Roads matching in a multi-scale spatial database using a least square line

With the rapid development of spatial data production, collection, and processing techniques large amounts of data, which cover the same geographical extent, are available. These huge amounts of data motivate the researchers in geospatial analysis domain to create multiscale database in order to update the datasets and perform the multi-scale analysis. In the core of any spatial multiscale database is the data matching process. In this paper, a new algorithm is proposed for data matching which is implemented and evaluated in two major steps. Firstly, after pre-processing, the search and identification of candidate objects in the two datasets are carried out based on the topological features. Secondly, the best pair of corresponding objects is selected using the least squares method and comparing the geometric features of linear objects between the candidate object selected in the previous step. The algorithm is evaluated on the roads maps of scales1:25000 and1:50000 produced by national cartography center of Iran (NCC).The results show that the algorithm is successful and 94.2% of the objects were correctly identified as the corresponding pairs.