Sealing ring processing quality monitoring system

Abstract

The invention discloses a sealing ring processing quality monitoring system, and relates to the technical field of sealing ring processing monitoring. The system comprises a polarization-dark field composite imaging module which is used for obtaining a target image which eliminates specular reflection and highlights scattering characteristics of tiny bubbles through the synergistic effect of polarization filtering and annular dark field illumination; the microbubble feature enhancement module is used for carrying out feature extraction and enhancement on the target image by adopting a deep learning network comprising a multi-scale cavity convolution unit and a self-adaptive attention mechanism unit; and the self-adaptive threshold segmentation module is used for processing the enhanced image in combination with segmentation strategies of the dual-threshold screening unit and the attention map guide segmentation unit so as to accurately segment a micro bubble region. Through cooperation of polarization-dark field composite imaging, a microbubble feature enhancement network and adaptive threshold segmentation, high-precision rapid detection of the microbubbles on the surface of the transparent silica gel sealing ring is realized.

Description

Technical Field

[0001] This invention relates to the field of sealing ring processing monitoring technology, specifically a sealing ring processing quality monitoring system. Background Technology

[0002] In fields with extremely high requirements for sealing and safety, such as medical infusion sets and food packaging, transparent silicone sealing rings are widely used due to their excellent biocompatibility and aging resistance. However, during their production, factors such as the vulcanization process can easily generate microbubbles on the surface of the sealing ring, especially in critical areas such as the lip and sealing surface. These microbubbles are typically extremely small in diameter, occupying only a few pixels in an image. Furthermore, due to the high light transmittance of transparent silicone (over 90%), traditional illumination methods such as ring light and coaxial light are prone to specular reflection, resulting in extremely low contrast between the microbubbles and the background, making it difficult for traditional machine vision systems to accurately identify and detect them. Meanwhile, while traditional deep learning models improve detection accuracy to some extent, their processing speed cannot meet the demands of high-speed, full-inspection production lines. Therefore, overcoming the technical bottlenecks in detecting microbubbles on the surface of transparent silicone sealing rings—such as light transmission interference, the weak characteristics of microbubbles, and the difficulty in balancing detection speed and accuracy—to achieve high-precision, rapid detection has become an urgent problem to be solved. Summary of the Invention

[0003] The purpose of this invention is to provide a sealing ring processing quality monitoring system to solve the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a sealing ring processing quality monitoring system, comprising:

[0005] The polarization-dark field composite imaging module is used to acquire target images that eliminate specular reflection and highlight the scattering characteristics of tiny bubbles through the synergistic effect of polarization filtering and ring dark field illumination, thus solving the problem of light transmission interference of transparent silicone material.

[0006] The microbubble feature enhancement module is used to extract and enhance features of the target image using a deep learning network containing multi-scale dilated convolutional units and adaptive attention mechanism units, thereby solving the problem of weak recognition of microbubble features.

[0007] The adaptive threshold segmentation module combines the segmentation strategy of the dual threshold screening unit and the attention map-guided segmentation unit to process the enhanced image to accurately segment the microbubble region, balancing detection accuracy and real-time requirements.

[0008] The dual-threshold filtering unit is used to filter images by a preset area threshold range and a preset roundness threshold.

[0009] The attention map-guided segmentation unit is used to perform local threshold adjustment on the image based on the attention map generated by the microbubble feature enhancement module.

[0010] Preferably, the polarization-dark field composite imaging module includes:

[0011] The polarization filter unit eliminates specular reflection by using a polarizer placed in front of the industrial camera lens. The polarization angle of the polarizer is an adjustable preset angle.

[0012] The annular dark field illumination unit uses an LED array in a ring structure to illuminate the sealing ring from the side at an adjustable preset incident angle, causing tiny bubbles to scatter light and create a bright area, while the flat surface forms a low-brightness background.

[0013] Preferably, the inner diameter and outer diameter of the annular LED array in the annular dark field lighting unit are in a preset ratio to ensure lighting uniformity and dark field effect.

[0014] Preferably, the multi-scale dilated convolution unit of the microbubble feature enhancement module expands the receptive field through dilated convolution with adjustable dilation rate to capture the contextual features of microbubbles; the adaptive attention mechanism unit generates a pixel-level attention map to weighted enhance the feature responses of typical bubble feature regions such as annular edges.

[0015] Preferably, the adaptive attention mechanism unit uses the combined effect of channel attention and spatial attention to make the feature response enhancement factor of the microbubble region a configurable parameter.

[0016] Preferably, the dual-threshold screening unit includes:

[0017] The area filtering subunit uses a configurable area threshold range to filter noise and large-sized impurities;

[0018] Shape filtering sub-unit, using the roundness formula Calculate the shape features of the connected region and distinguish between bubble and non-bubble defects by a configurable roundness threshold, where C is the roundness, A is the region area, and P is the region perimeter.

[0019] Preferably, the attention map-guided segmentation unit automatically adjusts the local segmentation threshold for low-contrast regions based on the attention map generated by the microbubble feature enhancement module.

[0020] Preferably, it also includes a result judgment module, which is used to determine whether the sealing ring is qualified based on the segmentation result of the adaptive threshold segmentation module and in combination with a preset quality standard, and record the detection data including the number, area, position and shape features of the bubbles.

[0021] Preferably, the result judgment module associates and stores the detection data with production batches and production process-related parameters to form a traceable quality monitoring database.

[0022] Compared with the prior art, the beneficial effects of the present invention are:

[0023] The polarization-dark-field composite imaging module utilizes the synergistic effect of polarization filtering and ring dark-field illumination to eliminate specular reflection, highlight the scattering characteristics of microbubbles, and solve the problem of light transmission interference from transparent silicone materials, achieving high-contrast image acquisition. The microbubble feature enhancement module employs multi-scale dilated convolutional units and adaptive attention mechanisms to extract and enhance features from the target image, solving the challenge of recognizing weak microbubble features and achieving accurate microbubble feature capture. The adaptive threshold segmentation module, combining dual-threshold filtering and attention map-guided segmentation strategies, processes the enhanced image, achieving accurate segmentation of microbubble regions and balancing detection accuracy with real-time requirements. The result judgment module associates and stores detection data with production batches and process-related parameters, forming a traceable quality monitoring database, enabling traceable management and efficient judgment of sealing ring quality. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of a sealing ring processing quality monitoring system provided in an embodiment of the present invention;

[0025] Figure 2 This is a schematic diagram of the structure of the polarization-dark field composite imaging module provided in an embodiment of the present invention;

[0026] Figure 3 This is a schematic diagram of the microbubble feature enhancement module provided in an embodiment of the present invention;

[0027] Figure 4 This is a schematic diagram of the adaptive threshold segmentation module provided in an embodiment of the present invention;

[0028] Figure 5 This is a schematic diagram of the structure of the dual threshold filtering unit provided in an embodiment of the present invention. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] Please see Figures 1-5This invention provides: a sealing ring processing quality monitoring system, comprising:

[0031] The polarization-dark field composite imaging module 11 is used to acquire a target image that eliminates specular reflection and highlights the scattering characteristics of tiny bubbles through the synergistic effect of polarization filtering and ring dark field illumination.

[0032] The microbubble feature enhancement module 12 is used to extract and enhance features of the target image using a deep learning network that includes multi-scale dilated convolutional units and adaptive attention mechanism units.

[0033] The adaptive threshold segmentation module 13 is used to combine the segmentation strategy of the dual threshold screening unit and the attention map guided segmentation unit to process the enhanced image to accurately segment the microbubble region.

[0034] The dual-threshold filtering unit is used to filter images by a preset area threshold range and a preset roundness threshold.

[0035] The attention map-guided segmentation unit is used to perform local threshold adjustment on the image based on the attention map generated by the microbubble feature enhancement module 12.

[0036] In an optional embodiment, the polarization-dark field composite imaging module 11 includes:

[0037] The polarization filter unit eliminates specular reflection by using a polarizer placed in front of the industrial camera lens. The polarization angle of the polarizer is an adjustable preset angle.

[0038] The annular dark field illumination unit uses an LED array in a ring structure to illuminate the sealing ring from the side at an adjustable preset incident angle, causing tiny bubbles to scatter light and create a bright area, while the flat surface forms a low-brightness background.

[0039] Based on the polarization characteristics of light, the polarization filter unit effectively suppresses specular reflection by adding a polarizer in front of the industrial camera lens. Specifically, the polarizer supports continuous angle adjustment from 0° to 180°. By dynamically adjusting the polarization angle parameter, it can accurately match the reflection characteristics of sealing ring surfaces made of different materials. In practical applications, the system can automatically call the preset optimal polarization angle according to the sealing ring material properties (such as rubber, silicone, etc.) to eliminate strong reflective interference and enhance image contrast.

[0040] The ring-shaped dark field illumination unit can be constructed using a high-brightness LED array in a ring structure. The brightness of the LED light source can be linearly adjusted from 0% to 100% through a programmable controller to adapt to different ambient light conditions. By continuously adjusting the incident angle from 0° to 45°, the angle of light illumination can be changed, which can optimize the scattering effect of tiny defects (such as bubbles and cracks) on the surface of the sealing ring. When the light is irradiated at a preset incident angle, the tiny bubbles on the surface of the sealing ring will form bright pixels due to the light scattering effect, while the flat areas will have low brightness because the light reflected by the mirror does not enter the camera's field of view, thus creating a high-contrast defect highlighting effect in the image.

[0041] In one optional embodiment, the inner diameter and outer diameter of the annular LED array of the annular dark field illumination unit are in a preset ratio. Based on the principle of optical reflection and imaging, high-contrast imaging of surface defects of the sealing ring can be achieved by precisely controlling the geometric parameters of the annular LED array.

[0042] Specifically, the incident angle of the ring LED array is 25°. Experimental verification has shown that this ensures light reaches the sealing ring surface at the optimal angle, effectively preventing direct reflection into the lens. The inner diameter is set to 30mm and the outer diameter to 80mm. This size ratio is suitable for the inspection needs of most standard sealing ring specifications, ensuring uniform illumination while avoiding blind spots caused by excessively large or small sizes. Simultaneously, the system employs a 45° polarizer. Utilizing the selective filtering characteristics of polarized light on reflected light, it effectively eliminates 90% of specular reflection from the silicone surface, achieving a high-contrast imaging effect of "white bubbles on a black background." This imaging effect greatly improves the identification of defects such as bubbles and cracks in the image, providing a high-quality image foundation for subsequent image analysis and quality assessment.

[0043] In an optional embodiment, the multi-scale dilated convolution unit of the microbubble feature enhancement module 12 expands the receptive field through dilated convolution with adjustable dilation rate to capture the contextual features of microbubbles; the adaptive attention mechanism unit generates a pixel-level attention map to weighted enhance the feature responses of typical bubble feature regions such as annular edges.

[0044] Specifically, the multi-scale dilated convolutional unit uses a dilated convolutional layer with an inflation rate of 2 to expand the receptive field of the convolutional kernel to 7×7 pixels.

[0045] The multi-scale dilated convolutional unit adopts a three-level cascaded structure: the first level is configured with a dilated convolutional layer with an inflation rate of 2, which expands the receptive field of the traditional 3×3 convolutional kernel to 7×7 pixels, and can capture 300% more background information than conventional convolution; the middle layer dynamically adjusts the inflation rate to 3, further expanding the receptive field to 13×13 pixels, and realizing feature coverage of different scale morphologies of bubbles; the final level uses a conventional convolution with an inflation rate of 1 for feature fusion. Through this multi-scale dilated convolutional design, not only can the fine structure of tiny bubbles be effectively captured, but also the correlation features between them and the surrounding medium can be obtained;

[0046] In the adaptive attention mechanism unit, based on the multi-scale feature map generated by spatial pyramid pooling, channel attention and spatial attention are calculated separately through a dual-channel attention network. Channel attention extracts channel-level statistical information of the feature map based on global average pooling and global max pooling, and generates channel weights using a multilayer perceptron. Spatial attention captures the spatial dependencies of the feature map based on convolution operations and generates pixel-level attention weights. Finally, the attention map generated by fusing the two is multiplied element-wise with the feature map to weight and enhance typical bubble feature regions such as ring edges and highlighted cores, effectively improving the accuracy and recognizability of feature representation.

[0047] In an optional embodiment, the adaptive attention mechanism unit enhances the feature response of the microbubble region by a configurable parameter through the combined effect of channel attention and spatial attention.

[0048] Specifically, the adaptive attention mechanism unit automatically generates a pixel-level attention map, which enhances the feature response of the microbubble region by 2.5 times.

[0049] The adaptive attention mechanism unit employs a collaborative architecture of channel attention and spatial attention, enhancing the features of microbubbles through an adjustable parameter configuration strategy. Specifically, the unit first calculates the importance weight of each channel based on the channel dimension information of the input image using global average pooling and a multilayer perceptron, forming a channel attention vector. Simultaneously, in the spatial dimension, it aggregates spatial features through max pooling and average pooling operations to generate a spatial attention matrix.

[0050] By weighted fusion of channel attention and spatial attention matrices, the adaptive attention mechanism unit can automatically generate a high-resolution pixel-level attention map. This map can accurately locate the region where microbubbles are located and enhance the feature response of the target region through preset enhancement parameters. In a typical application scenario, when the system detects microbubble defects on the surface of a rubber sealing ring, configuring the parameters to a 2.5-fold enhancement coefficient can effectively improve the feature contrast between the microbubbles and the background region, thereby improving the recognition accuracy of the subsequent classification model. This configurable enhancement mechanism can flexibly adjust the feature enhancement intensity according to different processing techniques and detection accuracy requirements, significantly improving the system's generalization ability.

[0051] In an optional embodiment, the dual-threshold filtering unit includes:

[0052] The area filtering subunit uses a configurable area threshold range to filter noise and large-sized impurities;

[0053] Shape filtering sub-unit, using the roundness formula Calculate the shape features of the connected region and distinguish between bubble and non-bubble defects by a configurable roundness threshold, where C is the roundness, A is the region area, and P is the region perimeter.

[0054] The area filtering subunit employs a configurable area threshold range to accurately filter noise and large-sized impurities. In the scenario of monitoring the processing quality of sealing rings, noise typically manifests as tiny, discrete pixels with an area much smaller than normal defect areas; while large-sized impurities may cover a large area, interfering with subsequent analysis. By setting a minimum area threshold, noise below that value can be eliminated; simultaneously, setting a maximum area threshold can exclude abnormally large impurities that exceed the actual defect size range of the sealing ring. Users can flexibly adjust the area threshold parameters according to different specifications of sealing rings and processing characteristics to adapt to diverse monitoring needs.

[0055] The shape screening subunit effectively distinguishes between bubbles and non-bubble defects based on shape features, specifically through the roundness formula. The shape characteristics of the connected region are calculated. The closer the value of C is to 1, the closer the region's shape is to a circle. In sealing ring defects, bubbles usually have near-circular characteristics, while other non-bubble defects (such as cracks, notches, etc.) have more irregular shapes. By setting a configurable roundness threshold, when the calculated roundness C is greater than the threshold, the connected region is determined to be a bubble defect; otherwise, it is classified as a non-bubble defect. Operators can dynamically adjust the roundness threshold according to common defect morphologies in actual production to ensure the accuracy and reliability of defect identification.

[0056] Specifically, the area filtering subunit sets the area threshold range to 3-500 pixels to filter noise and large impurities in the image; the shape filtering subunit sets the roundness threshold to be greater than 0.6 to distinguish between bubbles and non-circular defects such as scratches and burrs.

[0057] The area screening subunit, based on the actual acquisition resolution of the sealing ring image and the size characteristics of common noise and impurities, and through experiments and data analysis, precisely sets the area threshold range to 3-500 pixels. This threshold range can effectively filter out tiny noise generated by electronic noise from the equipment in the image, as well as large-sized impurities mixed in from the production environment, ensuring that only effective features related to sealing ring defects are retained. The shape screening subunit calculates the roundness index of the target area and sets the roundness threshold to be greater than 0.6. This threshold setting is based on the geometrical differences between bubbles and defects such as scratches and flash—bubbles usually present a near-circular outline in the image, while defects such as scratches and flash have irregular edges, thereby achieving efficient differentiation of different types of defects and laying the foundation for subsequent accurate identification.

[0058] In an optional embodiment, the attention map-guided segmentation unit automatically adjusts the local segmentation threshold for low-contrast regions based on the attention map generated by the microbubble feature enhancement module 12.

[0059] The attention map-guided segmentation unit employs a ResNet-18 improved network with multi-scale convolutional kernels to extract features from the attention map generated by the microbubble feature enhancement module 12. Through parallel processing of convolutional operations at three scales (3×3, 5×5, and 7×7), it achieves multi-dimensional feature capture of microbubbles of different sizes, accurately identifying the gray-level distribution differences between microbubbles and the background in low-contrast regions. Subsequently, based on the sliding window algorithm, it dynamically calculates the mean gray-level μ and standard deviation σ of the local region in 16×16 pixel units, and combines this with the inter-class variance maximization principle of the Otsu algorithm to construct an adaptive threshold model.

[0060] T=μ+k·σ

[0061] The dynamic adjustment coefficient k is automatically adjusted within the range of 0.8-1.2 according to the entropy change of the local area. In this way, the system can accurately adjust the local segmentation threshold of the low contrast area, effectively suppress the interference of complex texture background. In complex working conditions where the gray level difference between the bubble and the substrate material is less than 15 gray levels, this method significantly improves the segmentation accuracy of microbubble defects on the basis of the traditional threshold method, and significantly enhances the ability to identify microbubble defects on the surface of the sealing ring.

[0062] In an optional embodiment, the system includes a result judgment module, which is used to determine whether the sealing ring is qualified based on the segmentation result of the adaptive threshold segmentation module 13 and in combination with a preset quality standard, and to record detection data including the number, area, position and shape features of bubbles.

[0063] In the bubble count stage, the system, based on morphological analysis principles, employs mathematical morphological operations such as erosion and dilation to accurately segment and identify bubble regions in the sealing ring image. Through a connected component labeling algorithm, each independent bubble is counted individually, achieving precise bubble count. For area quantization, a pixel area calculation method is used to convert the actual physical size of each pixel in the image. By accumulating the number of pixels within the bubble region and combining this with a scale factor, the actual area of ​​the bubble is finally determined. Centroid localization utilizes a centroid algorithm, calculating the weighted average of the coordinates of pixels within the bubble region to quickly and accurately determine the specific coordinates of the bubble on the sealing ring.

[0064] In the shape feature extraction stage, ellipse fitting and Fourier descriptor techniques are used in combination. The ellipse fitting algorithm fits the ellipse that best fits the bubble shape based on the bubble contour data using the least squares method, and obtains key geometric parameters such as the major and minor semi-axes. The Fourier descriptor converts the two-dimensional coordinates of the bubble contour into frequency domain information and extracts shape feature vectors that are invariant to translation, rotation and scaling.

[0065] Finally, the system compares and analyzes the data on the number, area, location, and shape of the bubbles obtained from the above tests with the pre-set quality thresholds in multiple dimensions. When all the test indicators meet the threshold requirements, the sealing ring is automatically determined to be a qualified product; if any indicator exceeds the threshold range, it is determined to be a non-qualified product, and a corresponding judgment result report is generated.

[0066] In one optional embodiment, the result judgment module associates and stores the detection data with production batches and production process-related parameters to form a traceable quality monitoring database.

[0067] After completing the quality assessment, the module automatically associates the detection data, including the number, area, location, and shape characteristics of bubbles, with production process parameters such as production batch number, injection pressure, curing temperature, and mold number. By establishing a standardized data storage structure, the above information is integrated and entered into the quality monitoring database. This database supports multi-dimensional retrieval and traceability functions. Production managers can quickly retrieve the complete quality inspection files and production process parameters of the corresponding batch by entering the production batch or product number, providing comprehensive data support for subsequent quality analysis, process optimization, and defect tracing.

[0068] In this embodiment, the polarization-dark field composite imaging module 11 utilizes the synergistic effect of polarization filtering and annular dark field illumination to eliminate specular reflection, highlight the scattering characteristics of microbubbles, and solve the problem of light transmission interference from transparent silicone material, achieving a high-contrast image acquisition effect. The microbubble feature enhancement module 12 employs multi-scale dilated convolution units and adaptive attention mechanism units to extract and enhance features of the target image, solving the problem of weak microbubble feature recognition and achieving accurate microbubble feature capture. The adaptive threshold segmentation module 13 combines dual threshold screening and attention map-guided segmentation strategies to process the enhanced image, achieving accurate segmentation of the microbubble region and balancing detection accuracy and real-time requirements. The result judgment module associates and stores the detection data with production batches and production process-related parameters to form a traceable quality monitoring database, realizing traceable management and efficient judgment of sealing ring quality.

[0069] Furthermore, it should be noted that the combination of the various technical features in this case is not limited to the combination methods described in the claims of this case or the combination methods described in the specific embodiments. All technical features described in this case can be freely combined or combined in any way, unless they contradict each other.

[0070] It should be noted that the above examples are merely specific embodiments of the present invention, and the present invention is obviously not limited to the above embodiments, with many similar variations. All modifications that can be directly derived or conceived by those skilled in the art from the content disclosed in this invention should fall within the protection scope of this invention.

[0071] The above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A sealing ring processing quality monitoring system, characterized in that, include: The polarization-dark field composite imaging module is used to acquire target images that eliminate specular reflections and highlight the scattering characteristics of tiny bubbles through the synergistic effect of polarization filtering and ring dark field illumination. The microbubble feature enhancement module is used to extract and enhance features of the target image using a deep learning network that includes multi-scale dilated convolutional units and adaptive attention mechanism units. The adaptive threshold segmentation module is used to combine the segmentation strategy of the dual threshold filtering unit and the attention map-guided segmentation unit to process the enhanced image to accurately segment the microbubble region. The dual-threshold filtering unit is used to filter images by a preset area threshold range and a preset roundness threshold. The attention map-guided segmentation unit is used to perform local threshold adjustment on the image based on the attention map generated by the microbubble feature enhancement module.

2. The sealing ring processing quality monitoring system according to claim 1, characterized in that, The polarization-dark field composite imaging module includes: The polarization filter unit eliminates specular reflection by using a polarizer placed in front of the industrial camera lens. The polarization angle of the polarizer is an adjustable preset angle. The annular dark field illumination unit uses an LED array in a ring structure to illuminate the sealing ring from the side at an adjustable preset incident angle, causing tiny bubbles to scatter light and create a bright area, while the flat surface forms a low-brightness background.

3. The sealing ring processing quality monitoring system according to claim 2, characterized in that, The inner and outer diameters of the annular LED array in the annular dark field lighting unit are in a preset ratio.

4. The sealing ring processing quality monitoring system according to claim 1, characterized in that, The multi-scale dilated convolution unit of the microbubble feature enhancement module expands the receptive field through dilated convolution with adjustable dilation rate, capturing the contextual features of microbubbles; the adaptive attention mechanism unit generates pixel-level attention maps, which weight and enhance the feature responses of typical bubble feature regions such as annular edges.

5. The sealing ring processing quality monitoring system according to claim 4, characterized in that, The adaptive attention mechanism unit, through the combined effect of channel attention and spatial attention, makes the feature response enhancement factor of the microbubble region a configurable parameter.

6. The sealing ring processing quality monitoring system according to claim 1, characterized in that, The dual threshold filtering unit includes: The area filtering subunit uses a configurable area threshold range to filter noise and large-sized impurities; Shape filtering sub-unit, using the roundness formula Calculate the shape features of the connected region and distinguish between bubble and non-bubble defects by a configurable roundness threshold, where C is the roundness, A is the region area, and P is the region perimeter.

7. The sealing ring processing quality monitoring system according to claim 6, characterized in that, The attention map-guided segmentation unit automatically adjusts the local segmentation threshold for low-contrast regions based on the attention map generated by the microbubble feature enhancement module.

8. A sealing ring processing quality monitoring system according to any one of claims 1-7, characterized in that, It also includes a result judgment module, which is used to determine whether the sealing ring is qualified based on the segmentation result of the adaptive threshold segmentation module and in combination with the preset quality standard, and to record the detection data including the number, area, position and shape features of the air bubbles.

9. The sealing ring processing quality monitoring system according to claim 8, characterized in that, The result judgment module associates and stores the detection data with production batches and production process-related parameters to form a traceable quality monitoring database.

Images