Rapid Algorithm of Building Typification in Web Mapping

When comes to on-demand or real-time web mapping, difficulties is how to support adaptive and intelligent zooming and to generate web map that meeting user request rapidly. Solutions can be categorized into three approaches. However they do not fulfil the requirements for reason of performance or economy. The new strategy which intends to integrate multi-scale databases and on-thefly generalization is admired. The key point here is how to choose map scales in a multi-scale database and to develop effective online generalization operators to achieve balance between performance and economy. This paper mainly concerns building typification in the context of web mapping application. After pointing out the advantages of mesh simplification technique adapted by (Alessandro, 2003), further research shows that it is not a good delegate for the idea of integration. Thus an algorithm to typify buildings based on data matching is proposed by spatial partition and clustering. The matching process is composed of three stages, i.e. many-to-many matching by means of spatial partitioning, many-to-one matching by means of spatial clustering method, data enhancement and storage in specified data structure of building object. The typification procedure of buildings contains three steps. The first step is to determine the number of buildings on user request map using improved radical law. The second step is to determine the representation of new buildings in each cluster iteratively by reduce one building each time. The third step is to harmonize the size of buildings considering the minimal separate distance of building itself and the preservation of differences between buildings. Advantages and disadvantages of this method are detailed. At the end of this paper the conclusions,possible improvements and authors’ next research focuses are proposed. * Corresponding author.