MarchingCubeSDF (Using OpenGL)

I write this application for studying how to calculate SDF values from a mesh and render it with marching cube algorithm by OpenGL API. Therefore, this application is only for practice and is not optimized at all. The code style here is just for the fast implementation. So I just write codes here to make the algorithm work.

I support debugging rendering options for marching cube algorithm to see what's going on. However the debugging rendering code is also very slow.

When loading a obj file for the SDF values, adjust model_scale and grid_delta. grid_init2 or grid_init codes calculate the number of grid points with them. If you don't adjust, the number of grid points may become large and cause the application to crash.

As for calculating SDF values, I use a AABB tree whose leaf contains a triangle from a mesh. I query a closest triangle for a grid point through the BVH structure (AABB tree). After getting a closest triangle for a query (grid) point, you also know the closest point on the triangle from the query point. The vector from the closest point to the query point is used with the triangle normal to see whether the grid point is on the true plane of the triangle or not. If it's on the true plane, the query point is outside the mesh, which means the SDF value is positive. Otherwise, the SDF value is negative (inside). I am using my ThreadPool implementation to accelerate this process more.

There will be no updates on this repository. Enjoy your Graphics programming!

How to Build (tested only on Windows)

I am using CMake to build a Visual Studio Project for this.

On this repository, create a build folder and then move to the build folder:

mkdir build
cd build

Type a CMake command to create a Visual Studio project

cmake ../ -G "Visual Studio 17 2022" -A x64

Now you can see MarchingCubeSDF.sln file. After opening the project, you have to build INSTALL project first to move resource files (shader and obj files) to the executable file. After that, you can see the original bunny and the SDF bunny.

Control the application

I support a FPS camera on the application. You can use WASD to move around and use dragging to rotate the camera view. You can do whatever you want more at camera_update() function on camera.cpp.

I integrate the Dear ImGui library into my application so that you can control some options. You can interact with it on the application and do your things at app_gui() function on main.cpp.

Reference

• Matrin Cavarga's Blog and github repository: He wrote many blog posts about calculating SDF values with advanced algorithms. It was really helpful for me to study with his blog. He even provides his code on the github repository. I didn't implement all the advanced algorithms.

  • Implementing Your Own Metaballs and Meta-Objects
  • AABB Tree
  • Octree for Voxel Outlines of a Triangular Mesh
  • The Fast Sweeping Algorithm
  • Computing the Sign of Discrete Distance Fields
  • Repository

• Paul Bourke's Blog: my marching cubes code is just a copy of his blog code. He introduces the algorithm of marching cubes with his codes. He even gives you the table values necessary to implement the marching cubes. I just ported his code into my app with OpenGL API.

  • Polygonising a scalar field

• I additionally recommend to see libigl's tutorial and code. Their code also provides fast SDF calculation.

• Geometry algorithms from Real-Time Collision Detection by Christer Ericson.

  • triangle_closest_point
  • triangle_intersect_ray

• AABB tree construction from godot.

• glad

• glfw

• Dear ImGui

• tinyobjloader

• glm

• Stanford Bunny .obj model data from McGuire Computer Graphics Archive