Optical Motion Capture

Tracking markers on the body with multiple infrared cameras and triangulating them into 3D motion.

Optical motion capture is one of the two main approaches to motion capture: reflective markers (passive) or light-emitting LEDs (active) are attached to a person or object, surrounded by several calibrated infrared cameras; each camera sees the 2D position of the markers, and triangulation across cameras recovers 3D coordinates, which are then fit to a skeleton or body model to reconstruct motion. Leading vendors include Vicon and OptiTrack. It's highly accurate (active systems can reach about 0.1mm) with frame rates commonly above 120 fps, and is often treated as the “gold standard” for human motion data; the downsides are expensive equipment, the need for a dedicated space, and dropped markers when they're occluded. It's complementary to inertial motion capture (IMU-based, immune to occlusion but prone to drift) and markerless motion capture (estimating pose directly from video). Datasets such as AMASS and OMOMO were both captured with optical motion capture.

ExampleThe OMOMO dataset used 12 Vicon cameras at 120 frames per second to record subjects carrying objects, with 5 markers attached to each object so both the person's and the object's motion were tracked at once.

Also called

Marker-Based Motion Capture, Optical MoCap

Related

Motion Capture · Inertial Motion Capture · Markerless Motion Capture · Vicon · OptiTrack · AMASS (Archive of Motion Capture as Surface Shapes)

Sources

Motion capture - Wikipedia（Optical systems 一节） (Chinese)
Object Motion Guided Human Motion Synthesis（OMOMO 采集设置） (Chinese)