Screen printing equipment quality control process
1、System solution: distributed, modular and universal. The image acquisition system in the program has been modularized and can be configured according to the detection width, detection accuracy, detection speed, working distance, etc. The software framework supports distributed computing, and image processing modules can be deployed according to requirements.
2、 Detection and measurement system scheme
Set various parameters required for the detection area, positioning kernel, and detection algorithm. The positioning nuclei and detection zones can be automatically selected or manually selected.
2.2 Defect detection
General-purpose detection: detects various defects with relatively large grayscale or color differences and slightly larger areas.
By positioning correction of the real-time image and the template image position deviation, the corrected image, and the standard template comparison, to get the different point, the difference point connectivity analysis to get BLOB, BLOB area, duty cycle, energy, and other shape characteristics analysis to identify defects.
Targeted detection: drawing, shallow dirty, scattered spots, string color, such defects are difficult to check the general-purpose detection, the need for targeted extraction of its features to find out.
Drawing: generally only one or two-pixel width, but the length is longer, lighter color, based on point-to-point comparison of the general-purpose algorithm is difficult to detect such defects, the need to extract features from different color components, texture, contrast, longitudinal low-frequency transversal high-frequency and other dimensions for detection.
Shallow dirty: these defects are lighter in color and larger in the area but are converted to another color space, the defects will stand out, and these defects require special color conversion to find out.
Scattered spots: these defects are discontinuous, single spot defects are not large, but the joint is larger, for the scattered spots close to one or two pixels, it can be solved by the method of mathematical morphology, and the scattered spots with larger distance can be clustered by two dimensions of color and distance.
In the printing process due to different ink key positions on the ink volume, different areas of different plate pressure, different time temperature will lead to the printed product color and the standard sample have some differences.
The color conversion model from RGB to LAB space is obtained by calibrating the collection system with a standard whiteboard and proprietary color card, and the color difference △E is obtained by comparing the LAB values of different colors in different areas of the real-time printed products with the LAB values of the corresponding areas of the standard samples.
3、Integrated control interface
The integrated control interface mainly completes the functions of system initialization, hardware status monitoring, detection result display, real-time display of acquired images, real-time parameter adjustment, data storage, data query and browsing, printing reports, etc.
4 、Application scope
The image acquisition system, computer configuration, software framework, and algorithm module of this program are all modularized and can be flexibly configured for different widths, different operating speeds, different accuracy requirements, different printing contents, and materials. This solution can also be equipped with OCR and OCV modules according to actual needs, which can better meet production inspection requirements.