Analyzing complex and large data as generated in non-destructive testing (NDT) is a time-consuming and mentally demanding challenge. Such data is heterogeneous and integrates primary and secondary derived data from materials or material systems for spatial, spatio-temporal as well as high-dimensional data analysis. Currently, materials experts mainly rely on conventional desktop systems using standard 2D visualization techniques for this purpose. Our framework is a novel immersive visual analytics system, which supports the exploration of complex spatial structures and derived multidimensional abstract data in an augmented reality setting. It includes three novel visualization techniques: MDD-Glyphs, TimeScatter, and ChronoBins, each facilitating the interactive exploration and comparison of multidimensional distributions from multiple datasets and time steps. A qualitative evaluation conducted with materials experts and novices in a real-world case study demonstrated the benefits of the proposed visualization techniques. This evaluation also revealed that combining spatial and abstract data in an immersive environment improved their analytical capabilities and facilitated to better and faster identify patterns, anomalies, as well as changes over time.

翻译：分析无损检测（NDT）中产生的复杂大数据是一项耗时且脑力要求高的挑战。此类数据具有异质性，整合了来自材料或材料系统的原始及衍生数据，用于进行空间、时空以及高维数据分析。目前，材料专家主要依赖传统桌面系统，使用标准二维可视化技术为此类分析提供支持。我们的框架是一种新颖的沉浸式可视分析系统，能够在增强现实环境中支持对复杂空间结构和衍生多维抽象数据的探索。该系统包含三种新型可视化技术：MDD-Glyphs、TimeScatter和ChronoBins，每种技术均能促进对来自多个数据集和时间步长的多维分布进行交互式探索与比较。在真实案例研究中，通过对材料专家和新手进行的定性评估，证明了所提出可视化技术的优势。该评估还揭示，在沉浸式环境中结合空间数据与抽象数据，能够提升分析能力，并有助于更快速、更准确地识别模式、异常现象以及随时间的变化。