Automatic Generation of Digital Twin Based on Scanning and Object Recognition

Abstract

Digital Twin has been recognized as a strategic approach in the modern manufacturing industry to improve both the flexibility and the efficiency. To efficiently generate the Digital Twin of an existing real object in the factory, powerful methods are necessary. Hereby, a fast data acquisition including object recognition and model reconstruction methodology has been combined to resolve these issues. Such a data set often has to replace the missing original digital model. Subsequently, a model reconstruction plan has to be derived so that an editable CAD model, which fulfils process requirements, can be generated using standard geometry creation tools. Such a reverse-engineered CAD model preferably contains form-feature based design intent and can be easily modified due to new design and manufacturing constraints. The presented paper describes an industrial approach for a commercial service being in the implementation to generate the Digital Twin based on fast scanning on a factory.