SMP sintering regulation method and system based on image monitoring

Abstract

The embodiment of the application provides a kind of based on image monitoring's SMP sintering regulation method and system, belong to industrial process control technical field.The method includes: obtaining the time sequence image data of SMP blank in sintering process, and based on time sequence image data extraction is used to characterize the image state parameter of blank densification state;Based on image state parameter, construct synchronous densification evaluation model, and utilize synchronous densification evaluation model to calculate the synchronous densification index and mismatch risk level corresponding to each monitoring time;Based on synchronous densification index and mismatch risk level, combine current sintering stage parameter, determine the sintering regulation parameter matched with current blank densification state;Based on sintering regulation parameter, execute parameter revision to subsequent sintering process, and based on revised time sequence image data, update image state parameter and synchronous densification evaluation model.The application realizes the closed-loop regulation of sintering process based on image perception, reduces the mismatch risk of blank densification and improves structural consistency.

Description

Technical Field

[0001] This invention relates to the field of industrial process control technology, and specifically to an SMP sintering control method and system based on image monitoring. Background Technology

[0002] In the manufacturing process of RF connectors, some key structures are increasingly being completed using sinterable forming processes to reduce assembly errors and improve structural consistency. Especially with the trend towards integrated multi-component designs, achieving the coordinated forming of conductive and supporting structures through sintering has become a common technical approach. In actual production, the SMP (Sinterable Microwave Connector Part) blank undergoes multiple processes during the sintering stage, including heating, activation shrinkage, and densification stabilization. The sintering process is typically performed according to a preset temperature profile and atmospheric conditions.

[0003] Based on the on-site situation, existing control methods mainly rely on process parameters such as furnace temperature and atmosphere flow rate for adjustment. These parameters can reflect the overall environment inside the furnace, but cannot directly reflect the densification state of the billet itself. Especially during the critical densification stage, the shrinkage of different areas of the billet is not synchronized, easily resulting in the edges shrinking first and the center lagging behind in densification. This difference may not manifest as an obvious defect in the early stages, but it will gradually accumulate into dimensional deviations or structural asymmetries. For RF connectors, this will further cause impedance fluctuations and unstable contact performance.

[0004] Currently, there are also practices that monitor the sintering process using visual methods, but most of these focus on defect identification or post-processing inspection, primarily used for screening out substandard products, rather than participating in process control. Furthermore, image information is easily affected by atmospheric fluctuations, lighting changes, and slight deviations in the shape of the green body in the high-temperature sintering environment. Directly using this information for condition judgment can easily lead to misjudgments, thus affecting the accuracy of control.

[0005] Therefore, how to stably extract feature quantities that reflect the true densification state of the green body based on image information during the sintering process, and transform these feature quantities into control basis that can be used to adjust sintering parameters, so as to realize real-time closed-loop control of the sintering process, is a problem that needs to be solved. Summary of the Invention

[0006] The purpose of this invention is to provide an image-based SMP sintering control method and system, so as to at least solve the problem that existing sintering processes cannot accurately sense the densification state of the green body and adjust parameters accordingly, resulting in shrinkage mismatch and performance instability.

[0007] To achieve the above objectives, a first aspect of the present invention provides an SMP sintering control method based on image monitoring. The method includes: acquiring time-series image data of an SMP billet during the sintering process, and extracting image state parameters characterizing the densification state of the billet based on the time-series image data; constructing a synchronous densification evaluation model based on the image state parameters, and calculating a synchronous densification index and a mismatch risk level corresponding to each monitoring moment using the synchronous densification evaluation model; determining sintering control parameters matching the current densification state of the billet based on the synchronous densification index and the mismatch risk level, combined with parameters of the current sintering stage; performing parameter correction on subsequent sintering processes based on the sintering control parameters, and updating the image state parameters and the synchronous densification evaluation model based on the corrected time-series image data.

[0008] Preferably, acquiring time-series image data of the SMP billet during the sintering process includes: acquiring original image frames of the SMP billet at preset monitoring positions of the sintering equipment corresponding to different sintering stages, and continuously acquiring the original image frames based on preset sampling time intervals to form an original image sequence corresponding to the sintering process; performing timestamp calibration and stage identifier association processing on each original image frame in the original image sequence to establish a correspondence between each original image frame and the corresponding sintering time parameters and sintering stage parameters, thereby obtaining a time-series image sequence with stage identifiers.

[0009] Preferably, the extraction of image state parameters for characterizing the densification state of the billet based on the time-series image data includes: performing contour extraction processing on each image frame in the time-series image sequence with stage identifiers to obtain the billet contour boundary corresponding to each monitoring time, and determining the corresponding contour size parameter and contour area parameter based on the billet contour boundary; dividing each image frame into regions based on the billet contour boundary to obtain edge regions and center regions, and extracting grayscale distribution parameters of the edge regions and the center regions respectively to form grayscale feature parameters for characterizing regional densification differences; performing texture feature extraction processing on the target region defined by the billet contour boundary to obtain texture energy parameters and texture contrast parameters corresponding to each monitoring time, and constructing texture feature parameters for characterizing surface structure changes based on the texture energy parameters and the texture contrast parameters; and combining the contour size parameter, the contour area parameter, the grayscale feature parameter, and the texture feature parameter to construct image state parameters corresponding to each monitoring time.

[0010] Preferably, constructing a synchronous densification evaluation model based on the image state parameters includes: determining an directional imbalance parameter based on the contour size parameter and the contour area parameter; determining a densification difference parameter based on the grayscale feature parameter to characterize the densification difference between the edge region and the center region; determining a texture compression parameter based on the texture feature parameter to characterize the surface structure change; and constructing a synchronous densification evaluation model using the directional imbalance parameter, the densification difference parameter, and the texture compression parameter.

[0011] Preferably, determining the directional imbalance parameter based on the contour size parameter and the contour area parameter includes: determining the shrinkage ratio corresponding to different directions based on the contour size parameter, and constructing a directional deviation amount to characterize the shrinkage difference in each direction based on the shrinkage ratio, so as to determine the directional imbalance parameter based on the directional deviation amount; determining the densification difference parameter to characterize the densification difference between the edge region and the center region based on the grayscale feature parameter includes: determining the average grayscale value of the edge region and the center region respectively based on the grayscale feature parameter, and constructing a grayscale difference value to characterize the grayscale difference in the region based on the average grayscale value, so as to determine the densification difference parameter based on the grayscale difference; determining the texture compression parameter to characterize the surface structure change based on the texture feature parameter includes: determining the texture energy value and texture contrast value corresponding to each monitoring time based on the texture feature parameter, and constructing a texture change amount to characterize the degree of texture change based on the texture energy value and the texture contrast value, so as to determine the texture compression parameter based on the texture change amount.

[0012] Preferably, the synchronous densification index and mismatch risk level for each monitoring time are calculated using the synchronous densification evaluation model, including: introducing the directional imbalance parameter, the densification difference parameter, and the texture compression parameter as input parameters into the synchronous densification evaluation model, and performing normalization processing on the input parameters; performing weighted fusion processing on each normalized input parameter based on preset weights to obtain the synchronous densification index for each monitoring time, and using the synchronous densification index as an evaluation quantity characterizing the overall densification synchronization degree of the billet; determining the synchronous densification deviation for each monitoring time based on the deviation between the synchronous densification index and the preset synchronous densification interval, and performing a comprehensive judgment on the densification state of the current billet in combination with the changing trends of the directional imbalance parameter and the densification difference parameter; and classifying the densification state at the current monitoring time into different levels of mismatch risk based on the synchronous densification deviation and the comprehensive judgment result.

[0013] Preferably, based on the synchronous densification index and the mismatch risk level, and combined with the current sintering stage parameters, sintering control parameters matching the current compaction state of the green body are determined, including: determining the compaction deviation amount corresponding to the current monitoring time according to the deviation degree between the synchronous densification index and the preset synchronous densification interval, and performing a graded mapping process on the compaction deviation amount in combination with the mismatch risk level to obtain the corresponding control intensity parameter; calculating the corresponding heating rate adjustment amount, holding time adjustment amount, and atmosphere adjustment amount based on the control intensity parameter and the current sintering stage parameters, respectively, to form a set of sintering control parameters matching the current compaction state of the green body; and constraining the set of sintering control parameters with the preset sintering process parameter range to obtain sintering control parameters that meet the process constraint conditions.

[0014] Preferably, the parameter correction for subsequent sintering processes based on the sintering control parameters includes: inputting the sintering control parameters into the control interface of the sintering equipment, and parsing the sintering control parameters into temperature adjustment commands and atmosphere adjustment commands corresponding to the current sintering stage; correcting the heating rate and holding time in subsequent sintering processes based on the temperature adjustment commands, and correcting the atmosphere flow rate and atmosphere composition in subsequent sintering processes based on the atmosphere adjustment commands, to form corrected sintering operating parameters; and controlling subsequent sintering processes according to the corrected sintering operating parameters.

[0015] Preferably, updating the image state parameters and the synchronous density evaluation model based on the corrected time-series image data includes: re-extracting the contour size parameters, contour area parameters, grayscale feature parameters, and texture feature parameters corresponding to each monitoring time based on the corrected time-series image data, and performing normalization processing on each re-extracted parameter to form updated image state parameters; inputting the updated image state parameters into the synchronous density evaluation model, recalculating the synchronous density index corresponding to each monitoring time, and correcting the parameter weights in the synchronous density evaluation model based on the change relationship between the recalculated synchronous density index and the historical synchronous density index; and updating the synchronous density evaluation model based on the corrected model parameter weights.

[0016] A second aspect of the present invention provides an SMP sintering control system based on image monitoring. The system includes: an image acquisition unit, configured to acquire time-series image data of an SMP billet during the sintering process, and extract image state parameters characterizing the densification state of the billet based on the time-series image data; a risk level determination unit, configured to construct a synchronous densification evaluation model based on the image state parameters, and calculate the synchronous densification index and mismatch risk level corresponding to each monitoring time using the synchronous densification evaluation model; a control parameter generation unit, configured to determine sintering control parameters matching the current densification state of the billet based on the synchronous densification index and the mismatch risk level, combined with the parameters of the current sintering stage; and a control execution unit, configured to correct the execution parameters of subsequent sintering processes based on the sintering control parameters, and update the image state parameters and the synchronous densification evaluation model based on the corrected time-series image data.

[0017] Based on this technical embodiment, the present invention continuously acquires time-series image data during the SMP green body sintering process and extracts image state parameters that characterize the densification state of the green body, enabling direct perception of the structural evolution of the green body during sintering. Furthermore, a synchronous densification evaluation model is constructed based on the image state parameters to obtain a synchronous densification index and a mismatch risk level, thereby transforming the originally difficult-to-quantify densification process into a calculable evaluation quantity. Based on this evaluation quantity and sintering stage parameters, corresponding sintering control parameters are determined, and parameter corrections are performed on subsequent sintering processes, transforming the sintering process from a fixed process into a dynamic control process that matches the actual state of the green body. Simultaneously, the model is continuously updated using the corrected image data, forming a closed-loop control mechanism, thereby reducing the risk of green body shrinkage mismatch and improving the stability and forming consistency of the sintering process.

[0018] Other features and advantages of the embodiments of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0019] The accompanying drawings are provided to further illustrate embodiments of the present invention and form part of the specification. They are used together with the following detailed description to explain the embodiments of the present invention, but do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of the steps of the SMP sintering control method based on image monitoring provided by the present invention; Figure 2 This is a schematic diagram of SMP billet densification feature extraction based on time-series images provided by the present invention; Figure 3 This is a system structure diagram of the SMP sintering control system based on image monitoring provided by the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for illustration and explanation of the embodiments of this application and are not intended to limit the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0021] It should be noted that if the embodiments of this application involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicators will also change accordingly.

[0022] Furthermore, if the embodiments of this application involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed in this application.

[0023] Figure 1 This invention provides an image monitoring-based SMP sintering control method, comprising: Step S1: Obtain time-series image data of the SMP green body during the sintering process, and extract image state parameters to characterize the densification state of the green body based on the time-series image data.

[0024] Specifically, acquiring time-series image data of the SMP billet during the sintering process includes: acquiring original image frames of the SMP billet at preset monitoring positions in the sintering equipment corresponding to different sintering stages, and continuously acquiring the original image frames based on preset sampling time intervals to form an original image sequence corresponding to the sintering process; performing timestamp calibration and stage identifier association processing on each original image frame in the original image sequence to establish a correspondence between each original image frame and the corresponding sintering time parameters and sintering stage parameters, thereby obtaining a time-series image sequence with stage identifiers.

[0025] Furthermore, the extraction of image state parameters for characterizing the densification state of the billet based on the time-series image data includes: performing contour extraction processing on each image frame in the time-series image sequence with stage identifiers to obtain the billet contour boundary corresponding to each monitoring time, and determining the corresponding contour size parameter and contour area parameter based on the billet contour boundary; dividing each image frame into regions based on the billet contour boundary to obtain edge regions and center regions, and extracting grayscale distribution parameters of the edge regions and the center regions respectively to form grayscale feature parameters for characterizing regional densification differences; performing texture feature extraction processing on the target region defined by the billet contour boundary to obtain texture energy parameters and texture contrast parameters corresponding to each monitoring time, and constructing texture feature parameters for characterizing surface structure changes based on the texture energy parameters and the texture contrast parameters; and combining the contour size parameter, the contour area parameter, the grayscale feature parameter, and the texture feature parameter to construct image state parameters corresponding to each monitoring time.

[0026] In this embodiment of the invention, the SMP billet is placed inside the sintering equipment and undergoes preheating, activation shrinkage, and densification stabilization stages sequentially during the sintering process. Considering the significant differences in the magnitude and rate of morphological changes in the billet at different stages, an image acquisition device, such as a high-temperature observation window combined with an industrial camera, is arranged at a preset monitoring location within the sintering equipment. This preset monitoring location is typically selected in an area with a relatively stable and unobstructed temperature field, ensuring that the acquired images stably reflect the surface state of the billet. During the acquisition process, continuous high-speed sampling is not required; instead, discrete acquisition is performed based on a preset sampling time interval, for example, acquiring image frames at a fixed time interval of Δt. Δt can be set according to the sintering stage, taking a smaller value during the shrinkage stage and appropriately widening it during the stabilization stage.

[0027] Based on this, the original image sequence is formed. ,in, Indicates the first Image frames corresponding to each monitoring time. This represents the total number of sampled frames. To ensure the traceability of the image data, each frame is timestamped, and this time information is correlated with process parameters during sintering, such as the current furnace temperature, atmosphere, and sintering stage. Correspondingly, each frame can be expanded into a triplet format. ,in, Indicates time parameter, This indicates the sintering stage identifier. Through the above processing, the original image sequence is transformed into a time-series image sequence with stage identifiers, providing a basis for subsequent stage-by-stage feature extraction.

[0028] After obtaining the time-series image data, the image state parameter extraction process begins. This process constructs physically meaningful feature representations based on the compaction behavior of the billet. First, contour extraction is performed on each image frame. Contour extraction can employ a combination of edge detection and morphological processing to obtain closed contour boundaries. For example, edges can be extracted using the Canny operator, and gaps can be filled using a closing operation to obtain the billet contour set. in, Indicates the first The set of contour points at time. This represents the number of contour points. Based on this contour, geometric parameters such as the circumscribed rectangle size or equivalent diameter can be calculated to obtain the contour dimension parameters. And obtain the contour area parameters through integration: in, This represents the area enclosed by the contour. This type of parameter reflects the overall shrinkage behavior of the billet and is an important macroscopic characterization of densification.

[0029] After the contour is determined, the image is divided into regions based on this contour. Typically, the region within the contour is divided into edge regions and a central region. The edge regions are obtained by offsetting the contour inwards by a fixed pixel distance, while the central region is the remaining area. This division method can stably distinguish densification features at different locations without relying on complex partitioning algorithms. The gray-level distribution of each region is statistically analyzed, and the average gray-level value is calculated. in, and These represent the edge region and the central region, respectively. This corresponds to the number of pixels. This forms a grayscale feature parameter used to reflect the densification differences in different regions of the blank. In practice, it can be observed that as sintering proceeds, the edge regions often undergo densification changes first, and their grayscale response changes earlier than that of the central region. This difference is an important basis for subsequent evaluation.

[0030] After grayscale feature extraction, texture feature extraction is performed on the contour-defined target region. Texture feature extraction can employ methods such as gray-level co-occurrence matrix (GLCM) or local binary mode (LOM), aiming to characterize the changing trends of the microstructure on the surface of the blank. Taking GLCM as an example, texture energy and contrast can be calculated: in, These are elements in the gray-level co-occurrence matrix. Energy reflects the uniformity of the texture, while contrast reflects the drastic change in gray level. As the densification process progresses, the pores gradually close, and the texture distribution tends to become more uniform; the above parameters exhibit a identifiable pattern of change. Based on this pattern, texture feature parameters can be constructed to characterize the evolution of the surface structure from loose to dense.

[0031] After extracting contour features, grayscale features, and texture features, the various parameters are organized uniformly. Specifically, the contour size parameters are... Contour area parameters Gray-scale feature parameters and texture feature parameters Combine them to construct an image state parameter vector: This vector represents the image state parameters at the corresponding monitoring time. In this way, image information from different sources is uniformly mapped to the same parameter space, enabling subsequent model processing to be based on a unified input.

[0032] It should be noted that the above parameter selection is not the only limitation. In practical applications, the contour parameters, grayscale parameters, or texture parameters can be extended or replaced according to the specific structural characteristics of the SMP blank. For example, for connector structures with obvious directional characteristics, the directional projection length can be added as a supplementary parameter; for blanks with more complex surface structures, multi-scale texture description can be introduced. All of the above changes fall within the scope of protection of this application.

[0033] In one specific implementation, such as Figure 2 Images of the SMP billet are acquired at preset monitoring positions in the sintering equipment to obtain original billet images at each monitoring time. Based on these original images, contour extraction is first performed to obtain the outer boundary of the billet, and contour size and area parameters are calculated accordingly. The billet region is divided into edge and center regions using the contour boundary as a constraint, and the grayscale distribution in each region is extracted to calculate the corresponding average grayscale value, which reflects the densification differences in different regions. Simultaneously, a local area on the billet surface is selected to construct a texture extraction region, and texture feature analysis is performed on this region to obtain texture energy and contrast values ​​to characterize surface structure changes. Finally, the contour size, contour area, grayscale, and texture feature parameters are combined to construct the image state parameters corresponding to each monitoring time.

[0034] Step S2: Construct a synchronous density evaluation model based on the image state parameters, and use the synchronous density evaluation model to calculate the synchronous density index and mismatch risk level for each monitoring time.

[0035] Specifically, constructing a synchronous densification evaluation model based on the image state parameters includes: determining an directional imbalance parameter based on the contour size parameter and the contour area parameter; determining a densification difference parameter based on the grayscale feature parameter to characterize the densification difference between the edge region and the center region; determining a texture compression parameter based on the texture feature parameter to characterize the surface structure change; and constructing a synchronous densification evaluation model using the directional imbalance parameter, the densification difference parameter, and the texture compression parameter.

[0036] Further, determining directional imbalance parameters based on the contour size parameters and the contour area parameters includes: determining the shrinkage ratio corresponding to different directions based on the contour size parameters, and constructing a directional deviation amount to characterize the shrinkage difference in each direction based on the shrinkage ratio, so as to determine the directional imbalance parameters based on the directional deviation amount; determining a densification difference parameter to characterize the densification difference between the edge region and the center region based on the grayscale feature parameters includes: determining the average grayscale value of the edge region and the center region respectively based on the grayscale feature parameters, and constructing a grayscale difference value to characterize the grayscale difference in the region based on the average grayscale value, so as to determine the densification difference parameters based on the grayscale difference; determining a texture compression parameter to characterize the surface structure change based on the texture feature parameters includes: determining the texture energy value and texture contrast value corresponding to each monitoring time based on the texture feature parameters, and constructing a texture change amount to characterize the degree of texture change based on the texture energy value and the texture contrast value, so as to determine the texture compression parameter based on the texture change amount.

[0037] In this embodiment of the invention, the image state parameters have been constructed into vector form in step S1. Although these parameters are all related to the densification process, their expression dimensions are different. Directly using them for judgment is prone to bias. Therefore, it is necessary to further extract intermediate quantities with clear engineering significance.

[0038] Orientation imbalance parameters are constructed based on the contour dimension parameters and contour area parameters. During sintering, the shrinkage of the blank in different directions is not entirely uniform. Especially in RF connector structures, shrinkage deviation in one direction directly affects the dimensional stability of the coaxial structure or contact interface. Therefore, in this embodiment, the contour dimension parameters are decomposed into shrinkage ratios in different directions. Using the initial dimensions... For reference, regarding the first Normalizing the dimension parameters at each moment yields: in, and These represent the contraction ratios in two orthogonal directions. The directional deviation is constructed based on these contraction ratios. This directional deviation reflects the degree of consistency in contraction in different directions. When the value is close to zero, it indicates that the shrinkage of the billet in all directions is relatively synchronized; conversely, it indicates that there is a significant directional mismatch. Based on this, the directional deviation is used as the basic expression of the directional imbalance parameter, and it can be further normalized according to specific application scenarios, such as scaling by setting a maximum deviation threshold to form a directional imbalance parameter with uniform dimensions.

[0039] Subsequently, densification difference parameters are constructed based on grayscale feature parameters. Grayscale features typically correspond to the surface densification degree of the material during sintering. Due to differences in thermal conduction conditions and atmosphere contact states between the edge and center regions, their densification processes are not synchronized. Therefore, by comparing the grayscale changes between the edge and center regions, a quantitative expression of the densification difference between regions can be obtained. Specifically, the average grayscale values ​​are calculated for the edge and center regions respectively: in, and These represent the edge region and the central region, respectively. This corresponds to the number of pixels. A grayscale difference is constructed based on the above average grayscale value: This grayscale difference reflects the degree of densification variation between different regions. As sintering progresses, if the edge regions densify first, then... and A considerable difference will be generated between them. By normalizing this difference, a dense difference parameter can be obtained, which will have consistent dimensions with other parameters in the subsequent fusion process.

[0040] Furthermore, texture compression parameters are constructed based on texture feature parameters. Texture features mainly reflect the changes in the microstructure of the blank surface. During sintering, pores gradually close, and the internal structure of the material tends to become denser, which is usually represented in the image as a transformation of texture from rough to uniform. To quantify this change, texture change can be constructed based on texture energy and texture contrast. For example, it can be calculated using the gray-level co-occurrence matrix: in, This refers to the corresponding element in the gray-level co-occurrence matrix. Texture energy. Reflects the concentration of grayscale distribution and texture contrast. It reflects the degree of drastic change in grayscale. As densification progresses, it typically manifests as... Gradually increase Gradually decrease. Based on this, construct the texture variation, for example: in, This corresponds to the initial state value. This is a weighting coefficient. This texture variation is used to characterize the degree of texture compression and participates in subsequent model construction as a texture compression parameter.

[0041] After obtaining the directional imbalance parameter, density difference parameter, and texture compression parameter, the process of constructing the synchronous density evaluation model begins. These three parameters describe the billet state from three dimensions: macroscopic shrinkage consistency, regional density difference, and microstructural change. Their physical meanings are complementary, but their dimensions are different, thus requiring unified processing. In this embodiment, normalization is performed on each parameter to unify its value range to the [0,1] interval, denoted as... Normalization can be performed based on historical data or a preset range, for example, using linear scaling or piecewise functions. After normalization, a synchronous density index is formed through weighted fusion: in, For the weighting coefficients, satisfying In this expression, the orientation imbalance parameter and the density difference parameter are expressed as a subtraction of 1 because the larger their values, the more severe the mismatch. The texture compression parameter, on the other hand, directly reflects the degree of density, and the larger the value, the closer it is to a dense state.

[0042] Synchronous density index As a model output, it reflects the degree of synchronization of the overall densification of the billet at the current moment. The closer the value is to 1, the more consistent the changes in the structure of each region, direction, and surface are; a lower value indicates that there is a risk of mismatch to varying degrees.

[0043] Furthermore, the synchronous densification evaluation model is used to calculate the synchronous densification index and mismatch risk level for each monitoring time, including: introducing the directional imbalance parameter, the densification difference parameter, and the texture compression parameter as input parameters into the synchronous densification evaluation model, and performing normalization processing on the input parameters; performing weighted fusion processing on each normalized input parameter based on preset weights to obtain the synchronous densification index for each monitoring time, and using the synchronous densification index as an evaluation quantity characterizing the overall densification synchronization degree of the billet; determining the synchronous densification deviation for each monitoring time based on the deviation between the synchronous densification index and the preset synchronous densification interval, and performing a comprehensive judgment on the densification state of the current billet in combination with the changing trends of the directional imbalance parameter and the densification difference parameter; and classifying the densification state at the current monitoring time into different levels of mismatch risk based on the synchronous densification deviation and the comprehensive judgment result.

[0044] In this embodiment of the invention, the directional imbalance parameter, the density difference parameter, and the texture compression parameter are introduced as input parameters into the synchronous density evaluation model. Since these parameters originate from geometric contours, region grayscale, and texture structure, respectively, their dimensions and value ranges differ. If they are directly involved in the fusion calculation, it is easy for one type of parameter to dominate the result. Therefore, normalization is performed on each input parameter within the model to map them to a uniform scale range. Normalization can employ a linear scaling method, for example, for any input parameter... Its normalization result can be expressed as: in, and These are the minimum and maximum values ​​of the parameter within a preset range, respectively. Through the above processing, the normalized directional imbalance parameter is obtained. Dense difference parameters and texture compression parameters This eliminates the influence of different parameter dimensions.

[0045] After normalization, the input parameters are weighted and fused according to preset weights to obtain the synchronous compaction index. This process reflects the contribution of different features to compaction evaluation. Specifically, it can take the following form: in, Indicates the first Synchronous density index at each monitoring time, For the preset weighting coefficients, satisfy In the formula, for and The reverse processing is performed because the greater the directional imbalance and regional differences, the lower the degree of densification synchronization, while the texture compression parameters are positively correlated with the degree of densification. This fusion method allows multiple parameters with different physical meanings to be uniformly expressed as a continuously varying evaluation metric.

[0046] The synchronous densification index, used as an evaluation metric, reflects the degree of synchronous densification of the green body during the sintering process. In engineering applications, a synchronous densification range can be preset according to the specific material system and process conditions, for example, set as... When the synchronous densification index is within this range, it indicates that the shrinkage in all directions, densification in all regions, and changes in surface structure of the billet are in a relatively coordinated state at the current stage.

[0047] When the synchronization density index deviates from the above range, it is necessary to further quantify the degree of deviation in order to provide a basis for subsequent risk assessment. For this purpose, a synchronization density deviation can be defined as follows: in, This is used to characterize the degree of deviation of the synchronous compaction index from the target range. The larger the deviation, the more significant the difference between the current compaction state and the expected state.

[0048] Judging solely based on the synchronization compaction deviation may still be lagging in some cases. For example, when the directional imbalance parameter or compaction difference parameter continues to increase, but the synchronization compaction index has not yet significantly deviated from the target range, failure to differentiate may delay the identification of potential risks. Therefore, in this embodiment, the synchronization compaction deviation is jointly analyzed with the changing trends of the directional imbalance parameter and the compaction difference parameter. The changing trends can be obtained by parameter difference over continuous monitoring time, for example: when and When the value remains positive across multiple sampling periods, it indicates an increasing mismatch trend. Based on this trend information, a comprehensive assessment of the current compaction state of the billet is performed, thereby avoiding errors caused by judging with a single indicator.

[0049] After obtaining the synchronous compaction deviation and comprehensive judgment results, the compaction state at the current monitoring time is divided into different levels of mismatch risk. For example, based on the magnitude and trend of the deviation, the risk level can be divided into low risk, medium risk, and high risk. Low risk corresponds to the synchronous compaction index being stable within the target range or having only minor fluctuations; medium risk corresponds to a significant deviation or trend change in the synchronous compaction index; high risk corresponds to a large deviation in the synchronous compaction index with directional imbalance or a continuous expansion of regional differences. In specific implementation, the risk level classification threshold can be obtained through experimental data calibration or adjusted according to different types of SMP billets.

[0050] It should be noted that the calculation method of the synchronization density index and the risk level classification method described above are not the only limitations. In other embodiments, the weighting coefficients, normalization methods, or risk classification strategies can be adjusted according to actual needs, such as introducing nonlinear fusion functions or multi-level threshold systems. As long as it can effectively characterize the degree of densification synchronization and provide a basis for subsequent regulation, it should be considered within the scope of protection of this application.

[0051] Step S3: Based on the synchronous densification index and the mismatch risk level, and in combination with the current sintering stage parameters, determine the sintering control parameters that match the current densification state of the green body.

[0052] Specifically, based on the degree of deviation between the synchronous densification index and the preset synchronous densification interval, the densification deviation at the current monitoring time is determined, and a hierarchical mapping process is performed on the densification deviation in conjunction with the mismatch risk level to obtain the corresponding control intensity parameter; based on the control intensity parameter and the current sintering stage parameter, the corresponding heating rate adjustment, holding time adjustment, and atmosphere adjustment are calculated respectively to form a set of sintering control parameters that match the current densification state of the green body; the set of sintering control parameters is constrained and matched with the preset sintering process parameter range to obtain sintering control parameters that meet the process constraints.

[0053] In this embodiment of the invention, the synchronous density index... The aforementioned steps have already provided a continuous evaluation metric for the degree of densification synchronization. To enable its participation in regulation, it is necessary to further quantify its deviation from the target state. In this embodiment, a preset synchronization densification interval is used. Define the target state. When When deviating from this range, a densification deviation is introduced. Its definition can follow the aforementioned expression. This deviation reflects the gap between the current state and the desired state, but using only the deviation as the basis for regulation is still insufficient to cover different risk scenarios.

[0054] Therefore, mismatch risk level is introduced into the decision-making process. The risk level essentially reflects the nature and trend of the deviation; for example, for the same deviation magnitude, the control strategy should differ under stable fluctuation and continuous deterioration conditions. To this end, the densification deviation and mismatch risk level are jointly processed, and the control intensity parameter is obtained through hierarchical mapping. Specifically, a mapping function can be defined: in, Indicates the control intensity parameter. This indicates the level of mismatch risk. The function can be segmented; for example, at a low risk level, only larger deviations are amplified, while at a high risk level, even smaller deviations are given a higher control intensity. The function form is not limited and can be set according to different material systems and equipment characteristics.

[0055] After obtaining the intensity control parameters, it is necessary to calculate the specific process parameters in conjunction with the current sintering stage parameters. The sintering process is usually divided into different stages, such as the preheating stage, the activation and shrinkage stage, and the densification and stabilization stage. Different stages have different sensitivities to temperature changes and atmospheric conditions. Therefore, the control parameters must be coupled with the stage parameters when calculating the control amount. Taking the heating rate as an example, it can be adjusted in the following way: in, Given the current heating rate, This is the corrected heating rate. This is the stage correlation coefficient. During the contraction phase, Larger values ​​are chosen to enhance the regulatory effect; during the stable phase, The value is reduced to avoid excessive intervention.

[0056] Similarly, the adjustment amount of the heat preservation time can be calculated incrementally based on the control intensity parameter, for example: in, This is the preset insulation time for the current stage. This is the corrected heat preservation time. This is a coefficient related to material properties. By appropriately extending the insulation time, compensation can be provided for the densification of the internal regions.

[0057] The method for adjusting the atmosphere volume can be determined based on specific equipment conditions. For example, in a protective atmosphere environment, adjustment can be achieved by adjusting the atmosphere flow rate or component ratio. The adjustment method can be expressed as: in, This refers to the current atmosphere flow rate or concentration parameter. This is a correction value. This is the adjustment coefficient. By changing the atmospheric conditions, the reaction rate on the material surface can be affected to some extent, thereby indirectly regulating the densification process.

[0058] The aforementioned adjustments to the heating rate, holding time, and atmosphere constitute a set of sintering control parameters. Once formed, this set cannot be directly applied to the equipment; it must be matched with preset sintering process parameter ranges. This is because both the sintering equipment and the material system have safe operating boundaries, such as the maximum heating rate and the allowable atmosphere concentration range. If the control parameters exceed these ranges, new instability factors may be introduced.

[0059] In this embodiment, the set of control parameters is constrained within a preset range by performing constraint mapping. For example, the heating rate is truncated. in, These represent the upper and lower limits of the allowable range. A similar method can be applied to holding time and atmosphere parameters. Through this process, sintering control parameters that meet the process constraints are obtained.

[0060] In practical applications, different constraint ranges can be set according to different types of SMP preforms. For example, for conductive structural components, dimensional consistency is of greater concern, and the range of holding time adjustment can be appropriately expanded; for supporting structural components, overall strength is of greater concern, and the variation range of heating rate may be limited first. The above-mentioned differentiated settings are all variations of the implementation method of this application.

[0061] Step S4: Based on the sintering control parameters, perform parameter correction on the subsequent sintering process, and update the image state parameters and the synchronous compaction evaluation model based on the corrected time-series image data.

[0062] Specifically, the parameter correction for subsequent sintering processes based on the sintering control parameters includes: inputting the sintering control parameters into the control interface of the sintering equipment, and parsing the sintering control parameters into temperature adjustment commands and atmosphere adjustment commands corresponding to the current sintering stage; correcting the heating rate and holding time in subsequent sintering processes based on the temperature adjustment commands, and correcting the atmosphere flow rate and atmosphere composition in subsequent sintering processes based on the atmosphere adjustment commands, to form corrected sintering operating parameters; and controlling subsequent sintering processes according to the corrected sintering operating parameters.

[0063] Furthermore, updating the image state parameters and the synchronous density evaluation model based on the corrected time-series image data includes: re-extracting the contour size parameters, contour area parameters, grayscale feature parameters, and texture feature parameters corresponding to each monitoring time based on the corrected time-series image data, and performing normalization processing on each re-extracted parameter to form updated image state parameters; inputting the updated image state parameters into the synchronous density evaluation model, recalculating the synchronous density index corresponding to each monitoring time, and correcting the parameter weights in the synchronous density evaluation model based on the change relationship between the recalculated synchronous density index and the historical synchronous density index; and updating the synchronous density evaluation model based on the corrected model parameter weights.

[0064] In this embodiment of the invention, regarding parameter correction, the sintering control parameters are input to the control interface of the sintering equipment. This control interface can be a communication module connected to the furnace control system, or a control port that directly interacts with the temperature control unit and atmosphere conditioning unit. The input process requires parsing the control parameters into control commands recognizable by the equipment. Specifically, the sintering control parameters can be broken down into temperature conditioning commands and atmosphere conditioning commands, whereby the temperature conditioning commands control the heating rate and holding time, and the atmosphere conditioning commands control the atmosphere flow rate and gas composition ratio.

[0065] During the execution of the temperature regulation command, the temperature profile of the subsequent sintering stage is corrected. For example, when the control parameter indicates that the shrinkage rate needs to be reduced, the heating rate can be adjusted to reduce it from the original set value. Change to correction value This process can be achieved by controlling the heating power or adjusting the temperature gradient in the furnace zone. Corrections to the holding time are typically made by extending or shortening the residence time in a specific temperature zone to provide sufficient time for densification within the billet.

[0066] Atmosphere control commands apply to the gas conditions of the sintering environment. For sintering processes using a protective or reducing atmosphere, this can be achieved by adjusting the gas flow rate or changing the proportion of gas components. For example, increasing the inert gas flow rate can reduce the local reaction rate, thereby slowing down the surface densification process; adjusting the gas composition can alter the surface reaction environment of the material, making it more coordinated with the internal densification rhythm. The above control methods are not limited to specific implementation paths; any method that effectively controls the atmosphere conditions falls within the scope of this application.

[0067] After executing the temperature and atmosphere control commands, revised sintering operating parameters are generated. These parameters include revised heating rates, holding times, and atmosphere conditions, and are used to control subsequent sintering processes. It should be noted that this control remains in effect throughout the subsequent sintering process, allowing the green body to continue its densification evolution under the new process conditions.

[0068] As the sintering process continues under the corrected parameters, the actual state of the billet will change. To enable the control process to be adaptive, this change needs to be sensed again. In this embodiment, the billet state is re-analyzed by continuing to acquire corrected time-series image data. This process is consistent with the image acquisition and feature extraction process in step S1, but the input data is the new state after the control effect.

[0069] Based on the corrected time-series image data, contour size parameters, contour area parameters, grayscale feature parameters, and texture feature parameters are re-extracted at each monitoring time point. Since the billet has evolved under the new process conditions at this point, these parameters have different distributions compared to before the correction. To ensure the stability of subsequent calculations, the re-extracted parameters are normalized to maintain a unified dimensional system. The normalization method can be consistent with the aforementioned steps, such as using linear scaling or range mapping based on historical data.

[0070] After obtaining the updated image state parameters, they are input into the synchronous densification evaluation model to recalculate the synchronous densification index for each monitoring time. This index reflects the degree of synchronization in the densification of the billet under the current control conditions. By comparing the synchronous densification index with that corresponding to historical monitoring times, the changing trend of the densification state can be obtained. For example, the change in the synchronous densification index can be calculated: in, To synchronize the density index at the current moment, This value corresponds to the previous monitoring time. This change reflects the direction and magnitude of the impact of the control measures on the densification process.

[0071] Based on this, the parameter weights in the synchronous densification evaluation model are adjusted. The purpose of weight adjustment is to enable the model to adapt to changes in different sintering stages or different batches of green bodies. For example, when the texture compression parameter is found to contribute significantly to the change in the synchronous densification index, its weight can be appropriately increased; when the change in the directional imbalance parameter is more significant, its influence ratio in the model is increased. Weight adjustment can be achieved through simple proportional adjustment or through iterative updates based on multiple monitoring data.

[0072] The updated weighted model is used to replace the original synchronous densification evaluation model, thus completing the adaptive update of the model. At this point, the subsequent calculation of the synchronous densification index will be based on the new weighting system, allowing the model to gradually conform to the dynamic characteristics of the actual sintering process.

[0073] It should be noted that the above model update process does not require weight adjustments to be performed at every monitoring moment. In some implementations, an update cycle can be set, for example, performing a weight correction uniformly after several sampling periods to avoid frequent model fluctuations. Alternatively, an update mechanism can be triggered based on the mismatch risk level, performing model adjustments only when the risk level reaches a certain threshold.

[0074] In another implementation, after acquiring time-series image data, an adaptive partitioning process is performed on the billet region based on the gray-level change gradient field between consecutive frames, dividing the pixel set with similar gray-level change rates into several dynamic sub-regions. For each dynamic sub-region, its regional average gray-level change rate and area shrinkage rate are calculated, and a region splitting index is constructed to characterize the consistency of evolution between regions. The region splitting index is introduced as an additional input parameter into the synchronous compaction evaluation model, participating in the calculation of the synchronous compaction index together with the directional imbalance parameter, compaction difference parameter, and texture compression parameter.

[0075] During the control process, when a continuous increase in the regional splitting index is detected, the atmosphere control intensity is increased or the holding time of the local stage is extended to suppress the trend of sudden local densification. By introducing this dynamic regional splitting characteristic parameter, the model can identify fine-grained local evolution inconsistencies, thereby improving the response accuracy of sintering control.

[0076] like Figure 3 One embodiment of the present invention provides an SMP sintering control system based on image monitoring. The system includes: an image acquisition unit, used to acquire time-series image data of the SMP billet during the sintering process, and extract image state parameters to characterize the densification state of the billet based on the time-series image data; a risk level determination unit, used to construct a synchronous densification evaluation model based on the image state parameters, and use the synchronous densification evaluation model to calculate the synchronous densification index and mismatch risk level corresponding to each monitoring time; a control parameter generation unit, used to determine sintering control parameters matching the current densification state of the billet based on the synchronous densification index and the mismatch risk level, combined with the current sintering stage parameters; and a control execution unit, used to correct the execution parameters of subsequent sintering processes based on the sintering control parameters, and update the image state parameters and the synchronous densification evaluation model based on the corrected time-series image data.

[0077] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a microcontroller, chip, or processor to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0078] The optional embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details described above. Within the scope of the technical concept of the embodiments of the present invention, various simple modifications can be made to the technical solutions of the embodiments of the present invention, and these simple modifications all fall within the protection scope of the embodiments of the present invention. It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the embodiments of the present invention will not further describe the various possible combinations.

[0079] Furthermore, various different embodiments of the present invention can be combined in any way, as long as they do not violate the spirit of the embodiments of the present invention, they should also be regarded as the content disclosed by the embodiments of the present invention.

Claims

1. A method for controlling SMP sintering based on image monitoring, characterized in that, The method includes: Acquire time-series image data of SMP green bodies during the sintering process, and extract image state parameters to characterize the densification state of the green bodies based on the time-series image data; A synchronous density evaluation model is constructed based on the image state parameters, and the synchronous density index and mismatch risk level are calculated for each monitoring time using the synchronous density evaluation model. Based on the synchronous densification index and the mismatch risk level, and in combination with the current sintering stage parameters, determine the sintering control parameters that match the current densification state of the green body. Based on the sintering control parameters, the parameters for subsequent sintering processes are corrected, and the image state parameters and the synchronous compaction evaluation model are updated based on the corrected time-series image data.

2. The SMP sintering control method based on image monitoring according to claim 1, characterized in that, Acquire timing image data of the SMP preform during the sintering process, including: Original image frames of the SMP billet at different sintering stages are acquired at preset monitoring positions in the sintering equipment, and the original image frames are continuously acquired based on preset sampling time intervals to form an original image sequence corresponding to the sintering process. Timestamp calibration and stage identifier association processing are performed on each original image frame in the original image sequence to establish a correspondence between each original image frame and the corresponding sintering time parameter and sintering stage parameter, so as to obtain a time-series image sequence with stage identifier.

3. The SMP sintering control method based on image monitoring according to claim 2, characterized in that, Based on the time-series image data, image state parameters for characterizing the densification state of the billet are extracted, including: Contour extraction processing is performed on each image frame in the time-series image sequence with stage identifiers to obtain the blank contour boundary corresponding to each monitoring time, and the corresponding contour size parameters and contour area parameters are determined based on the blank contour boundary. Based on the outline boundary of the blank, each image frame is divided into regions to obtain edge regions and center regions, and gray-level distribution parameters of the edge regions and the center regions are extracted respectively to form gray-level feature parameters for characterizing the difference in regional densification. Texture feature extraction processing is performed on the target area defined by the outline boundary of the blank to obtain the texture energy parameters and texture contrast parameters corresponding to each monitoring time, and texture feature parameters for characterizing surface structure changes are constructed based on the texture energy parameters and texture contrast parameters. The contour size parameter, the contour area parameter, the grayscale feature parameter, and the texture feature parameter are combined to construct the image state parameters corresponding to each monitoring time.

4. The SMP sintering control method based on image monitoring according to claim 3, characterized in that, Based on the image state parameters, a synchronous compactness evaluation model is constructed, including: The directional imbalance parameters are determined based on the contour dimension parameters and the contour area parameters; Based on the gray-scale feature parameters, a density difference parameter is determined to characterize the density difference between the edge region and the center region; Based on the texture feature parameters, texture compression parameters for characterizing surface structure changes are determined; A synchronous density evaluation model is constructed for the directional imbalance parameter, the density difference parameter, and the texture compression parameter.

5. The SMP sintering control method based on image monitoring according to claim 4, characterized in that, Determining directional imbalance parameters based on the contour dimension parameters and the contour area parameters includes: Based on the contour size parameters, the shrinkage ratios corresponding to different directions are determined, and based on the shrinkage ratios, a directional deviation amount is constructed to characterize the shrinkage differences in each direction, so as to determine the directional imbalance parameters based on the directional deviation amount. Based on the grayscale feature parameters, a density difference parameter is determined to characterize the density difference between the edge region and the center region, including: The average gray values ​​of the edge region and the center region are determined based on the gray-scale feature parameters, and a gray-scale difference value is constructed based on the average gray-scale value to characterize the gray-scale difference of the region, so as to determine the density difference parameter based on the gray-scale difference value. Based on the texture feature parameters, texture compression parameters for characterizing surface structure changes are determined, including: Based on the texture feature parameters, the texture energy value and texture contrast value corresponding to each monitoring time are determined respectively. Based on the texture energy value and the texture contrast value, a texture change quantity is constructed to characterize the degree of texture change, and the texture compression parameters are determined based on the texture change quantity.

6. The SMP sintering control method based on image monitoring according to claim 4, characterized in that, The synchronous compactness evaluation model is used to calculate the synchronous compactness index and mismatch risk level for each monitoring time, including: The directional imbalance parameter, the density difference parameter, and the texture compression parameter are introduced as input parameters into the synchronous density evaluation model, and the input parameters are normalized. Based on preset weights, the normalized input parameters are subjected to weighted fusion processing to obtain the synchronous densification index corresponding to each monitoring time, and the synchronous densification index is used as an evaluation quantity to characterize the overall densification synchronization degree of the billet. Based on the degree of deviation between the synchronous densification index and the preset synchronous densification interval, the synchronous densification deviation at each monitoring time is determined, and combined with the changing trends of the directional imbalance parameter and the densification difference parameter, a comprehensive judgment is made on the densification state of the current billet. Based on the synchronous compaction deviation and the comprehensive judgment result, the compaction state at the current monitoring time is divided into different levels of mismatch risk.

7. The SMP sintering control method based on image monitoring according to claim 1, characterized in that, Based on the synchronous densification index and the mismatch risk level, and in conjunction with the current sintering stage parameters, sintering control parameters matching the current densification state of the green body are determined, including: Based on the degree of deviation between the synchronous densification index and the preset synchronous densification interval, the densification deviation corresponding to the current monitoring time is determined, and the densification deviation is subjected to hierarchical mapping processing in combination with the mismatch risk level to obtain the corresponding control intensity parameter. Based on the control intensity parameters and the current sintering stage parameters, the corresponding heating rate adjustment, holding time adjustment, and atmosphere adjustment are calculated respectively to form a set of sintering control parameters that match the current densification state of the green body. The set of sintering control parameters is matched with the preset range of sintering process parameters to obtain sintering control parameters that meet the process constraints.

8. The SMP sintering control method based on image monitoring according to claim 7, characterized in that, Based on the sintering control parameters, the parameters for subsequent sintering processes are adjusted, including: The sintering control parameters are input into the control interface of the sintering equipment, and the sintering control parameters are parsed into temperature adjustment commands and atmosphere adjustment commands corresponding to the current sintering stage. Based on the temperature adjustment command, the heating rate and holding time in the subsequent sintering process are corrected, and based on the atmosphere adjustment command, the atmosphere flow rate and atmosphere composition in the subsequent sintering process are corrected to form the corrected sintering operating parameters. The subsequent sintering process is controlled according to the revised sintering operating parameters.

9. The SMP sintering control method based on image monitoring according to claim 8, characterized in that, Updating the image state parameters and the synchronous compactness evaluation model based on the corrected time-series image data includes: Based on the corrected time-series image data, the contour size parameters, contour area parameters, grayscale feature parameters, and texture feature parameters corresponding to each monitoring time are re-extracted, and the re-extracted parameters are normalized to form updated image state parameters. The updated image state parameters are input into the synchronous density evaluation model to recalculate the synchronous density index for each monitoring time. Based on the relationship between the recalculated synchronous density index and the historical synchronous density index, the parameter weights in the synchronous density evaluation model are corrected. The synchronous compactness evaluation model is updated based on the corrected model parameter weights.

10. An SMP sintering control system based on image monitoring, characterized in that, The system includes: The image acquisition unit is used to acquire time-series image data of the SMP billet during the sintering process, and extract image state parameters to characterize the densification state of the billet based on the time-series image data. The risk level determination unit is used to construct a synchronous density evaluation model based on the image state parameters, and to use the synchronous density evaluation model to calculate the synchronous density index and mismatch risk level for each monitoring time. The control parameter generation unit is used to determine sintering control parameters that match the current compaction state of the green body based on the synchronous densification index and the mismatch risk level, combined with the current sintering stage parameters. The control and execution unit is used to correct the execution parameters of the subsequent sintering process based on the sintering control parameters, and to update the image state parameters and the synchronous densification evaluation model based on the corrected time-series image data.

Images