Collaborative SLAM and AR-guided navigation for floor layout inspection
- PDF / 2,058,637 Bytes
- 13 Pages / 595.276 x 790.866 pts Page_size
- 70 Downloads / 189 Views
ORIGINAL ARTICLE
Collaborative SLAM and AR-guided navigation for floor layout inspection Kihyun Yu1 · JeongHyeon Ahn1 · Jongmin Lee1 · Myungho Kim2 · JungHyun Han1
© Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract This paper presents how visual SLAM and AR interfaces are integrated for inspecting the floor layouts, which is a standard stage in building a semiconductor fabrication plant (commonly called a fab). The proposed system supports multiple inspectors working in a large fab floor through client–server communications. The inspectors/clients scan the fab floor using smartphones to construct the floor map collaboratively, and then, its discrepancy from the given floor plan is checked by the server. The inspection results are broadcast to all clients and visualized via AR to help the inspectors move to unscanned areas. The experiments made with the proposed system show its accuracy and efficiency. By solving a real-world industrial problem, the proposed system proves the benefits that can be brought by integrating collaborative SLAM technique and AR-guided navigation interface. Keywords SLAM · Augmented reality · Inspection · Discrepancy check
1 Introduction In industrial domains, augmented reality (AR) has been explored for visual inspection. The video capturing of an object is overlaid with the pre-defined requirement of the object, which is typically specified in its CAD model, so that the human inspectors determine whether there exists discrepancy between them. If the object’s geometric model can be created on the fly through 3D reconstruction or simultaneous localization and mapping (SLAM), however, the inspection can be made automatically.
B
JungHyun Han [email protected] Kihyun Yu [email protected] JeongHyeon Ahn [email protected] Jongmin Lee [email protected] Myungho Kim [email protected]
1
Dept. of Computer Science and Engineering, Korea University, Seoul, South Korea
2
Iaan Co., Ltd, Bucheon, South Korea
Consider a semiconductor fabrication plant or simply a fab, which refers to a factory where integrated circuits are manufactured. Equipment arrangement planning in a large fab floor goes through many steps, one of which is layout inspection. Once the fab floor’s layout (also called the floor plan) is fixed, a large number of markers for machine locations are glued onto the floor. Then, the human inspectors measure the marker positions to test if they are correctly placed. The layout inspection is not only labor-intensive but also time-consuming. In order to resolve such problems, this paper proposes automatic layout inspection based on visual SLAM (vSLAM). As illustrated in Fig. 1, multiple inspectors scan the fab floor using smartphones while communicating with the server. The floor map composed of markers is constructed progressively, and the server checks discrepancy between the floor map and the given floor plan, i.e., each marker is either approved or disapproved by the server. The decisions are immediately broadcast to all inspectors and are vis
Data Loading...
