Vacuum sealer control method, device and equipment and storage medium

Abstract

The invention relates to the field of sealing control, in particular to a vacuum sealing machine control method, device and equipment and a storage medium, and the method comprises the following steps: collecting a full-wave band spectral image of a to-be-packaged object based on a spectral sensor; performing spectral analysis according to the full-band spectral image to generate a spectral feature set; performing similarity calculation on a standard material spectrum template in a pre-stored database based on the spectrum feature set, and determining an actual material type; packaging parameter analysis is carried out based on the actual material type, and real-time packaging process parameters are output; generating a control instruction based on the real-time packaging process parameters; and the sealing machine is controlled in real time according to the control instruction, and the sealing process is completed. The sealing quality and the production efficiency are improved, and the energy consumption and the defective rate of the number sealing machine are reduced.

Description

Technical Field

[0001] This invention relates to the field of sealing control, and more particularly to a vacuum sealing machine control method, apparatus, equipment, and storage medium. Background Technology

[0002] With the widespread application of vacuum sealing machines in high-frequency, high-speed production scenarios, the issues of intelligent control and sealing quality stability have become increasingly prominent. In actual production, due to differences in the material, thickness, and moisture content of the items to be sealed, as well as real-time changes in process parameters such as sealing temperature, pressure, and speed, a lack of accurate material identification and dynamic process adjustment can easily lead to problems such as insecure sealing, air leakage, scorching, or packaging deformation, thus affecting product quality and production efficiency. Traditional sealing machine control methods mainly rely on empirical parameter settings or simple fixed-value control systems, making it difficult to achieve dynamic adaptation to different materials and sealing conditions. While these methods can accomplish basic sealing tasks to a certain extent, they often suffer from unstable sealing quality, frequent manual intervention, and low sensitivity to material differences. Existing control systems lack the ability to identify and warn of sealing anomalies, making it difficult to achieve intelligent, real-time sealing quality management. Summary of the Invention

[0003] To address the aforementioned technical problems, this invention proposes a vacuum sealing machine control method, apparatus, equipment, and storage medium, thereby resolving at least one of the aforementioned technical problems.

[0004] To achieve the above objectives, the present invention provides a vacuum sealing machine control method, comprising the following steps: Step S1: Acquire full-band spectral images of the item to be packaged based on a spectral sensor; perform spectral analysis based on the full-band spectral images to generate a spectral feature set; Step S2: Calculate the similarity of standard material spectral templates in the pre-stored database based on the spectral feature set to determine the actual material type; Step S3: Analyze the packaging parameters based on the actual material type and output the real-time packaging process parameters; Step S4: Generate control instructions based on the real-time packaging process parameters; perform real-time control of the sealing machine according to the control instructions to complete the sealing process.

[0005] This specification provides a vacuum sealing machine control device for performing the method described above, comprising: The spectral analysis unit is used to acquire full-band spectral images of the item to be packaged based on a spectral sensor; and to perform spectral analysis based on the full-band spectral images to generate a spectral feature set. The material comparison unit is used to calculate the similarity of standard material spectral templates in a pre-stored database based on the spectral feature set to determine the actual material type. The parameter calculation unit is used to analyze the packaging parameters based on the actual material type and output real-time packaging process parameters. The packaging control unit is used to generate control commands based on the real-time packaging process parameters; and to perform real-time control of the sealing machine according to the control commands to complete the sealing process.

[0006] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the vacuum sealing machine control method described in any of the above claims.

[0007] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the vacuum sealing machine control method described in any of the preceding claims.

[0008] The beneficial effects of this invention are as follows: By acquiring full-band spectroscopic data, spectral information of the item to be packaged in the visible, near-infrared, and even mid-infrared bands can be obtained, accurately reflecting the chemical composition and physical properties of the item. Spectral analysis does not require direct contact with the item or damage to its structure to obtain material information, ensuring the item remains intact before packaging. Analysis of full-band spectral images allows for the extraction of multi-dimensional spectral features (such as absorption peaks, reflectance distribution, and band feature ratios), providing a reliable data foundation for subsequent material identification and process adjustments. Similarity calculations match the spectral features of the actual item with standard material templates in the database, accurately determining key characteristics such as material type, moisture content, and thickness of the item to be packaged. Traditional sealing processes may rely on manual judgment of material properties, which is prone to errors; automated spectral comparison improves the consistency and reliability of the judgment. The optimal packaging temperature, pressure, and sealing time are automatically calculated based on actual material characteristics (such as material thickness, material type, and sealing temperature requirements), enabling real-time adjustments. Real-time calculated process parameters ensure a firm seal and good airtightness, reducing the risk of poor sealing, leakage, or material scorching. The sealing machine automatically adjusts temperature, pressure, and sealing time based on generated control commands, achieving fully intelligent operation and reducing manual intervention. Precise and real-time control commands ensure consistent sealing quality for each item, reducing the defect rate. Sealing parameters can be fine-tuned based on real-time sensor monitoring results, achieving closed-loop control and improving sealing accuracy and reliability. Attached Figure Description

[0009] Figure 1 This is a schematic diagram of the steps of a vacuum sealing machine control method according to the present invention; Figure 2This is a detailed flowchart illustrating the implementation steps of step S1. Figure 3 This is a flowchart illustrating the detailed implementation steps of step S2. Detailed Implementation

[0010] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0011] This application provides a vacuum sealing machine control method, apparatus, device, and storage medium. The execution entities of the vacuum sealing machine control method, apparatus, device, and storage medium include, but are not limited to, mechanical equipment, data processing platforms, cloud server nodes, and network upload devices that can be considered as general computing nodes in this application. The data processing platform includes, but is not limited to, at least one of an audio-visual management system, an information management system, and a cloud-based data management system.

[0012] Please see Figures 1 to 3 This invention provides a vacuum sealing machine control method, comprising the following steps: Step S1: Acquire full-band spectral images of the item to be packaged based on a spectral sensor; perform spectral analysis based on the full-band spectral images to generate a spectral feature set; Step S2: Calculate the similarity of standard material spectral templates in the pre-stored database based on the spectral feature set to determine the actual material type; Step S3: Analyze the packaging parameters based on the actual material type and output the real-time packaging process parameters; Step S4: Generate control instructions based on the real-time packaging process parameters; perform real-time control of the sealing machine according to the control instructions to complete the sealing process.

[0013] In the embodiments of the present invention, see Figure 1 The diagram below illustrates the steps of a vacuum sealing machine control method according to the present invention. In this example, the steps of the vacuum sealing machine control method include: Step S1: Acquire full-band spectral images of the item to be packaged based on a spectral sensor; perform spectral analysis based on the full-band spectral images to generate a spectral feature set; In this embodiment, before packaging, full-band spectral image acquisition of the item to be packaged is required to obtain the spectral information of the material. The spectral sensor can cover the visible to near-infrared or short-wave infrared band (400–2500 nm), acquiring the material's reflectance or absorptivity distribution by acquiring images band by band. During acquisition, to avoid ambient light interference, uniform lighting conditions and a fixed acquisition angle should be used, and whiteboard and dark field corrections should be performed to eliminate the effects of light source inhomogeneity and sensor noise. The acquired full-band images form a three-dimensional data matrix in both space and spectrum, reflecting the surface and possible internal structural features of the material, providing basic information for identifying the material type. After acquiring the spectral images, spectral analysis is required to generate a spectral feature set. The analysis process typically includes first- and second-order difference analysis of the band reflectance curves, absorption peak position identification, key band ratio calculation, spectral texture analysis, and principal component extraction. The feature set can reflect the material's overall spectral shape, local fluctuations, texture roughness, directionality, and contrast, compressing redundant dimensions.

[0014] Step S2: Calculate the similarity of standard material spectral templates in the pre-stored database based on the spectral feature set to determine the actual material type; In this embodiment, after obtaining the spectral feature set, it needs to be matched with pre-stored standard material spectral templates to determine the actual material type of the item to be packaged. The standard material spectral templates cover common packaging materials such as PE, PP, PET, PA, and multilayer composite films. The spectral features in the templates are normalized and processed by principal component analysis to facilitate similarity calculation with the collected feature set. Similarity calculation can use methods such as cosine similarity, Euclidean distance, or Mahalanobis distance to compare the feature vector of the material to be tested with the feature vector of each template. The comparison results yield the similarity distribution of various materials, which are then sorted according to the magnitude of similarity. Typically, the material type is initially determined based on the principle of maximum similarity, and the reliability of multi-material matching is analyzed by combining the similarity difference. When the difference between the first and second matching results is significant, the material identification can be considered to have a high confidence level; if the difference is small, it can be further verified through key bands or texture features to improve the identification accuracy.

[0015] Step S3: Analyze the packaging parameters based on the actual material type and output the real-time packaging process parameters; In this embodiment, after clarifying the material properties, encapsulation parameter analysis is required to generate real-time encapsulation process parameters suitable for the material. Encapsulation parameter analysis mainly involves a comprehensive consideration of the material's thermal properties, mechanical properties, and sealing region geometric parameters. Material thermal properties include melting temperature range, specific heat capacity, and thermal conductivity, used to calculate heating temperature distribution and heating time; mechanical properties include yield strength and elastic modulus, used to determine pressure intensity and holding time; sealing region geometric parameters such as length, width, and area are used to calculate heat distribution and pressure application method. By combining these factors, complete real-time encapsulation process parameters can be generated, including target heating temperature, heating rate, heating duration, sealing pressure, pressure holding time, and vacuum degree setpoint. Process parameters can be finely allocated in time and space; for example, local temperature compensation and pressure gradients can be set for long seals or composite film seals.

[0016] Step S4: Generate control instructions based on the real-time packaging process parameters; perform real-time control of the sealing machine according to the control instructions to complete the sealing process.

[0017] In this embodiment, after obtaining the real-time packaging process parameters, they need to be converted into control instructions executable by the sealing machine. These control instructions include the temperature curve of the heating unit, the pressure value and pressure application time of the pressurizing device, the vacuum pump parameters, and the execution sequence of the entire sealing process. The instruction generation process requires structuring parameters such as temperature, pressure, time, and vacuum level, and issuing them in stages to ensure coordinated actions at each stage and avoid conflicts or delays between heating, pressurization, and vacuuming. The sealing machine performs real-time control according to the control instructions, executing the sealing operation. The target vacuum level is achieved, followed by heating the sealing area to the target temperature, applying the designed pressure, and maintaining it for the specified time. Throughout the sealing process, sensors monitor temperature and pressure in real time to ensure that the material in the sealing area is fully fused in a thermally softened state. After heating and pressurization are completed, the sealing machine releases pressure and stops heating, allowing the sealing layer to cool and solidify, forming a stable and uniform seal.

[0018] In this embodiment, see Figure 2 The diagram below illustrates the detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include: Full-band spectral images of the items to be packaged are acquired using a spectral sensor; Wavelet denoising and sharpening filtering are performed on the full-band spectral image to obtain a sharpened and enhanced spectral image; Spectral texture fluctuation analysis is performed on the sharpened and enhanced spectral image to generate texture fluctuation features; the texture fluctuation features include texture roughness, directionality and contrast parameters. Perform band-by-band difference operations on the sharpened and enhanced spectral image to generate regional gradient characteristics; Based on the sharpened and enhanced spectral image, the ratio of specular reflection to diffuse reflection at different incident angles is calculated to obtain the spectral reflectance. Principal component analysis was performed on spectral reflectance, regional gradation characteristics, and texture fluctuation features to generate a spectral feature set.

[0019] In this embodiment, after the item to be packaged enters the spectral imaging area, spectral image data covering the entire wavelength band is acquired by a spectral imaging device installed in the vacuum sealing machine's identification module. The imaging device typically has a wide spectral response range of 400–2500 nm, used to capture surface color information in the visible light band, material absorption characteristics in the near-infrared band, and moisture content, density changes, and surface structure differences in the short-wave infrared band. To obtain comparable spectral images, the sensor undergoes dark-field and white-board calibration before spectral acquisition. Dark-field images are used to compensate for electronic noise, and a standard reflective white board is used to correct uneven light source illumination, ensuring high spectral consistency in the output spectral values. The light source can be halogen illumination with continuous spectral characteristics or a broadband LED array, maintaining a combination of 45° incident and forward reception to reduce saturation problems caused by strong specular reflection. To adapt to the fast-moving flow of items during the sealing process, a trigger synchronization method can be used to ensure that the spectral acquisition time is consistent with the item's position in the imaging area, avoiding spatial misalignment or band mismatch in spectral data. The acquired spectral images contain random noise introduced by light source instability, circuit readout errors, and spatial scattering. Therefore, multi-scale wavelet methods are needed for noise reduction, combined with sharpening algorithms to enhance spectral edge details. The wavelet denoising process treats each spectral band as a processing object, decomposing it into low-frequency approximation components and high-frequency detail components at different scales through discrete wavelet transform. Commonly used wavelet bases include Daubechies and Symlets, used to adapt to multi-scale texture structures. A soft thresholding function is used to suppress high-frequency noise components, preserving as much realistic structural detail as possible, making the image smoother and less distorted in spatial texture. After reconstruction, Laplacian sharpening, nonlinear contrast enhancement, or multi-scale Retinex-like methods are used to strengthen gradient changes in edge regions, thereby improving the discernibility of material textures, membrane boundaries, and wrinkles in sealing areas in the spectral image. Throughout the processing, consistent parameters are maintained for all bands to ensure no artificial bias is introduced in the spectral dimension, keeping the shape of the spectral curve natural and continuous. The spectral image after wavelet denoising and sharpening enhancement presents a more stable spectral gradient, clearer spatial texture, and a higher signal-to-noise ratio.

[0020] By sharpening and enhancing the spectral image, a more detailed analysis of the texture state of the object surface and sealing film area can be performed. The role of spectral texture fluctuation analysis is to characterize the stability and change trend of the surface structure, avoiding the impact of factors such as film wrinkles, surface particles, and local indentations on sealing reliability. Texture roughness can be calculated through spectral gradient changes, spectral energy coefficients, and local variance in local windows to reflect the surface undulation. Directionality is evaluated using indicators such as energy, entropy, and correlation in multiple directions (e.g., 0°, 45°, 90°, 135°) of the gray-level co-occurrence matrix to assess whether there is obvious structural directionality in the texture, often used to determine whether linear wrinkles or stress concentrations have appeared in the sealing film. Contrast parameters are used to represent the brightness differences and structural layers of the texture area; higher contrast often indicates the presence of local indentations, stretching, or abrupt changes in the material's refractive index. To adapt to the characteristic of spectral imaging containing a large number of continuous bands, the synchronicity of texture fluctuations can also be analyzed across multiple bands. Cross-band consistency characteristics can be used to determine whether there are stable structural changes in the material state, rather than random brightness interference. Band-by-band differential analysis is used to capture the changing trends of spectral curves along band directions, thereby revealing the gradual changes in the internal state or surface structure of materials. First-order difference measures the rate of change between adjacent bands, while second-order difference identifies the curvature of change. These differential responses can amplify subtle spectral differences and local structural changes. To avoid amplifying noise during the differential process, a smoothing filter method (such as Savitzky-Golay filtering based on polynomial fitting) can be used to smooth the spectral curve, making it more stable while maintaining its shape. After differential analysis, the mean, variance, gradient direction, and differential energy of the difference values ​​are statistically analyzed through local windows to create a spatially continuous structural description of the regional gradual changes. If the sealing film material experiences abnormal local stress due to uneven pressure, uneven heat sealing temperature, or the shape of the object during vacuum sealing, the spectral curve often shows abnormal changes in certain bands. Differential analysis can effectively amplify and represent these changes.

[0021] Spectral data includes specular reflection and diffuse reflection, both of which affect the shape of the spectral curve. Therefore, it is necessary to separate the two reflectance components to extract a spectral reflectance ratio that more closely approximates the intrinsic properties of the material. To achieve this separation, multi-angle illumination can be used to make the object exhibit different reflection responses at different incident angles. By comparing the spectral values ​​obtained at incident angles such as 30°, 45°, and 60°, the specular and diffuse reflectance coefficients can be derived based on the influence of changes in the incident direction of the light source on the brightness component. Specular reflection typically manifests as a more pronounced high-brightness peak with varying angles, while diffuse reflection represents the true spectral information of the material's surface and internal structure. Therefore, the ratio between the two can be determined through curve fitting. Spectral normalization and incident angle cosine correction can be used to offset the influence of differences in light source energy on the spectrum, making the calculated spectral reflectance ratio more valuable. Once the true reflectance is obtained, it can effectively reflect the intrinsic properties of the material, such as changes in film thickness, differences in surface microstructure, and stress state in the sealing area. Spectral reflectance, regional gradation characteristics, and texture fluctuation features are all high-dimensional data. Directly using them for intelligent recognition and control of sealing machines would involve significant computational loads and potential feature redundancy. Therefore, Principal Component Analysis (PCA) is needed to map multidimensional features to a lower-dimensional principal component space. All features are normalized to ensure comparability of features from different sources and with different dimensions. Subsequently, a covariance matrix is ​​constructed, and principal component loadings are obtained through eigenvalue decomposition. Several principal components representing the main information are selected based on their contribution rates, ensuring that the overall feature space retains the main trends of the original information even with reduced dimensionality. When processing spectral reflectance, the first few principal components extracted by PCA typically correspond to trends in spectral intensity or material absorption characteristics; when processing regional gradation characteristics, principal components reflect structural features such as changes in spectral slope and curvature; and when processing texture features, principal components can correspond to structural factors such as changes in roughness and texture directionality.

[0022] In this embodiment, see Figure 3 The diagram below illustrates the detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include: Obtain standard material spectral templates from the pre-stored database; Extract the feature vector of the standard material spectral template; Similarity calculations are performed on the feature vectors based on the spectral feature set to extract similarity between multiple standard materials; Material confidence is determined by assessing the similarity of multiple standard materials to identify the actual material type.

[0023] In this embodiment, to accurately identify the material properties of the items to be packaged, a pre-established standard material spectral template database needs to be accessed. This database typically contains full-band spectral templates for various common packaging materials, covering 12 categories of packaging materials, including polyethylene (PE), polypropylene (PP), polyester (PET), nylon (PA), aluminum-plastic composite film, paper-plastic composite material, EVOH barrier film, PVDC coating material, biodegradable film, multilayer co-extruded film, metallized film, and silicon oxide coated film. Each template is obtained under uniform illumination conditions, uniform acquisition angle, and uniform spectral correction method, and undergoes dark-field and white-board correction and spectral normalization to ensure high spectral representativeness and stability. Spectral templates generally record reflectance or absorptivity curves in the spectral range of 400–2500 nm and may include key band annotations, such as the absorption peak distribution of plastic materials at 1200 nm, 1450 nm, and 1720 nm. To improve retrieval efficiency, the database is categorized according to material type, film structure, thickness, surface treatment method, and production batch, so that the corresponding template can be quickly located when material identification is required. During the template reading process, the sealing machine prioritizes potentially relevant spectral templates based on background information of the current recognition task, such as packaging bag type, sealing film material category, and item properties, to reduce time loss caused by irrelevant templates participating in the calculation. After loading the standard material spectral template, the spectral curve needs to be converted into a feature vector for similarity comparison. The feature vector typically encompasses four aspects: the overall shape of the spectral curve, key band features, differential response, and spectral texture features. The original reflectance curve of the spectral template is analyzed, and basic features describing the spectral variation trend are formed by calculating parameters such as first-order difference, second-order difference, local slope, and spectral peak and valley positions. For example, absorption peaks at positions such as 970 nm, 1200 nm, 1450 nm, and 1720 nm are often used to distinguish different types of plastic materials, especially in distinguishing between PET and PP. Secondly, by calculating the normalized area of ​​the spectral curve, the ratio of characteristic bands (such as the 1450 / 1200 nm absorptivity), and spectral continuity index, the feature vector can express the overall spectral shape of the material. If the template stores spectral texture feature information, texture roughness, directionality, and cross-band consistency parameters can be added to make the feature vector more comprehensive. To facilitate matching, all features are usually normalized so that they participate in the calculation under the same dimensions.

[0024] After obtaining the feature vectors of the standard materials, it is necessary to calculate their similarity with the spectral feature set of the item to be packaged to determine the degree of matching. Similarity calculation can employ various distance metrics and correlation algorithms based on the characteristics of spectral recognition, such as Euclidean distance, cosine similarity, Mahalanobis distance, and Pearson correlation coefficient. Since spectral features are often processed by PCA, similarity calculations in principal component space are often more stable and can effectively avoid the curse of dimensionality that easily occurs in high-dimensional spaces. In PCA space, features such as spectral curve shape changes, absorption peak shifts, and local band response differences are condensed into a small number of principal components. Therefore, calculating the distance between these principal component vectors can more accurately reflect the true differences between materials. Furthermore, different weights can be set between different feature types; for example, spectral shape can be given a higher weight, while texture parameters can be used as auxiliary factors to improve the overall robustness of the matching. After obtaining the similarity values ​​corresponding to multiple standard materials, it is necessary to further determine the confidence level of these values ​​to identify the material properties that best match reality. The confidence level determination can be based on the principle of maximum similarity, that is, the material template with the highest similarity is used as the initial identification result. However, to avoid misjudgments due to extreme cases, it is necessary to further calculate the similarity gap, such as the ratio between the first and second matching results. If the gap is greater than a preset threshold (e.g., 15% or 20%), the identification can be considered to have high credibility. If the similarity of multiple materials is relatively close, it is necessary to further verify using key parameters with high discriminative power in the feature dimension, such as absorption peak position shift, differential fluctuation curvature, and characteristic band reflectance, to determine a more accurate material property and type. To make the confidence score more interpretable, the similarity can be used to generate a material confidence score through Softmax or normalized probability transformation, allowing the sealing machine to automatically adjust the sealing parameters according to the confidence score. For example, PE materials usually require a lower sealing temperature, while PET materials require a higher temperature. When the material confidence score exceeds a set threshold, the sealing parameters of the corresponding material can be directly referenced. If the confidence score is lower than the threshold, it can trigger re-acquisition or indicate a film material abnormality.

[0025] In this embodiment, step S3 includes the following steps: The sealing location region is identified based on full-band spectral images; the sealing length, sealing width, and sealing area of ​​the sealing location region are calculated to obtain the geometric parameters of the sealing region; Analyze the thermal properties of the material based on the actual material type; calculate the heating time and temperature distribution based on the thermal properties of the material, and generate sealing temperature parameters; Extract the yield strength and elastic modulus of the material based on the actual material type; The sealing time and sealing pressure are determined based on the yield strength and elastic modulus of the material and the geometric parameters of the sealing area, thus obtaining the sealing pressure parameters; Identify the type of the item to be packaged; adjust the adaptive vacuum level according to the item type to obtain the adaptive vacuum level; Real-time packaging process parameters are output based on adaptive vacuum degree, sealing pressure parameters, and sealing temperature parameters.

[0026] In this embodiment, after obtaining the full-band spectral image of the item to be packaged, it is necessary to identify the sealing location region to provide basic spatial information for subsequent sealing thermal parameters, pressure parameters, and vacuum adjustment. In the spectral image, the sealing area typically exhibits unique spectral characteristics due to differences in material layers, structural thickness variations, and surface spectral reflectance properties. For example, it may show stronger absorption or reflection differences in the near-infrared region of 900–1300 nm. By performing multi-band fusion analysis on these differences, the approximate boundary of the sealing area can be accurately identified. Identification methods typically include spectral difference analysis, spectral angle matching, and region enhancement based on band ratios. For example, in the 1200 nm vs. 1450 nm band ratio map, the material overlap region often exhibits a significant brightness shift, which can serve as a basis for determining the sealing boundary. After completing the sealing region identification, it is necessary to calculate the geometric parameters of this region, including the sealing length, sealing width, and sealing area. The sealing length can be calculated by accumulating pixels along the sealing direction, with spectral accuracy typically on the order of 0.1 mm. The sealing width is obtained by locating the boundary in the vertical direction. The sealing area can be calculated by multiplying the two dimensions, or by integrating pixels one by one to obtain a more refined area estimate. In practical applications, to ensure the stability of geometric parameters, a binary image fused from multi-band features is used for morphological processing to avoid edge noise affecting the measurement of the area. Based on the material properties determined in the preceding steps, the thermal characteristics of the material need to be analyzed to determine the appropriate sealing temperature parameters. Different packaging materials have significantly different thermal characteristics. For example, the melting range of PE is mostly between 110–140°C, the melting range of PP is usually between 150–170°C, while PET has a higher heat softening temperature and a narrower heat-sealing window. Therefore, when calculating the sealing temperature, it is necessary to estimate based on parameters such as the material's specific heat capacity, thermal conductivity, thermal diffusivity, and melting temperature range width. To ensure a secure seal and prevent excessive melting of the material, a material heat conduction model needs to be established to estimate the temperature propagation path and temperature gradient curve after heat is applied to the heating plate or heating wire. The heat diffusion equation is typically used to calculate the heating time, with material thickness, layer structure, and thermal conductivity as key inputs. For example, in the heat sealing process of typical thin films (40–70 μm thick), the heating time is usually between 0.2 and 1.0 seconds; if the material is a multilayer composite film, the heat conduction path is more complex, and the heating time may need to be appropriately extended. In the temperature distribution calculation, the temperature gradient within the sealing area also needs to be considered. Because the contact between the heating plate and the film material is not completely uniform, temperature deviations will occur in some areas; therefore, it is necessary to calculate the temperature uniformity in both the transverse and longitudinal directions.

[0027] After identifying the material type, it is necessary to further extract the yield strength and elastic modulus of the material to provide a mechanical basis for subsequent calculations of sealing pressure and sealing time. Different packaging materials exhibit significant differences in mechanical properties during heating and pressurization. Yield strength reflects the critical point at which a material undergoes plastic deformation under stress, while elastic modulus reflects the material's ability to resist deformation. For example, PE materials typically have a lower elastic modulus and higher ductility, while PET exhibits a higher elastic modulus and stronger tensile strength. This results in completely different pressure and holding times required for sealing the two types of materials. The yield strength of the material can be directly extracted based on pre-stored information in the material database combined with material identification results; while the elastic modulus can be further estimated by combining spectral texture characteristics and film thickness information. In many film packaging scenarios, the yield strength is typically in the range of 8–30 MPa, while the elastic modulus may fluctuate in the range of 300–2500 MPa. If the material is a composite film, its equivalent elastic modulus needs to be calculated based on a multi-layer material mechanical superposition model. By accurately obtaining these mechanical parameters, it is possible to further predict the amount of material deformation, the melting depth between heat-sealing layers, and the degree of adhesion during the compression process. Only after fully understanding the mechanical behavior of the material can the sealing pressure and sealing time be accurately determined, ensuring that the sealing process not only guarantees a strong bond between the materials but also avoids indentation, deformation, or even cracking due to excessive pressure.

[0028] After obtaining the sealing geometry, yield strength, and elastic modulus, the sealing pressure and sealing time can be accurately calculated. The sealing pressure must ensure that the heat-sealing layer adheres tightly under reasonable pressure after the material reaches a thermally softened state; therefore, the pressure range must match the material's mechanical properties. Generally, the calculation of sealing pressure needs to consider the stress-strain relationship of the material under heating conditions. The upper limit of pressure is determined by the yield strength, the amount of deformation by the elastic modulus, and the pressure distribution is determined by combining the sealing width and sealing area. For example, soft materials such as PE and PP require lower sealing pressures, generally around 0.1–0.3 MPa; while PET or multilayer composite film materials may require 0.2–0.5 MPa of pressure to achieve good adhesion. The calculation of sealing time is also closely related to the material's heat conduction rate, melt layer thickness, and sealing area. If the sealing area is large, the pressure needs to remain stable for a longer period to ensure that the heat-sealing layer completes full adhesion at the thermal softening temperature. For example, for a sealing area 50 mm long and 6 mm wide, in typical food packaging scenarios, the sealing time for soft materials is usually 0.5–1.0 seconds, while for composite films it may require 1.0–2.5 seconds. The generated sealing pressure parameters include: target pressure value, pressure application time, pressure stability requirements, and pressure compensation value. These parameters will serve as direct inputs to the sealing machine's actuator, ensuring that the sealing action precisely matches the material properties and sealing area characteristics, thereby guaranteeing sealing strength and product qualification rate.

[0029] In addition to material properties, it is also necessary to identify the type of item to be packaged in order to determine an appropriate vacuum level. Item type identification typically relies on input information from the packaging task, weight distribution, appearance characteristics, and even spectral auxiliary features to determine the type, such as distinguishing between solid foods, powders, liquids, and soft items. Different items have significantly different vacuum requirements. For example, crispy biscuits and puffed foods require lower vacuum levels to avoid crushing; powders require higher vacuum levels to reduce residual air; liquids may require limited vacuum levels to prevent liquid from being drawn out; and for flexible ingredients, excessive vacuuming should be avoided to prevent deformation. Therefore, item type is a core reference for determining vacuum adjustment strategies. Adaptive vacuum levels can be generated based on item type characteristics, such as volume compressibility, structural brittleness, and air permeability. Vacuum levels are usually expressed in kPa and can be adjusted from light vacuum (30–50 kPa) to deep vacuum (below 10 kPa). To prevent damage to items during the vacuuming process, a vacuum curve can be set, which involves reducing pressure in stages. For example, rapid vacuuming can be performed initially, followed by a slower vacuuming process as the target vacuum level approaches, thus improving stability. The generated adaptive vacuum parameters will serve as direct input to the vacuum pump control module, ensuring that the packaging process guarantees the air content inside the bag meets preservation requirements without causing pressure damage to the items.

[0030] After calculating vacuum level, pressure, and temperature, these three key packaging process parameters need to be integrated and output in real time to enable the sealing machine to perform precise and coordinated packaging actions. The generation of real-time process parameters must consider the timing relationship between the vacuuming, heating, and pressurizing stages to ensure that these three parameters do not interfere with each other during execution. For example, heating should not be initiated before vacuuming is complete to avoid material deformation due to heat; and pressure must be applied at appropriate times after heating to ensure the material is bonded in its optimal thermal softening state. The resulting real-time packaging process parameters include: target vacuum level and vacuuming time, sealing temperature and heating rate, sealing pressure and pressure holding time, and timing control throughout the entire sealing cycle. For example, for flexible food packaging bags, the following parameter combination may be formed: vacuum degree 40 kPa, sealing temperature 130°C, pressure 0.2 MPa, and pressure holding for 0.7 seconds; while for high-barrier composite films, a combination such as vacuum degree 10 kPa, sealing temperature 155°C, pressure 0.35 MPa, and pressure holding for 2 seconds may be formed.

[0031] In this embodiment, step S4 includes the following steps: Control commands are generated based on the real-time packaging process parameters; The sealing machine is controlled in real time according to the control instructions, and the temperature parameters of the sealing area are monitored in real time. Set a sealing safety temperature threshold; monitor the temperature parameters of the sealing area throughout the process according to the sealing safety temperature threshold; when the temperature parameters of the sealing area do not exceed the sealing safety temperature threshold, continue to perform the sealing operation to complete the sealing process; When the temperature parameter of the sealing area is detected to be higher than the sealing safety temperature threshold, the protection mechanism is triggered; Based on the aforementioned protection mechanism, the sealing operation is stopped and allowed to cool naturally. Once the temperature of the sealing area is detected to be below the sealing safety temperature threshold, the sealing operation is performed again to complete the sealing process. Acquire an overall image of the packaged material; perform defect identification on the overall material image and output a defect distribution report; Based on the defect distribution report, optimize the packaging parameters and perform secondary packaging.

[0032] In this embodiment, after obtaining the real-time packaging process parameters, these parameters need to be converted into control commands executable by the sealing machine to drive the coordinated operation of the vacuum unit, heating unit, sealing unit, cooling unit, and other execution components. The real-time packaging process parameters include adaptive vacuum level, sealing temperature, heating rate, pressure applied, pressure applied, and the overall timing plan of the sealing process. To ensure the stability of command execution, these parameters need to be structured. For example, in the heating stage, the target temperature should be subdivided into temperature curves for the heating stage, the isothermal stage, and the temperature compensation stage to avoid temperature jumps in the heating elements; in the pressure application stage, the pressure parameters should be converted into specific execution information such as the pressure head pressing speed, the maximum pressure holding time, and the pressure release time. The generation of control commands also needs to consider the geometric parameters of the sealing area, such as the sealing length and sealing width, to determine the heating area and pressure distribution. For example, for a wider sealing area, heating compensation commands are needed to maintain a consistent temperature between the heating elements on both sides to avoid uneven sealing due to temperature differences; for a longer sealing area, the heating duration needs to be extended to achieve a stable heat distribution. In addition, control commands are usually issued in the form of time-series queues to ensure the logical relationship between vacuum, temperature and pressure. For example, the heating stage is prohibited before the vacuum stage is completed, so as to avoid the material from deforming due to softening during the vacuuming process.

[0033] After the control command is issued, the sealing machine immediately begins to perform actions such as vacuuming, heating, pressurizing, and cooling. To ensure the stability of the sealing process, the temperature of the sealing area needs to be monitored in real time during execution. This is typically done using temperature sensors, such as thermocouples, thermistors, or infrared temperature detection elements, placed near the sealing heating components or at the projection location of the sealing area. The time resolution of temperature monitoring is generally controlled within 5–20 ms, which meets the requirement of capturing rapid temperature fluctuations during the sealing process. The purpose of real-time temperature monitoring is to ensure that the material remains within a suitable thermal softening range during heat sealing. For example, PE materials typically require a temperature range of 120–140°C to achieve optimal sealing strength, while PET has a higher and narrower heat-sealing window, thus requiring precise temperature control to prevent material melt-through or incomplete sealing. Temperature monitoring forms a closed-loop relationship with the control command: if the temperature is below the target temperature, the heating unit will automatically increase its power; if the temperature gradually approaches the upper limit of the safe temperature, a cooling strategy will be triggered in advance. During real-time control, it is also necessary to simultaneously monitor whether the applied pressure is stable, whether the pressurization holding time meets the requirements, and whether the vacuum degree reaches the target value. Temperature data at each time point is recorded to dynamically determine whether the sealing process remains within a safe range.

[0034] To ensure that the material does not suffer thermal damage, melt-through, excessive adhesion, or structural failure due to overheating during the sealing process, a safe sealing temperature threshold needs to be pre-set. This temperature threshold is typically determined based on factors such as the material's melting temperature, thermal softening temperature, upper heat resistance limit, and sealing window width. For example, the safe temperature threshold for PE can be set at around 150°C, while for PET it can be set at around 210°C. The safe temperature threshold needs to have an appropriate margin to ensure that the material is not damaged by random temperature fluctuations or environmental factors. During the sealing process, the real-time temperature monitoring module continuously records the temperature of the sealing area and dynamically compares it with the safe temperature threshold. If the temperature remains below the safe temperature threshold, the sealing operation continues according to the established process, including stages such as heating, temperature control, pressurization, holding, and cooling. This continuous monitoring process can be considered a full-process temperature protection mechanism, meaning that each temperature sampling result contributes to the judgment of the sealing status. If the temperature of the sealing area remains within a controllable range, it indicates that the sealing action is safe, the material will not suffer thermal damage, and the sealing process can be completed smoothly.

[0035] If the temperature of the sealing area exceeds the safe temperature threshold during the sealing process, the protection mechanism must be triggered immediately to prevent irreversible damage to the material due to overheating. The protection mechanism consists of three main parts: stopping heating, releasing pressure, and interrupting the current sealing process. Stopping heating prevents the temperature from continuing to rise, while releasing pressure avoids the combined effect of pressure and high temperature causing the film material to puncture. Interrupting the sealing process is a comprehensive safety measure to ensure that subsequent actions are not carried out at dangerous temperatures. Temperature exceeding limits is usually caused by various factors, such as a short-term power surge in the heating element, excessively high ambient temperature, insufficient material thickness, and temperature sensor misalignment. Therefore, the protection mechanism needs sufficient sensitivity and execution speed to promptly prevent potential material damage. After triggering the protection, the sealing machine immediately enters a safety management state and uses parameters such as the current temperature, the extent of exceeding the threshold, and the heating rate to subsequently determine whether there is a hardware fault or material abnormality. Temperature over-limit protection not only prevents sealing failure but is also a crucial measure to prevent overheating damage to the sealing machine, material adhesion to the heating plate, and charring of the sealing opening. By triggering the protection mechanism in a timely manner, the sealing process can be ensured to always be within a safe range.

[0036] Upon triggering the protection mechanism, the sealing operation will immediately pause and enter a natural cooling phase. Natural cooling typically involves reducing the heating element power to zero and keeping the sealing machine's top cover and pressurizing mechanism open or partially open, allowing air convection to quickly remove heat. During this phase, the temperature of the sealing area will be continuously monitored to determine if the temperature drop trend is normal. If the material temperature remains high for a short period, the cooling time may need to be extended, or supplementary cooling methods such as air cooling may be used. Once the temperature drops below the sealing safety temperature threshold, the sealing machine will automatically exit the protection state and restart the sealing process. At this time, real-time sealing process parameters will be reloaded for reheating, pressurization, and sealing. To prevent the temperature from exceeding the limit again, the initial heating power may be appropriately reduced or the heating rate adjusted during resuming sealing, making the sealing process safer and more stable. If the protection mechanism is triggered multiple times, the sealing machine can use internal strategies to determine if there is material incompatibility, abnormal temperature sensing, or abnormal heating element, and will prompt for inspection. After the temperature returns to normal and the sealing action is re-executed, the sealing process can be completed smoothly under safe conditions, thereby ensuring sealing quality and material integrity.

[0037] In this embodiment, the specific steps for acquiring an overall image of the packaged material, identifying defects in the overall material image, and outputting a defect distribution report are as follows: Acquire an overall image of the packaged material; perform global gamma correction on the overall material image and extract a brightness-optimized image; Perform depth visual recognition on the brightness-optimized image to mark the location of sealing defects; The location of the sealing defect is classified at the pixel level to generate a defect type; the defect type includes internal unfused areas, microcracks and air bubbles. The number and location distribution of the defects are counted, and a defect distribution report is output.

[0038] In this embodiment, after packaging, a comprehensive image of the sealed material needs to be acquired for subsequent sealing defect analysis. The acquired image covers the sealing area and its surrounding material to ensure complete capture of sealing quality characteristics. A high-resolution industrial camera is typically used, paired with a diffuse light source or a ring-shaped LED array to uniformly illuminate the material surface, avoiding areas that are too bright or too dark and reducing specular reflection interference. During shooting, the camera maintains a fixed distance from the material to ensure consistent image scale, allowing the sealing length, width, and area distribution to be accurately mapped in the image. To enhance the reliability of defect detection, multi-band acquisition, including visible light and near-infrared spectroscopy, can be performed, which helps distinguish low-contrast defects such as microcracks, unfused areas, and air bubbles. The acquired images are saved in a two-dimensional matrix format, containing complete pixel brightness and spatial resolution information, providing basic data for subsequent image processing, brightness optimization, and depth vision recognition. The acquired overall material image often suffers from uneven brightness, insufficient contrast, or loss of local details, which affects the accuracy of subsequent defect identification. Therefore, global gamma correction is required to optimize brightness and enhance visual contrast. Gamma correction is a non-linear brightness transformation method that adjusts pixel grayscale values ​​according to the gamma index to enhance details in dark areas while avoiding overexposure in bright areas. For example, the gamma index can be set in the range of 0.8–1.2, selecting the most suitable index value based on the material's surface reflectivity to achieve a more uniform overall brightness distribution in the image. After gamma correction, the brightness contrast of sealing edges, micro-cracks, unfused areas, and bubble locations in the image is significantly enhanced, allowing subsequent depth vision algorithms to more accurately identify subtle defects. Global correction ensures consistent brightness across the entire sealing area, avoiding recognition errors caused by localized overexposure or underexposure.

[0039] Based on a brightness-optimized image, a deep vision recognition method is used to detect defects in the sealing area. Deep vision recognition typically employs convolutional neural networks (CNNs), multi-scale feature extraction or segmentation networks, such as U-Net and Mask R-CNN, to identify abnormal structures in the sealing area. The model input is the brightness-optimized image, which extracts local texture and shape features through convolutional layers and achieves sensitive capture of minute defects through multi-layer feature fusion. During the recognition process, the network automatically generates a defect heatmap or binary mask to mark potentially defective areas. For example, unfused areas typically appear as regions with low local brightness and smooth boundaries, microcracks appear as thin, elongated linear structures, and air bubbles appear as circular or elliptical bright areas. Through deep vision recognition, the location of sealing defects in the image can be accurately located, providing coordinate and regional information for subsequent pixel-level classification and defect type identification.

[0040] After marking the defect locations, pixel-level classification is required to determine the defect type. Defect types mainly include internal unfused regions, microcracks, and inclusions, each with different image characteristics. Internal unfused regions typically appear as low-brightness patches with strong regional connectivity; microcracks present as continuous or discontinuous fine line structures; and inclusions are mostly circular or elliptical bright spots. Pixel-level classification can be achieved using a multi-channel convolutional network, mapping each pixel to a specific defect category to obtain an accurate defect type map. During classification, color, brightness gradient, texture directionality, edge intensity, and morphological features can be used as auxiliary discrimination parameters to improve the recognition accuracy of microcracks and detailed edges. For example, the width of a microcrack may only be 2–5 pixels; pixel-level classification can prevent it from being ignored in coarse-grained detection.

[0041] After completing pixel-level defect classification, it is necessary to perform quantity statistics and spatial distribution analysis on the defects to generate a complete defect distribution report. The statistics include the quantity, area, length, width, and coordinate position of each defect type within the sealing area. For internal unfused areas, the area of ​​each patch can be calculated; for microcracks, the length and direction distribution can be calculated; and for inclusions, the quantity and diameter distribution can be calculated. Location distribution analysis maps defects to a two-dimensional coordinate system of the sealing area, generating heat maps or distribution density maps to show areas of concentrated or sparse defects. The defect distribution report is not only used for sealing quality assessment but also provides a reference for optimizing sealing process parameters. For example, if unfused areas are concentrated at the sealing edge, it may indicate insufficient heating temperature or uneven pressurization; if microcracks are concentrated in the central area, it may be necessary to optimize pressurization time or pressure uniformity; an excessive distribution of inclusions may reflect insufficient vacuum or gas content within the material.

[0042] In this embodiment, the specific steps for optimizing packaging parameters and performing secondary packaging based on the defect distribution report are as follows: Based on the defect distribution report, a correlation analysis between defect type and packaging parameters is performed to obtain defect correlation data; Based on defect correlation data, encapsulation parameters are tuned to obtain adaptive optimization parameters; Secondary encapsulation processing is performed based on adaptive optimization parameters.

[0043] In this embodiment, after obtaining the defect distribution report, it is necessary to perform correlation analysis between the sealing defect types and packaging process parameters to identify the potential causes of the defects. Defect correlation analysis mainly involves statistically comparing the quantity, area, length, and location distribution of different types of defects (such as internal unfused areas, microcracks, and embedded bubbles) with the initial packaging process parameters (sealing temperature, sealing pressure, heating time, pressure holding time, vacuum level, etc.). By analyzing the spatial distribution patterns and typological characteristics of defects in the sealing area, the correlation between specific defects and specific parameter anomalies can be inferred. For example, unfused areas are usually concentrated at the sealing edge, which may be related to insufficient temperature at the edge of the heating plate or uneven pressure; microcracks are mostly distributed along the sealing centerline, which may be related to excessively fast heating speed or insufficient pressure holding time; embedded bubbles are concentrated in the local area of ​​the sealing area, which may reflect insufficient vacuum or insufficient gas removal from the membrane material. After obtaining the defect correlation data, it is necessary to perform targeted optimization of the packaging parameters to reduce the occurrence of defects and improve sealing quality. The parameter tuning process analyzes the sensitivity of different defect types to temperature, pressure, heating time, pressure holding time, and vacuum level to determine the direction and magnitude of adjustments. For example, for defects with unfused edges, the heating temperature at the sealing edge can be appropriately increased or the heating time extended; for areas with frequent microcracks, pressure uniformity or pressure holding time can be adjusted; and for bubble inclusion problems, optimization may be achieved by increasing the vacuum level or extending the vacuuming time. Optimization parameters are typically generated adaptively, combining defect type, quantity, and distribution location to calculate adjustment values ​​for each sealing region, allowing for dynamic fine-tuning of each packaging process under specific conditions. These adaptive optimization parameters include not only sealing temperature, pressure, heating time, and vacuum level, but also local temperature compensation, pressure gradient control, and sealing sequence optimization, thus forming a closed-loop control strategy.

[0044] After obtaining the adaptive optimization parameters, a secondary encapsulation process is required to repair or reduce defects generated during the initial encapsulation. The execution method of the secondary encapsulation depends on the specific content of the optimization parameters, including temperature increases, pressure adjustments, extended heating times, and local vacuum optimization for specific sealing areas. During the process, the sealing equipment readjusts the heating plate temperature distribution, pressure, and holding time according to the adaptive optimization parameters to ensure the sealing layer material is fully fused in a thermally softened state, eliminating unfused areas. Appropriate adjustments to the vacuum level can eliminate air bubbles and reduce internal voids in the material. The key to the secondary encapsulation process lies in targeted repair of the location and type of defects, rather than simply repeating the initial encapsulation process. This approach ensures a significant improvement in sealing quality and avoids damage to the material caused by overheating or excessive pressure. After the secondary encapsulation is completed, the sealed area can be verified by acquiring overall images and performing defect detection again, thus forming a complete closed-loop defect repair process.

[0045] In this embodiment, a vacuum sealing machine control device is provided for performing the method described above, including: The spectral analysis unit is used to acquire full-band spectral images of the item to be packaged based on a spectral sensor; and to perform spectral analysis based on the full-band spectral images to generate a spectral feature set. The material comparison unit is used to calculate the similarity of standard material spectral templates in a pre-stored database based on the spectral feature set to determine the actual material type. The parameter calculation unit is used to analyze the packaging parameters based on the actual material type and output real-time packaging process parameters. The packaging control unit is used to generate control commands based on the real-time packaging process parameters; and to perform real-time control of the sealing machine according to the control commands to complete the sealing process.

[0046] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the vacuum sealing machine control method described in any of the above claims.

[0047] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the vacuum sealing machine control method described in any of the preceding claims.

[0048] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0049] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement it. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein are implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A control method for a vacuum sealing machine, characterized in that, Includes the following steps: Step S1: Acquire full-band spectral images of the item to be packaged based on a spectral sensor; perform spectral analysis based on the full-band spectral images to generate a spectral feature set; Step S2: Calculate the similarity of standard material spectral templates in the pre-stored database based on the spectral feature set to determine the actual material type; Step S3: Analyze the packaging parameters based on the actual material type and output the real-time packaging process parameters; Step S4: Generate control instructions based on the real-time packaging process parameters; perform real-time control of the sealing machine according to the control instructions to complete the sealing process.

2. The vacuum sealing machine control method according to claim 1, characterized in that, The specific steps of step S1 are as follows: Full-band spectral images of the items to be packaged are acquired using a spectral sensor; Wavelet denoising and sharpening filtering are performed on the full-band spectral image to obtain a sharpened and enhanced spectral image; Perform spectral texture fluctuation analysis on the sharpened and enhanced spectral image to generate texture fluctuation features; The texture fluctuation features include texture roughness, directionality, and contrast parameters; Perform band-by-band difference operations on the sharpened and enhanced spectral image to generate regional gradient characteristics; Based on the sharpened and enhanced spectral image, the ratio of specular reflection to diffuse reflection at different incident angles is calculated to obtain the spectral reflectance. Principal component analysis was performed on spectral reflectance, regional gradation characteristics, and texture fluctuation features to generate a spectral feature set.

3. The vacuum sealing machine control method according to claim 1, characterized in that, The specific steps of step S2 are as follows: Obtain standard material spectral templates from the pre-stored database; Extract the feature vector of the standard material spectral template; Similarity calculations are performed on the feature vectors based on the spectral feature set to extract similarity between multiple standard materials; Material confidence is determined by assessing the similarity of multiple standard materials to identify the actual material type.

4. The vacuum sealing machine control method according to claim 1, characterized in that, Step S3 is as follows: The sealing location region is identified based on full-band spectral images; the sealing length, sealing width, and sealing area of ​​the sealing location region are calculated to obtain the geometric parameters of the sealing region; Analyze the thermal properties of the material based on the actual material type; calculate the heating time and temperature distribution based on the thermal properties of the material, and generate sealing temperature parameters; Extract the yield strength and elastic modulus of the material based on the actual material type; The sealing time and sealing pressure are determined based on the yield strength and elastic modulus of the material and the geometric parameters of the sealing area, thus obtaining the sealing pressure parameters; Identify the type of the item to be packaged; adjust the adaptive vacuum level according to the item type to obtain the adaptive vacuum level; Real-time packaging process parameters are output based on adaptive vacuum degree, sealing pressure parameters, and sealing temperature parameters.

5. The vacuum sealing machine control method according to claim 1, characterized in that, The specific steps of step S4 are as follows: Control commands are generated based on the real-time packaging process parameters; The sealing machine is controlled in real time according to the control instructions, and the temperature parameters of the sealing area are monitored in real time. Set a safe temperature threshold for sealing; The temperature parameters of the sealing area are monitored throughout the process according to the sealing safety temperature threshold. When the temperature parameters of the sealing area do not exceed the sealing safety temperature threshold, the sealing operation continues to be performed to complete the sealing process. When the temperature parameter of the sealing area is detected to be higher than the sealing safety temperature threshold, the protection mechanism is triggered; Based on the aforementioned protection mechanism, the sealing operation is stopped and allowed to cool naturally. Once the temperature of the sealing area is detected to be below the sealing safety temperature threshold, the sealing operation is performed again to complete the sealing process. Acquire an overall image of the packaged material; perform defect identification on the overall material image and output a defect distribution report; Based on the defect distribution report, optimize the packaging parameters and perform secondary packaging.

6. The vacuum sealing machine control method according to claim 5, characterized in that, The specific steps for acquiring and packaging the overall image of the material, identifying defects in the overall image of the material, and outputting a defect distribution report are as follows: Acquire an overall image of the packaged material; perform global gamma correction on the overall material image and extract a brightness-optimized image; Perform depth visual recognition on the brightness-optimized image to mark the location of sealing defects; The location of the sealing defect is classified at the pixel level to generate a defect type; the defect type includes internal unfused areas, microcracks and air bubbles. The number and location distribution of the defects are counted, and a defect distribution report is output.

7. The vacuum sealing machine control method according to claim 6, characterized in that, The specific steps for optimizing packaging parameters and performing secondary packaging based on the defect distribution report are as follows: Based on the defect distribution report, a correlation analysis between defect type and packaging parameters is performed to obtain defect correlation data; Based on defect correlation data, encapsulation parameters are tuned to obtain adaptive optimization parameters; Secondary encapsulation processing is performed based on adaptive optimization parameters.

8. A control device for a vacuum sealing machine, characterized in that, For performing the vacuum sealing machine control method as described in claim 1, comprising: The spectral analysis unit is used to acquire full-band spectral images of the item to be packaged based on a spectral sensor; and to perform spectral analysis based on the full-band spectral images to generate a spectral feature set. The material comparison unit is used to calculate the similarity of standard material spectral templates in a pre-stored database based on the spectral feature set to determine the actual material type. The parameter calculation unit is used to analyze the packaging parameters based on the actual material type and output real-time packaging process parameters. The packaging control unit is used to generate control commands based on the real-time packaging process parameters; and to perform real-time control of the sealing machine according to the control commands to complete the sealing process.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the vacuum sealing machine control method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the vacuum sealing machine control method according to any one of claims 1 to 7.

Images