ABSTRACTIn order to keep up with the economic and safety demands of modern tunnel construction projects, especially in urban areas, there is a need to detect safety threats in real time while the construction is advancing. Tunnel prediction methods accompanying the drilling process can help to correlate and update a priori information on expected geological structures with their actual spatial location or even existence ahead of the tunnel face. We recently presented a seismic prediction approach using tunnel surface waves, which has already proven its potential during field surveys. However, common tunnel seismic data interpretation, regardless of the prediction method, requires human interaction. Either a specially trained field technician have to be present at the construction site or the data has to be uploaded to an office for further interpretation. In this work we present a simple but stable approach to automatically detect major geological structures ahead of the tunnel face. We focus on the accurate determination of the distance of fault zones or lithological boundaries from the tunnel face without any a priori information. By 3D seismic finite difference modelling we simulated a synthetic tunnel seismic survey that includes typical features encountered in tunnel construction. The developed prediction sequence was tested on these data and later successfully applied to two different tunnel data sets acquired at the Gotthard Base Tunnel (Switzerland) and during the construction of the ‘Neuer Schlüchterner Tunnel’ close to Fulda (Germany).