Relative Pose from Deep Learned Depth and a Single Affine Correspondence

TL;DR

This paper introduces a fast, robust method for estimating relative camera pose using deep-learned monocular depth and a single affine correspondence, significantly speeding up large-scale structure-from-motion tasks.

Contribution

It presents a novel 1AC+D solver that combines deep depth and affine features for efficient pose estimation within 1-point RANSAC, enabling faster large-scale 3D reconstruction.

Findings

Achieves similar accuracy to traditional methods

Significantly reduces computation time for pose estimation

Effective in large-scale SfM pipelines

Abstract

We propose a new approach for combining deep-learned non-metric monocular depth with affine correspondences (ACs) to estimate the relative pose of two calibrated cameras from a single correspondence. Considering the depth information and affine features, two new constraints on the camera pose are derived. The proposed solver is usable within 1-point RANSAC approaches. Thus, the processing time of the robust estimation is linear in the number of correspondences and, therefore, orders of magnitude faster than by using traditional approaches. The proposed 1AC+D solver is tested both on synthetic data and on 110395 publicly available real image pairs where we used an off-the-shelf monocular depth network to provide up-to-scale depth per pixel. The proposed 1AC+D leads to similar accuracy as traditional approaches while being significantly faster. When solving large-scale problems, e.g., pose-graph initialization for Structure-from-Motion (SfM) pipelines, the overhead of obtaining ACs and monocular depth is negligible compared to the speed-up gained in the pairwise geometric verification, i.e., relative pose estimation. This is demonstrated on scenes from the 1DSfM dataset using a state-of-the-art global SfM algorithm. Source code: https://github.com/eivan/one-ac-pose