While 3D scanning technology provides high-precision and efficient model data for building reality capture, a critical challenge remains unresolved in the reverse modeling phase: how to overcome the bottleneck of human dependency in reconstructing non-standard architectural components and achieve fully automated, reusable BIM digital reconstruction. This paper proposes an innovative method that deeply integrates 3D scanned point cloud data with the Dynamo visual programming platform to construct an automated reverse modeling workflow. The method focuses on the rapid identification and modeling of irregular components and building contours to build a reusable digital parametric component library. This approach significantly enhances the efficiency and automation level of reverse modeling, enabling the rapid reconstruction of complex-shaped components. The established parametric family library supports convenient retrieval and subsequent modifications in future projects, providing an effective tool for improving the efficiency, quality, and standardization of architectural design and construction. It demonstrates significant practical engineering value.