平板显示器件Mura缺陷的计算摄像检测研究

The flat panel display nowadays has a 170-million-square-meter annual output, which is an important spatial light modulator in computational photography study as well as the core component of the consumer electronic products. The process and structure of flat panel display are complex, and there are low contrast and edge fuzzy Mura defects. The mechanism study of Mura production could not make a breakthrough because of the lack of high-precision and high-efficiency testing technology, so it has restricted the industrial development over a long period. This project is aimed at Mura detection which is an international difficulty of weak characteristic light information reconstruction. Fully considering the intrinsic characteristics of flat panel display as a spatial light modulator (SLM), under the guidance of computational photography, this project creates a modulating image database by acquiring coupled spectrum modulating images, extracts the modulating intrinsic properties of Mura defects, learns the natural mechanism of weak characteristic detection, summarizes the appearance model of Mura, as well as establish new method of rapidly extracting and detecting weak characteristic defects in non-spatial/frequency-domain filtering. This project also builds a verification platform for feedback and optimization, which changes the current Mura detection system in both basic principle and system framework. This project strives to raise a new understanding of weak-characteristic defect detecting rules in scientific principles, as well as guide the practice of industrial inspection to support the mechanism study of Mura production.

平板显示器件年产量1.7亿平方米，是计算摄像研究中的重要空间光调制器件，更是消费类电子产品的核心部件。平板显示器件工艺和结构复杂，存在低对比、边缘模糊的Mura缺陷。Mura产生机理研究因高精高效检测技术的缺乏而无法取得突破性进展，长期制约了产业发展。本项目针对Mura检测这一是弱特征光信息重建国际难题，充分考虑平板显示器件的空间光调制器本质，在计算摄像学思想指导下，耦合采集光谱调制图像，形成调制图像数据库，挖掘Mura缺陷在调制能力上的本质特性，学习理解弱特征探测的自然规律，总结归纳Mura的表征模型，构建非空/频域滤波的弱特征缺陷快速提取和判别方法。同时搭建验证平台实现反馈优化，从基本原理和系统构架两个层面改变目前平板显示器件Mura缺陷检测的体系。力争在科学原理方面促成对弱特征缺陷检测规律的新认识；指导工业检测实践，为Mura缺陷产生机理研究提供支撑。

平板显示器件年产量1.7亿平方米，是计算摄像研究中的重要空间光调制器件，更是消费类电子产品的核心部件。平板显示器件工艺和结构复杂，存在低对比、边缘模糊的Mura缺陷。 Mura产生机理研究因高精高效检测技术的缺乏而无法取得突破性进展，长期制约了产业发展。 本研究通过挖掘平板显示器件的空间光调制器本质，开展了弱特征缺陷的多光谱耦合采集和重建研究，搭建验证平台并形成了缺陷数据库；从基本原理层面探索弱特征缺陷检测规律的新认识，在系统架构层面指导工业检测实践。 本项目研究成果充分挖掘空间光调制机理和特性，提出了调制深度缺陷表征模型，刻画了缺陷与非缺陷的本质区别；发明了缺陷区域直接提取算法，抛弃多尺度边缘滤波提高效率；设计了空域欠缺陷检测架构，实现三维缺陷检测精度自适应提高。在项目研究中，课题组成员在国内外重要学术期刊上发表了6篇高质量论文，获得7项发明专利授权，缺陷检测技术应用并保障了自然基金委重大仪器的顺利研制。综上所述，课题组已完成了预期的研究目标，在理论和关键技术研究上取得了突破，为后续的相关研究奠定了坚实的基础。