Industrial Augmented Reality (iAR) has demonstrated its advantages to communicate technical information in the fields of maintenance, assembly, and training. However, literature is scattered among different visual assets (i.e., AR visual user interface elements associated with a real scene). In this work, we present a systematic literature review of visual assets used in these industrial fields. We searched five databases, initially finding 1757 papers. Then, we selected 122 iAR papers from 1997 to 2019 and extracted 348 visual assets. We propose a classification for visual assets according to (i) what is displayed, (ii) how it conveys information (frame of reference, color coding, animation), and, (iii) why it is used. Our review shows that product models, text and auxiliary models are, in order, the most common, with each most often used to support operating, checking and locating tasks respectively. Other visual assets are scarcely used. Product and auxiliary models are commonly rendered world-fixed, color coding is not used as often as expected, while animations are limited to product and auxiliary model. This survey provides a snapshot of over 20 years of literature in iAR, useful to understand established practices to orientate in iAR interface design and to present future research directions.

中文翻译：

---

在工业增强现实中使用什么、如何以及为什么视觉资产？维护、装配和培训的系统审查和分类（从 1997 年到 2019 年）

工业增强现实 (iAR) 已证明其在维护、装配和培训领域的技术信息交流方面的优势。然而，文献分散在不同的视觉资产（即与真实场景相关联的 AR 视觉用户界面元素）中。在这项工作中，我们对这些工业领域中使用的视觉资产进行了系统的文献综述。我们搜索了五个数据库，最初找到了 1757 篇论文。然后，我们选择了 1997 年到 2019 年的 122 篇 iAR 论文，并提取了 348 个视觉资产。我们根据（i）显示的内容，（ii）它如何传达信息（参考框架、颜色编码、动画）以及（iii）使用它的原因建议对视觉资产进行分类。我们的审查表明，产品模型、文本和辅助模型依次是最常见的，最常用于分别支持操作、检查和定位任务。其他视觉资产很少使用。产品和辅助模型通常是世界固定的，颜色编码没有像预期的那样经常使用，而动画仅限于产品和辅助模型。该调查提供了 20 多年 iAR 文献的快照，有助于了解确定 iAR 界面设计方向和提出未来研究方向的既定做法。