3D Geometry on Stereoids

A practical, notebook‑first tour of core 3‑D geometry building blocks for LiDAR/camera perception. Each notebook is self‑contained and runnable.

Completed Modules

Concept	What you'll see	Notebook
Bird’s‑Eye View (BEV)	Top‑down projection of point clouds into fixed‑resolution grids (occupancy / intensity).	notebooks/BEV.ipynb
3‑D Bounding Box (7‑DoF)	Parameterization (x,y,z,h,w,l,θ), corner extraction, yaw handling, basic transforms.	notebooks/BoundingBox_7DOF.ipynb
DBSCAN	Density‑based clustering on point clouds; eps / minPts intuition; labeled clusters.	notebooks/DBSCAN.ipynb
k‑d Tree / Octree	Spatial indexing for fast nearest‑neighbor and range queries; complexity notes.	notebooks/kd_oct_tree.ipynb
Range Image	Spherical projection (azimuth/elevation → row/col) with index mappings back to 3‑D points.	notebooks/range_image.ipynb
RANSAC Plane Fitting	Robust ground plane extraction with inlier/outlier masking.	notebooks/Ransac.ipynb

Quickstart

git clone https://github.com/hiteshhedwig/3d-geometry-on-stereoids
cd 3d-geometry-on-stereoids
jupyter lab  # or: jupyter notebook

Open any notebook from the notebooks/ folder and run all cells.

Repository Layout

.
├── notebooks/
│   ├── BEV.ipynb
│   ├── BoundingBox_7DOF.ipynb
│   ├── DBSCAN.ipynb
│   ├── kd_oct_tree.ipynb
│   ├── range_image.ipynb
│   └── Ransac.ipynb
└── README.md