AI眼镜与透明光场屏深度结合文博场景
在第十一届中国博物馆及相关产品与技术博览会上,“衍射光波导”技术成为焦点。该技术通过AI眼镜与透明光场屏等形式,实现虚拟画面与文博场景的深度结合,改变了观众看文物讲解及三维影像的方式。
AI-generated summary
第十一届中国博物馆及相关产品与技术博览会在海口开幕,展示了多项文博科技创新成果。
中新网海口11月13日电 (记者 王晓斌)一种叫“衍射光波导”的技术,正在改变中国博物馆里看文物的方式。记者在13日开幕的第十一届中国博物馆及相关产品与技术博览会(下称“博博会”)上看到,这项“魔法光”技术已经和文博场景深度结合——观众可以选择戴上AI眼镜看文物讲解,也可以什么都不戴,看三维影像悬浮在一块透明玻璃中。
专家介绍,衍射光波导的原理可以用一句话概括:光在刻有纳米级光栅的薄玻璃或镜片里会“拐弯”,被“导”着沿设计好的路径传播,最后在需要的位置射出。射出的位置不同,产品就分成了两类:一类是近眼显示,图像光经波导耦合进入瞳孔,在眼前叠加一层可交互的虚拟画面,做成AI眼镜;另一类是空间成像,光在屏幕前方的自由空间中汇聚成像,观众裸眼像看实物一样观看悬浮的影像,做成透明光场屏。
博博会现场展示的乐奇Rokid文娱行业版全彩AI眼镜重55克,以海南骑楼老街、海南科技馆、海大南门夜市等场景串联体验内容,支持AI讲解、AI问答、智能导览和多语种讲解。灵伴科技(杭州)股份有限公司的工作人员介绍,这副眼镜可结合时间、位置与游览偏好提供个性化路线建议,面向游客提供多达13种语言的讲解与互动问答。
AI眼镜在博物馆里的赋能价值,已经在实际运营中显现。2026年6月,良渚博物院上线全馆AI眼镜智能导览,观众驻足玉琮前,眼镜自动触发讲解;长按镜腿追问“它在良渚文化中有什么功能”,眼镜在眼前给出答案,并在视野中叠加显示玉琮‘内圆外方’的图文示意。据灵伴科技方面披露的数据,其AI+AR智能眼镜已覆盖200余家博物馆和150余处古镇景区,累计接待游客500万人次,人均体验时长突破90分钟。
如果说AI眼镜是一对一的个性化导览,透明光场屏就是一对多的集体观看。观众不戴任何设备,一群人站在同一位置,就能同时看到三维影像悬浮在半空,背后的真实景观不被遮挡。博博会现场,苏州苏大维格科技集团股份有限公司的工作人员告诉记者,这套基于大幅面纳米光场屏的光调控技术的透明屏,影像看起来就像悬浮在十几米外的半空中,画面清晰到看不出像素颗粒。
透明光场屏落地样本是苏州丝绸博物馆“未来馆”。装置设在落地窗前,窗外是千年北寺塔实景,屏中悬浮着动态演示采桑制丝的古装采桑女,千年古塔与三维影像交叠在一起。这套“面向文物数字孪生的跨媒介虚实叠加显示系统”入选了工业和信息化部等五部门公示的2025年度视听系统典型案例。2026年9月,苏大维格又与苏州御窑金砖博物馆合作,将元和塘古窑遗址的发掘过程和传统金砖烧制工序以三维动态形式悬浮呈现在实体沙盘之上,观众裸眼即可俯瞰古窑群落的整体格局。
AI outlook — possibilities, not facts
AI眼镜与透明光场屏将在更多博物馆景区落地应用
Likely · Within months
As the 33rd APEC Leaders’ Informal Meeting is about to be held in Shenzhen, innovative cooperation between artificial intelligence and small and medium-sized enterprises has become the focus of all parties. Shenzhen has attracted a large number of global makers through its open supply chain and open source ecosystem, effectively narrowing the intelligence gap and promoting cooperation in the Asia-Pacific region.
Through primate experiments, a team of researchers from the Institute of Automation, Chinese Academy of Sciences, Yu Shan, revealed that there is an explicit representation of hand position in three-dimensional space in the motor cortex of macaque brains. The brain maintains a dynamic three-dimensional hand position map similar to GPS, providing underlying neuroscience principles for the development of efficient brain-computer interfaces.

Taoyuan MRT Corporation announced that it will launch three major highlights in optimizing ride services starting from December 8, including increasing the number of international credit card organizations from 4 to 6, expanding electronic payment to 17, and gradually expanding mobile payment gates to all gates on the entire line to improve the convenience of passengers and Taoyuan citizens.

The Taipei City Legal Affairs Bureau has invited major domestic e-commerce platforms and central units to hold a fraud prevention meeting to review current fraud prevention measures, discuss the introduction of AI monitoring to enhance high-risk message identification and real-time warning mechanisms, and promote the inclusion of friendly e-commerce assessment projects.

Google DeepMind launched the Gemini 4 Hydrogen model on November 12, focusing on faster speed and lower cost. The new model's task speed is 2.5 times faster than the previous generation, errors are reduced by five times, and it supports millions of token contexts and multi-model collaboration. It has been gradually rolled out to Gemini Apps and developers around the world.
The 2026 World Chinese Language Conference opened in Beijing on November 13. At the conference, the Yifei International Chinese Education Vertical Model was released to provide intelligent services to Chinese learners, teachers and educational institutions around the world, and to promote the digital and intelligent development of international Chinese education.