Learning Physics through Computer Vision

Tracks physical projectile, pendular, and rotational motions in video streams with sub-pixel spatial accuracy.

Automated regression solver extracting empirical velocity, acceleration, and damping coefficients from observed video.

Allows students to overlay theoretical differential equations directly on top of real-world recorded motion videos.

Surfaces real-world aerodynamic drag and frictional losses, teaching students why physical experiments diverge from ideal models.