People experience and memorize space primarily with the help of landmarks. These landmarks have structural salience, besides visual and semantic salience. When people move in urban space they perceive first the street network as structuring this space. Therefore, streets are a good candidate for investigating structural salience. This paper investigates different structural representations of the urban fabric, and measures to describe the structural salience especially of elements of the street network and dependent elements. The measures are taken from topology and network analysis. The goal is to identify a generic model of structural salience for urban elements that favors the automatic identification of references for route directions. The proposed model is illustrated by a case study applied to a small city in northern France.