Abstract
This paper proposes an identification model of buildings on different spatial information. This model identifies buildings by ratio of overlapped area of building shapes. In this model, we extract the overlapped area of building shapes by using the method of Boolean Operation of Overlapping Polygons. This method can also be applied to overlap polygons with holes. Moreover, we verify this model by comparing other existing identification methods through case studies in five regions of Tokyo. Furthermore, we present a method that considers spatial information with systematic positional errors, which must be dealt before applying overlapping building shapes. To do so, we first use the local search optimization and then adjust the position of building shape. Therefore, this method improves the accuracy of identification compared to methods without the local search optimization.
