Error matching detection and robust estimation adjustment approach for map conflation

The conflation of geographic datasets is one of the key technologies in the realm of spatial data capture and integration in geographic information system (GIS). Map conflation is a complex process of matching and merging spatial data. Due to various reasons such as errors in original data related to map data discrepancies, a great amount of uncertainties exists during the process and this will result in errors in featuring matching, especially point feature. Thus, it is vital to develop the method to detect the errors in feature matching and further the conflation results will not be affected. In this paper, error matching detection and robust estimation adjustment methods are proposed for map conflation. The characteristics of errors in feature matching are first analyzed, then a new approach for map conflation based on the least-squares adjustment is presented, and a robust estimation adjustment method is further proposed to detect and process matching errors. The results of the map conflation test show that the proposed method not only determines the errors in feature matching, but also obtains the optimal merging results in map conflation.