PoissonRecon

PoissonRecon is an open-source software tool for 3D surface reconstruction from oriented point clouds using the Poisson surface reconstruction approach. It implements the Poisson reconstruction method introduced by Kazhdan, Bolitho, and Hoppe, which estimates a watertight surface that best explains the observed points and their normals.

The tool takes as input a set of 3D points with associated normals (and optionally confidence or color). It builds an adaptive octree representation of space and solves a Poisson equation to recover a scalar field whose gradient approximates the input normals. An isosurface of this field is then extracted, typically by marching cubes, producing a triangulated mesh.

Key features include the use of a multi-resolution octree to handle large point sets efficiently, control over surface resolution via the octree depth parameter, and robustness to noise and incomplete data. The produced surface tends to be smooth and watertight, making PoissonRecon well suited for applications in cultural heritage, reverse engineering, and 3D scanning pipelines that require clean meshes. It can fuse multiple scans and fill holes in the data, though the quality depends on the normal field accuracy and sampling density.

PoissonRecon is cross-platform and distributed as an open-source executable and library, with common output formats such as PLY and OBJ. It is widely cited in the 3D reconstruction community as a practical implementation of the Poisson surface reconstruction paradigm.