Case Study: The Dolly (Carrello)

Project Target & Shot Conditions

The primary focus of this matchmove session was to achieve a robust 3D camera solve on handheld DSLR footage while tracking a moving dolly (carrello). The footage was captured on a Nikon D3200 at 59.94 frames per second to reduce motion scan lines and capture smooth motion arcs.

To provide stable anchors for the solver under dynamic handheld motion, high-contrast red and white tracking markers were applied manually to the dolly and key points across the static background environment.

Adopted Workflow

1. Manual Marker Calibration: Placing reference tags on the physical carriage dolly to capture structural rotation.
2. Feature Tracking: Tracking high-contrast markers manually to eliminate coordinate slip under fast motion.
3. Camera Path Solving: Initializing focal estimates and running solver iterations to reconstruct the physical lens path.
4. Object Reconstruction: Modeling the spatial coordinates of the dolly geometry relative to the camera solver plane.

Solver Challenges & Solutions

Handheld Jitter & Rolling Shutter: Handheld DSLR sensors introduce rolling shutter scan delay, which can warp coordinate lines during quick panning. For this track, the rolling shutter was not corrected directly in the pre-pass. However, the manual trackers kept the final solver deviation under 1 pixel. If rolling shutter correction were applied, the deviation index could be further reduced by approximately 0.4 pixels.

Marker Occlusion: As the dolly moves across the set, background points are temporarily hidden. This was resolved by creating overlapping manual tracking links, ensuring that the solver maintains continuous camera telemetry.