Nowadays, data for digital maps are collected using a small fleet of specialised expensive vehicles. This approach is not sufficient for daily updates of all roads. Daily or even faster updates, needed for modern driver assistance systems, can be made by combining the, individually less accurate, sensor data of standard cars. In this paper we propose a framework which comprises a parallel system to apply use case specific iterative algorithms to incoming data. Fast adaptation to changed situations is achieved by using a graph structure for representing sequences of groups of detected objects. A case study on maintaining a map of traffic signs has been performed to prove the capabilities of the framework.