Objectives  Buildings can be divided into physical cluster and property right cluster from different perspectives, and the latter attached to the former. The existing methods can construct physical cluster and property right cluster of the same building automatically, but the two resulting clusters are independent of each other. This not only increases the cost of modeling, but also is not conducive to the update and maintenance of model data in the later period.

  Methods  For the problem, we study the relationship between physical cluster and property right cluster of apartment buildings, and find that the hierarchy of connected boundaries determines the property right solids aggregated by cells, and present a method to transform physical clustering into property right clustering automatically. With existing physical clusters, the method transforms the cells of physical clusters into dual points, and the connected boundaries between cells into semantic edges, and the whole physical cluster into node relation graph with Poincaré duality transformation. A segmenting algorithm is designed for node relation graph, which can divide node relation graph into sub node relation graph representing proprietary and co-owned property according to the semantic information of the edges. Furthermore, the non-common boundary surfaces of the cell set corresponded by sub node relation graph are extracted to construct the property right solids, and the aggregation of the property right solids forms the property right cluster.

  Results  Instead of building two separate clusters, the proposed method only builds a physical cluster and property right cluster is generated by the transformation.

  Conclusions  The results show that the proposed method can identify the property right solids and construct property right cluster in the existing physical cluster automatically. It saves the modeling cost and facilitates the update and maintenance of the later data.