3D shape measurement of static and moving objects with adaptive spatiotemporal correlation

Area of Science:

• Photogrammetry
• Computer Vision
• 3D Reconstruction

Background:

• Traditional photogrammetric 3D shape measurement methods rely on either temporal or spatial features.
• Temporal methods offer high accuracy but are restricted to static objects.
• Spatial methods handle moving objects but yield lower accuracy.

Purpose of the Study:

• To develop a unified approach for accurate 3D shape measurement in dynamic scenes containing both static and moving objects.
• To overcome the limitations of existing temporal and spatial photogrammetric techniques.

Main Methods:

• A novel spatiotemporal correlation technique is proposed, adapting temporal and spatial support based on local object motion.
• Motion compensation is achieved through temporal warping of correlated image regions.
• The method utilizes structured illumination with random patterns, enabling high frame rates suitable for dynamic scenes.

Main Results:

• The developed spatiotemporal correlation successfully measures dynamic scenes with both static and moving objects.
• The technique demonstrates adaptability to varying object motion.
• Validation performed using both simulated data and real-world measurements.

Conclusions:

• The proposed spatiotemporal correlation method effectively combines the strengths of temporal and spatial techniques for dynamic 3D shape measurement.
• This approach offers a significant advancement for accurately capturing complex dynamic environments.