Flexible Panel Repair Using AI Defect Classification

Resolve Bottlenecks,
Find Innovative Solutions
Generate 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

VSEngineering 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

Engineering Contradiction:
Improvedefect classification accuracyVSAvoidrepair efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

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

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

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

Inventive Principle:
Principle #35Parameter changes

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

Engineering Contradiction:
Improveprocess simplicityVSAvoidrepair quality
Core Design Contradiction:
Ease of manufactureVSReliability

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

Inventive Principle:
Principle #3Local quality

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

Inventive Principle:
Principle #10Preliminary action

3Reliability

If extended baking time is used for repair, then repair completeness is improved, but production efficiency and throughput are reduced

Engineering Contradiction:
Improverepair completenessVSAvoidproduction throughput
Core Design Contradiction:
ReliabilityVSProductivity

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

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4068202B1Method and repair apparatus for repairing flexible panel, device and storage medium
Publication Date: 2026.03.18 BEIJING XIAOMI MOBILE SOFTWARE CO LTD
  • EP4068202B1 patent drawingFigure 1~2
  • EP4068202B1 patent drawingFigure 3~4
  • EP4068202B1 patent drawingFigure 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.