A method and system for detecting defects in screens during the pharmaceutical production process.

Abstract

This invention belongs to the field of pharmaceutical manufacturing technology, specifically disclosing a method and system for detecting defects in screens during the pharmaceutical production process. The method includes: minimizing the verification area based on the common overlapping area of ​​the screen surface images of normal screens; establishing minor damage and severe damage standards by constructing a defect type calibration disk and setting parameter thresholds; classifying screens after determining them to be verifiable, marking verifiable screens as target diagnostic screens or target replacement screens, and remarking target diagnostic screens that meet preset replacement conditions as target replacement screens; and differentiating screens with different degrees of damage by correlating the accuracy difference with economic indicators. This invention uses a pre-screening mechanism to exclude screens that cannot be effectively visually analyzed, avoiding invalid or erroneous defect interpretations on unqualified images, and binding defect severity to specific parameter values ​​to ensure the validity of the detection results.

Description

Technical Field

[0001] This invention belongs to the field of pharmaceutical manufacturing technology, specifically relating to a method and system for detecting defects in screens during the pharmaceutical production process. Background Technology

[0002] In modern pharmaceutical production lines, sieving is a fundamental process for raw material pretreatment and finished product homogenization. The sieves used in the sieving process are key equipment to ensure the purity and uniformity of particle size distribution of drugs. The integrity of the sieves is directly related to whether unqualified particles and foreign impurities can be effectively removed. Damage or blockage of the sieves may lead to quality defects in the final drug.

[0003] In existing technologies, the detection of screen condition during the pharmaceutical production process relies mainly on manual offline visual inspection. Offline inspection requires stopping the production line or conducting inspections during batch changes, resulting in low inspection efficiency and reduced production efficiency. Furthermore, if the screen is damaged, there is no real-time response, leading to quality risks for the entire batch of drugs. In addition, due to the limitations of manual observation, it is difficult to identify and address hidden defects such as tiny cracks or initial blockages, making it easy to miss or misjudge screens.

[0004] Based on the above-mentioned technical problems, this application provides a method and system for detecting defects in screens during the pharmaceutical production process in a pharmaceutical manufacturing line. Summary of the Invention

[0005] In order to solve the technical problems in the prior art, the purpose of this invention is to provide a method and system for detecting defects in screens during the pharmaceutical production process in a pharmaceutical manufacturing line.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A method for detecting defects in screens during the pharmaceutical manufacturing process in a pharmaceutical production line includes the following steps:

[0008] Acquire images of the area to be tested on the screen surface and identify the target defects from the screen surface images;

[0009] When the screen containing the target defect is determined to be a verifiable screen, the screen handling procedure is executed.

[0010] The screen handling process includes: classifying verifiable screens based on preset defect standards to mark them as target diagnostic screens or target replacement screens; performing screen diagnosis and upgrading on target diagnostic screens, including performing accuracy testing on the target diagnostic screens, and remarking the target diagnostic screens as target replacement screens when the accuracy test results meet preset replacement conditions; and performing screen matching and replacement on screens marked as target replacement screens.

[0011] Preferably, the step of determining the screen to be a verifiable screen includes:

[0012] When the area to be tested is greater than or equal to the preset minimum verification area, and the edge characteristics of the screen meet the preset flatness conditions, the screen is determined to be a verifiable screen.

[0013] Preferably, determining the minimum verification area includes:

[0014] Edge denoising is performed on the surface images of multiple normal screens to obtain their respective contour information; for each contour information, the offset value of its geometric center relative to the center of a preset fixed axis point is extracted; the normal screen with the largest offset value is determined as the reference screen; based on the contour information of the reference screen, the common overlapping area defined by the contour information of all normal screens is calculated, and the area of ​​the common overlapping area is determined as the minimum verification area.

[0015] Preferably, the preset defect criteria include:

[0016] Multiple defect type calibration disks are constructed, each corresponding to a standard defect image; a preset offset is added to the standard defect image to generate a defect area range map; a calibration object is defined within the defect area range map, and the parameter values ​​of the calibration object are obtained; based on the comparison relationship between the parameter values ​​and the first threshold and the second threshold, a minor damage standard and a severe damage standard are defined.

[0017] Preferably, in the screen diagnosis and upgrade process, the step of remarking the target diagnostic screen as the target replacement screen includes:

[0018] Obtain the rated accuracy and current actual accuracy of the target diagnostic sieve; calculate the diagnostic value based on the accuracy difference between the rated accuracy and the current actual accuracy, as well as the preset product total quantity ratio parameter; if the diagnostic value is less than the preset total amount of powder in the test batch, then remark the target diagnostic sieve as the target replacement sieve.

[0019] A system for detecting defects in screens during the pharmaceutical production process in a pharmaceutical manufacturing line includes the following modules:

[0020] The defect acquisition module is used to acquire an image of the screen surface of the area to be tested, and to identify the target defect from the screen surface image to generate a defect trigger signal.

[0021] The screen diagnosis and classification module is used to respond to defect trigger signals and to: determine the screen containing the target defect as a verifiable screen; classify the verifiable screen based on preset defect standards to mark the verifiable screen as a target diagnostic screen or a target replacement screen; perform accuracy testing on the target diagnostic screen, and when the accuracy test result meets the preset replacement conditions, re-mark the target diagnostic screen as a target replacement screen;

[0022] The screen replacement module is used to perform screen matching and replacement on screens that have been marked as target replacement screens by the screen diagnosis and classification module.

[0023] Preferably, the screen diagnosis and classification module determines a screen to be a verifiable screen by including:

[0024] When the area to be tested is greater than or equal to the preset minimum verification area, and the edge characteristics of the screen meet the preset flatness conditions, the screen is determined to be a verifiable screen.

[0025] Preferably, the system further includes:

[0026] The minimum verification area determination module processes the surface images of multiple normal screens to obtain their respective contour information, determines the reference screen based on the contour information, calculates the common overlapping area defined by the contour information of all normal screens, determines the area of ​​the common overlapping area as the minimum verification area, and provides the minimum verification area to the screen diagnosis and classification module.

[0027] Preferably, the preset defect criteria include:

[0028] Multiple defect type calibration disks are constructed, each corresponding to a standard defect image; a preset offset is added to the standard defect image to generate a defect area range map; a calibration object is defined within the defect area range map, and the parameter values ​​of the calibration object are obtained; based on the comparison relationship between the parameter values ​​and the first threshold and the second threshold, a minor damage standard and a severe damage standard are defined.

[0029] Preferably, remarking the target diagnostic screen as the target replacement screen includes:

[0030] Obtain the rated accuracy and current actual accuracy of the target diagnostic sieve; calculate the diagnostic value based on the accuracy difference between the rated accuracy and the current actual accuracy, as well as the preset product total quantity ratio parameter; if the diagnostic value is less than the preset total amount of powder in the test batch, then remark the target diagnostic sieve as the target replacement sieve.

[0031] Beneficial effects

[0032] This invention constructs a defect type calibration disk and sets parameter thresholds to establish standards for minor damage and severe damage. During the detection process, the target defect in the area to be tested is compared with the preset defect standards to achieve screen classification. This can bind the severity of defects to specific parameter values, eliminate the uncertainty caused by human factors, and thus ensure the accuracy and reliability of subsequent classification, diagnosis and replacement decisions.

[0033] This invention determines the minimum verification area based on the common overlapping area of ​​the surface images of multiple normal screens. This minimum verification area, together with the flatness condition of the area to be tested, is used as the admission condition for determining whether a screen is a verifiable screen. The pre-screening mechanism excludes screens that cannot be effectively visually analyzed due to installation misalignment, deformation, or damage areas that are too small, thus avoiding invalid or erroneous defect interpretations on unqualified images and ensuring the validity of the test results.

[0034] This invention employs a graded diagnosis and upgrade screen management strategy. After initially classifying target diagnostic screens and target replacement screens, for target diagnostic screens with only minor damage, the current actual accuracy is detected and a diagnostic value is calculated in conjunction with production parameters to determine whether it needs to be upgraded to a target replacement screen. By linking the accuracy difference with economic indicators, screens with different degrees of damage are treated differently, thereby extending screen life and reducing operating costs while ensuring quality. Attached Figure Description

[0035] Figure 1 This is an overall flowchart of the present invention;

[0036] Figure 2 This is a block diagram of the implementation system of the present invention. Detailed Implementation

[0037] In the following description, certain specific details are set forth in order to provide a thorough understanding of different embodiments of the invention, and those skilled in the art will understand after reading this disclosure that the invention can be practiced without many of these details.

[0038] Example 1

[0039] Please refer to Figure 1 This embodiment discloses a method for detecting defects in screens during the pharmaceutical manufacturing process in a pharmaceutical production line, including the following steps:

[0040] In pharmaceutical production lines, structured light illumination or multi-angle ring light sources are used to illuminate the screen surface to enhance the image contrast of the fine texture and pore structure of the screen, thereby overcoming interference such as uneven lighting or screen reflection in the production environment; and the screen surface image is continuously monitored and acquired through high-resolution industrial cameras and other image acquisition equipment.

[0041] In the image of the screen surface, the entire screen imaging area is pre-divided into several analysis units. The image data of each analysis unit is compared with the standard pore size, shape, surface gray distribution and texture pattern of a normal screen that meets the quality standard. When the difference between the image data of a certain analysis unit and the reference features exceeds the preset normal fluctuation range, the unit and its adjacent area are defined as the area to be tested.

[0042] Among them, non-standard hole shapes, foreign object attachments, structural damage, or abnormal color black spots in image data are usually caused by local color changes caused by high-temperature sintering of materials, chemical corrosion, or metal fatigue.

[0043] Then, the specific abnormal features in the test area are located and parameters are extracted. The extracted structured data containing information such as location, geometric size, shape and grayscale statistics are defined as target defects. All identified target defects are assigned a unique code and entered into the defect dataset.

[0044] Furthermore, by constructing and utilizing a defect type calibration disk, defect standards for distinguishing the severity of defects are set. Specifically, each defect type calibration disk is prefabricated with a specific type of standard defect through precision machining. During the calibration process of the defect type calibration disk, images of standard defects are acquired to form standard defect images. Then, a geometric offset is added to the defect contour in each standard defect image to generate a defect area map with a wider coverage, so as to simulate the stress influence zone or expansion trend that defects may generate in actual use.

[0045] The standard defect can be a hole with a precise diameter, a scratch with a specific depth and width, a rough surface that simulates uniform wear, or a block that simulates the adhesion of impurities, in order to define the standard defect.

[0046] Then, a calibration object for parameter extraction is defined within the defect area map, and image processing is performed on the calibration object. Image processing includes: separating the defect from the background through binarization, extracting the defect contour, and calculating the area, perimeter, circularity of the shape, boundary complexity, or average gray value within the defect area based on the contour to form the parameter values ​​of the calibration object; based on statistical analysis of parameter values ​​of a large number of standard defect samples and combined with empirical data of production processes, a first threshold and a second threshold greater than the first threshold are set to define a preset defect standard: when the parameter value of a target defect is greater than the first threshold and less than or equal to the second threshold, its severity is judged to meet the minor damage standard; when its parameter value is greater than the second threshold, it is judged to meet the severe damage standard.

[0047] Furthermore, multiple brand-new and standard-compliant screen surface images were collected and recorded. Gaussian filtering was used to smooth the screen surface images and perform edge noise reduction processing to eliminate the influence of minor differences in manufacturing tolerances or installation positioning on the measurement.

[0048] By calculating the gray-level gradient between adjacent pixels and identifying the set of pixels whose gradient values ​​exceed a certain threshold, clear and continuous contour information is extracted. A fixed coordinate reference is set as the preset fixed axis point center. This fixed axis point center can be the optical center of the image sensor or the preset reference point of the screen mounting fixture. The measurement error caused by the inconsistent placement of the screen in the image can be eliminated by using this fixed axis point center.

[0049] For the contour information of each screen, calculate the two-dimensional coordinate offset value of its geometric center relative to the center of the fixed axis point to quantify the installation eccentricity of each screen. Then, among all normal screens, the screen with the largest offset value is determined as the reference screen, and its contour information is used as the alignment reference to ensure that all normal tolerance ranges are covered. Then, through digital image transformation, the contour information of all normal screens is aligned and superimposed on the coordinate system of the reference screen, and the common overlapping area covered by the contour information is calculated. The area of ​​the common overlapping area is then determined as the minimum verification area.

[0050] Furthermore, a verifiability determination is performed on the screen to determine whether it is a verifiable screen. The verifiability determination includes two parallel criteria: Criterion 1, obtaining the area of ​​the test area and comparing it with the minimum verification area; Criterion 2, obtaining the edge features of the screen where the test area is located to evaluate the flatness of its overall structure. When the area of ​​the test area is not less than the minimum verification area, and the flatness test result satisfies the condition that the maximum distance is less than a specific millimeter value, the screen is determined to be a verifiable screen, indicating that although the screen has local defects, the main body still has repair or diagnostic value. Conversely, if either condition is not met, the screen is determined to be an unverifiable screen and is classified into the replacement category.

[0051] The edge features of the screen are obtained through a laser profile scanner or structured light 3D imaging. By obtaining the 3D coordinate data of a series of points on the edge of the screen or its mounting frame, the maximum distance or root mean square error from all edge points to the best-fit plane is calculated, thereby obtaining a quantified flatness detection result.

[0052] Furthermore, for a screen that is determined to be a verifiable screen, each target defect on the verifiable screen is traversed, and its parameter value is compared with the preset defect standard one by one to classify the screen according to the severity of the target defect, and then decide whether to perform in-depth diagnosis or direct replacement.

[0053] If all target defects on the screen meet the minor damage criteria, it means that these defects do not currently pose a structural threat, but may affect screening accuracy or pose a risk of deterioration. In this case, the verifiable screen is marked as a target diagnostic screen.

[0054] If any target defect on the screen meets the criteria for severe damage, it means that the defect has substantially damaged the structural integrity or core screening function of the screen, and the screen is no longer safe to use. In this case, the verifiable screen is directly marked as the target replacement screen, thereby achieving the diversion of screens in different states and optimizing the allocation of processing resources.

[0055] Furthermore, for all screens marked as target diagnostic screens, a performance evaluation is conducted by correlating their visually minor damage with actual sieving performance degradation to determine whether they can continue to serve, thereby diagnosing and upgrading the screens. Specific performance evaluations include: retrieving the rated accuracy represented by the screen's nominal mesh size or corresponding micron-level pore size from the equipment file or production management database; conducting a trial sieving experiment on the target diagnostic screen using standard particulate material; and then using a laser particle size analyzer to detect the particle size distribution of the powder passing through the screen for accuracy testing, thereby obtaining the statistical distribution of the target diagnostic screen's actual retention efficiency and effective pore size to determine the current actual accuracy; finally, the accuracy difference between the rated accuracy and the current actual accuracy is calculated to quantify the degree of performance degradation caused by minor damage to the screen.

[0056] Specifically, by introducing a product risk weighting coefficient related to the value or quality sensitivity of the current product, this accuracy difference is transformed into a decision-making basis related to production risk: the higher the value of the product, the larger the product risk weighting coefficient, indicating a lower tolerance for accuracy deviation. The accuracy difference is multiplied by the product risk weighting coefficient to calculate a diagnostic value to quantify the production risk caused by the decline in screen performance. This diagnostic value is compared with a performance benchmark threshold representing the lowest acceptable performance level. If the calculated diagnostic value is less than the performance benchmark threshold, it indicates that although the damage to the screen is visually minor, its impact on the actual screening performance has exceeded the acceptable range, posing a risk of batch product non-conforming. Therefore, the status of the target diagnostic screen is upgraded, i.e., it is re-marked as the target replacement screen.

[0057] Furthermore, for all screens that are ultimately marked as target replacement screens, a screen matching and replacement process is performed: based on the specifications such as model, size, material, and rated accuracy of each target replacement screen, at least one available target matching screen with completely identical specifications is retrieved from the spare parts inventory database. Then, a replacement work order containing the replacement location, spare parts information, and operating procedures is generated and sent to the mobile terminal of the maintenance manager or field operator to instruct the maintenance personnel to perform the physical replacement operation.

[0058] After the replacement operation is completed, maintenance personnel confirm the completion, record the replacement log, update the equipment status file, and deduct the spare parts inventory information. This forms a complete management process from defect detection to problem resolution. The above management process is repeated periodically to ensure that all screens in the pharmaceutical production process are always maintained in a validated, safe, and efficient working state, thereby improving the quality control level of the drug production process.

[0059] Example 2

[0060] Please refer to Figure 2 This embodiment discloses a screen defect detection system for a pharmaceutical production line. The system communicates and controls image acquisition devices such as industrial cameras and automated actuators such as robotic arms and alarm devices on the production line via a data interface. The system includes the following modules:

[0061] The defect acquisition module uses a combination of a high-resolution camera and a special light source to acquire high-contrast images of the screen surface. By comparing the real-time images with the baseline features of a normal screen, it identifies the area to be tested, extracts the structured data of abnormal features as the target defect, and then records the target defect into the defect dataset. After successfully identifying the target defect, a defect trigger signal is generated, and the signal, along with information containing the location and features of the target defect, is sent to the screen diagnosis and classification module.

[0062] The minimum verification area determination module performs alignment and overlay analysis on the contour information of multiple normal screens during the system initialization or calibration phase, calculates the common overlapping area, and determines the area of ​​the common overlapping area as the minimum verification area.

[0063] Specifically, the minimum verification area determination module determines the minimum verification area by: acquiring multiple intact and unused normal screen surface images; performing edge denoising processing on the screen surface images using Gaussian filtering or median filtering to eliminate the interference of image noise on edge recognition; extracting the screen contour information; calculating the geometric center of each contour information; calculating the two-dimensional coordinate offset value of the geometric center relative to the preset fixed axis point center; determining the normal screen with the largest two-dimensional coordinate offset value as the reference screen to represent the case with the largest positional deviation under normal installation conditions; using the contour information of the reference screen as a reference, calculating the common overlapping area covered by the contour information of all normal screens in the coordinate system; and determining the area of ​​the common overlapping area as the minimum verification area.

[0064] The screen diagnosis and classification module is used to respond to the defect trigger signal generated by the defect acquisition module and to diagnose and classify the screens with target defects in order to perform a verifiability determination on the screens.

[0065] Specifically, the verifiability determination includes: first, obtaining the area of ​​the current test area and comparing it with the minimum verification area pre-calculated by the minimum verification area determination module; then, obtaining the edge features of the screen and determining whether the continuity of the edge contour and the curvature change are within the allowable range, thereby determining whether the structural stability condition is met, so as to exclude screens that cannot be effectively analyzed due to severe deformation or improper installation; specifically, when the area of ​​the test area is greater than or equal to the minimum verification area, and the flatness test result simultaneously meets the structural stability condition, the screen is determined to be a verifiable screen.

[0066] The screen diagnosis and classification module is also used to classify screens that have been determined to be verifiable screens based on preset defect standards. The process of setting preset defect standards includes: constructing multiple defect type calibration disks with defects of known size and type, and acquiring their standard defect images; attaching preset geometric offsets to the standard defect images to generate a defect area range map to simulate detection errors; defining calibration objects within this range and obtaining their defect area, length, and other parameter values; defining minor damage standards and severe damage standards to distinguish different damage levels by comparing the relationship between these parameter values ​​and a first threshold and a second threshold; comparing the parameter values ​​of the identified target defects with these two standards; and if the target defect meets the criteria... If the target defect meets the severe damage standard, the verifiable screen is marked as the target replacement screen. If the target defect meets the minor damage standard, it is marked as the target diagnostic screen. The screen diagnosis classification module is also used to perform screen diagnosis and upgrade process for screens marked as target diagnostic screens: the average aperture and aperture distribution of the screen are calculated by image analysis algorithm to perform accuracy detection on the target diagnostic screen to obtain its current actual accuracy. Then, it is determined whether the result of the accuracy detection meets the preset replacement conditions. By obtaining the rated accuracy in the factory calibration value of the screen and the detected current actual accuracy, the accuracy difference between the two is calculated. Based on the accuracy difference and the product risk weight coefficient, the diagnostic value is calculated.

[0067] If the diagnostic value is less than the performance benchmark threshold, it means that the potential drug loss caused by the decrease in the screen's accuracy has exceeded the acceptable range. In this case, the preset replacement conditions are considered to be met, and the status of this target diagnostic screen is upgraded, and the screen is remarked as the target replacement screen.

[0068] The screen replacement module is used to perform screen matching and replacement for all screens that are ultimately marked as target replacement screens. When the screen diagnosis and classification module marks or remarks a screen as a target replacement screen, the screen replacement module queries and matches the target matching screen in the spare parts inventory database according to the identification information such as the model and specifications of the screen to be replaced. After a successful match, a replacement work order containing the replacement location, spare parts information and operating procedures is generated and sent to the maintenance management system or the mobile terminal of the field operator to instruct the maintenance personnel to perform the physical replacement operation.

[0069] By detecting and addressing screen defects, screen defects can be identified in a timely manner. Furthermore, intelligent decision-making based on multi-dimensional information can differentiate between screens in different states, thereby ensuring drug quality and preventing foreign object contamination.

[0070] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for detecting defects in screens during the pharmaceutical manufacturing process in a pharmaceutical production line, characterized in that, Includes the following steps: Acquire images of the area to be tested on the screen surface and identify the target defects from the screen surface images; When the screen containing the target defect is determined to be a verifiable screen, the screen handling procedure is executed. The steps for determining a screen as a verifiable screen include: when the area of ​​the area to be tested is greater than or equal to the preset minimum verification area, and the edge features of the screen meet the preset flatness conditions, the screen is determined to be a verifiable screen. The screen handling process includes: classifying verifiable screens based on preset defect standards to mark them as target diagnostic screens or target replacement screens; performing screen diagnosis and upgrading on target diagnostic screens, including performing accuracy testing on the target diagnostic screens, and remarking the target diagnostic screens as target replacement screens when the accuracy test results meet preset replacement conditions; and performing screen matching and replacement on screens marked as target replacement screens. The steps for classifying verifiable screens based on preset defect criteria include: based on statistical analysis of parameter values ​​from a large number of standard defect samples and combined with empirical data from production processes, a first threshold and a second threshold greater than the first threshold are set. When the parameter value of a target defect is greater than the first threshold and less than or equal to the second threshold, its severity is determined to meet the minor damage standard, and the verifiable screen is marked as a target diagnostic screen; when its parameter value is greater than the second threshold, it is determined to meet the severe damage standard, and the verifiable screen is directly marked as a target replacement screen. The steps for performing screen diagnosis and upgrade on the target diagnostic screen include: obtaining the rated accuracy and current actual accuracy of the target diagnostic screen, as well as the product risk weight coefficient related to the value or quality sensitivity of the current product variety. The higher the value of the product, the larger the product risk weight coefficient value; calculating the accuracy difference between the rated accuracy and the current actual accuracy; multiplying the accuracy difference by the product risk weight coefficient to calculate the diagnostic value; comparing this diagnostic value with the performance benchmark threshold representing the minimum acceptable performance level; if the calculated diagnostic value is less than the performance benchmark threshold, upgrading the status of the target diagnostic screen, i.e., remarking it as the target replacement screen.

2. The method for detecting defects in screens during the pharmaceutical manufacturing process according to claim 1, characterized in that, Determining the minimum verification area includes: Edge denoising is performed on the surface images of multiple normal screens to obtain their respective contour information; for each contour information, the offset value of its geometric center relative to the center of a preset fixed axis point is extracted; the normal screen with the largest offset value is determined as the reference screen; based on the contour information of the reference screen, the common overlapping area defined by the contour information of all normal screens is calculated, and the area of ​​the common overlapping area is determined as the minimum verification area.

3. The method for detecting defects in screens during the pharmaceutical manufacturing process according to claim 1, characterized in that, The preset defect criteria include: Multiple defect type calibration disks are constructed, each corresponding to a standard defect image; a preset offset is added to the standard defect image to generate a defect area range map; a calibration object is defined within the defect area range map, and the parameter values ​​of the calibration object are obtained; based on the comparison relationship between the parameter values ​​and the first threshold and the second threshold, a minor damage standard and a severe damage standard are defined.

4. A system for detecting defects in screens during the pharmaceutical manufacturing process in a pharmaceutical production line, using the method for detecting defects in screens during the pharmaceutical manufacturing process as described in any one of claims 1-3, characterized in that, Includes the following modules: The defect acquisition module is used to acquire an image of the screen surface of the area to be tested, and to identify the target defect from the screen surface image to generate a defect trigger signal. The screen diagnosis and classification module is used to respond to defect trigger signals and to: determine the screen where the target defect is located as a verifiable screen; classify the verifiable screens based on preset defect standards, so as to mark the verifiable screens as target diagnostic screens or target replacement screens; The target diagnostic screen is subjected to accuracy testing, and when the accuracy test results meet the preset replacement conditions, the target diagnostic screen is remarked as the target replacement screen. The steps for determining a screen as a verifiable screen include: when the area of ​​the area to be tested is greater than or equal to the preset minimum verification area, and the edge features of the screen meet the preset flatness conditions, the screen is determined to be a verifiable screen. The steps for classifying verifiable screens based on preset defect criteria include: based on statistical analysis of parameter values ​​from a large number of standard defect samples and combined with empirical data from production processes, a first threshold and a second threshold greater than the first threshold are set. When the parameter value of a target defect is greater than the first threshold and less than or equal to the second threshold, its severity is determined to meet the minor damage standard, and the verifiable screen is marked as a target diagnostic screen; when its parameter value is greater than the second threshold, it is determined to meet the severe damage standard, and the verifiable screen is directly marked as a target replacement screen. The steps for performing screen diagnosis and upgrade on the target diagnostic screen include: obtaining the rated accuracy and current actual accuracy of the target diagnostic screen, as well as the product risk weight coefficient related to the value or quality sensitivity of the current product variety. The higher the value of the product, the larger the product risk weight coefficient value; calculating the accuracy difference between the rated accuracy and the current actual accuracy; multiplying the accuracy difference by the product risk weight coefficient to calculate the diagnostic value; comparing this diagnostic value with the performance benchmark threshold representing the lowest acceptable performance level; if the calculated diagnostic value is less than the performance benchmark threshold, upgrading the status of the target diagnostic screen, i.e., remarking it as the target replacement screen. The screen replacement module is used to perform screen matching and replacement on screens that have been marked as target replacement screens by the screen diagnosis and classification module.

5. A pharmaceutical production line screen defect detection system according to claim 4, characterized in that, The system also includes: The minimum verification area determination module processes the surface images of multiple normal screens to obtain their respective contour information, determines the reference screen based on the contour information, calculates the common overlapping area defined by the contour information of all normal screens, determines the area of ​​the common overlapping area as the minimum verification area, and provides the minimum verification area to the screen diagnosis and classification module.

6. A pharmaceutical production line screen defect detection system according to claim 4, characterized in that, The preset defect criteria include: Multiple defect type calibration disks are constructed, each corresponding to a standard defect image; a preset offset is added to the standard defect image to generate a defect area range map; a calibration object is defined within the defect area range map, and the parameter values ​​of the calibration object are obtained; based on the comparison relationship between the parameter values ​​and the first threshold and the second threshold, a minor damage standard and a severe damage standard are defined.

Images