Flexible Panel Repair Using AI Defect Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing flexible panel repair methods rely heavily on manual experience, leading to inefficiencies and low production yield and efficiency due to inaccurate defect classification and uniform repair processes.
Innovation Solution
A method utilizing a deep learning model to classify flexible panel defects and determine optimal repair parameters based on image information, eliminating the need for manual intervention and enabling intelligent, tailored repair processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual experience is used to determine repair parameters, then flexibility in handling different defects is maintained, but classification accuracy and repair efficiency are reduced
Solution Approach 1:
The patent replaces the manual mechanical inspection and decision-making system with an automated optical inspection system coupled with machine learning algorithms. The system captures images of panel defects, automatically classifies them using trained models, and determines optimal repair parameters without human intervention, thereby improving both classification accuracy and repair efficiency simultaneously
Solution Approach 2:
The patent changes the operational parameters of the repair system from fixed manual settings to dynamically optimized parameters based on defect classification. Different defect types (Sandy Mura, black masses, etc.) are matched with specific repair parameters such as baking temperature, time, and atmosphere, enabling precise and efficient repair for each defect category
2Ease of manufacture
If uniform repair process is applied to all panels, then process simplicity is maintained, but repair quality and yield rate are reduced
Solution Approach 1:
The patent implements local quality by tailoring repair parameters to specific defect characteristics. Instead of applying a uniform repair process to all panels, the system classifies defects into categories (Sandy Mura, black masses, etc.) and applies customized repair parameters to each category, ensuring optimal repair quality for each defect type while maintaining automated process simplicity
Solution Approach 2:
The patent performs preliminary classification and parameter optimization before the actual repair process. By pre-training machine learning models with historical defect data and repair outcomes, the system determines the optimal repair parameters in advance for each defect type, ensuring high repair quality without compromising process simplicity during actual production
3Reliability
If extended baking time is used for repair, then repair completeness is improved, but production efficiency and throughput are reduced
Solution Approach 1:
The patent introduces dynamic adjustment of repair parameters based on real-time defect classification. Instead of using fixed extended baking times for all panels, the system dynamically optimizes repair duration and temperature based on the specific defect type and severity, achieving complete repair for each defect category while minimizing unnecessary processing time and maintaining high production throughput
Data Source
Figure 1~2
Figure 3~4
Figure 5
AI summary
The present disclosure relates to a method and an apparatus for repairing a flexible panel, a device and a storage medium. The method includes: obtaining (101) first image information of a flexible panel to be repaired; determining (102) a first classification tag of the flexible panel to be repaired based on a deep learning model and the first image information; and determining (103) an operating parameter of a repair device according to the first classification tag, and repairing the flexible panel to be repaired based on the operating parameter.