智能视觉检测技术

智能视觉检测技术

  • 智能视觉检测技术
    ART/201CP
    20151001 - 20170630
    11936

    蔡振荣博士
    1. Develop mixed characteristic surface visual inspection technologies, including (1) Develop human eye bionic illumination optical system: a) light distribution angle: ≤20 degree ; b) Brightness: >500 LUX (2) Develop image enhancement algorithm for projection image and capture image compensation (3) Develop large size 3D measurement methodology including pattern generation system and point cloud generation algorithm: a) FOV: >150mm, b) detected size 0.1~1mm (diameter) (4) Develop machine learning based illumination optimization & defect detection methodology (5) Develop anti-glare image capture optical design with locally tunable dynamic range: a) Micro-display panel resolution: <1024x720; b) Image sensor resolution: >2.0M pixels, c) F-number: >3.0 (6) Develop mixed characteristic surface visual inspection system: • Object size: <350x350mm • Defect size: >0.15mm • Operation speed: <10s/station 2. Develop multiple layers transparent surface visual inspection technologies, including: (1) Develop all in focus image capture system, including: optical/mechanical/electrical designs: a) Sensor resolution: ≥1.0M pixels; b) FOV: <20x20mm; c) Optical depth: <50um; d) F-number: >3.0; e) Working distance: >5mm (2) Develop blur image comparison methodology for defect detection. (3) Develop multiple layer transparent surface visual inspection system: • Detected thickness: 50um~800um • Object size: 3~8” • Depth of focus(DOF): <50um • Defect size: >0.1mm (point) • Operation speed: <5 second/single surface 3. Develop curve transparent surface visual inspection technologies, including: (1) Develop curve transparent surface image capture technology, including: optical/mechanical/electrical designs: a) Resolution of spatial light modulator: < 1024x720 b) Sensor resolution: ≥2.0M pixels; c) FOV: >6x6mm; d) Optical depth: >0.5mm; e) F-number: >3.0; f) Working distance: >10mm (2) Develop curve transparent surface visual inspection system: • Object size: 2~20mm (in diameter) • Depth of focus(DOF): >0.5mm • Defect size: >0.01mm • Operation speed: <5s/object 4. Contract service deliverables for Customer (1), including: (1) One set of prototype for touch panel (OGS (One Glass Solution, 一體化觸控) without projection film) defects inspection: • Measurement area: front/back panel • Object size: 3~8” OGS panel • Defect size: >0.1mm (point) • False detected rate: <1% • Fail detected rate: <0.1‰ • Operation speed: <5 second/single surface (2) One set of prototype for touch panel (OGS (One Glass Solution, 一體化觸控) with projection film) defects inspection: • Measurement area: front/back panel • Object size: 3~8” OGS panel • Depth of focus(DOF): <50um • Defect size: >0.1mm (point) • False detected rate: <2% • Fail detected rate: <0.5‰ • Operation speed: <5 second/single surface 5. Industrial implementation of the project deliverables: Industrial trial run of project deliverables in Customer (1)'s factory or other industrial partner’s production line.
    Mr Pei Chin, Ryan Chung Mr Kin Lung, Kenny Chan Ms Ying, Anna Liu Dr Xiu Ling Zhu Mr Derek Dehu Liu Mr Wei-ping Tang Mr Chang-li Wu Dr Chi Yong Mr Chun Yip Wong Dr Xueyan Tang Ms Wei, Wendy Zhang Mr Jie KOU Dr Zhao, Fiona Wang Mr Chuanyun Deng Mr Chi Kin Wong Dr Yan Wang Mr Yuk lung, Alex Cheung Mr Chak Wai Chau Mr Kwok Chung Chan Mr Hon Lung Cheung Ms jie, Fion Lu Ms Siqi, Cathy Li Mr Weiwen, Wilman Zou Ms Hui, Tracy Hu Ms Ka Man Liu Mr Wai Kin Luk
    K&K CORPORATION LTD K&K Corporation Ltd. (Licensing) [赞助机构] Kopin (HK) Limited [赞助机构] Truly Opto-Electronics Limited [赞助机构] Truly Opto-Electronics Limited (Contract Service) [赞助机构] Truly Opto-Electronics Limited (Licensing) [赞助机构]

    在生产製造过程中,表面品质对终端产品的外观和零件功能都是非常重要的评定 参数. 以一个 iPhone 的代工厂为例,在 iPhone 的製造流程中,大於 30%的人工 均消耗在手机的外观检测上,每年需支付给检测员的人工高达 48 亿人民币。 然而随著近几年中国每年大於 10%的工资增长速度,低劳动成本在中国已不复 存在。因此削减人工成本已成为各製造企业最为重要和急需解决的问题。基於智 慧图像采集和影像处理技术的自动化机器视觉检测系统是解决这一问题的有效 方案。它可以全面提高生产速度和品质进而取代昂贵的劳动力。 本项目主要目的是开发一套针对不同表面特性的系统级智能线上机器视觉表面 检测技术平臺。为开发此技术,本项目将专注下列有专利保护的平臺技术的开发, 包括 1). 多层透明表面瑕疵检测 2)弯曲透明表面瑕疵检测 3)混合表面瑕疵检测。 上述技术可广泛应用於手机、平板电脑、笔记本和桌上型电脑等各种消费类电子 在生产过程中的表面品质检测,该类产品超过 60%均在中国生产製造。据 IDC 预测,2014 年,全世界智慧手机的出货量将达到 12 亿部,平板电脑出货量达 2.6 亿部,笔记本和桌式电脑出货量则达到 2.9 亿部,且该趋势在未来几年将不断上涨。这意味著对检测设备的出货需求量将超过 20000 台/年。按每台设备 100 万人民币估值,该技术平臺将给业界每年带来超过 200 亿人民币的产值。