面向复杂装备AR辅助维修的混合跟踪技术研究

装备维修是生产制造、作业运转的基础,特别是在船舶、航空等领域。由于在维修过程中装备结构复杂、维修人员经验技术不足和缺少直观维修手段导致容易导致维修效率低下和操作出错。因此,通过研究基于增强现实的装备维修方案,直观交互地进行维修作业,能够有效地提升维修的质量和效率。由于装备体型较大,结构复杂,给跟踪定位带来了新的困难,针对实际的工程需求,创新性地将主流方法进行结合,提出了多Marker和ORB-SLAM混合跟踪的方法。通过二次调整多Marker定位结果的动态权值,提高局部定位维修的精度和可靠性。分析比较现有的特征点算法,选择ORB-SLAM作为方案,实现全局的定位和导航,对维修过程给以指导,并设计开发装备诱导维修系统,为大型复杂设备地维修提供了通用解决方案。

Research on hybrid tracking technology for AR aided maintenanceof complex equipment

Equipment maintenance is the basis of production and operation, especially in the fields of ship, aviation, etc. Due to the complexity of equipment structure, the lack of experience and technology of maintenance personnel and the lack of intuitive maintenance means in the maintenance process, it is easy to lead to low maintenance efficiency and operation errors. Therefore, through the research of equipment maintenance scheme based on augmented reality, the maintenance work can be carried out intuitively and interactively, which can effectively improve the quality and efficiency of maintenance. Due to the large size and complex structure of the equipment, it brings new difficulties to the tracking and positioning. According to the actual engineering requirements, the mainstream methods are creatively combined, and a hybrid tracking method of multi marker and ORB-SLAM is proposed. The accuracy and reliability of local location maintenance can be improved by adjusting the dynamic weight of multi marker location results twice. This paper analyzes and compares the existing feature point algorithms, selects ORB-SLAM as the scheme, realizes the global positioning and navigation, guides the maintenance process, and designs and develops the equipment induction maintenance system, which provides a general solution for the maintenance of large and complex equipment.