This paper addresses the use of non-destructive testing (NDT) methods to assess indicators for both the concrete durability (porosity, degree of saturation) and mechanical properties (elasticity modulus, compressive strength) of reinforced concrete structures. NDT results, called 'observables', are obtained by means of ultrasonic or electromagnetic methods and then correlated with these mechanical and durability indicators. The conversion model used to transform observables into indicators depends on the actual concrete mix design. If this conversion model is unavailable for the reinforced structure under study, then the evaluation may be inadequate due to high uncertainty on the results. This paper proposes a calibration methodology to derive a conversion model appropriate for the structure by use of a minimum number of cores in order to improve the on-site evaluation. A motorway bridge is tested and characterized by NDT, after which some cores are extracted for calibration and others for validation. The cores are subsequently non-destructively characterized in the laboratory and/or used to determine indicators by means of standardized destructive methods. The non-destructive calibration protocol on cores is presented first. Next, NDT results recorded in situ and on the corresponding core are compared. Also, durability indicators deduced from on-site NDT measurements in addition to calibration are compared with reference durability indicators that have been independently determined by standard destructive methods. Results obtained by analyzing more than 1600 data fully validate the tested calibration methodology.