A safer mobility is a key concern for the strategy of the European governments and automotive industries. At the moment, a new generation of Intelligent Integrated Transport Safety Systems (IITSS) offers a promising potential to address this issue. Such co-operative systems will enable safety-concerning information to be exchanged between the vehicles or between the vehicles and the infrastructure. IITSS mainly involves sensors for detection and interpretation of the driving environment. One of the most significant related methods is the so-called Lane Departure Warning (LDW) since the unintended cross of the lane boundary is one of the largest cause of highway fatalities. Implementation of LDW requires an accurate survey of the lane layout is necessary. The Geodetic Engineering Laboratory of the Swiss Federal Institute of Technology in Lausanne has designed a mobile mapping system (Photobus) that provides fast and accurate acquisition of the road geometry. Recent surveys carried out by Photobus in Switzerland provide good basis in establishing the optimal relation between the required point sampling and the model of the road geometry. This paper focuses on methods for an accurate modelling of the road geometry and the derived geometric features for Advanced Driver Assistance Systems (ADAS). This offers new perspectives for Lane Departure Warning since precise positioning provides sufficient information for computing the offset of a vehicle from the road centreline, or from the boundaries of the lane.