Procedural generation and real-time rendering of plant-like fractals using L-Systems.
The project combines a Python grammar engine (rule expansion & geometry export) with a custom C++ OpenGL renderer.
As a bonus, it includes mesh optimisation (CGAL) and reproducible experiments (CSV logs, scripts).
- L-Systems: standard, stochastic, and parametric production rules
- Geometry export: generate meshes and save as
.obj - C++/OpenGL renderer: GLFW • GLAD • GLM • ImGui; custom shaders (GLSL)
- Optimisation (optional): CGAL-based mesh simplification with controllable ratio
- Experiments: batch scripts + CSV metrics for repeatable tests
lsysGrammar/ # L-System rules, Python generation, notebooks/scripts
src/ # C++ renderer (loading, camera, UI, draw loop)
shaders/ # GLSL vertex/fragment shaders
results/ # generated meshes / optimisation outputs / CSV logs
plotting/ # small analysis/visualisation utilities (Python)
tests/ # experiments and validation
vendor/ # third-party deps (headers/libs)
optimize_all.py # batch optimisation helper
CMakeLists.txt # C++ build configuration
Python (generator, experiments) • C++17 / OpenGL (real-time rendering) • GLSL (shaders) • CGAL (mesh simplification) • CMake
# from the folder that contains main.py, grammar.py, generator.py
cd lsysGrammar
python3 main.pyRunning this command will:
- expand the L-system(s) defined in
main.py- save the intermediate expansions as
.txtinGoldenOaktree/- convert them to
.objmeshes inGoldenOaktree_obj/
cd lsysGrammar
python3 -c 'from generator import process_lsystems_in_folder; process_lsystems_in_folder("GoldenOaktree","GoldenOaktree_obj")'mkdir -p build && cd build
cmake ..
make -j
# run the viewer on a generated mesh
./FractalEngine ../lsysGrammar/GoldenOaktree_obj/GoldenOaktree__parametric__iter_7.obj- Python 3.x with
numpy,open3d - C++17 toolchain with CMake, OpenGL, GLFW, GLAD, GLM, ImGui
- Experiments log faces/vertices before/after, execution time, and chosen simplification ratio.
- CSV samples live under
results/(e.g.,render_*.csv). - The renderer is used to visually inspect fidelity after simplification.
| Model | Faces → After | Reduction | Time (s) |
|---|---|---|---|
| fern.obj | 120k → 60k | -50% | 1.9 |
| tree.obj | 180k → 90k | -50% | 2.7 |
- L-System expansion (Python) → apply production rules (incl. stochastic/parametric).
- 3D turtle interpretation → build vertices/faces and export
.obj. - OpenGL renderer (C++) → load mesh, camera controls, shader pipeline, ImGui UI.
- (Optional) CGAL simplification → reduce complexity while preserving visual quality; record metrics.
Bachelor’s thesis — University POLITEHNICA of Bucharest, Faculty of Automatic Control and Computers (2025)
Topics: procedural generation, computational geometry, real-time graphics, mesh optimisation.
- Clear separation of generation ↔ rendering ↔ optimisation, showcasing pipeline engineering
- Mix of Python scripting and C++ real-time systems
- Solid base for research demos or production-style tooling (exportable assets, benchmarks)