While Vehicle-to-Everything (V2X) messages from connected vehicles can act as reliable data sources in real-time road infrastructure monitoring, their low penetration rates cannot guarantee a continuous monitoring in practice. Here, Collective Perception (CP), especially from infrastructure units, can help to bridge the gap. Unfortunately, objects detected by infrastructure are inherently more noisy and, often, a complete trajectory over an intersection cannot be obtained. To overcome this problem, we present an approach to augment trajectories collected by CP with data gathered from connected vehicles, thereby allowing a continuous monitoring.