Automated Indoor Data Acquisition with Stretch Robot for Photogrammetry and Immersive Virtual Reality

• Published in: International Conference on Automation, Control and Robotics Engineering (Online) pp. 107 - 113
• Main Authors: Alzyout, Mohammad S., Tikkisetty, Yashwanth Naidu, Alawneh, Shadi, Brudnak, Mark J., Louie, Wing-Yue Geoffrey
• Format: Conference Proceeding
• Language: English
• Published: IEEE 16.07.2025
• Subjects: Automated data acquisition; Cameras; Data acquisition; Human-robot interaction; Indoor environment; Manipulators; Photogrammetry; Reality Capture; Robot vision systems; Solid modeling; Stretch Robot; Three-dimensional displays; Virtual reality
• ISSN: 2997-6278
• DOI: 10.1109/CACRE66141.2025.11119559

An automated system for indoor data acquisition and image-based 3D reconstruction using a Stretch robot is presented. This system addresses the challenges of manual data acquisition in cluttered indoor environments, which are critical for applications such as immersive VR. The system setup includes a high-resolution camera, polarization filter, custom 3D-printed mount, and an adjustable LED lighting system for capturing dense, detailed images in both open and cluttered spaces. The robot's mobility, combined with its multi-axis arm and camera system, allows precise control over camera angles and positioning, ensuring comprehensive coverage of surfaces. We present the first integration of a mobile manipulator - Stretch 3 - with autonomous viewpoint scheduling and adjustable multi height camera control for photogrammetric 3D reconstruction in occluded indoor environments. Using Reality Capture software for photogrammetry, we demonstrate its effectiveness in generating high-resolution, textured 3D models for VR applicati-ons. Experimental evaluations reveal that our system achieves a 92% successful image alignment rate in cluttered environments and an 98% rate in open spaces, with optimized image overlap and efficient navigation between capture positions. This research advances automated photogrammetry for VR, providing a scalable, high-resolution solution applicable to various sectors, including training in Human-Robot Interaction (HRI), cultural heritage preservation, and architectural modeling.