基于粒度计算的空间大数据增强现实可视分析方法研究

Big Data is everywhere in your daily life, even at your fingertips. A picture is worth about thousands of words, when the big data was been displayed in graphically and intuitively, people can discern the information behind the data and convert knowledge at a glance. On the other side, augmented reality is becoming the new generation computing platforms and interactive interface. However, the current augmented reality applications which with a low viewing angle, narrow viewing field, and under near-Earth space scales, is under the condition of enhanced data fusion position excessive accumulation and clutter information level confusion, lack of decision-making knowledge, make the interaction complex, low application bonding degree, that make the AR lacks of killer apps. This application is aiming at big data integration technology, information extraction technology and visual analytic technology of knowledge analysis in decision-making methods and strategies under the big geospatial data environment, which with a multi-source, heterogeneous, dynamic and multi-modal spatial data. We introduction of granular computing theory into the field. After distinguishing the professional geospatial service and public map service applications, we study the visualization method driven by multi-granularity information fusion and augmented reality interactive display, to achieve the skills and visual effects which result to “Visual that is required, needed to make visible”. This application is intended to explore augmented reality typical application scenarios and interaction model to improve the augmented reality display visual information analysis and decision-making capabilities for applications, a large space for the data processing and analysis provide new ideas and research point of view of the system.

大数据无处不在、触手可及；一幅图胜过千言万语，当大数据以图形直观展示时，人们能一眼洞悉数据背后的信息并转化为知识；另一方面，增强现实正在成为新一代计算平台与交互接口。然而当前低视角、窄视域、近地空间尺度下增强现实应用中，融合的增强数据位置过度堆集与杂乱、信息层次混淆、决策知识不足，使得交互复杂，应用黏合度低，缺乏“杀手锏”级应用。本申请针对大数据环境下多源、异构、动态、多模态的空间数据，研究数据的集成管理、信息提取与可视分析知识决策过程的方法与交互策略，通过引入粒度计算理论，区分专业地理信息服务应用与大众地图服务应用的不同交互模式，研究应用驱动的多粒度信息融合与增强现实交互展示的可视化方法，达到“可视即所需、所需尽可视”的分析能力与可视效果。本申请旨在探索增强现实的典型应用场景与交互模式，提高面向应用的增强现实可视分析与决策能力，为空间大数据处理与分析提供新的系统的研究角度与思路。

可视化分析与应用是进行深层次数据挖掘与知识发现的重要方法。本研究首先研究了空间环境的多源、异构与多模态数据的数据融合，进行空间环境的快速构建与集成，立足空间环境的增强现实可视分析与应用，以点云数据出发，研究校园空间环境的建筑提取，分析增强现实注册策略，研发贴合建筑表面显示方法，构建了基于校园导航的移动增强现实应用；在此基础上，扩展到教育教学环境、线上学习等具体应用场景，发展地图可视分析的理论与方法，结合校园生活大数据，研究基于活动行为轨迹的学习行为分析，同时提出了学科知识的地图空间化理论与方法，构建了基于地图的学科知识可视化表达与基于地图浏览的在线学习方式；结合混合现实技术，面向特色课程构建了具身认知的大学生在学伴，有效探索了增强现实与校园环境、学习深度结合的应用案例，对推动校园空间增强现实应用，提供有效参考。