Automated reconstruction of multiple objects with individual movement based on PSP

Area of Science:

• Optical Metrology
• 3D Reconstruction
• Computer Vision

Background:

• Phase shifting profilometry is widely used for 3D object reconstruction.
• Existing methods struggle with reconstructing multiple objects exhibiting independent movements, leading to errors.
• Accurate reconstruction of independently moving objects is crucial for various applications.

Purpose of the Study:

• To develop an automated method for tracking and reconstructing multiple objects with individual movements.
• To address the limitations of current phase shifting profilometry techniques in handling complex object dynamics.
• To improve the accuracy and robustness of 3D reconstruction for dynamic scenes.

Main Methods:

• Automatic object identification and bounding box generation.
• Kernelized Correlation Filters (KCF) algorithm for object tracking across fringe patterns.
• Scale-Invariant Feature Transform (SIFT) for describing individual object motion (rotation matrix and translation vector).
• Reconstruction of multiple objects utilizing their distinct movement information.

Main Results:

• Successful automated identification and tracking of multiple objects with individual movements.
• Accurate estimation of individual object motion parameters (rotation and translation).
• Effective 3D reconstruction of multiple objects based on their unique movement data.
• Experimental validation demonstrating the method's effectiveness.

Conclusions:

• The proposed automated method effectively reconstructs multiple moving objects with individual movements.
• This approach overcomes previous limitations in phase shifting profilometry for dynamic scenes.
• The technique offers a robust solution for 3D reconstruction in complex, multi-object scenarios.