Roads are important objects for many applications of topographic data. They are often acquired manually and as this entails significanteffort, automation is highly desirable. Deficits in the automatic extraction hindering a wide-scale practical use haveled to the ideaof setting-up a EuroSDR test comparing different approaches for automatic road extraction. The goal is to show the potential of thestate-of-the-art approaches as well as to identify promising directions for research and development. After describing the data and theevaluation criteria used, we present the approaches of a number of groups which have submitted results and give a detailed discussionof the outcome of the evaluation of the submitted results. Wefinally present a summary and conclusions.