Camera Calibration for 3D Reconstruction and Gaussian Splatting

In the first phase, this project delves into the intricate process of camera calibration with the ultimate goal of achieving robust 3D reconstruction. The fundamental premise involves utilizing two fixed cameras, strategically positioned to observe the same scene, from which approximately 100 images are captured by each camera. The core objective is to accurately estimate both the extrinsic and intrinsic parameters of these cameras through the application of Structure from Motion (SfM) techniques.

For a deeper understanding of SfM and related methodologies, the COLMAP project on GitHub offers a comprehensive resource:
 https://github.com/colmap/colmap

The second phase involves building Gaussian splatting after the camera calibration.

For more information about the Gaussian splatting, please see:

https://repo-sam.inria.fr/fungraph/3d-gaussian-splatting/.