Anti-foam interference millimeter wave radar curdling solidification state monitoring method and system

Abstract

The invention discloses an anti-foam interference millimeter wave radar curdling solidification state monitoring method and system. The method comprises the following steps: transmitting a signal by using a millimeter wave radar, processing a received intermediate frequency signal, and generating a range profile sequence; creating a combined atom dictionary capable of simultaneously representing the solidification process and the foam interference; for the range profile sequence, solving a joint sparse optimization problem based on a joint atom dictionary and with morphological prior; solving a joint sparse optimization problem by applying an alternating direction multiplier method; the range profile sequence is decomposed into a pure solidification signal component and a foam interference signal component, different solidification stages are identified based on the pure solidification signal component, a solidification end point is quantitatively judged, and foam interference energy is monitored and alarmed; and when the curdling reaches a solidification end point, the subsequent process is automatically triggered, and the foam interference coefficient energy is monitored in real time. By implementing the method provided by the invention, the problem of interference of foam on millimeter wave radar monitoring can be solved, and the robustness and environmental adaptability of the system are enhanced.

Description

Technical Field

[0001] This invention relates to the field of food processing monitoring technology, and in particular to a millimeter-wave radar method and system for monitoring the solidification state of slurry with resistance to foam interference. Background Technology

[0002] A key step in tofu production is mixing soy milk with a coagulant (commonly known as coagulation) to coagulate the soy milk into tofu curd. Accurately determining the coagulation endpoint is crucial for ensuring the final product's yield, texture, and taste. Traditionally, this judgment relies on operator experience, a method that, while intuitive, has significant drawbacks, including high subjectivity, inefficiency, and difficulty in standardizing operations. With technological advancements, non-contact sensing technologies are increasingly being applied in food processing, showing great potential, particularly in monitoring the coagulation state of soy milk. Millimeter-wave radar technology has garnered attention due to its non-contact characteristics, strong penetrating power, and good resistance to environmental interference. By analyzing changes in radar echo signals, such as monitoring the attenuation of the peak amplitude of the echo from the liquid surface, the degree of soy milk coagulation can be inferred.

[0003] However, in practical applications, the large amount and unstable foam generated during the coagulation process poses a challenge to millimeter-wave radar. The foam layer reflects and absorbs electromagnetic waves, which not only weakens the energy of the actual soy milk surface echo signal but may also completely mask these signals, affecting measurement accuracy. Because foam has different dielectric properties than air and soy milk, it forms a random dielectric layer that generates complex and constantly changing reflection peaks. These peaks may be mistaken for the location of the soy milk surface, leading to incorrect location judgments. The presence of foam severely distorts the one-dimensional range image of the radar because the foam reflection is superimposed on the reflection from the actual liquid surface, thus undermining the effectiveness of traditional judgment methods based on single features. Therefore, effectively suppressing or separating the interference caused by foam and accurately extracting the true signal related to the coagulation state of soy milk has become a key challenge for the practical application of millimeter-wave radar technology in tofu production. Solving this problem is of great significance for improving the automation level of tofu production and enhancing the consistency and stability of product quality.

[0004] Therefore, it is necessary to design a new method to accurately separate the real coagulation signal and the foam interference signal from the mixed radar signal. This not only solves the problem of foam interference with millimeter-wave radar monitoring, but also enhances the robustness and environmental adaptability of the system. At the same time, it provides rich process monitoring dimensions and solid physical and mathematical model support, which significantly improves the automation and quality control level of tofu production. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a millimeter-wave radar method and system for monitoring the solidification state of slurry with resistance to foam interference.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a millimeter-wave radar method for monitoring the solidification state of slurry with resistance to foam interference, comprising: A millimeter-wave radar installed above a solidification tank transmits a linear frequency modulated signal, and the received intermediate frequency signal is processed to generate a range image sequence. Create a joint atomic dictionary that can simultaneously characterize the solidification process and foam interference; For the distance image sequence, solve the joint sparse optimization problem based on the joint atom dictionary and with morphological priors; The joint sparse optimization problem is solved by applying the alternating direction multiplier method. The distance image sequence is decomposed into a pure solidification signal component and a foam interference signal component. Based on the pure solidification signal component, different stages of solidification are identified and the solidification endpoint is quantified. The foam interference energy is monitored and alarmed. Once the slurry reaches the solidification endpoint, the subsequent processes are automatically triggered, and the foam interference coefficient energy is monitored in real time.

[0007] The further technical solution is as follows: The method of using a millimeter-wave radar installed above the solidification tank to transmit a linear frequency modulated signal and processing the received intermediate frequency signal to generate a range image sequence includes: The FMCW millimeter-wave radar installed above the coagulation tank transmits linear frequency modulated chirp signals, and the received intermediate frequency signals are processed by FFT to generate a range image sequence containing information on the liquid surface, foam, the interior of the soy milk, and the reflection from the bottom of the tank.

[0008] The further technical solution is as follows: the joint atomic dictionary includes a coagulation process sub-dictionary and a foam interference sub-dictionary, wherein the coagulation process sub-dictionary is composed of multiple atoms representing typical radar response modes of soy milk at different coagulation stages; the foam interference sub-dictionary contains atoms representing different radar echo patterns caused by foam interference. In the offline phase, distance image sequences under different working conditions are collected, and a dictionary learning algorithm is applied to train two sub-dictionaries that can clearly distinguish between the solidification process and foam interference, so as to form a joint atomic dictionary.

[0009] The further technical solution is as follows: the joint sparse optimization problem is expressed as: ,in, It is the joint sparse coefficient vector to be solved; It is an L1 norm; It is an introduced morphological prior regularization term. Process smoothing term Penalty solidification process coefficient The drastic time variation forces its solution to be smooth and continuous in time, consistent with the physical fact that the solidification process changes slowly. Disturbance mutation term. Encourage the foam interference coefficient It exhibits mutation characteristics; Functions that promote time sparsity; This is a dictionary of union atoms.

[0010] The further technical solution is as follows: the application of the alternating direction multiplier method to solve the joint sparse optimization problem includes: By introducing auxiliary variables, the joint sparse optimization problem is transformed into a more solvable form. Dual variables are introduced, and an augmented Lagrangian function is constructed to obtain the processed problem. The ADMM method is used to iteratively solve the processed problem.

[0011] The further technical solution is as follows: the iterative solution of the processed problem using the ADMM method includes: The sparse representation of the problem at the current moment is updated after processing, and information fusion is achieved by combining time-varying information and historical data; Update the representation of the foam interference part for the processed problem; The smoothing part of the updated problem process after the processing; The interference mutation part is updated after the problem is processed to capture sudden changes or anomalies; The dual variable of the processed problem is updated so that the processed problem converges to the global optimum.

[0012] The further technical solution is as follows: The distance image sequence is decomposed into a pure solidification signal component and a foam interference signal component; different stages of solidification are identified and the solidification endpoint is quantified based on the pure solidification signal component; and the foam interference energy is monitored and alarmed, including: The distance image sequence is decomposed into pure coagulation signal components and foam interference signal components. Based on the pure coagulation signal components, the different stages of soy milk coagulation are identified by analyzing their changes over time. When the atoms representing the stable state at the end of coagulation become the main components and the energy ratio exceeds a preset threshold, and the amount of state change is continuously lower than the stable threshold, the coagulation process is determined to have reached the end of coagulation. Monitor and alarm on the energy of foam interference.

[0013] The further technical solution is as follows: the monitoring and alarm of foam interference energy includes: The severity and changes of foam are reflected by real-time monitoring and plotting the foam interference energy curve. At the same time, an alarm is triggered when the energy is continuously higher than the danger threshold or the rate of change is abnormal, prompting the operator to check the raw materials or adjust the process parameters.

[0014] The further technical solution is as follows: when the slurry reaches the solidification endpoint, the subsequent process is automatically triggered, and the foam interference coefficient energy is monitored in real time, including: When the solidification endpoint is reached, an output signal is generated to control subsequent processes. The energy value of the foam interference coefficient is monitored in real time, and an alarm is triggered when it is too high for an extended period of time to prompt the operator to check or adjust. At the same time, the evolution curve of the sparsity coefficient in the process is stored in the database as a digital fingerprint for quality traceability.

[0015] This invention also provides a millimeter-wave radar solidification state monitoring system resistant to foam interference, comprising: The sequence generation unit is used to transmit linear frequency modulated signals using a millimeter-wave radar installed above the solidification tank, and to process the received intermediate frequency signals to generate a range image sequence. A dictionary creation unit is used to create a joint atom dictionary that can simultaneously characterize the solidification process and foam interference, wherein each atom in the dictionary is a vector with the same length as the distance image sequence; The first solving unit is used to solve a joint sparse optimization problem based on the joint atomic dictionary and with morphological priors for the distance image sequence. The second solution unit is used to solve the joint sparse optimization problem by applying the alternating direction multiplier method. The identification unit is used to decompose the distance image sequence into a pure solidification signal component and a foam interference signal component, identify different stages of solidification based on the pure solidification signal component, quantify and determine the solidification endpoint, and monitor and alarm the foam interference energy. The monitoring unit is used to automatically trigger subsequent processes when the slurry reaches the solidification endpoint, and to monitor the foam interference coefficient energy in real time.

[0016] The beneficial effects of this invention compared with the prior art are as follows: By creating a joint atom dictionary and applying the alternating direction multiplier method to solve the joint sparse optimization problem with morphological priors, this invention accurately separates the pure coagulation signal and foam interference signal from the mixed radar signal. This not only effectively solves the interference of foam on millimeter-wave radar monitoring and enhances the robustness and environmental adaptability of the system, but also provides rich process insights through quantitative monitoring of foam interference. Based on solid physical and mathematical model support, it significantly improves the automation level and quality control capability of tofu production.

[0017] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A schematic flowchart of a millimeter-wave radar method for monitoring the solidification state of slurry with anti-foam interference provided in an embodiment of the present invention; Figure 2 A schematic block diagram of a millimeter-wave radar spot slurry solidification state monitoring system 300 with anti-foam interference provided in an embodiment of the present invention; Figure 3 A schematic block diagram of a computer device provided for an embodiment of the present invention. Detailed Implementation

[0020] 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, not all, of the embodiments of the present invention. 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.

[0021] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0022] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0023] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0024] Please see Figure 1 , Figure 1This is a flowchart illustrating the anti-foaming interference millimeter-wave radar method for monitoring the coagulation state of soybean milk provided in this embodiment of the invention. This method is applied in a server. The server interacts with the millimeter-wave radar, transmitting a linear frequency modulated signal via the radar mounted above the coagulation tank. The received intermediate frequency signal is processed to generate a range image sequence, creating a joint atomic dictionary capable of simultaneously characterizing the coagulation process and foaming interference. Using this dictionary, a joint sparse optimization problem with morphological priors is solved, and the alternating direction multiplier method is applied to decompose the pure coagulation signal component and the foaming interference signal component, thereby accurately identifying different stages of soybean milk coagulation and quantifying the coagulation endpoint. Furthermore, the system can monitor the energy of foaming interference in real time, triggering an alarm when the value is too high to ensure timely adjustments by the operator. The entire process not only effectively solves the interference of foaming interference on monitoring, enhancing the system's robustness and environmental adaptability, but also provides powerful quality traceability by storing the sparse coefficient evolution curve as a "digital fingerprint" in the database, significantly improving the automation level and quality control capabilities of tofu production. This method combines a solid physical foundation with advanced mathematical models, providing multi-dimensional support for monitoring the tofu production process.

[0025] Figure 1 This is a schematic flowchart of the millimeter-wave radar method for monitoring the solidification state of slurry with anti-foam interference provided in an embodiment of the present invention. Figure 1 As shown, the method includes the following steps S110 to S160.

[0026] S110. A millimeter-wave radar installed above the solidification tank transmits a linear frequency modulated signal and processes the received intermediate frequency signal to generate a range image sequence.

[0027] In this embodiment, the range image sequence refers to the data sequence obtained by processing the linear frequency modulated signal transmitted by the millimeter-wave radar and the intermediate frequency signal received, which reflects the evolution of the comprehensive echo signal intensity of the target reflector at different distances over time.

[0028] Specifically, the FMCW millimeter-wave radar installed above the coagulation tank transmits a linear frequency modulated Chirp signal, and the received intermediate frequency signal is processed by FFT to generate a range image sequence containing information on the liquid surface, foam, the interior of the soy milk, and the reflection from the bottom of the tank.

[0029] This is a series of data reflecting the target distance and the intensity of its reflected signal obtained by processing signals transmitted and received by a frequency-modulated continuous wave (FMCW) millimeter-wave radar installed above the solidification tank. Specifically, the radar preferably transmits a chirp signal whose frequency varies linearly with time, and within each frame period, at the... Frame, transmitted signal It can be represented as: ;in It is the fast time within a single frequency sweep. It is the amplitude of the transmitted signal. It is the starting frequency. It is the sweep bandwidth. It is the frequency sweep cycle.

[0030] When the intermediate frequency (IF) signal is extracted from the received signal through mixing and filtering, for a single stationary target at a certain distance from the radar, the frequency of the IF signal is linearly proportional to the distance to the target.

[0031] Next, from the intermediate frequency signal Data sampled in the medium-fast time dimension undergoes a Fast Fourier Transform (FFT) to transform it into the frequency domain. This process yields a one-dimensional range image sequence. The x-axis is directly mapped to the target distance. For distance is A single stationary target, the frequency of the intermediate frequency signal Distance from target The distance and reflection intensity are linearly proportional. The vertical axis represents the combined echo signal intensity of all reflectors (such as the liquid surface, foam, the interior of the soy milk, and the bottom of the can) at this distance. Therefore, the range image sequence essentially provides a method to transform radar echo information in the time domain into an intuitive range-reflection intensity map, enabling us to monitor and analyze the state and changes of different components during the coagulation process. This not only effectively captures the key stages of the soy milk coagulation process but also allows for real-time monitoring of foam interference that may affect measurement accuracy.

[0032] S120. Create a joint atomic dictionary that can simultaneously characterize the solidification process and foam interference.

[0033] In this embodiment, the joint atom dictionary includes a coagulation process sub-dictionary and a foam interference sub-dictionary. The coagulation process sub-dictionary consists of multiple atoms representing typical radar response patterns of soy milk at different coagulation stages. The foam interference sub-dictionary contains atoms representing different radar echo patterns caused by foam interference. Each atom in the dictionary is associated with a one-dimensional distance image. Vectors of the same length have shapes that simulate the echo characteristics under specific physical conditions.

[0034] In the offline phase, distance image sequences under different working conditions are collected, and a dictionary learning algorithm is applied to train two sub-dictionaries that can clearly distinguish between the solidification process and foam interference, so as to form a joint atomic dictionary.

[0035] In this embodiment, the union dictionary It is composed of two sub-dictionaries: in This is a sub-dictionary of the solidification process. The atoms in this dictionary represent typical "range-image morphological primitives" at different stages of the solidification process, physically corresponding to specific radar response patterns in the evolution of the composite dielectric constant of soybean milk. Its atoms include the following categories: Initial soybean milk state atoms Before coagulation, soy milk is a homogeneous liquid. Electromagnetic waves undergo primary reflection at the liquid surface, and some energy attenuates in the homogeneous medium after penetrating. This atom represents a sharp primary peak reflecting clearly from the liquid surface; after the primary peak, the signal intensity follows an exponential trend. Rapid decay, among which It is the dielectric loss coefficient of soy milk. It refers to the penetration depth. ;in It refers to the liquid level. The sharpness of the peak. It is a step function, and Peak() is a peak shape function used to simulate the shape of reflection peaks in radar echoes. Based on the actual system model, this invention preferably defines it as a Gaussian function, with the specific expression as follows: ;in, For amplitude parameters, The parameter representing the center position of the peak. These are key parameters for controlling peak width and sharpness. Adjustment is used... It can simulate sharp reflections from liquid surfaces (small) Value) and broad, gentle foam or inhomogeneous body reflection (large) value).

[0036] Protein aggregation phase atoms In the initial stage of coagulation, protein molecules begin to aggregate into tiny particles, making the medium non-uniform and enhancing the scattering and absorption of electromagnetic waves. Compared to... The main peak of this atom shows a slight broadening. > The peak value may decrease slightly. More importantly, the decay after the main peak is no longer a smooth exponential decay, but rather exhibits slight fluctuations and faster energy decay, simulating the scattering effect of internal micro-inhomogeneities. ;in This is the decay coefficient during the aggregation period. > This indicates an accelerated decline. It is a decay function with random fluctuations, used to simulate the incoherent scattering effect induced by tiny scatterers inside a medium, and is defined as a stochastic process. ;in It is a narrowband Gaussian random process that simulates the random fluctuations of the scattered signal; The effective attenuation caused by scattering is simulated; A' is the scattering intensity coefficient.

[0037] Atoms during gel network formation The formation of a three-dimensional gel network leads to a drastic, macroscopic change in the dielectric constant, fundamentally altering the propagation characteristics of radar signals. This is a transitional and highly dynamic class of atoms. Its morphological characteristics include a significantly broadened main peak with a decreased amplitude, and the appearance of new, relatively broad secondary reflection peaks or absorption valleys behind the liquid surface. This represents a strong reflection or absorption of electromagnetic waves by an emerging macroscopic layered structure. This indicates a significant broadening of the peak width. This represents a complex response caused by newly formed internal structures. It is obtained through a combination of physical models of electromagnetic wave propagation in layered media and feature learning from experimental data. Specifically: Physical modeling: Treat it as the bottom interface of the emerging gel layer (depth) The result of the superposition interference of the reflection from the liquid surface and the reflection from the liquid surface can be approximately expressed as: ;in This indicates the secondary reflection peak of the gel structure. is the attenuation coefficient of the gel layer.

[0038] By employing machine learning algorithms such as clustering, common echo morphological features of the gel formation stage are extracted from a large amount of labeled experimental data, and these are used as the specific implementation of the function. Ultimately, the dictionary... atoms in It includes a series generated by the above methods. The primitive is used to comprehensively describe the dynamic process of gel network formation.

[0039] Stable atoms at the solidification endpoint The gel network stabilizes, and the overall dielectric properties of the tofu pudding tend towards a new, homogeneous, or layered stable state. The morphology of the atoms also tends towards a new stable mode. This may manifest as the main peak at the liquid surface becoming relatively stable again (but possibly wider than the initial state), and remaining stable at a specific depth. A stable and clear characteristic peak or characteristic attenuation gradient forms at the solidification endpoint, which becomes the "fingerprint" characteristic of the solidification endpoint. ;where the function For defining the stable radar echo characteristics formed at the solidification endpoint, the present invention preferably uses the following deterministic model: ;in For characteristic amplitude, The position of the characteristic peak ( > ), This refers to the width parameter of the characteristic peak. The specific shape and parameters of this function were determined through analysis and averaging of a large amount of endpoint sample data, ultimately forming a sub-dictionary. The atom representing the end point of solidification .

[0040] Another major category of dictionaries is the foam interference sub-dictionary. The atoms in this dictionary represent typical "distance-image morphological primitives" caused by foam interference. Its atoms include the following categories: Thin-layer / sparse foam atoms : at the actual liquid level peak Previously, there were one or more small, irregular peaks with low amplitude and random locations, and the amplitude of the actual liquid surface peak was somewhat attenuated. ;in The amplitude after attenuation. For the sharp peak width, the function The specific definition is to simulate the real liquid surface. The expressions for the previously observed weak and disordered reflection peaks are as follows: ;in The number of foam reflection peaks is a random variable that follows a Poisson distribution. It is the first The amplitude of the foam reflection. It is the first The distance (depth) position of each foam reflection peak satisfies and in the interval Uniformly distributed inside This indicates the maximum possible thickness of the foam layer.

[0041] Thick / Dense Foam Atoms : It manifests as a location located The reflection peaks at the surface are broad and strong, while the actual liquid surface peaks are greatly attenuated or submerged. ;in The width of the foam peak.

[0042] Dynamic foam atoms This is a set of primitives representing "change patterns," capturing the rapid changes in the position, shape, and amplitude of bubble peaks. Its time-varying characteristics are reflected in the regularization terms of the joint sparse decomposition.

[0043] During the offline phase, distance image sequences were acquired under various operating conditions, including pure solidification process, foam-only process, and foam-containing solidification process. Dictionary learning algorithms (such as K-SVD) are used to train a dictionary of separated data with clear physical meaning. and .

[0044] S130. For the distance image sequence, solve the joint sparse optimization problem based on the joint atomic dictionary and with morphological priors.

[0045] In this embodiment, the joint sparse optimization problem refers to solving an optimization problem with morphological priors when processing the real-time acquired distance image vector, thereby simultaneously suppressing foam interference and accurately analyzing the solidification state. The optimization problem utilizes the L1 norm to drive the sparsity of the solution and designs specific time dynamic regularization terms to distinguish and process the smooth continuous component representing the real solidification process and the abrupt component representing foam interference.

[0046] The joint sparse optimization problem is expressed as: ,in, It is the joint sparse coefficient vector to be solved; Is The coefficients on the graph represent the actual solidification process components; Is The coefficient on the figure represents the foam interference component. It is the L1 norm, used to drive the sparsity of the solution, that is, to represent the signal with the fewest possible combinations of atoms; This is an introduced morphological prior regularization term that leverages the fundamental difference in the temporal dynamics between the solidification process and foam interference to enhance the separation effect. Specifically, this regularization term can be designed as follows: Process smoothing term Penalty solidification process coefficient The drastic time variation forces its solution to be smooth and continuous in time, consistent with the physical fact that the solidification process changes slowly. Disturbance mutation term. Encourage the foam interference coefficient It exhibits mutation characteristics; In this invention, the preferred function is one that promotes temporal sparsity, such as... ,encourage It remains unchanged most of the time, but undergoes large jumps when a few bubbles are formed or burst. This is a dictionary of union atoms.

[0047] Specifically, for each point in time acquired in real time Distance vector (Right now (A time slice), by solving a joint sparse optimization problem with morphological priors, simultaneously suppressing foam interference and analyzing the solidification state.

[0048] S140. Solve the joint sparse optimization problem by applying the alternating direction multiplier method.

[0049] In one embodiment, step S140 described above may include steps S141 to S142.

[0050] S141. Introduce auxiliary variables to transform the joint sparse optimization problem into a more solvable form. Introduce dual variables and construct an augmented Lagrangian function to obtain the processed problem.

[0051] S142. The ADMM method is used to iteratively solve the processed problem.

[0052] In one embodiment, step S142 described above may include steps S1421 to S1425.

[0053] S1421. Update the sparse representation of the processed problem at the current moment, and combine time-varying information and historical data to achieve information fusion; S1422. Update the representation of the foam interference part of the processed problem; S1423, Smoothing part of the updated problem process after processing; S1424. Update the interference mutation part of the processed problem to capture sudden changes or abnormal situations; S1425. Update the dual variables of the processed problem so that the processed problem converges to the global optimal solution.

[0054] In this embodiment, the above joint optimization problem is solved using the Alternating Direction Multiplier Method (ADMM), which can be used to solve the original mixed signal. Decomposed into two parts: pure solidification signal component With foam interference signal components The specific steps are as follows: Introducing auxiliary variables and Optimization problem Equivalent to: ; Introducing dual variables and Constructing the augmented Lagrange function : ;in It is a penalty parameter.

[0055] ADMM iterative solution, at the... In each iteration, variables are updated alternately. and dual variables :renew , This step integrates the sparse representation of the current moment, time-varying information, and historical information. Next, the update... , ;renew (Process smoothing section) ;renew (Interference mutation part) Update dual variables .

[0056] To ensure real-time performance in industrial settings, this embodiment employs the following strategy: The previous moment convergent solution As of the present moment The initial values ​​used for solving can significantly reduce the number of iterations.

[0057] Parameter adaptive adjustment penalty parameter The choice of has a significant impact on the convergence speed, and an adaptive adjustment strategy can be adopted to accelerate convergence.

[0058] S150. Decompose the distance image sequence into a pure solidification signal component and a foam interference signal component. Based on the pure solidification signal component, identify different stages of solidification and quantify the solidification endpoint. Monitor and alarm the foam interference energy.

[0059] In one embodiment, step S150 described above may include steps S151 to S152.

[0060] S151. Decompose the distance image sequence into a pure coagulation signal component and a foam interference signal component. Based on the pure coagulation signal component, analyze its changes over time to identify different stages of soy milk coagulation. When the atoms representing the stable state at the end of coagulation become the main components and the energy ratio exceeds a preset threshold, and the state change amount is continuously lower than the stable threshold, determine that the coagulation process has reached the end of coagulation.

[0061] In this embodiment, based on the isolated, foam-independent coagulation process sparsity coefficient... This greatly improves the accuracy and robustness of the judgment. Analysis The zero coefficient of China-Africa relations over time "relay" process identifies each stage from non-solidification, aggregation, network formation to stability.

[0062] Endpoint quantification includes: Condition A (mode-dominated): Represents the "solidification endpoint steady state". Atoms in When the energy proportion of the corresponding coefficient exceeds a preset threshold, the energy becomes dominant. .

[0063] Condition B (State Stability): Defines state changes based on purity coefficients. .when For a continuous period of time, the value is less than the stable threshold. At that point, it was determined that the macroscopic characteristics of the solidification process had ceased to change.

[0064] When conditions A and B are both met, the optimal solidification endpoint is determined to have been reached.

[0065] S152. Monitor and alarm on the energy of foam interference.

[0066] Specifically, the severity and changes of foam are reflected by real-time monitoring and plotting the foam interference energy curve. At the same time, an alarm is triggered when the energy is continuously higher than the danger threshold or the rate of change is abnormal, prompting the operator to check the raw materials or adjust the process parameters.

[0067] Define foam interference energy To achieve: real-time drawing on the human-computer interaction interface The curve visually reflects the severity and dynamic changes of the bubble. If... Continuously exceeding the preset danger threshold or its rate of change If an anomaly is detected, an alarm is triggered, prompting the operator to check the raw materials (such as the defoaming of soy milk) or process parameters (such as stirring speed). This provides crucial data for achieving closed-loop control of the production process.

[0068] S160. When the slurry reaches the solidification endpoint, the subsequent process is automatically triggered, and the foam interference coefficient energy is monitored in real time.

[0069] In this embodiment, a signal is output when the solidification endpoint is reached to control subsequent processes. The energy value of the foam interference coefficient is monitored in real time, and an alarm is triggered when it is too high for an extended period of time to prompt the operator to check or adjust. At the same time, the evolution curve of the sparsity coefficient during the process is stored in the database as a digital fingerprint for quality traceability.

[0070] Once the solidification endpoint is determined, the system outputs a signal to control subsequent processes (such as cutting and pressing). Real-time monitoring of the foam interference coefficient is also implemented. energy If this value remains too high for an extended period, an alarm may be triggered, prompting the operator to check the raw materials or process parameters. The sparsity coefficient evolution curve for the entire process will be displayed. The data is stored in the database as a "digital fingerprint" of this batch of products for quality analysis and traceability.

[0071] This embodiment of the method is the first to model foam interference as part of a sparse representation problem. By constructing a joint dictionary and applying morphological priors for joint sparsity optimization, it achieves "blind source separation" of the real solidification process signal and foam interference signal from mixed radar signals. This method not only fundamentally solves the foam interference problem, but also provides a more accurate and effective solution compared to simple filtering or thresholding.

[0072] Since the state judgment is based on the pure solidification signal coefficient after separation The method in this embodiment is highly robust to dynamically changing foams that are unavoidable in actual production. Regardless of changes in foam thickness and density, the system can adaptively remove their effects, thereby ensuring reliability and stability under various complex operating conditions.

[0073] In addition to accurately determining the solidification endpoint, the method in this embodiment also provides additional insights into the production process through quantitative monitoring of the foam interference coefficient. For example, abnormal foaming conditions can be detected in a timely manner, which not only helps in process optimization but also provides new data support for quality control, enhancing the controllability and transparency of the entire production process.

[0074] The method in this embodiment is based on the signal processing model of "signal = process component + interference component" and provides a rigorous mathematical solution using sparse representation and optimization theory. The physical meanings of atomic morphology and regularization terms are clearly defined, making the technical solution logically sound and highly interpretable. This design not only ensures the effectiveness of the method but also enhances its scientific basis and credibility.

[0075] The method in this embodiment retains all the advantages of the original solution, such as its reliance on dielectric property evolution, utilization of complete distance image information, and intelligent judgment. Furthermore, by addressing the critical bottleneck of foam interference, it further enhances the overall performance and practicality of the technical solution. This makes the solution not only more promising for industrial application but also establishes a stronger technological barrier, laying a solid foundation for technological innovation and development in related fields.

[0076] In summary, the method of this embodiment introduces an innovative approach to address the foam interference problem, which not only improves the accuracy of solidification process monitoring but also greatly enhances the robustness and adaptability of the system and provides more valuable information for production process monitoring, demonstrating its significant advantages in both technology and application.

[0077] The aforementioned millimeter-wave radar solidification state monitoring method for resisting foam interference, by creating a joint atomic dictionary and applying the alternating direction multiplier method to solve a joint sparse optimization problem with morphological priors, accurately separates the pure solidification signal and foam interference signal from the mixed radar signal. This not only effectively solves the interference of foam on millimeter-wave radar monitoring and enhances the robustness and environmental adaptability of the system, but also provides rich process insights through quantitative monitoring of foam interference. Based on solid physical and mathematical model support, it significantly improves the automation level and quality control capability of tofu production.

[0078] Figure 2This is a schematic block diagram of a millimeter-wave radar slurry solidification state monitoring system 300 with anti-foam interference provided in an embodiment of the present invention. Figure 2 As shown, corresponding to the above-described millimeter-wave radar slurry solidification state monitoring method against foam interference, the present invention also provides a millimeter-wave radar slurry solidification state monitoring system 300 against foam interference. This millimeter-wave radar slurry solidification state monitoring system 300 against foam interference includes a unit for performing the above-described millimeter-wave radar slurry solidification state monitoring method against foam interference, and the system can be configured in a server. Specifically, please refer to... Figure 2 The anti-foaming interference millimeter-wave radar spot slurry solidification state monitoring system 300 includes a sequence generation unit 301, a dictionary creation unit 302, a first solution unit 303, a second solution unit 304, an identification unit 305, and a monitoring unit 306.

[0079] The sequence generation unit 301 is used to transmit a linear frequency modulated signal using a millimeter-wave radar installed above the coagulation tank and process the received intermediate frequency signal to generate a range image sequence; the dictionary creation unit 302 is used to create a joint atom dictionary that can simultaneously characterize the coagulation process and foam interference, wherein each atom in the dictionary is a vector with the same length as the range image sequence; the first solution unit 303 is used to solve a joint sparse optimization problem based on the joint atom dictionary and with morphological priors for the range image sequence; the second solution unit 304 is used to solve the joint sparse optimization problem using the alternating direction multiplier method; the identification unit 305 is used to decompose the range image sequence into a pure coagulation signal component and a foam interference signal component, identify different stages of coagulation based on the pure coagulation signal component, quantify and determine the coagulation endpoint, and monitor and alarm the foam interference energy; the monitoring unit 306 is used to automatically trigger subsequent processes when the slurry reaches the coagulation endpoint and monitor the foam interference coefficient energy in real time.

[0080] In one embodiment, the sequence generation unit 301 is used to transmit a linear frequency modulated Chirp signal using an FMCW millimeter-wave radar installed above the coagulation tank, and process the received intermediate frequency signal through FFT to generate a range image sequence containing information on the liquid surface, foam, the interior of the soy milk, and the reflection from the bottom of the tank.

[0081] In one embodiment, the first solving unit 303 is used to introduce auxiliary variables to transform the joint sparse optimization problem into a more easily solvable form, introduce dual variables, and construct an augmented Lagrangian function to obtain the processed problem; and use the ADMM method to iteratively solve the processed problem.

[0082] In one embodiment, the first solving unit 303 is used to update the sparse representation of the processed problem at the current time, and combine time-varying information and historical data to achieve information fusion; update the representation of the bubble interference part of the processed problem; update the process smoothing part of the processed problem; update the interference mutation part of the processed problem to capture sudden changes or abnormal situations; and update the dual variable of the processed problem so that the processed problem converges to the global optimal solution.

[0083] In one embodiment, the identification unit 305 is used to decompose the distance image sequence into a pure coagulation signal component and a foam interference signal component. Based on the pure coagulation signal component, it analyzes its changes over time to identify different stages of soy milk coagulation. When the atoms representing the stable state at the end of coagulation become the main components and the energy proportion exceeds a preset threshold, and the state change amount is continuously lower than the stable threshold, it is determined that the coagulation process has reached the end of coagulation. The foam interference energy is monitored and alarmed.

[0084] In one embodiment, the identification unit 305 is used to reflect the severity and changes of foam by real-time monitoring and plotting the foam interference energy curve, and to trigger an alarm when the energy is continuously higher than the danger threshold or the rate of change is abnormal, prompting the operator to check the raw materials or adjust the process parameters.

[0085] In one embodiment, the monitoring unit 306 is used to output a signal to control subsequent processes when the solidification endpoint is determined, and to trigger an alarm to prompt the operator to check or adjust when the energy value of the foam interference coefficient is too high for a long time by real-time monitoring. At the same time, the evolution curve of the sparsity coefficient in the process is stored in the database as a digital fingerprint for quality traceability.

[0086] It should be noted that those skilled in the art can clearly understand that the specific implementation process of the above-mentioned anti-foam interference millimeter-wave radar spot slurry solidification state monitoring system 300 and each unit can be referred to the corresponding description in the foregoing method embodiments. For the sake of convenience and brevity, it will not be repeated here.

[0087] The aforementioned millimeter-wave radar solidification state monitoring system 300, which is resistant to foam interference, can be implemented as a computer program. This computer program can be used in various applications such as... Figure 3 It runs on the computer device shown.

[0088] Please see Figure 3 , Figure 3 This is a schematic block diagram of a computer device provided in an embodiment of this application. The computer device 500 can be a server, wherein the server can be a standalone server or a server cluster composed of multiple servers.

[0089] See Figure 3 The computer device 500 includes a processor 502, a memory, and a network interface 505 connected via a system bus 501. The memory may include a non-volatile storage medium 503 and internal memory 504.

[0090] The non-volatile storage medium 503 may store an operating system 5031 and a computer program 5032. The computer program 5032 includes program instructions that, when executed, cause the processor 502 to perform a millimeter-wave radar solidification state monitoring method resistant to foam interference.

[0091] The processor 502 provides computing and control capabilities to support the operation of the entire computer device 500.

[0092] The internal memory 504 provides an environment for the operation of the computer program 5032 in the non-volatile storage medium 503. When the computer program 5032 is executed by the processor 502, the processor 502 can execute a millimeter-wave radar solidification state monitoring method that is resistant to foam interference.

[0093] This network interface 505 is used for network communication with other devices. Those skilled in the art will understand that... Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device 500 to which the present application is applied. The specific computer device 500 may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0094] The processor 502 is used to run the computer program 5032 stored in the memory to implement all the steps of the millimeter-wave radar spot slurry solidification state monitoring method for resisting foam interference.

[0095] It should be understood that in the embodiments of this application, the processor 502 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0096] It will be understood by those skilled in the art that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program includes program instructions and can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiments of the above methods.

[0097] Therefore, the present invention also provides a storage medium. This storage medium can be a computer-readable storage medium. The storage medium stores a computer program, wherein when executed by a processor, the computer program causes the processor to perform all steps of the millimeter-wave radar solidification state monitoring method for resisting foam interference.

[0098] The storage medium can be any computer-readable storage medium capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory (ROM), magnetic disk, or optical disk.

[0099] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0100] In the embodiments provided by this invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the division of each unit is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0101] The steps in the method of this invention can be adjusted, merged, or reduced in order according to actual needs. The units in the system of this invention can be merged, divided, or reduced according to actual needs. Furthermore, the functional units in the various embodiments of this invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0102] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part 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 to cause a computer device (which may be a personal computer, a terminal, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.

[0103] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method of millimeter wave radar point cloud coagulation state monitoring against foam interference, characterized by, The method comprises the following steps: A millimeter wave radar is installed above the coagulation tank to emit a linear frequency modulation signal, and the received intermediate frequency signal is processed to generate a range image sequence; A joint atom dictionary is created to represent the coagulation process and foam interference at the same time; A joint sparse optimization problem based on the joint atom dictionary and with a morphology prior is solved for the range image sequence; An alternating direction multiplier method is applied to solve the joint sparse optimization problem; The range image sequence is decomposed into a pure coagulation signal component and a foam interference signal component, different stages of coagulation are identified based on the pure coagulation signal component, the coagulation endpoint is quantitatively judged, and the foam interference energy is monitored and alarmed; When the point slurry reaches the coagulation endpoint, the subsequent process is automatically triggered, and the foam interference coefficient energy is monitored in real time.

2. The anti-foam interference millimeter wave radar point coagulation state monitoring method according to claim 1, characterized by, The method comprises the following steps: A FMCW millimeter wave radar is installed above the coagulation tank to emit a linear frequency modulation Chirp signal, and the received intermediate frequency signal is processed by FFT to generate a range image sequence containing liquid surface, foam, soybean milk internal and tank bottom reflection information.

3. The anti-foam interference millimeter wave radar point coagulation state monitoring method according to claim 1, characterized by, The joint atom dictionary comprises a coagulation process sub-dictionary and a foam interference sub-dictionary, wherein the coagulation process sub-dictionary is composed of a plurality of atoms representing typical radar response patterns of soybean milk at different coagulation stages; the foam interference sub-dictionary contains atoms representing different radar echo patterns caused by foam interference; In the offline stage, two sub-dictionaries capable of clearly distinguishing between coagulation process and foam interference are obtained by collecting range image sequences under different working conditions and applying dictionary learning algorithm to training, to form a joint atom dictionary.

4. The anti-foam interference millimeter wave radar point coagulation state monitoring method according to claim 1, characterized by, The joint sparse optimization problem is represented as where, is the joint sparse coefficient vector to be solved; is the L1 norm; is the introduced morphological prior regularization term, the process smoothness term penalizes the drastic temporal variation of the solidification process coefficients forcing their solution to be smoothly continuous in time, in accordance with the physical fact that the solidification process varies slowly. The disturbance jump term encourages the foam disturbance coefficients to exhibit a jump characteristic; is a function that promotes temporal sparsity; is the joint atom dictionary.

5. The anti-foam interference millimeter wave radar point coagulation state monitoring method according to claim 4, characterized by, The method comprises the following steps: An auxiliary variable is introduced to convert the joint sparse optimization problem into a more easily solvable form, a dual variable is introduced, and a augmented Lagrangian function is constructed to obtain a processed problem; The ADMM method is used to iteratively solve the processed problem.

6. The anti-foam interference millimeter wave radar point coagulation state monitoring method according to claim 5, characterized by, The method comprises the following steps: The sparse representation of the processed problem at the current time is updated, and information fusion is realized by combining time-varying information and historical data; The representation of the foam interference part of the processed problem is updated; The process smoothing part of the processed problem is updated; The interference mutation part of the processed problem is updated to capture sudden changes or abnormal situations; The dual variable of the processed problem is updated to make the processed problem converge to a global optimal solution.

7. The anti-foam interference millimeter wave radar point coagulation state monitoring method of claim 1, wherein, The method comprises the following steps: The distance image sequence is decomposed into a pure coagulation signal component and a foam interference signal component, different stages of soybean milk coagulation are identified by analyzing the change of the pure coagulation signal component over time, and when the energy ratio of the atom representing the coagulation end point stable state becomes the main component and exceeds the preset threshold, and the state change quantity is continuously lower than the stable threshold, it is determined that the coagulation process reaches the coagulation end point; The foam interference energy is monitored and alarmed.

8. The anti-foam interference millimeter wave radar point coagulation state monitoring method according to claim 7, characterized by, The monitoring and alarming of the foam interference energy includes: By monitoring and drawing the foam interference energy curve in real time, the severity and change of the foam are reflected, and when the energy is continuously higher than the danger threshold or the change rate is abnormal, an alarm is triggered to prompt the operator to check the raw materials or adjust the process parameters.

9. The anti-foam interference millimeter wave radar point coagulation state monitoring method of claim 1, wherein, After the soybean milk reaches the coagulation end point, the subsequent process is automatically triggered, and the foam interference coefficient energy is monitored in real time, including: When the coagulation end point is determined, an output signal is output to control the subsequent process, and by monitoring the energy value of the foam interference coefficient in real time, an alarm is triggered when the energy value is abnormally high for a long time to prompt the operator to check or adjust, and the sparse coefficient evolution curve in the process is stored as a digital fingerprint in the database for quality traceability.

10. A millimeter wave radar point cloud coagulation state monitoring system against foam interference, characterized by, It includes: A sequence generation unit for transmitting a linear frequency modulation signal using a millimeter wave radar installed above the coagulation tank and processing the received intermediate frequency signal to generate a distance image sequence; A dictionary creation unit for creating a joint atom dictionary capable of representing both the coagulation process and the foam interference, wherein each atom in the dictionary is a vector with the same length as the distance image sequence; A first solving unit for solving a joint sparse optimization problem based on the joint atom dictionary with a morphology prior for the distance image sequence; A second solving unit for solving the joint sparse optimization problem using an alternating direction multiplier method; An identification unit for decomposing the distance image sequence into a pure coagulation signal component and a foam interference signal component, identifying different stages of coagulation based on the pure coagulation signal component and quantitatively determining the coagulation end point, and monitoring and alarming the foam interference energy; A monitoring unit for automatically triggering the subsequent process when the soybean milk reaches the coagulation end point, and monitoring the foam interference coefficient energy in real time.

Images