Rice grain quality detection method and system adopting ultrasonic analysis

Abstract

The invention discloses a rice grain quality detection method and system adopting ultrasonic analysis, and relates to the technical field of ultrasonic non-destructive detection and data modeling, and the method comprises the following steps: preprocessing a rice sample, and describing sample distribution and positioning errors by using a Gibbs point process and an anisotropic Gaussian noise model; detecting the internal structure of the rice through a pulse signal, and establishing an ultrasonic propagation and scattering model to extract parameters in the signal; and on the basis of a hierarchical Bayesian regression model, introducing a random effect between samples to predict continuous quality indexes of the rice grains. The method fully overcomes the problems of inaccurate sample positioning, single signal feature extraction and insufficient quality evaluation model precision in the prior art, improves the accuracy, stability and real-time performance of rice grain quality detection, and has a relatively high industrial application prospect.

Description

Technical Field

[0001] The present invention relates to the technical field of ultrasonic non-destructive testing and data modeling, and in particular to a rice grain quality detection method and system using ultrasonic analysis. Background Art

[0002] In recent years, with the increasing demand for food safety and quality monitoring, non-destructive testing technologies have gradually become an important means of assessing grain and food quality. Ultrasonic testing, as an efficient, real-time analytical method, has been widely used in fields such as internal material structure detection, industrial non-destructive testing, and medical imaging. In particular, research on detecting internal defects and physical properties using pulsed ultrasonic signals continues to deepen, while signal processing and data modeling technologies are becoming increasingly mature, providing theoretical and technical support for online detection and high-precision evaluation.

[0003] However, the existing technology still has many shortcomings in rice grain quality detection. First, the traditional method is relatively rough in processing the distribution of samples on the detection platform, often using a simple uniform arrangement, and lacks quantitative analysis of the spatial distribution characteristics and positioning errors of the samples, resulting in the detection results being greatly affected by the uneven arrangement of samples. Secondly, in terms of ultrasonic signal processing, most existing technologies only use basic time domain and frequency domain analysis methods, ignoring complex physical phenomena such as multipath propagation and signal attenuation, making it difficult to accurately extract key parameters reflecting the internal structure of rice. Thirdly, existing quality assessment methods mostly rely on simple linear regression or average statistical models, failing to fully introduce the random effects between samples, resulting in insufficient prediction accuracy for continuous quality indicators of rice grains. Therefore, how to use advanced mathematical models to accurately describe sample distribution and positioning errors, extract signal parameters by establishing a high-precision ultrasonic propagation and scattering model, and use a hierarchical Bayesian regression model to comprehensively consider sample randomness to achieve accurate prediction of continuous indicators of rice grain quality has become a key technical problem that needs to be solved urgently. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problems solved by the present invention are: the existing ultrasonic analysis method has problems such as the inability to accurately quantify sample distribution, the signal processing model is too simple, and the quality assessment fails to fully consider the random effects of samples, and how to use advanced mathematical models to accurately predict continuous indicators of rice grain quality.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: a method for detecting rice grain quality using ultrasonic analysis, comprising pre-treating a rice sample, conveying the pre-treated rice sample to a detection platform using an automatic conveyor belt or a vibrating plate, and automatically counting the rice sample using a photoelectric sensor and a weighing device, constructing a Gibbs point process model to describe the ideal spatial distribution of the sample in the detection area, combining an anisotropic Gaussian noise model to quantitatively describe the positioning error caused by mechanical vibration factors, and correcting the sample position through real-time feedback; using a signal generator to generate pulsed ultrasonic waves, The signal is transmitted into the rice through the transducer and transferred through the coupling medium. Receiving transducers are arranged at different positions of the sample to collect echo signals. The collected analog signals are converted into digital signals at high speed, amplified, low-pass filtered and smoothed in the time domain. An ultrasonic propagation and scattering model is established to extract echo delay, signal amplitude, attenuation, center frequency and bandwidth parameters. The signal features from different angles are combined into a comprehensive feature vector through the weighted covariance fusion method, and the comprehensive feature vector is used as input to construct a hierarchical Bayesian regression model to predict the continuous quality indicators of rice grains through the random effects between samples.

[0007] As a preferred embodiment of the rice grain quality detection method using ultrasonic analysis of the present invention, the pretreatment of the rice sample includes washing away attachments to the rice sample and separating impurities using a sieve.

[0008] As a preferred embodiment of the rice grain quality detection method using ultrasonic analysis described in the present invention, wherein: constructing a Gibbs point process model to describe the ideal spatial distribution of the sample in the detection area, combining an anisotropic Gaussian noise model to quantitatively describe the positioning error caused by mechanical vibration factors, and correcting the sample position through real-time feedback includes delivering the pretreated rice sample into the detection area; Automatically count the number of samples; The pretreated rice samples were sent to the detection area, and a Gibbs point process model was constructed to quantitatively describe the spatial distribution and positioning error of the samples on the detection platform. The interaction between the rice samples was also described by the Gibbs point process model. Anisotropic Gaussian noise is used to quantify the positioning error, and a positioning error and dynamic transportation model is constructed.

[0009] As a preferred embodiment of the rice grain quality detection method using ultrasonic analysis of the present invention, the method further comprises: generating a pulsed ultrasonic wave by a signal generator, emitting the pulsed ultrasonic wave into the rice via a transducer, transmitting the signal via a coupling medium, and arranging receiving transducers at different positions of the sample to collect the echo signal, comprising generating a pulsed ultrasonic wave by a signal generator, emitting the ultrasonic wave signal via a transducer, and transmitting the ultrasonic wave signal to the sample via a coupling medium; Arrange receiving transducers at different positions of the sample to obtain echo signals; Set the sampling rate and resolution of the analog-to-digital converter; The collected echo signal is amplified, low-pass filtered and time-domain smoothed to remove high-frequency noise and interference.

[0010] As a preferred embodiment of the rice grain quality detection method using ultrasonic analysis described in the present invention, the method includes: performing high-speed analog-to-digital conversion on the collected analog signal, amplifying, low-pass filtering, and time-domain smoothing, establishing an ultrasonic propagation and scattering model, and extracting echo delay, signal amplitude, attenuation, center frequency, and bandwidth parameters. The method includes introducing an attenuation term to construct a wave equation to describe the propagation of sound waves in a medium, introducing an absorption term to simulate energy attenuation in an actual medium, and constructing a received signal model based on the wave equation with the attenuation term. Perform Fourier transform on the received signal model to output the spectrum and obtain the center frequency and bandwidth; Continuous wavelet transform is used for analysis to capture the time and frequency domain information of the ultrasound signal; After collecting the ultrasonic signal, the features are extracted and the weighted covariance fusion method is used to obtain the comprehensive feature vector.

[0011] As a preferred embodiment of the rice grain quality detection method using ultrasonic analysis of the present invention, the method comprises: predicting the continuous quality index of the rice grains includes introducing random effects between samples, outputting a posterior distribution using a Bayesian method, and predicting the rice quality; If there is nonlinearity in the relationship between the quality index and the extracted features, a multi-layer neural network is used to construct a nonlinear mapping model, and a three-layer neural network is constructed through the fused feature vectors; Rice quality is divided into discrete grades, and a conditional random field model with spatial smoothness constraints is constructed.

[0012] As a preferred embodiment of the rice grain quality detection method using ultrasonic analysis of the present invention, the method comprises: predicting continuous rice grain quality indicators includes using a real-time data transmission and database storage mechanism, and displaying various indicators and change trends using line graphs and bar graphs; Set a quality threshold and trigger an alarm if the test result is abnormal; at the same time, compare the abnormal data with historical data, and re-calibrate the nonlinear mapping model parameters through the feedback mechanism.

[0013] Another object of the present invention is to provide a rice grain quality detection system using ultrasonic analysis. The system can detect the internal structure of rice through pulse signals, establish an ultrasonic propagation and scattering model to extract parameters in the signal, and solve the problem of the current ultrasonic analysis method that the sample distribution cannot be accurately quantified.

[0014] As a preferred embodiment of the rice grain quality detection system using ultrasonic analysis described in the present invention, the system includes: a sample processing and positioning and conveying module, an ultrasonic detection and feature extraction module, and a data modeling and quality assessment module; the sample processing and positioning and conveying module is used for cleaning, grading, drying pretreatment, automatic conveying, precise positioning, and fixing of rice samples; the ultrasonic detection and feature extraction module is used for converting physical information inside rice grains into time domain and frequency domain data; and the data modeling and quality assessment module is used for establishing a mathematical model using the extracted comprehensive features.

[0015] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of a rice grain quality detection method using ultrasonic analysis.

[0016] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a method for detecting rice grain quality using ultrasonic analysis.

[0017] Beneficial effects of the present invention: The rice grain quality detection method using ultrasonic analysis provided by the present invention achieves uniform distribution and stable fixation of the sample by pre-processing, precise positioning and dynamic transportation of the rice sample, and uses the Gibbs point process and anisotropic Gaussian noise model to quantitatively describe the spatial distribution and positioning error of the sample on the detection platform, thereby effectively eliminating the signal interference caused by uneven sample arrangement. Through the emission, reception and signal acquisition of pulsed ultrasonic waves, the internal structure of rice is detected, and an ultrasonic propagation and scattering model with attenuation terms is established. Fourier transform and continuous wavelet transform are used to extract multi-dimensional parameters such as echo delay, amplitude, attenuation coefficient, center frequency, bandwidth, etc., and weighted covariance fusion technology is used to integrate multi-angle data, effectively improving the stability and robustness of signal characteristics. Furthermore, based on a hierarchical Bayesian regression model, random effects between samples are introduced to accurately predict continuous rice grain quality indicators. When the relationship between quality indicators and signal characteristics is nonlinear, a multi-layer neural network is used to construct a nonlinear mapping. A conditional random field model is also used to discretely grade rice quality. Through real-time data transmission, dynamic feedback, and an alarm mechanism, automatic storage of test data and adaptive model optimization are achieved. This overall solution fully overcomes the existing problems of inaccurate sample positioning, single-source signal feature extraction, and insufficiently accurate quality assessment models. It improves the accuracy, stability, and real-time performance of rice grain quality testing and has great potential for industrial application. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 This is an overall flow chart of a method for detecting rice grain quality using ultrasonic analysis provided in the first embodiment of the present invention.

[0020] Figure 2 This is an overall flow chart of a rice grain quality detection system using ultrasonic analysis provided in the third embodiment of the present invention. DETAILED DESCRIPTION

[0021] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0022] Example 1, with reference to Figure 1 , as one embodiment of the present invention, provides a method for detecting rice grain quality using ultrasonic analysis, comprising: S1: Pre-treat the rice sample and transport it to the detection platform using an automatic conveyor belt or vibrating plate. Automatically count the rice using a photoelectric sensor and a weighing device. A Gibbs point process model is constructed to describe the ideal spatial distribution of the sample within the detection area. The anisotropic Gaussian noise model is used to quantify the positioning error caused by mechanical vibration, and the sample position is corrected through real-time feedback.

[0023] Furthermore, to ensure the quality of signal transmission during ultrasonic testing, the raw rice is thoroughly cleaned to remove surface sand, dust, and other impurities. The raw rice is placed in a dedicated cleaning machine, where high-pressure water and mechanical agitation rinse away any impurities. Multiple layers of screens are then used to separate impurities of varying sizes and densities, ensuring a pure and uniform rice sample for the next step of testing.

[0024] It should be noted that sample uniformity significantly impacts ultrasonic testing results. Different particle sizes and moisture levels can cause variations in ultrasonic propagation velocity and attenuation, necessitating sample grading and proper drying. Rice should be sorted by particle size using optical or mechanical grading devices. For samples with high moisture content, hot air drying (controlled at 40-60°C) is used to reduce the moisture content to the preset standard, ensuring consistent sample condition throughout the testing process.

[0025] Furthermore, large-scale testing requires that samples be automatically transported to the testing platform and evenly distributed across the surface to prevent local density variations from affecting ultrasonic signal acquisition. Pre-treated rice samples are delivered to the testing area using an automated conveyor belt or vibrating plate system. Sensors (such as photoelectric or weighing devices) automatically count the samples, ensuring that the number and distribution density of samples tested each time meet pre-set standards.

[0026] It should be noted that to ensure stable sample positioning during ultrasonic testing and reduce signal interference caused by sample movement or overlap, the sample must be precisely positioned and secured. This is accomplished using an automatically adjustable positioning device or robotic arm to precisely position the sample on the testing platform. Laser positioning or real-time camera monitoring systems are combined with adsorption or electrostatic fixation techniques to ensure the sample remains stationary during testing.

[0027] Furthermore, in order to quantitatively describe the spatial distribution and positioning error of the sample on the detection platform, a spatial point process model is constructed, and the detection platform area is set as , the ideal position of rice grains constitutes a point set

[0028] in, represents the ideal position set of rice grains on the detection platform, Indicates the representative The specific position of each rice grain in space.

[0029] The Gibbs point process is introduced to describe the interaction between samples, and its joint probability density function is:

[0030] in, Represents a given sample configuration The probability density function of (i.e., the set of all sample positions), It means proportional to, which means the probability density is proportional to the expression on the right. represents the natural exponential function, For all samples (in ) and, Indicates that for each pair of samples, the Euclidean distance is calculated Then, the pairing potential function is applied , Indicates sample and samples The Euclidean distance between .

[0031] The pairing potential function is defined as:

[0032] in, represents the distance between the two samples, Indicates the minimum allowed distance. When the distance is less than When , the potential function is set to infinity to prevent overlap. Indicates the adjustment parameter, which is used to control the The repulsion strength, represents an exponential decay term.

[0033] It should be noted that during the actual transportation process, the actual position of each rice grain Will be affected by system noise and mechanical vibration, and the ideal position There is a bias, modeled using anisotropic Gaussian noise:

[0034] in, Indicates sample The actual (detection) position of sample The ideal or preset position of represents the position error vector, Representation error The mean is 0 and the covariance matrix is Two-dimensional normal distribution.

[0035] The covariance matrix is expressed as:

[0036] in, represents the variance of the lateral (x-direction) error, represents the variance of the longitudinal (y-direction) error, are the standard deviations in the x and y directions, express and Correlation coefficient between directional errors.

[0037] In order to optimize the overall transportation effect, the following objective function is designed:

[0038] in, represents the objective function, which is used to evaluate the overall quality of the sample distribution. Indicates sample The square of the positioning error, that is, the deviation between the actual position and the ideal position, Represents the weight coefficient, which is used to balance the influence between positioning error and repulsive potential between samples. It means that for each pair of samples, the inverse of the square of the distance is used as the repulsive potential; Represents a small constant that prevents division by zero.

[0039] The first term is the accumulation of positioning errors, and the second term uses repulsive potential to encourage samples to maintain a reasonable distance between each other. and The two parameters are the balancing parameter and the constant to prevent the denominator from being zero. Through real-time feedback (such as Kalman filter correction), the system continuously iterates and optimizes to achieve the optimal delivery state.

[0040] S2: Use a signal generator to generate pulsed ultrasonic waves, which are emitted into the rice through a transducer. The signal is transmitted through a coupling medium, and receiving transducers are arranged at different positions of the sample to collect echo signals. The collected analog signals are converted into digital signals at high speed, and then amplified, low-pass filtered, and smoothed in the time domain. An ultrasonic propagation and scattering model is established to extract the echo delay, signal amplitude, attenuation, center frequency, and bandwidth parameters.

[0041] Furthermore, ultrasonic testing technology uses pulsed signals to probe the internal structure of rice. Ultrasonic waves can non-destructively penetrate rice, revealing internal material distribution and defects. A signal generator generates pulsed ultrasonic waves (frequency range 0.5-10 MHz). The signal is transmitted by a transducer and transferred to the sample via a coupling medium (such as water or a special coupling adhesive). Multiple receiving transducers are placed at different locations on the sample to obtain echo signals from multiple angles.

[0042] It should be noted that the received ultrasonic signal is an analog signal and needs to be converted to a digital signal using a high-speed analog-to-digital converter. After filtering, amplification, and noise reduction, the signal data must be suitable for subsequent analysis. The sampling rate of the analog-to-digital converter should be set to 2 to 5 times the highest frequency of the signal (for example, if the signal frequency is 5 MHz, the sampling rate should be no less than 10 to 25 MHz), and the resolution should be 12 bits or higher. The collected signal should be amplified, low-pass filtered, and time-smoothed to remove high-frequency noise and interference.

[0043] Furthermore, when ultrasound waves propagate through rice, they produce multipath reflections and attenuation due to the uneven internal structure, varying moisture content, and the presence of scatterers. Establishing a corresponding physical model will help extract key parameters from the signal.

[0044] It should be noted that based on the wave equation with attenuation term, it can be expressed as:

[0045] in, Indicates the location and time The sound pressure, represents the second-order derivative of the sound pressure with respect to time, that is, the acceleration term, Represents the speed of sound, which characterizes the speed of wave propagation in the medium; The Laplace operator reflects the diffusion characteristics of the signal in space. Represents the absorption coefficient, reflecting the energy attenuation rate, represents the first-order derivative of the sound pressure with respect to time, Represents the signal source term, which describes the excitation of ultrasonic waves.

[0046] The received signal is represented as:

[0047] in, Indicates time The received ultrasonic signal represents the number of scattering paths that exist, Indicates the The reflection intensity of the path, Indicates the The attenuation coefficient of each path, Indicates the The propagation delay of each path, Represents the unit pulse function (or step function), reflecting the starting time of the signal. Represents background noise.

[0048] Furthermore, in order to fully capture the time and frequency domain information of the ultrasonic signal, traditional methods and continuous wavelet transform (CWT) were used for analysis, respectively.

[0049] It should be noted that the time of launch is recorded and the moment of echo maximum The time difference is used to construct the echo delay :

[0050] Get the maximum amplitude of the echo signal and construct the signal amplitude Expressed as:

[0051] Assuming that the signal envelope decays exponentially, construct the attenuation coefficient Expressed as:

[0052] It was obtained by linear regression fitting after logarithmic transformation.

[0053] right Perform Fourier transform, expressed as:

[0054] in, Indicates signal In frequency On the spectrum, represents a frequency variable, represents the complex exponential kernel.

[0055] After obtaining the spectrum, calculate: Center frequency :

[0056] bandwidth :

[0057] in, represents the center frequency, that is, the energy-weighted average frequency, represents the power spectral density of the signal in the frequency domain, represents bandwidth and represents the standard deviation of the spectrum distribution.

[0058] In order to capture the local time-frequency characteristics of the signal, CWT is used to express it as:

[0059] in, Represents the wavelet coefficient, which represents the scale and pan The signal characteristics of Represents the scale parameter, reflecting the frequency information, Indicates the translation parameter, reflecting the time positioning information, Represents the mother wavelet function, describing the basis function of wavelet transform, represents the conjugate complex number of the mother wavelet function, represents the normalization factor.

[0060] scale parameter and translation parameters Used to extract local energy distribution And main frequency trajectory information.

[0061] Furthermore, in order to eliminate the error effect of a single detection angle, the system collects ultrasonic signals from multiple angles and extracts features respectively. The eigenvector obtained by the angle is expressed as

[0062] In order to obtain a more robust comprehensive feature vector, the weighted covariance fusion method is used:

[0063] in, Indicates the The feature vector extracted from the detection angle is Indicates the The weight of each angle is determined based on the signal-to-noise ratio or stability. Indicates the The outer product of angles constitutes the covariance information, Express The weighted sum of the angles is Represents matrix inversion, used for normalized fusion.

[0064] Weight Pre-calibrated through experiments based on the signal-to-noise ratio or detection stability at each angle.

[0065] S3: Signal features from different angles are fused into a comprehensive feature vector using the weighted covariance fusion method. The comprehensive feature vector is then used as input to construct a hierarchical Bayesian regression model to predict the continuous quality indicators of rice grains through random effects between samples.

[0066] Furthermore, to predict the continuous quality indicators of rice grains (such as moisture content, breakage rate or intrinsic defect score), a hierarchical Bayesian-based regression model was constructed, which not only considered the linear relationship between global features and quality, but also introduced random effects between samples.

[0067] Set up the first The quality index of the samples is , the model establishment is expressed as:

[0068] in, Indicates the The quality indicators of each sample (such as moisture content), express The mean is , the variance is The normal distribution of Indicates the The expected quality indicators of each sample, represents the intercept of the regression model, Indicates the The regression coefficient corresponding to the feature, Indicates the The sample in The value of the feature, represents a between-sample random effect, capturing unmodeled individual differences.

[0069] It should be noted that if the relationship between the quality index and the extracted features is highly nonlinear, a multi-layer neural network is used to construct a nonlinear mapping model.

[0070] Assume that the fused feature vector is , construct a three-layer neural network:

[0071] parameter Through back propagation algorithm training, the goal is to minimize the mean square error:

[0072] in, represents the fused input feature vector, represents the weight matrix of each layer, represents the bias vector of each layer, Represents an activation function, such as ReLU or tanh, which gives nonlinear mapping capabilities. and represents the output of the first and second hidden layers, The predicted quality indicators obtained by the output layer are shown.

[0073] Furthermore, for applications that require classifying rice quality into discrete grades (such as excellent, good, fair, and poor), a conditional random field (CRF) model with spatial smoothing constraints is constructed. This model leverages the spatial proximity of samples to smooth detection results and improve classification accuracy.

[0074] It should be noted that, if the sample Quality level , define the conditional probability:

[0075] in, Represents the conditional probability, given a sample feature and its neighborhood labels Under the condition, the sample belongs to the category The probability of Representation and Category The associated parameter vector, represents the inner product, which indicates the similarity between the sample features and the category parameters. represents the smoothing parameter, which regulates the influence of the neighborhood. Indicates sample The neighborhood sample set of shows the indicator function, when the neighborhood samples Category and If they are the same, the value is 1; otherwise, it is 0.

[0076] Furthermore, the detection system displays the analyzed rice quality results in real time on a graphical user interface (GUI), making it easy for operators to monitor. The system also stores raw signals, pre-processed data, characteristic parameters, and analysis results, providing data support for subsequent quality tracking and model optimization.

[0077] It should be noted that the system utilizes real-time data transmission and database storage, visually displaying various indicators and trends using line and bar charts. To ensure the robustness of the detection system, when test results exceed preset quality standards, the system automatically triggers an alarm and records the abnormal data. Multiple quality thresholds are set, and abnormal test results trigger an alarm module. Abnormal data is also compared with historical data, and a feedback mechanism is used to recalibrate model parameters to continuously optimize detection accuracy.

[0078] Example 2, reference Figure 2 , as an embodiment of the present invention, provides a rice grain quality detection system using ultrasonic analysis, including a sample processing and positioning transportation module, an ultrasonic detection and feature extraction module, and a data modeling and quality assessment module.

[0079] The sample processing and positioning and transportation module is used for cleaning, grading, drying pretreatment, automatic transportation, precise positioning and fixation of rice samples. The ultrasonic detection and feature extraction module is used to convert the physical information inside the rice grains into time domain and frequency domain data. The data modeling and quality assessment module is used to establish a mathematical model using the extracted comprehensive features.

[0080] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of the present invention. The aforementioned storage medium includes various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0081] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0082] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting, or processing it in another suitable manner as necessary, and then storing it in a computer memory.

[0083] It should be understood that various aspects of the present invention may be implemented using hardware, software, firmware, or a combination thereof. In the aforementioned embodiments, multiple steps or methods may be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, it may be implemented using any one or a combination of the following technologies known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc. It should be noted that the above embodiments are intended only to illustrate the technical solutions of the present invention and are not intended to be limiting. While the present invention has been described in detail with reference to preferred embodiments, those skilled in the art will appreciate that modifications or equivalent substitutions may be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and such modifications are intended to be encompassed by the claims of the present invention.

[0084] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for detecting rice grain quality using ultrasonic analysis, characterized in that: include: Rice samples are pretreated and transported to the testing platform using an automatic conveyor belt or vibrating plate. Automatic counting is performed using a photoelectric sensor and weighing device. A Gibbs point process model is constructed to describe the ideal spatial distribution of the sample within the testing area. An anisotropic Gaussian noise model is used to quantify positioning errors caused by mechanical vibration, and sample position is corrected using real-time feedback. A signal generator generates pulsed ultrasonic waves, which are transmitted into the rice through a transducer. The signals are then transmitted through a coupling medium. Receiving transducers are placed at different locations on the sample to collect echo signals. The collected analog signals undergo high-speed analog-to-digital conversion, amplification, low-pass filtering, and time-domain smoothing. An ultrasonic propagation and scattering model is established to extract echo delay, signal amplitude, attenuation, center frequency, and bandwidth parameters. The signal features from different angles are combined into a comprehensive feature vector through the weighted covariance fusion method. The comprehensive feature vector is used as input to construct a hierarchical Bayesian regression model to predict the continuous quality indicators of rice grains through the random effects between samples.

2. The method for detecting rice grain quality using ultrasonic analysis according to claim 1, wherein: The pre-processing of the rice sample comprises washing away the attachments of the rice sample and separating the impurities by using a sieve.

3. The method for detecting rice grain quality using ultrasonic analysis according to claim 2, wherein: The Gibbs point process model is constructed to describe the ideal spatial distribution of the sample in the detection area, the anisotropic Gaussian noise model is combined to quantitatively describe the positioning error caused by mechanical vibration factors, and the sample position is corrected through real-time feedback, including sending the pretreated rice sample into the detection area; Automatically count the number of samples; The pretreated rice samples were sent to the detection area, and a Gibbs point process model was constructed to quantitatively describe the spatial distribution and positioning error of the samples on the detection platform. The interaction between the rice samples was also described by the Gibbs point process model. Anisotropic Gaussian noise is used to quantify the positioning error, and a positioning error and dynamic transportation model is constructed.

4. The method for detecting rice grain quality using ultrasonic analysis according to claim 3, wherein: The method comprises utilizing a signal generator to generate pulsed ultrasonic waves, emitting the pulsed ultrasonic waves into the rice via a transducer, transmitting the signals via a coupling medium, and arranging receiving transducers at different positions of the sample to collect echo signals, comprising utilizing a signal generator to generate pulsed ultrasonic waves, emitting the ultrasonic waves via a transducer, and transmitting the signals to the sample via a coupling medium; Arrange receiving transducers at different positions of the sample to obtain echo signals; Set the sampling rate and resolution of the analog-to-digital converter; The collected echo signal is amplified, low-pass filtered and time-domain smoothed to remove high-frequency noise and interference.

5. The method for detecting rice grain quality using ultrasonic analysis according to claim 4, wherein: The collected analog signal is subjected to high-speed analog-to-digital conversion, and then amplified, low-pass filtered, and time-domain smoothing, and an ultrasonic propagation and scattering model is established to extract echo delay, signal amplitude, attenuation, center frequency, and bandwidth parameters. This includes introducing an attenuation term to construct a wave equation to describe the propagation of sound waves in the medium, and introducing an absorption term to simulate energy attenuation in the actual medium. A received signal model is constructed based on the wave equation with the attenuation term; Perform Fourier transform on the received signal model to output the spectrum and obtain the center frequency and bandwidth; Continuous wavelet transform is used for analysis to capture the time and frequency domain information of the ultrasound signal; After collecting the ultrasonic signal, the features are extracted and the weighted covariance fusion method is used to obtain the comprehensive feature vector.

6. The method for detecting rice grain quality using ultrasonic analysis according to claim 5, wherein: The method of predicting continuous quality index of rice grains includes introducing random effects between samples, outputting posterior distribution using Bayesian method, and predicting rice quality; If there is nonlinearity in the relationship between the quality index and the extracted features, a multi-layer neural network is used to construct a nonlinear mapping model, and a three-layer neural network is constructed through the fused feature vectors; Rice quality is divided into discrete grades, and a conditional random field model with spatial smoothness constraints is constructed.

7. The method for detecting rice grain quality using ultrasonic analysis according to claim 6, wherein: The method of predicting continuous quality indexes of rice grains includes adopting a real-time data transmission and database storage mechanism, and using a line graph and a bar graph to display various indicators and change trends; Set a quality threshold and trigger an alarm if the test result is abnormal; at the same time, compare the abnormal data with historical data, and re-calibrate the nonlinear mapping model parameters through the feedback mechanism.

8. A system using the rice grain quality detection method using ultrasonic analysis according to any one of claims 1 to 7, characterized in that: It includes sample processing and positioning transportation module, ultrasonic detection and feature extraction module, data modeling and quality assessment module; The sample processing and positioning and conveying module is used for cleaning, grading, drying and pre-processing of rice samples, as well as automatic conveying, precise positioning and fixing; The ultrasonic detection and feature extraction module is used to convert the physical information inside the rice grain into time domain and frequency domain data; The data modeling and quality assessment module is used to establish a mathematical model using the extracted comprehensive features.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the rice grain quality detection method using ultrasonic analysis according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the rice grain quality detection method using ultrasonic analysis according to any one of claims 1 to 7 are implemented.

Images