VoxelHashing

Why we included this project

This is the reference implementation of voxel hashing, the real-time 3D reconstruction technique from the Siggraph Asia 2013 paper by Nießner and colleagues. It shows how a spatial hash table keeps surface data compact and streams it in and out as the sensor moves, which makes it a solid baseline for anyone studying depth fusion or volumetric mapping. The code walks through the whole pipeline, from depth pre-processing and camera pose estimation to fusion and surface rendering, on commodity GPUs, so it works well as a teaching example of how consumer depth cameras become dense 3D models. Keep in mind it is a research prototype, not a library you drop into a project: it expects older toolchains like DirectX June 2010, the Kinect SDK, and CUDA 6.5, and ships under a non-commercial Creative Commons license. For study and experimentation it is still the clearest way to see the algorithmic ideas in action.