Importing a YOLO model or wrapping a Hugging Face pipeline can get a computer vision demo running, but it leaves you stuck the moment lighting, camera angle, or environment changes. This course builds perception systems from first principles instead, starting with the mathematics of how a 2D image relates to 3D space.
Core topics
- How points in real space map to pixels in an image
- Camera intrinsics and extrinsics
- The projection model that turns an image into spatial data
- Optical flow, feature matching, triangulation, and stereo vision for robotics
Building a full pipeline
You work through the complete perception pipeline, camera calibration, lens distortion correction, and generating depth maps and point clouds, turning raw pixels into usable spatial information.
Who it's for
Aimed at engineers who want to go beyond off-the-shelf libraries and understand the underlying geometry, particularly those designing perception systems for robotics, autonomous devices, and other intelligent machines.
