The intensity and structure of traffic is a key information for road asset management. The traffic data are collected by a variety of measuring devices, either mounted permanently in one spot or periodically transferred between different locations. These measurements are often imperfect, because of missing data, time-shifts, implausible or even wrong values. The paper discusses organisational and algorithmic methods for ensuring the quality of traffic data at various levels: monitoring of individual sensors, scheduling of measurements within road networks, and verifying the plausibility of the obtained results. The focus is not on particular measurement technologies but rather of the whole data collection process. The presented methods are illustrated with application examples and practical experiences from Poland and Germany.