Multistage SFM: A Coarse-to-Fine Approach for 3D Reconstruction

Abstract

Several methods have been proposed for large-scale 3D reconstruction from large, unorganized image collections. A large reconstruction problem is typically divided into multiple components which are reconstructed independently using structure from motion (SFM) and later merged together. Incremental SFM methods are most popular for the basic structure recovery of a single component. They are robust and effective but strictly sequential in nature. We present a multistage approach for SFM reconstruction of a single component that breaks the sequential nature of the incremental SFM methods. Our approach begins with quickly building a coarse 3D model using only a fraction of features from given images. The coarse model is then enriched by localizing remaining images and matching and triangulating remaining features in subsequent stages. The geometric information available in form of the coarse model allows us to make these stages effective, efficient, and highly parallel. We show that our method produces similar quality models as compared to standard SFM methods while being notably fast and parallel.

A Novel Approach to Landslide Monitoring Based on Unmanned Aerial System Photogrammetry

“…Before starting the LOOCV processes, it is possible to filter datasets to make these processes more efficient in terms of the necessary processing power and time. Guided by cognition that the features' reliability increases with the number of images on which they can be found (Shah et al, 2015; URL 5, 2022), displacement vectors can be filtered based on the number of images on which each feature point is found.…”

Hierarchical Segment-based Optimization for SLAM

“…One line of research [14], [15], [16], [17], [18], [19], [20] uses the method of incremental BA, where the new frames are incrementally merged to the existing map. However, the performance of the system suffers from poor initialization and accumulated error, especially for long sequences.…”