameters.

[0041] Preprocessing methods include: noise removal from the acquired S-parameters. ; in, This represents the parameter data after denoising. This indicates repeated data collection at the same measurement point. n The amplitude of the transmission coefficient. m =1,2,..., n Subsequently, the denoised parameter data from different measurement batches were standardized: ; in, This represents the standardized parameter data. μk Indicates the first in the training data k The average amplitude of each frequency point σk Indicates the first in the training data k The standard deviation of the amplitude at each frequency point is used to apply a smoothing filter to the standardized parameter data to further reduce isolated noise points and obtain the preprocessed parameters.

[0042] The feature extraction unit is used to extract feature vectors from the preprocessed parameters to characterize the moisture content of the sample.

[0043] Specifically, for each sample or measurement point, select several representative frequency points that are sensitive to moisture content. f 1, ..., fd And construct the feature vector as follows: ; in, fd This represents the representative frequency point sensitive to moisture content selected for each sample, and T indicates the transpose symbol.

[0044] In one preferred implementation, in a preset frequency band Internal calculations include statistical characteristics such as integral, slope, and half-power bandwidth, for example: ; ; in, Ii , j Indicates the preset frequency band Integrals within, Ki , j Indicates the preset frequency band Two reference frequencies selected within fa and fb The slope characteristics of the transmission coefficient amplitude can be combined with other features to form a high-dimensional feature vector. The dimensionality can be reduced by methods such as principal component analysis (PCA) to reduce redundancy and retain dominant change information.

[0045] The discrimination and classification unit constructs a moisture content prediction model based on feature vectors, and obtains the moisture content estimate of the tested sample based on the moisture content prediction model.

[0046] Moisture content prediction models include: linear regression model, SVR model or AdaBoost-SVR model.

[0047] In this embodiment, during the offline calibration stage, the true moisture content of several representative wheat sample groups is obtained using a standard drying method. Let the total mass of a certain training sample group be... m 0, after drying, the mass is md The true moisture content of this sample group is... M It can be calculated using the following formula: ; For each training sample or representative measurement region, the training feature vector is calculated using the feature vector construction method. xp and the corresponding actual moisture content Mp Composition of the training sample set: ; The moisture content prediction model adopts a multiple linear regression form: ; in, Let represent the regression weight vector, and b represent the bias term. The regression parameters are solved using the least squares criterion. ; in, p This indicates the number of the training sample or representative measurement area.

[0048] In another implementation, the moisture content prediction model employs a nonlinear regression model based on support vector regression (SVR). Let the kernel function be... The regression function obtained through training can then be expressed as: in, and The Lagrange multiplier represents the optimization problem of SVR. The kernel function can be a radial basis kernel, a linear kernel, or a polynomial kernel, etc.

[0049] In a further preferred implementation, the moisture content prediction model can adopt an SVR model based on the AdaBoost framework, that is, a model consisting of several SVR sub-models. As a weak learner, the weighted ensemble regression model is obtained by iteratively adjusting the sample weights: ; in, H Indicates the number of sub-models. γh These are the weight coefficients for each sub-model. The specific AdaBoost-SVR training process can be implemented using existing methods disclosed in the art, and this embodiment does not limit its internal optimization details.

[0050] The results visualization unit generates a moisture content distribution map and a marker of moisture content anomalies based on the estimated moisture content.

[0051] Will Input the calibrated moisture content prediction model (linear regression model, SVR model, or AdaBoost-SVR model) to obtain the moisture content estimate: ; in, This represents the estimated moisture content. F (·) indicates a moisture content prediction model.

[0052] To facilitate safety assessment, upper and lower thresholds can be set according to warehouse safety standards. Mmin , Mmax ,like Beyond the safe zone Mmin , Mmax If the value is ], then the location is marked as an area with abnormal moisture content.

[0053] The results visualization unit displays the estimated moisture content. With two-dimensional spatial coordinates ( i , j Correspondingly, a moisture content distribution map or a pseudo-color map of abnormal areas is generated. Preferably, the moisture content can be divided into several grade intervals, displayed in different colors or contour lines, and indicators such as the area of ​​abnormal areas and the size of the largest continuous abnormal area can be statistically analyzed to achieve intuitive display and quantitative assessment of the moisture content status inside the grain pile.

[0054] The aforementioned automatic detection system enables automatic estimation of wheat moisture content and identification of abnormal areas, eliminating the need for operators to observe each point individually. S The parameter curve can obtain the spatial distribution of water content of a large area sample, which improves detection efficiency and result consistency, while reserving an interface for extended applications based on more complex machine learning models.

[0055] Extended applications: Based on the above embodiments, the system of the present invention can also be extended to detect the microwave properties of other media materials, such as measuring the moisture content of other grains like corn and rice, or testing the moisture content or dielectric constant of soil and cement specimens. Different media can be detected by changing the type of microwave probe, adjusting the frequency range, and calibrating the model. By adding a third linear motion unit or lifting platform, multi-layer scanning can also be achieved to acquire three-dimensional data. S Parameter distribution data, thereby enabling three-dimensional imaging of dielectric properties or water content.

[0056] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A grain moisture content acquisition and detection system based on microwave detection, characterized in that, include: Vector network analyzer, microwave probe, two-dimensional motor drive mechanism, motor drive unit, LabVIEW measurement and control module, MATLAB main control module and data processing and automatic detection module; The vector network analyzer is used to transmit microwave signals to the sample under test in a preset frequency band and receive reflected or transmitted signals in order to measure the S-parameters of the sample under test. The microwave probe is connected to the vector network analyzer and is used for electromagnetic coupling with the sample under test; The two-dimensional motor drive mechanism is connected to the microwave probe and includes two linear motion units arranged along a first and a second direction that are perpendicular to each other, for driving the microwave probe to move relative to the sample under test in a two-dimensional plane. The motor drive unit is connected to the stepper motor in the two-dimensional motor drive mechanism, and is used to receive motion control commands and drive the stepper motor to run; The LabVIEW measurement and control module is connected to the motor drive unit and is used to execute low-level motor control commands based on the received parameters. The MATLAB main control module is connected to the LabVIEW measurement and control module and the vector network analyzer, and is used to collect and save the S-parameters and spatial coordinate information; The data processing and automatic detection module is integrated into the MATLAB main control module, and generates a moisture content distribution map and a mark result of abnormal moisture content areas based on the S parameters.

2. The grain moisture content acquisition and detection system based on microwave detection according to claim 1, characterized in that, Each linear motion unit of the two-dimensional motor drive mechanism includes a stepper motor, a lead screw transmission mechanism, and a linear guide rail. The stepper motor drives the slider mounted on the linear guide rail to move through the lead screw transmission mechanism, thereby driving the microwave probe to move in the corresponding direction.

3. The grain moisture content acquisition and detection system based on microwave detection according to claim 1, characterized in that, The workflow of the MATLAB main control module includes: It receives user-defined parameters such as scan area size, scan step size, scan path type, measurement frequency band, and detection mode. Generate a 2D scanning path and a list of measurement tasks; The LabVIEW measurement and control module controls the two-dimensional motor drive mechanism, and triggers the vector network analyzer to collect the S-parameters when the microwave probe reaches each measurement point. Acquire and save the S-parameters and spatial coordinate information of each measurement point.

4. The grain moisture content acquisition and detection system based on microwave detection according to claim 1, characterized in that, The data processing and automatic detection module includes: a data preprocessing unit, a feature extraction unit, a discrimination and classification unit, and a result visualization unit; The data preprocessing unit is used to preprocess the S-parameters to obtain preprocessed parameters; The feature extraction unit is used to extract a feature vector from the preprocessed parameters to characterize the moisture content of the sample. The discrimination and classification unit constructs a moisture content prediction model based on the feature vector, and obtains the moisture content estimate of the tested sample based on the moisture content prediction model; The result visualization unit generates the moisture content distribution map and the marking results of the moisture content anomaly areas based on the moisture content estimate.

5. The grain moisture content acquisition and detection system based on microwave detection according to claim 4, characterized in that, The preprocessing method includes: Noise removal is performed on the collected S-parameters: ； in, This represents the parameter data after denoising. This indicates repeated data collection at the same measurement point. n The amplitude of the transmission coefficient. m =1,2,..., n ; The denoised parameter data from different measurement batches were then standardized: ； in, This represents the standardized parameter data. μk Indicates the first in the training data k The average amplitude of each frequency point σ k Indicates the first in the training data k Standard deviation of amplitude at each frequency point; A smoothing filter is applied to the standardized parameter data to obtain the preprocessed parameters.

6. The grain moisture content acquisition and detection system based on microwave detection according to claim 4, characterized in that, The feature vector extracted from the preprocessed parameters is: ； in, fd This represents the representative frequency point sensitive to moisture content selected for each sample, and T indicates the transpose symbol.

7. The grain moisture content acquisition and detection system based on microwave detection according to claim 4, characterized in that, The moisture content prediction model includes: linear regression model, SVR model or AdaBoost-SVR model.

8. The grain moisture content acquisition and detection system based on microwave detection according to claim 4, characterized in that, The methods for obtaining the moisture content estimate include: ； in, This represents the estimated moisture content. F (·) indicates a moisture content prediction model.

Images