A control method and system for soft candy production equipment in food processing

Abstract

This application provides a control method for a gummy candy production equipment used in food processing. Applied to the field of food processing technology, it directly acquires real-time spectral data of the material within the processing container through a spectral detection unit. Based on the real-time spectral data, real-time characteristic indicators representing the material's activation state are obtained. Then, the deviation between the real-time characteristic indicators and preset target characteristic indicators is calculated, accurately reflecting the degree of deviation between the material's activation state and the ideal state. Finally, a final execution compensation value is determined based on the deviation value and superimposed on the basic heating control command to adjust the output power of the heating components. This achieves closed-loop feedback control of the material's activation state, ensuring that the novel gelling agent achieves a precise activation level during the cooking process, thus guaranteeing the consistency of the gummy candy product in terms of texture, elasticity, and shaping effect.

Description

Technical Field

[0001] This application relates to the field of food processing technology, and in particular to a control method and system for a soft candy production equipment for food processing. Background Technology

[0002] In the food processing industry, gummy candy production demands extremely precise control of process parameters. Temperature control during the cooking stage is crucial in determining the final product quality. Traditional gummy candy production lines typically rely on high-precision temperature sensors to monitor the real-time temperature of the syrup in the cooking pot, converting the collected analog signals into digital signals, which are then transmitted to the central control unit of the equipment. The control unit adjusts the output power of the heating device in real time based on preset process formula parameters and advanced control strategies. The core objective of this closed-loop feedback control mechanism is to ensure that the actual temperature of the syrup is stably and accurately maintained within the range required by the process flow. This provides qualified semi-finished syrup for subsequent stages such as concentration, pouring, cooling, and drying, ensuring the consistency of the final gummy candy product in terms of shape, taste, and stability.

[0003] However, in long-term, continuous food production operations, especially when handling materials with high sugar content, high viscosity, and operating at high temperatures, irregularly distributed sugar deposits easily form on the surface of the sensor probe. From a thermophysical perspective, this sugar crystal layer on the sensor probe surface is equivalent to introducing an additional thermal resistance between the sensor's temperature-sensing element and the measured syrup. Therefore, the temperature measurement value transmitted by the sensor to the control unit can no longer completely reflect the true temperature of the syrup in real time, but rather carries a certain time lag and numerical deviation.

[0004] During the production of gummies, a gelling agent is added to the syrup. To fully activate the gelling potential of this agent and form an ideal elastic structure, the syrup temperature must be precisely and rapidly increased from a lower stable temperature to a higher peak temperature within a very short time. Furthermore, the duration of this peak temperature must be strictly controlled within a very narrow window. If the temperature rise is too slow, causing the syrup to remain at the peak temperature for too long, or if the duration at the peak temperature is too long or too short, the molecular structure of the gelling agent will not cross-link as expected. This will severely affect the texture, elasticity, transparency, and shaping effect of the gummies, potentially resulting in a series of quality defects such as a hard texture, stickiness, excessive softness, brittleness, or a rough surface.

[0005] Therefore, in the cooking process of gummy candy production equipment for food processing, the gradual measurement delay and deviation caused by the slow accumulation of sugar crystals on the surface of the temperature sensor probe, and the fact that this deviation fails to trigger conventional fault alarms, present a complex working condition. How to achieve high-precision, real-time, and self-adjustable temperature curve control for the syrup cooking process using a new type of gelling agent, so as to ensure that the texture and shaping effect of the gummy candy product remain highly consistent during the extremely short peak temperature maintenance time, has become a pressing technical problem to be solved. Summary of the Invention

[0006] In view of the shortcomings of the prior art, this application provides a control method and system for a gummy candy production equipment for food processing, which overcomes the measurement error and response lag caused by sugar residue accumulation in traditional temperature sensors, and realizes direct perception and precise control of the activation state of materials, thereby ensuring the consistency of gummy candy product quality.

[0007] A first aspect includes a method for controlling a gummy candy production equipment for food processing, the equipment comprising a processing container, a heating assembly, and a spectral detection unit, the method comprising the steps of:

[0008] S1: Obtain real-time spectral data of the material inside the processing container through the spectral detection unit;

[0009] S2: Obtain real-time characteristic indicators representing the activation state of the material based on the real-time spectral data analysis; wherein, the analysis includes: calling a preset activation state mapping model, which represents the correspondence between spectral characteristic data and the degree of thermal activation of the gelling agent, obtaining the current value of the activation degree of the gelling agent molecules in the material, and determining the value of the activation degree of the gelling agent molecules as the real-time characteristic indicator; obtaining the real-time temperature data of the material in the processing container, and generating basic heating control instructions based on the real-time temperature data to drive the heating component to operate according to the preset process temperature;

[0010] S3: Calculate the deviation between the real-time feature index and the preset target feature index;

[0011] S4: Calculate the temperature compensation value or the maintenance time compensation value based on the deviation value, calculate the temperature change rate of the real-time temperature data within the preset sliding time window and the index change rate of the real-time characteristic index within the preset sliding time window, call the preset physical consistency logic table, determine the feedback gain coefficient based on the matching relationship between the temperature change rate and the index change rate, and use the feedback gain coefficient to weight the temperature compensation value or the maintenance time compensation value to obtain the final execution compensation value, and add the final execution compensation value to the basic heating control command to adjust the output power of the heating component.

[0012] Furthermore, step S1 includes:

[0013] S11: Control the spectral detection unit to emit a light signal, so that the light signal penetrates the material inside the processing container;

[0014] S12: Receive the transmitted light signal after penetrating the material;

[0015] S13: Convert the transmitted light signal into an electrical signal, and perform noise reduction processing on the electrical signal to generate the real-time spectral data.

[0016] Furthermore, in step S2, the parsing also includes:

[0017] S21: Extract the characteristic band data from the real-time spectral data, input the characteristic band data into the activation state mapping model, and calculate the activation degree of the gelling agent molecules of the current material.

[0018] Furthermore, in step S2, invoking the preset activation state mapping model includes the following steps:

[0019] S221: Prepare several groups of calibration samples with different activation levels under controlled temperature and time conditions;

[0020] S222: Perform physical performance tests on each group of calibration samples to quantify the corresponding actual activation degree value;

[0021] S223: Collect the sample spectral data for each set of calibration samples;

[0022] S224: Analyze the correlation between the sample spectral data and the actual activation degree value, and select the absorbance or absorption peak intensity ratio at a specific wavelength that is associated with the actual activation degree value as the model input feature;

[0023] S225: Fit and establish the functional relationship between the model input features and the actual activation level values, generate the activation state mapping model, and call the activation state mapping model.

[0024] Furthermore, step S3 includes the following:

[0025] S41: Determine the activation state trend represented by the deviation value;

[0026] S42: When the deviation value represents that the material activation state is lagging behind the preset ideal activation state, calculate the positive power compensation value according to the magnitude of the deviation value, and control the heating component to increase the positive power compensation value based on the current power.

[0027] S43: When the deviation value represents that the material activation state is ahead of the preset ideal activation state, calculate the negative power correction value according to the magnitude of the deviation value, and control the heating component to reduce the negative power correction value based on the current power.

[0028] Furthermore, the temperature compensation value is a compensation value for the preset process temperature, and the holding time compensation value is a holding time compensation value for the heating duration.

[0029] Furthermore, the physical consistency logic table is used to increase the feedback gain coefficient when the temperature change rate matches the index change rate, and to decrease the feedback gain coefficient when the temperature change rate does not match the index change rate, so as to attenuate or shield the influence of the temperature compensation value or the holding time compensation value on the heating component.

[0030] Furthermore, in step S4, determining the feedback gain coefficient includes the following steps:

[0031] S44: When the rate of temperature change is positive and greater than the preset heating threshold, determine whether the rate of change of the index is within the preset reasonable growth range.

[0032] Further, in step S4, the temperature compensation value or the duration compensation value is weighted using the feedback gain coefficient to obtain the final execution compensation value, and the final execution compensation value is superimposed on the basic heating control command to adjust the output power of the heating component, including:

[0033] S45: If the rate of change of the index is within the reasonable growth range, the spectral data is determined to be valid, and the feedback gain coefficient is set to a first value. The first value is used to maintain or amplify the influence of the temperature compensation value.

[0034] S46: If the rate of change of the index exceeds the reasonable growth range or shows negative growth, it is determined that the spectral data is subject to physical interference. The feedback gain coefficient is set to a second value, which is less than the first value, to attenuate or shield the influence of the temperature compensation value, so that the heating component mainly operates according to the basic heating control command.

[0035] Secondly, a control system for a soft candy production equipment for food processing, the system comprising:

[0036] Acquisition module: Acquires real-time spectral data of the material inside the processing container through the spectral detection unit;

[0037] The analysis module analyzes the real-time spectral data to obtain real-time characteristic indicators representing the activation state of the material. The analysis includes: calling a preset activation state mapping model, which represents the correspondence between spectral characteristic data and the thermal activation degree of the gelling agent, to obtain the activation degree value of the gelling agent molecules in the current material, and determining the activation degree value of the gelling agent molecules as the real-time characteristic indicator; acquiring real-time temperature data of the material in the processing container, and generating basic heating control commands based on the real-time temperature data to drive the heating components to operate according to the preset process temperature.

[0038] Calculation module: Calculates the deviation between the real-time feature index and the preset target feature index;

[0039] Adjustment module: Calculates temperature compensation value or maintenance time compensation value based on the deviation value, calculates the temperature change rate of the real-time temperature data within a preset sliding time window and the index change rate of the real-time characteristic index within the preset sliding time window, calls a preset physical consistency logic table, determines the feedback gain coefficient based on the matching relationship between the temperature change rate and the index change rate, and uses the feedback gain coefficient to weight the temperature compensation value or the maintenance time compensation value to obtain the final execution compensation value, and superimposes the final execution compensation value into the basic heating control command to adjust the output power of the heating component.

[0040] Beneficial Effects: The control method and system for producing soft candy in this application directly acquires real-time spectral data of the material inside the processing container through a spectral detection unit, avoiding the measurement errors and response lag problems caused by sugar residue accumulation in traditional temperature sensors. Next, real-time characteristic indicators representing the activation state of the material are obtained based on the real-time spectral data, achieving direct and quantitative perception of the activation degree of the gelling agent molecules, rather than indirect temperature inference. Then, the deviation value between the real-time characteristic indicator and the preset target characteristic indicator is calculated, accurately reflecting the degree of deviation between the material's activation state and the ideal state. Finally, the final execution compensation value is determined based on the deviation value and superimposed on the basic heating control command to adjust the output power of the heating component, thereby achieving closed-loop feedback control of the material's activation state. This ensures that the novel gelling agent achieves a precise activation degree during the cooking process, guaranteeing the consistency of the soft candy product in terms of texture, elasticity, and shaping effect. Attached Figure Description

[0041] Figure 1 This is a flowchart of a control method for a soft candy production equipment for food processing proposed in this application.

[0042] Figure 2This is a structural diagram of a control system for a soft candy production equipment proposed in this application.

[0043] Figure 3 This is a schematic diagram of a control system for a soft candy production equipment proposed in this application.

[0044] Attached reference numerals: 201, Acquisition module; 202, Parsing module; 203, Calculation module; 204, Adjustment module. Detailed Implementation

[0045] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. The components of the embodiments of this application described and marked in the accompanying drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0046] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0047] Please refer to Figure 1 This application proposes a control method for a gummy candy production equipment used in food processing. The gummy candy production equipment includes a processing container, a heating component, and a spectral detection unit. The method includes the following steps:

[0048] S1: Obtain real-time spectral data of the material inside the processing container through the spectral detection unit;

[0049] S2: Obtain real-time characteristic indicators representing the activation state of materials based on real-time spectral data analysis; wherein, analysis includes: calling a preset activation state mapping model, which represents the correspondence between spectral characteristic data and the degree of thermal activation of the gelling agent, obtaining the activation degree value of the gelling agent molecules of the current material, and determining the activation degree value of the gelling agent molecules as the real-time characteristic indicator; obtaining real-time temperature data of the material in the processing container, and generating basic heating control instructions based on the real-time temperature data to drive the heating components to operate according to the preset process temperature;

[0050] S3: Calculate the deviation between the real-time feature indicators and the preset target feature indicators;

[0051] S4: Calculate the temperature compensation value or the maintenance time compensation value based on the deviation value, calculate the temperature change rate of real-time temperature data within the preset sliding time window and the index change rate of real-time characteristic indicators within the preset sliding time window, call the preset physical consistency logic table, determine the feedback gain coefficient based on the matching relationship between the temperature change rate and the index change rate, and use the feedback gain coefficient to weight the temperature compensation value or the maintenance time compensation value to obtain the final execution compensation value, and add the final execution compensation value to the basic heating control command to adjust the output power of the heating component.

[0052] In the cooking process of gummy candy production, traditional control methods heavily rely on temperature sensors. However, during long-term production, especially when handling high-sugar, high-viscosity materials, a layer of sugar crystals inevitably forms on the surface of the temperature sensor probe. This crystallized layer has a significant thermal resistance effect, acting like an insulating layer for the sensor, causing a noticeable delay and deviation in the sensor's perception of the material's true temperature. This deviation accumulates gradually at a very slow rate, making it difficult for conventional equipment fault alarm systems to detect effectively. When new gelling agents with extremely high requirements for temperature profiles are introduced into the production process, such as those requiring rapid heating and precise maintenance of peak temperatures within a very short time, control based on lagging and inaccurate temperature data cannot guarantee proper thermal activation of the gelling agent molecules. This leads to severe batch-to-batch variations and inconsistent quality in the final product's texture, elasticity, and molding properties.

[0053] To fundamentally solve this technical challenge, the method proposed in this application no longer relies on indirect and easily interfered-prone temperature measurements, but instead introduces a novel control dimension that directly observes the internal molecular state of materials. Its core working principle lies in using spectroscopic detection technology to monitor the molecular-level changes of materials within the processing container in real time and non-contactly. The activation of gelling agents during heating is essentially a complex physicochemical process involving changes in the conformation of their macromolecular chains, the exposure of functional groups, and alterations in their interaction with water molecules. These changes in microstructure are directly reflected in the absorption characteristics of the material for specific wavelengths of light. Therefore, by analyzing the spectral data of the material, the degree of gelling agent activation can be directly quantified, thus bypassing the measurement uncertainties caused by scaling of temperature sensors.

[0054] Specifically, the entire control process constitutes a sophisticated closed-loop feedback system. First, the spectral detection unit deployed on the gummy candy production equipment periodically collects real-time spectral data of the material flowing through or contained within the processing container. This raw spectral data is a series of information reflecting the absorption intensity of light at different wavelengths. Subsequently, a core analytical step in the control method processes this complex spectral data, extracting the spectral features most relevant to the gelling agent's activation state, and converting it into a single, physically meaningful value—the real-time characteristic index—based on a pre-established mathematical relationship. This index can be understood as a quantitative score of the activation degree of gelling agent molecules in the current material. Specifically, since the raw real-time spectral data contains information from hundreds or thousands of wavelengths, much of this information may be irrelevant to the gelling agent's activation state, or even redundant. Directly using full-spectrum data for analysis is not only computationally intensive but also easily affected by irrelevant factors. Therefore, converting the spectral data into a concise and accurate activation state index requires an intelligent analytical process.

[0055] The analysis process invokes a pre-defined activation state mapping model. This model is the core of the entire analysis process, serving as a bridge connecting spectral language and physical states. Essentially, this model is a mathematical function or algorithm that internally establishes the quantitative relationship between spectral feature data and the actual thermal activation degree of the gelling agent. This model was obtained through extensive offline calibration experiments and machine learning training. The activation state mapping model characterizes the correspondence between spectral feature data and the thermal activation degree of the gelling agent, ensuring the accuracy of the analysis.

[0056] The extracted characteristic wavelength data is then input into this activation state mapping model for calculation. For example, absorbance values ​​at 1450 nm and 1680 nm are used as input variables and substituted into the model's function. The model outputs a single numerical value based on its internal weighting coefficients and computational logic. This output value represents the activation level of the gelling agent molecules in the current material. For example, the model might output a value between 0 and 100, where zero represents that the gelling agent is completely unactivated, and 100 represents reaching the ideal optimal activation state. This value is directly determined as a real-time characteristic indicator for subsequent closed-loop control. In this way, complex spectral information is successfully decoded into an intuitive and quantifiable process control parameter.

[0057] Furthermore, to construct a more robust and reliable control system, this application proposes a hybrid control strategy that combines innovative spectral activation state control with traditional temperature control. In this strategy, traditional temperature control plays the role of providing a basic thermal platform and macroscopic process monitoring.

[0058] Specifically, even with the introduction of spectral detection, the existing temperature sensors on the gummy candy production equipment, such as platinum resistance thermometers or thermocouples, are retained and continue to operate. The control system synchronously acquires real-time temperature data of the materials within the processing container. Although this data may have the aforementioned delays and biases, it still provides a macroscopic, rough indication of the overall thermodynamic state of the materials, especially during the start-up and stabilization phases of the process, where temperature information remains of significant reference value.

[0059] Based on this real-time temperature data, the control system operates a basic temperature control loop. This is typically a standard feedforward-feedback control algorithm, such as proportional-integral-derivative (PID) control. The algorithm calculates the temperature deviation based on a preset process temperature curve, for example, requiring a linear temperature increase from 80 degrees Celsius to 120 degrees Celsius within five minutes and maintaining this temperature for two minutes. The control algorithm compares the real-time temperature measurement with the target temperature value at the current moment, calculates the temperature deviation, and generates a basic heating control command. This command directly drives the heating components to operate according to the preset process temperature target. The purpose of this basic control loop is to establish and maintain a roughly correct production thermal environment, ensuring that the material temperature macroscopically conforms to process requirements, providing a stable and reliable operating platform for subsequent spectral-based fine-tuning.

[0060] Next, the control method compares this real-time acquired characteristic indicator with a preset target characteristic indicator. This target characteristic indicator represents the ideal activation state that the material should achieve at the current process stage, and is an optimal value preset based on a large amount of experimental data and product quality requirements. The difference between the two, i.e., the deviation value, accurately reflects the gap between the current production process and the ideal state. This deviation value is the core signal driving the entire feedback control system to make adjustments.

[0061] Finally, the control method can dynamically calculate the required power adjustment for the heating component based on the sign and magnitude of this deviation value. If the deviation value is negative, it indicates insufficient material activation, requiring increased heat input, and a positive power adjustment command will be generated. Conversely, if the deviation value is positive, it indicates over-activation of the material, requiring reduced heat input, and a negative power adjustment command will be generated. This adjustment command directly acts on the heating component, such as adjusting the opening of the steam valve or the power supply current of the electric heating element, thereby precisely adjusting the heat applied to the material and bringing the material's activation state back to the preset target trajectory in real time.

[0062] Alongside this approach, to further enhance the intelligence level of the control system and its robustness in handling complex operating conditions, this application also introduces a self-verification and adaptive adjustment mechanism to adjust the output power of the heating component. The core idea of ​​this mechanism is to perform a rapid, physical-law-based cross-verification of the reliability of the spectral data before executing correction commands derived from the spectral data. The problem it aims to solve is that if the spectral signal experiences abnormal fluctuations due to accidental factors, such as a large number of bubbles instantaneously generated in the material or the optical window being momentarily blocked by contaminants, the system should be able to identify such anomalies and avoid making incorrect control decisions based on erroneous information.

[0063] To achieve this mechanism, once the spectral detection system calculates the deviation value of the activated state, the control logic does not simply adjust the heating power directly, but transforms this deviation, which represents the molecular-level state, into a correction of macroscopic process parameters, namely temperature or time.

[0064] In one embodiment, temperature compensation is employed. If the calculated activation state deviation is negative, for example, lagging by 10%, the control system infers, based on a preset correlation model, that this is likely because the actual temperature of the material is lower than the temperature sensor reading. Therefore, a positive temperature compensation value, such as 0.5 degrees Celsius, is calculated. This compensation value is then superimposed on the target setpoint of the base temperature control loop. If the original target temperature was 120 degrees Celsius, the corrected new target temperature becomes 120.5 degrees Celsius. The base temperature control loop drives the heating components based on this new, higher target, thereby applying more heat to compensate for the insufficient heating caused by the sensor deviation.

[0065] In another embodiment, particularly in process stages requiring precise control of holding time, a holding time compensation approach can be employed. For example, the process requires holding at the peak temperature for forty seconds. If, during the first twenty seconds of the holding phase, spectral detection reveals a persistent lag in the activation state, the system calculates a holding time compensation value, such as an additional five seconds. This compensation value is added to the total holding time plan, extending the actual holding time to forty-five seconds to ensure the gelling agent has sufficient time to complete full activation.

[0066] By converting spectral information into dynamic corrections to temperature or time setpoints and superimposing these correction values ​​into the basic heating control command, this application implements a dual-protection control mechanism. Basic temperature control ensures the macroscopic stability of the process, while spectral activation state control acts like a precise calibrator, correcting errors caused by the physical limitations of sensors in real time. Ultimately, this allows the corrected heating command to more accurately reflect the intrinsic needs of the material, achieving unprecedented control precision.

[0067] Then, the control system continuously calculates the rate of change of real-time temperature data within a short, preset sliding time window, such as the last five seconds, i.e., the speed at which the temperature rises or falls. At the same time, it also calculates the rate of change of real-time characteristic indicators within the same time window, i.e., the speed of evolution of the active state.

[0068] Then, a pre-defined physical consistency logic table is invoked to compare the two rates of change. This logic table contains built-in judgment rules based on fundamental physicochemical principles. A core rule is that during the heating stage, the increase in material temperature should be accompanied by an increase in the degree of gelling agent activation. Therefore, the temperature change rate and the index change rate should show a positive correlation. The logic table dynamically outputs a feedback gain coefficient based on the current matching relationship between the two. This coefficient can be understood as a quantitative assessment of the system's reliability of the current spectral data, and its value is typically between 0 and 1, or higher.

[0069] The previously calculated temperature compensation or duration compensation value is weighted using this feedback gain coefficient, resulting in a final execution compensation value equal to the original compensation value multiplied by the feedback gain coefficient. For example, if the temperature and activation index trends are highly consistent and meet physical expectations, the logic table might output a gain coefficient equal to 1. In this case, the final execution compensation value equals the original compensation value, and the spectral correction is fully adopted. However, if an abnormal situation occurs where the temperature rises rapidly while the activation index stagnates or even declines, the logic table will determine that the spectral data may be disturbed, and will output a very small gain coefficient, such as 0.1. Thus, after multiplying the original compensation value by 0.1, its impact on the final control command is greatly weakened.

[0070] Finally, this weighted, confidence-adjusted final execution compensation value is superimposed on the basic heating control command to adjust the output power of the heating components. Through this dynamically weighted adaptive mechanism, the control system can fully utilize the accuracy of spectral data when its reliability is trusted, and intelligently reduce its impact when its reliability is questioned, temporarily relying more on the relatively stable basic temperature control. This effectively prevents control instability caused by momentary sensor anomalies, greatly enhancing the stability and safety of the entire system.

[0071] Through the above control method, this application can effectively overcome the lag and deviation problems of traditional temperature control, and achieve unprecedented precise control of the gelling agent activation process. It is especially suitable for the production of new gelling agents with extremely demanding process window requirements, thereby significantly improving the quality consistency and production stability of gummy candy products.

[0072] Furthermore, in some embodiments, the step of acquiring real-time spectral data of the material inside the processing container specifically includes:

[0073] S11: Control the spectral detection unit to emit light signals so that the light signals penetrate the material inside the processing container;

[0074] S12: Receives the transmitted light signal after penetrating the material;

[0075] S13: Convert the transmitted light signal into an electrical signal, and perform noise reduction processing on the electrical signal to generate real-time spectral data.

[0076] To obtain spectral information that truly reflects the overall state of the material, rather than merely surface information, this application preferably employs transmission spectroscopy. Specifically, the spectral detection unit is configured as an active detection device. This unit includes a light source and a light signal receiver, respectively installed on both sides of the processing container or material flow pipeline, ensuring that there are windows through which light can pass.

[0077] In one specific embodiment, the spectral detection unit can be a near-infrared spectrometer. The step of controlling the spectral detection unit to emit a light signal specifically refers to controlling the broadband light source built into the spectrometer, such as a tungsten halogen lamp, to emit a light signal covering the near-infrared band, such as the range of 900 nanometers to 1700 nanometers. This light signal passes through an optical collimation system to form a parallel beam of light, which is perpendicularly incident on an optical window set on the wall of the processing container, for example, a window made of sapphire or high-strength quartz glass, a material with high light transmittance in the near-infrared band and resistance to high temperature and pressure. After passing through the window, the light beam enters and penetrates the syrup material of a certain thickness flowing inside the container.

[0078] During the process of light signals penetrating materials, various molecules in the material, such as water molecules, sugar molecules, and gelling agent macromolecules, selectively absorb light energy of specific wavelengths based on their own molecular vibrational characteristics. For example, the stretching vibration of the hydroxyl groups in water molecules produces a strong absorption peak at a specific wavelength. During the thermal activation process, changes such as the unfolding of molecular chains and the reconstruction of hydrogen bond networks in gelling agent molecules also lead to subtle but measurable changes in their absorption characteristics for near-infrared light.

[0079] On the other side of the processing container, directly opposite the light source, is a light signal receiver. This receiver is responsible for receiving the transmitted light signal after it has penetrated the material. The intensity of the light signal after penetrating the material attenuates at different wavelengths; the degree of attenuation records the absorption information of the material. The receiver can be an indium gallium arsenide array detector, capable of simultaneously and rapidly capturing spectral information across the entire near-infrared band.

[0080] The light signal captured by the receiver is an analog light intensity signal, which needs to be converted into a digital electrical signal before it can be understood by the subsequent processor. The process of converting the transmitted light signal into an electrical signal and then denoising it begins with an analog-to-digital converter (ADC) performing the digitization. However, in industrial production environments, factors such as electromagnetic interference, light source fluctuations, and detector thermal noise can introduce noise into the electrical signal. To ensure the purity and reliability of the spectral data, denoising is essential. Denoising can be implemented using digital signal processing algorithms. For example, the Savitzky-Gore smoothing filter can be applied, effectively filtering out high-frequency random noise by performing polynomial least-squares fitting on the data points within a moving window, while preserving the true shape and position information of the spectral peaks to the maximum extent. Alternatively, moving average filtering or wavelet transform can also be used for noise reduction. After this series of processing steps, the final generated data is a clean and accurate real-time spectral data frame, providing high-quality input for subsequent activation state analysis.

[0081] Furthermore, in some embodiments, the parsing step further includes:

[0082] S21: Extract the characteristic band data from the real-time spectral data, input the characteristic band data into the activation state mapping model, and calculate the activation degree of the gelling agent molecules in the current material.

[0083] This process occurs before invoking a pre-defined activation state mapping model. The purpose of this step is to sift through massive amounts of data to extract the most sensitive and representative information regarding changes in the activation state of the gelling agent. Preliminary experimental analysis reveals that during thermal activation, the vibrational modes of specific chemical bonds within the gelling agent molecules, such as hydroxyl, carbonyl, or amide bonds, change. These changes are concentrated at certain wavelengths or bands in the spectrum. Therefore, these characteristic bands can be pre-determined. For example, analysis shows that the absorption peak intensity near 1450 nm is highly correlated with the degree of hydration of the gelling agent molecular chains, while another absorption peak at 1680 nm is related to the breaking and recombination of intramolecular hydrogen bonds. In real-time processing, the analysis program precisely extracts the absorbance values ​​of these two wavelengths or their adjacent small bands from each frame of spectral data as input for subsequent analysis. This feature extraction significantly reduces data dimensionality, improves computational efficiency, and enhances model robustness.

[0084] Furthermore, in step S2, invoking the preset activation state mapping model includes the following steps:

[0085] S221: Prepare several groups of calibration samples with different activation levels under controlled temperature and time conditions;

[0086] S222: Perform physical performance tests on each set of calibration samples and quantify the corresponding actual activation degree value;

[0087] S223: Collect the sample spectral data for each set of calibration samples;

[0088] S224: Analyze the correlation between the sample spectral data and the actual activation degree values, and select the absorbance or absorption peak intensity ratio at a specific wavelength that is associated with the actual activation degree values ​​as the model input features;

[0089] S225: Fit and establish the functional relationship between the model input features and the actual activation level values, generate the activation state mapping model, and call the activation state mapping model.

[0090] The accuracy and reliability of the activation state mapping model directly determine the performance of the entire control system. Therefore, building a high-quality model is crucial. This modeling process is a systematic, experimental data-based process, typically completed offline before production line deployment.

[0091] Several sets of calibration samples with different activation levels are prepared under controlled temperature and time conditions. This is to create a series of physical samples with known activation states as a benchmark for model training. For example, dozens of identical gummy candy syrup raw materials can be prepared. These samples are then subjected to different degrees of heating in a laboratory-grade precision cooking apparatus. For instance, the first set of samples is heated at 120 degrees Celsius for 30 seconds, the second for 35 seconds, the third for 40 seconds, and so on, or different peak temperatures are used to systematically prepare a series of gradient samples from under-activated to properly activated to over-activated.

[0092] Physical performance tests were performed on each set of calibrated samples to quantify the corresponding actual activation level. This was to assign an objective, quantified activation level label to each physical sample, i.e., the target value or true value for model training. Since the activation level of the gelling agent directly determines the textural properties of the final gummy candy, the activation level can be inferred by measuring these properties.

[0093] In one specific embodiment, a texture analyzer can be used to perform texture profile analysis on the cooled and molded sample. By measuring multiple parameters such as the sample's hardness, elasticity, adhesiveness, and chewiness, and designing a weighted scoring formula based on product requirements, a comprehensive texture score is calculated. This score can then be used as the numerical value for the actual degree of activation. For example, the texture score for the product with the best taste is defined as 100 points.

[0094] In another embodiment, a rheometer can be used to measure the storage modulus and loss modulus of the sample at a specific temperature. These two parameters directly reflect the viscoelastic properties and gel network structure strength of the material, and their values ​​or ratios can very accurately quantify the degree of activation and crosslinking of the gelling agent, thereby obtaining a more fundamental physical value for the actual degree of activation.

[0095] Using the same spectral detection unit as that on the future online production line, the spectra of each group of samples in different activation states are measured and the corresponding spectral data are recorded. In this way, the spectral fingerprint of each sample and its corresponding actual activation level value are obtained, forming a one-to-one data pair.

[0096] The spectral data matrix of all samples and the corresponding numerical vector of actual activation levels are input into the chemometric analysis software. Multivariate statistical methods such as partial least squares regression and principal component analysis can be used. These algorithms can analyze the entire spectral range to identify the wavelengths or bands most strongly correlated with the trend of actual activation level changes. For example, the analysis results may show that the absorbance at a specific wavelength, or the ratio of the heights of two different absorption peaks, exhibits a strong linear or nonlinear relationship with the change in the sample's texture score. These selected spectral variables constitute the input features of the model, ensuring that the model focuses only on the most effective information.

[0097] Using the model input features selected in the previous step and the corresponding actual activation level values, mathematical modeling is performed. A suitable regression algorithm can be selected based on the data characteristics, such as multiple linear regression, support vector machine regression, or artificial neural networks. The algorithm learns and fits the optimal functional relationship between the two. For example, a simple multiple linear regression model might take the form: Activation level = a × Absorbance (λ1) + b × Absorbance (λ2) + c, where a, b, and c are coefficients obtained through model training. This established functional relationship, along with all its parameters, is fixed, forming the final activation state mapping model. In actual production, the control system only needs to call this model to quickly and accurately calculate the activation state of the material based on real-time spectral data.

[0098] Furthermore, in some embodiments, step S3 is followed by:

[0099] S41: Determine the trend of the activation state represented by the deviation value;

[0100] S42: When the deviation value represents that the material activation state is lagging behind the preset ideal activation state, calculate the positive power compensation value according to the magnitude of the deviation value, and control the heating component to increase the positive power compensation value based on the current power.

[0101] S43: When the deviation value represents that the material activation state is ahead of the preset ideal activation state, calculate the negative power correction value according to the magnitude of the deviation value, and control the heating component to reduce the negative power correction value based on the current power.

[0102] The first step in this process is to determine the activation trend represented by the deviation value. This can be accomplished through a simple mathematical comparison. The deviation value is obtained by subtracting the target characteristic index from the real-time characteristic index. If the deviation value is negative, it means the real-time index is less than the target index, indicating that the current activation state of the material is lagging behind the process requirements, i.e., activation is delayed. If the deviation value is positive, it indicates that the activation state is ahead of the process requirements. If the deviation value is close to zero, it means the activation state meets the requirements. This trend determination guides subsequent control decisions.

[0103] When a lag in material activation is detected, the control logic enters a branch to increase or enhance heating. At this point, a positive power compensation value needs to be calculated based on the magnitude of the deviation. The larger the absolute value of the deviation, the more severe the lag, and the greater the required power compensation. This can be achieved using a proportional control strategy. For example, the positive power compensation value can be calculated as: Compensation value = Kp × |Deviation value|, where Kp is a pre-set proportional gain constant used to adjust the control sensitivity. After calculating the compensation value, the control command will drive the heating component to increase this compensation value based on the current output power. For example, if the current output power of the heating component is 70%, and the calculated positive power compensation value is 5%, then the new output power will be set to 75%.

[0104] Conversely, when it is determined that the material activation state is ahead of schedule, the control logic enters a branch to cool down or reduce heating. At this time, a negative power correction value also needs to be calculated based on the magnitude of the deviation. The larger the deviation value, the more severe the pre-activation, and the more power needs to be reduced. The calculation method is similar, for example: Correction value = Kp × Deviation value. Then, the control command will drive the heating component to reduce this correction value from the current power. For example, if the current power is 70% and the calculated correction value is 8%, the new output power will be set to 62%.

[0105] In this way, the adjustment of heating power is no longer a rigid on or off action, but a continuous, proportional adjustment closely linked to the actual activation state of the material. This precise control ensures that the activation process of the material always closely follows the preset optimal path, effectively avoiding problems of overheating or underheating.

[0106] Furthermore, the temperature compensation value is the compensation value for the preset process temperature, and the holding time compensation value is the holding time compensation value for the heating duration.

[0107] The temperature compensation value refers to the value used to adjust the output power of the heating component so that the actual temperature of the material is closer to the preset process temperature. Specifically, it can be calculated using a PID controller based on the deviation value, or obtained by consulting a preset compensation lookup table. For example, when the real-time temperature is lower than the preset process temperature, the temperature compensation value can be a positive value to increase the heating power; when the real-time temperature is higher than the preset process temperature, the temperature compensation value can be a negative value to reduce the heating power.

[0108] The preset process temperature refers to the target temperature set during the gummy production process to achieve a specific activation state. This temperature can be determined through experimentation or experience, depending on factors such as the type and concentration of the gelling agent and the desired gummy texture.

[0109] The holding time compensation value refers to the numerical value used to adjust the duration for which the heating component maintains heating of the material at a specific temperature. Specifically, a time relay or PLC program can be used to control the operating time of the heating component. For example, when the material activation state is delayed, the holding time compensation value can be a positive value to extend the heating time; when the material activation state is advanced, the holding time compensation value can be a negative value to shorten the heating time.

[0110] The heating time refers to the duration for which the material is heated at a specific temperature. This can be determined based on the activation characteristics of the gelling agent and the production process requirements of the gummies.

[0111] Furthermore, the physical consistency logic table is used to increase the feedback gain coefficient when the temperature change rate matches the index change rate, and to decrease the feedback gain coefficient when the temperature change rate does not match the index change rate, so as to attenuate or shield the influence of the temperature compensation value or the duration compensation value on the heating component.

[0112] The physical consistency logic table refers to a pre-defined set of judgment rules used to dynamically adjust the feedback gain coefficient based on the relationship between the temperature change rate and the indicator change rate. Specifically, it can be implemented using lookup tables, conditional statements (such as if-else structures), or rule-based expert systems. For example, the logic table can be defined as follows: when the temperature change rate is positive and the indicator change rate is within a reasonable range, the feedback gain coefficient is set to a high value; when the temperature change rate is positive but the indicator change rate is abnormal or negative, the feedback gain coefficient is set to a low value.

[0113] This solution introduces a physical consistency logic table to intelligently adjust the feedback gain coefficient, thereby optimizing the control effect of the heating components and ensuring the stability of the gummy candy production process and product quality.

[0114] Furthermore, in step S4, determining the feedback gain coefficient includes the following steps:

[0115] S44: When the rate of temperature change is positive and greater than the preset heating threshold, determine whether the rate of change of the index is within the preset reasonable growth range.

[0116] The trigger condition for this logic is set to occur when the rate of temperature change is positive and exceeds a preset heating threshold. For example, the logic is only activated when the temperature rise rate exceeds 0.1 degrees Celsius per second. The purpose of this setting is to focus the verification on critical periods when the material state is undergoing significant changes. This is because small fluctuations in the signal are not significant during the isothermal or slow-change phases, while the synchronization of the two signals is crucial during rapid heating.

[0117] Once the triggering conditions are met, the control system immediately determines whether the rate of change of the real-time characteristic index is within a preset reasonable growth range. This reasonable growth range is pre-calibrated experimentally based on the activation kinetics of the gelling agent at a specific temperature range. For example, experiments show that during the heating process from 110 degrees Celsius to 120 degrees Celsius, the normal growth rate of the activation index should be between 0.5 units and 2 units per second.

[0118] Furthermore, in step S4, the temperature compensation value or the holding time compensation value is weighted using the feedback gain coefficient to obtain the final execution compensation value. This final execution compensation value is then superimposed onto the basic heating control command to adjust the output power of the heating component, including:

[0119] S45: If the rate of change of the index is within a reasonable growth range, the spectral data is deemed valid, and the feedback gain coefficient is set to the first value. The first value is used to maintain or amplify the influence of the temperature compensation value.

[0120] S46: If the rate of change of the index exceeds the reasonable growth range or shows negative growth, it is determined that the spectral data is subject to physical interference. The feedback gain coefficient is set to the second value, which is less than the first value. This is used to attenuate or shield the influence of the temperature compensation value, so that the heating component mainly operates according to the basic heating control command.

[0121] If the detected rate of change of the index falls within this reasonable growth range, such as 1.2 units per second, it indicates that the spectral signal is highly consistent with the physical process driven by the temperature change. At this point, the system determines that the spectral data is valid and reliable. Accordingly, the feedback gain coefficient is set to a relatively high initial value. This initial value can be equal to 1, indicating complete trust and adoption of the spectral correction; in some scenarios requiring rapid response, it can even be set to a value greater than 1, such as 1.2, to moderately amplify the impact of the compensation value and accelerate the system's convergence to the target state.

[0122] Conversely, if the detected rate of change of the indicator exceeds this reasonable growth range—for example, increasing too rapidly to 3 units per second, or increasing too slowly to 0.1 units per second, or even showing negative growth (i.e., temperature rising while activation indicators decrease)—these situations strongly suggest that the spectral signal has been physically interfered with. In this case, the system will determine that the spectral data is temporarily unreliable. To avoid erroneous control, the feedback gain coefficient will be immediately set to a second value much smaller than the first value, such as 0.1, or even set to 0. This lower gain coefficient will greatly attenuate or even completely shield the influence of the temperature compensation value calculated from the abnormal spectral data. In this situation, the control system will operate primarily or even entirely based on the basic temperature control commands within the current control cycle, temporarily ignoring the fine-tuning of the spectrum, thereby ensuring basic production safety and process stability. Once the spectral signal returns to normal, the gain coefficient will automatically recover, and the system will return to high-precision control mode.

[0123] Please refer to Figure 2 , Figure 3 This application also provides a control system for a soft candy production equipment for food processing, the system comprising:

[0124] Acquisition module 201: Acquires real-time spectral data of the material inside the processing container through the spectral detection unit;

[0125] Analysis module 202: Analyzes real-time spectral data to obtain real-time characteristic indicators representing the activation state of the material; the analysis includes: calling a preset activation state mapping model, which represents the correspondence between spectral characteristic data and the thermal activation degree of the gelling agent, obtaining the activation degree value of the gelling agent molecules of the current material, and determining the activation degree value of the gelling agent molecules as the real-time characteristic indicator; acquiring real-time temperature data of the material in the processing container, and generating basic heating control instructions based on the real-time temperature data to drive the heating components to operate according to the preset process temperature;

[0126] Calculation module 203: Calculates the deviation between real-time feature indicators and preset target feature indicators;

[0127] Adjustment module 204: Calculates the temperature compensation value or the maintenance time compensation value based on the deviation value, calculates the temperature change rate of real-time temperature data within the preset sliding time window and the index change rate of real-time characteristic indicators within the preset sliding time window, calls the preset physical consistency logic table, determines the feedback gain coefficient based on the matching relationship between the temperature change rate and the index change rate, and uses the feedback gain coefficient to weight the temperature compensation value or the maintenance time compensation value to obtain the final execution compensation value, and superimposes the final execution compensation value into the basic heating control command to adjust the output power of the heating component.

[0128] Specifically, the acquisition module 201 is the system's data acquisition front end. In terms of hardware, it includes the spectral detection unit itself, such as the light source, detector, fiber optic probe, and optical window mounted on the processing container of the near-infrared spectrometer. In terms of software, it includes the low-level driver program that drives these hardware components to perform periodic data acquisition, the analog-to-digital converter circuit responsible for converting the raw optical signals into digital signals, and related firmware. The function of this module is to stably and reliably provide the entire system with a high-quality stream of raw spectral data.

[0129] The analysis module 202 is the core of the system's intelligent analysis, typically implemented in software within an industrial computer or programmable logic controller. This module internally embeds the aforementioned activation state mapping model. Upon receiving real-time spectral data from the acquisition module, the analysis module executes a feature band extraction algorithm, inputs the extracted feature data into the mapping model for computation, and ultimately outputs quantified real-time feature indicators.

[0130] The calculation module 203 is a relatively simple logic operation unit, also implemented in software within the controller. Its responsibility is to receive real-time characteristic indicators from the analysis module, read the target characteristic indicators for the current stage from the process parameter database, perform a subtraction operation to obtain the deviation value between the two, and then pass this deviation value to the adjustment module.

[0131] The adjustment module 204 is the system's decision-making and execution center. This module integrates complex control algorithm logic, including a basic temperature control loop, as well as logic for calculating compensation values ​​based on activation state deviation, physical consistency verification logic, and dynamic adjustment logic for the feedback gain coefficient. The adjustment module receives deviation values ​​and real-time temperature data, makes a comprehensive judgment, and generates a final, corrected, and weighted heating control command. This command is ultimately output to the actuators of the heating components, such as the controllers of steam proportional control valves or solid-state relays, in the form of a standard industrial signal, such as a 4-20 mA current signal or a pulse width modulation signal, thereby achieving precise, closed-loop adjustment of the heating power.

[0132] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A control method for a gummy candy production equipment used in food processing, for controlling the gummy candy production equipment, the gummy candy production equipment comprising a processing container, a heating component, and a spectral detection unit, characterized in that, The method includes the following steps: S1: Obtain real-time spectral data of the material inside the processing container through the spectral detection unit; S2: Obtain real-time characteristic indicators representing the activation state of the material based on the real-time spectral data analysis; wherein, the analysis includes: calling a preset activation state mapping model, which represents the correspondence between spectral characteristic data and the degree of thermal activation of the gelling agent, obtaining the current value of the activation degree of the gelling agent molecules in the material, and determining the value of the activation degree of the gelling agent molecules as the real-time characteristic indicator; obtaining the real-time temperature data of the material in the processing container, and generating basic heating control instructions based on the real-time temperature data to drive the heating component to operate according to the preset process temperature; S3: Calculate the deviation between the real-time feature index and the preset target feature index; S41: Determine the activation state trend represented by the deviation value; S42: When the deviation value represents that the material activation state is lagging behind the preset ideal activation state, calculate the positive power compensation value according to the magnitude of the deviation value, and control the heating component to increase the positive power compensation value based on the current power. S43: When the deviation value represents that the material activation state is ahead of the preset ideal activation state, calculate the negative power correction value according to the magnitude of the deviation value, and control the heating component to reduce the negative power correction value based on the current power. S4: Calculate the temperature compensation value or the maintenance time compensation value based on the deviation value, calculate the temperature change rate of the real-time temperature data within the preset sliding time window and the index change rate of the real-time characteristic index within the preset sliding time window, call the preset physical consistency logic table, determine the feedback gain coefficient based on the matching relationship between the temperature change rate and the index change rate, and use the feedback gain coefficient to weight the temperature compensation value or the maintenance time compensation value to obtain the final execution compensation value, and add the final execution compensation value to the basic heating control command to adjust the output power of the heating component; Determining the feedback gain coefficient includes the following steps: S44: When the rate of temperature change is positive and greater than the preset heating threshold, determine whether the rate of change of the index is within the preset reasonable growth range; The process of weighting the temperature compensation value or the duration compensation value using the feedback gain coefficient to obtain the final execution compensation value, and then adding the final execution compensation value to the basic heating control command to adjust the output power of the heating component includes: S45: If the rate of change of the index is within the reasonable growth range, the spectral data is determined to be valid, and the feedback gain coefficient is set to a first value. The first value is used to maintain or amplify the influence of the temperature compensation value. S46: If the rate of change of the index exceeds the reasonable growth range or shows negative growth, it is determined that the spectral data is subject to physical interference. The feedback gain coefficient is set to a second value, which is less than the first value, to attenuate or shield the influence of the temperature compensation value, so that the heating component mainly operates according to the basic heating control command.

2. The control method for a soft candy production equipment for food processing according to claim 1, characterized in that, Step S1 includes: S11: Control the spectral detection unit to emit a light signal, so that the light signal penetrates the material inside the processing container; S12: Receive the transmitted light signal after penetrating the material; S13: Convert the transmitted light signal into an electrical signal and perform noise reduction processing on the electrical signal to generate the real-time spectral data.

3. The control method for a soft candy production equipment for food processing according to claim 1, characterized in that, In step S2, the parsing further includes: S21: Extract the characteristic band data from the real-time spectral data, input the characteristic band data into the activation state mapping model, and calculate the activation degree of the gelling agent molecules of the current material.

4. The control method for a soft candy production equipment for food processing according to claim 3, characterized in that, In step S2, invoking the preset activation state mapping model includes the following steps: S221: Prepare several groups of calibration samples with different activation levels under controlled temperature and time conditions; S222: Perform physical performance tests on each group of calibration samples to quantify the corresponding actual activation degree value; S223: Collect the sample spectral data for each set of calibration samples; S224: Analyze the correlation between the sample spectral data and the actual activation degree value, and select the absorbance or absorption peak intensity ratio at a specific wavelength that is associated with the actual activation degree value as the model input feature; S225: Fit and establish the functional relationship between the model input features and the actual activation level values, generate the activation state mapping model, and call the activation state mapping model.

5. The control method for a soft candy production equipment for food processing according to claim 1, characterized in that, The temperature compensation value is a compensation value for the preset process temperature, and the holding time compensation value is a holding time compensation value for the heating duration.

6. The control method for a soft candy production equipment for food processing according to claim 1, characterized in that, The physical consistency logic table is used to increase the feedback gain coefficient when the temperature change rate matches the index change rate, and decrease the feedback gain coefficient when the temperature change rate does not match the index change rate, so as to attenuate or shield the influence of the temperature compensation value or the duration compensation value on the heating component.

7. A control system for a soft candy production equipment for food processing, characterized in that, The system employs a control method for a soft candy production equipment for food processing as described in any one of claims 1-6, and includes the following modules: Acquisition module: Acquires real-time spectral data of materials inside the processing container through the spectral detection unit; The analysis module: analyzes the real-time spectral data to obtain real-time characteristic indicators representing the activation state of the material; wherein, the analysis includes: calling a preset activation state mapping model, which represents the correspondence between spectral characteristic data and the thermal activation degree of the gelling agent, obtaining the activation degree value of the gelling agent molecules of the current material, and determining the activation degree value of the gelling agent molecules as the real-time characteristic indicator; acquiring the real-time temperature data of the material in the processing container, and generating basic heating control commands based on the real-time temperature data to drive the heating components to operate according to the preset process temperature; Calculation module: Calculates the deviation between the real-time feature index and the preset target feature index; Adjustment module: Calculates temperature compensation value or maintenance time compensation value based on the deviation value, calculates the temperature change rate of the real-time temperature data within a preset sliding time window and the index change rate of the real-time characteristic index within the preset sliding time window, calls a preset physical consistency logic table, determines the feedback gain coefficient based on the matching relationship between the temperature change rate and the index change rate, and uses the feedback gain coefficient to weight the temperature compensation value or the maintenance time compensation value to obtain the final execution compensation value, and superimposes the final execution compensation value into the basic heating control command to adjust the output power of the heating component.

Images