Unconstrained Large-scale 3D Reconstruction and Rendering across Altitudes

TL;DR

This paper introduces a new benchmark dataset for large-scale 3D reconstruction and rendering from images captured at various altitudes, addressing real-world challenges like limited images and heterogeneous camera setups.

Contribution

It provides the first public dataset for 3D reconstruction across different altitudes, along with evaluation benchmarks and baseline performance analysis.

Findings

Benchmark dataset reveals real-world challenges in 3D reconstruction.

Baseline methods show limitations under diverse conditions.

Evaluation metrics highlight areas for future research.

Abstract

Production of photorealistic, navigable 3D site models requires a large volume of carefully collected images that are often unavailable to first responders for disaster relief or law enforcement. Real-world challenges include limited numbers of images, heterogeneous unposed cameras, inconsistent lighting, and extreme viewpoint differences for images collected from varying altitudes. To promote research aimed at addressing these challenges, we have developed the first public benchmark dataset for 3D reconstruction and novel view synthesis based on multiple calibrated ground-level, security-level, and airborne cameras. We present datasets that pose real-world challenges, independently evaluate calibration of unposed cameras and quality of novel rendered views, demonstrate baseline performance using recent state-of-practice methods, and identify challenges for further research.