Digitization of critical infrastructure structures using multi-data fusion. The rail infrastructure is highly significant for the economy and society in the Federal Republic of Germany and the European Union. On average, rail infrastructure bridges are around 75 years old. Accordingly, the inventory documentation of the structures is also historical or outdated. For Deutsche Bahn AG, the inventory, updating and provision of more than 4.6 million as-built plans represents an enormous challenge. Furthermore, the as-built documents are managed decentralised and in heterogeneous data formats at the plant operator. This poses a serious problem for a qualitative and efficient maintenance management. Within the research project mdfBIM, a holistic, semi-automated method for the creation of object-based, digital models of as-built bridges by means of multi-data fusion is being developed. For this method, a holistic process model is developed and the possibilities of a semi-automated data acquisition and processing towards a BIM model as a central data basis are explored. The approach was subsequently validated on a real demonstrator. The article successfully demonstrates a partially automated approach using laser scanning, artificial intelligence and digital building models to digitize bridge structures and create a consistent database for maintenance management.