Moving objects like pedestrians and vehicles change their state and location over time. Recent advances in sensors, such as Lidar and RGB-D sensors, have made it easier to track their movement in real-time. In the real world, multiple sensors individually monitor a specific area and send data streams about moving objects’ locations to a cloud server. However, the areas covered by independent sensors typically overlap, resulting in duplicate location information for moving objects. To understand the trajectory of each object, the server needs to find these clusters of overlapping locations and incorporate them into the trajectory for the time window of the streaming process, i.e., distance-based self-join. This paper proposes an efficient approach to integrating location stream data into trajectory data using a discrete global grid system. Geocode-based matching can reduce the cost of distance computation among the stream data. In experimental results, our method outperforms distance-based processing for clustering by a geocode-based index used to remove redundant trajectory information.