Industrial Layout Mapping by Human-Centered Approach and Computer Vision

Digital twins and immersive mixed reality have been improving the human-centered digital transformation within the automotive industry context. In computer vision, photogrammetry-based techniques for reconstructing 3D environments make it possible to create virtual models of real scenery based solely on a set of digital images. This paper addresses the problem of projecting and updating shopfloor layouts through 3D mapping of the real environment, corresponding to its digital twin. Within industrial plants, we encounter objects of various sizes, shapes, and layouts, which would typically require a significant amount of time to create a digital twin model using manual measurement devices such as tape measures and clipboards to annotate length, distances, and angles. Through photogrammetry using images, we can recreate these layouts and objects with accuracy, speed, and rich details, eliminating the need for hand-drawing them in CAD software. Therefore, we propose an automated method for digitizing shop floor layouts and creating digital twins where humans can interact and make changes to the modeled equipment structure to simulate different arrangements before implementing actual changes on the shop floor in an industrial setting. This text presents the innovations to digital environments technologies, discusses the feasibility, and evaluates the expectation of the proposed method.