Mask Detection Method and Apparatus

CN122089738APending Publication Date: 2026-05-26CHENGDU CNS VISION TECH CO LTD
View PDF 6 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-04-24
Publication Date
2026-05-26

AI Technical Summary

Technical Problem

Existing technologies cannot meet the high-precision, high-efficiency, and fully automated mask defect detection requirements of OLED production lines. Manual visual inspection is inefficient and prone to missed or false detections, while traditional machine vision inspection is insufficient in detecting minute defects and cannot be adapted to different models and mesh parameters.

Method used

A mask detection method is adopted. By constructing a defect dataset, feature extraction and pyramid network fusion are performed to build a defect detection model. The model is optimized by combining Focal Loss, DIoU Loss and Dice Loss to achieve automated detection. Based on the detection results, process adjustment instructions are generated and fed back to the processing equipment.

Benefits of technology

It enables automated detection of mask defects, reduces personnel workload, improves production efficiency, corrects process deviations in a timely manner, avoids the generation of batch defective products, and reduces production costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122089738A_ABST
    Figure CN122089738A_ABST
Patent Text Reader

Abstract

This application discloses a mask inspection method and apparatus for detecting defects in a mask after screen stretching. The method includes acquiring an initial image of the mask to be inspected; standardizing the initial image to obtain a standardized image; inputting the standardized image into a defect detection model for inspection to obtain a detection result; and generating a process adjustment instruction based on the detection result (if the result is not satisfactory), and feeding the process adjustment instruction back to the mask processing equipment to adjust the processing parameters of the processing equipment. According to the mask inspection method of this invention, process adjustment instructions can be generated in real time based on the detection result and fed back to the mask processing equipment, realizing a closed-loop linkage between the processing steps and defect detection. This allows for timely warning and correction of processing deviations, avoiding the generation of batch defective products and further reducing manufacturing costs.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Netting device and netting method for mask template

    CN106086784A

  • Net expansion equipment

    CN107732012A

  • Deep learning-based tomato disease and insect pest detection and identification method

    CN116630803A

  • Mask plate defect detection method, system and equipment based on deep learning and medium

    CN120807506A

  • Greenhouse tomato fruit maturity instance segmentation method based on dual-backbone network fusion

    CN121789210A