AR HMD Guidance for Controlled Hand-Held 3D Acquisition.

Area of Science:

• Computer Vision
• 3D Reconstruction
• Augmented Reality

Background:

• Photogrammetry enables high-quality 3D reconstruction from images but requires dense, overlapping views.
• Ensuring sufficient coverage during data acquisition is challenging for operators, often leading to reconstruction failures.
• Manual repair or re-acquisition of data is time-consuming and frequently impractical.

Purpose of the Study:

• To develop an augmented reality system for guided photogrammetric data acquisition.
• To improve the efficiency and success rate of 3D model reconstruction for novice users.
• To provide real-time feedback for optimal camera positioning during image capture.

Main Methods:

• An augmented reality head-mounted display (AR HMD) system was developed for interactive photogrammetric guidance.
• The system generates a coarse scene model and suggests optimal camera poses for complete coverage.
• Real-time visual feedback guides the operator, and images are automatically captured when poses are achieved.

Main Results:

• Participants using the AR HMD guidance system produced improved 3D reconstructions compared to unguided novices.
• The guided acquisition method successfully captured challenging scenes without reconstruction failures observed in the control group.
• The system facilitated rapid data acquisition, capturing necessary images in minutes.

Conclusions:

• The AR HMD-based approach offers a self-contained and portable solution for effective photogrammetric data capture.
• Real-time, geometry-aware guidance significantly enhances reconstruction quality and reduces operator error.
• This method democratizes high-quality 3D reconstruction by simplifying the acquisition process for non-experts.