ImLoc: Revisiting Visual Localization with Image-based Representation

TL;DR

This paper introduces ImLoc, a visual localization method that uses image-based representations augmented with estimated depth maps, achieving high accuracy, efficiency, and flexibility in challenging conditions.

Contribution

ImLoc revisits 2D image-based localization by integrating depth maps, enabling high accuracy and efficiency while maintaining ease of building and updating the map.

Findings

Achieves state-of-the-art accuracy on standard benchmarks.

Outperforms existing memory-efficient methods at similar map sizes.

Efficient GPU-accelerated implementation with flexible accuracy-memory trade-offs.

Abstract

Existing visual localization methods are typically either 2D image-based, which are easy to build and maintain but limited in effective geometric reasoning, or 3D structure-based, which achieve high accuracy but require a centralized reconstruction and are difficult to update. In this work, we revisit visual localization with a 2D image-based representation and propose to augment each image with estimated depth maps to capture the geometric structure. Supported by the effective use of dense matchers, this representation is not only easy to build and maintain, but achieves highest accuracy in challenging conditions. With compact compression and a GPU-accelerated LO-RANSAC implementation, the whole pipeline is efficient in both storage and computation and allows for a flexible trade-off between accuracy and highest memory efficiency. Our method achieves a new state-of-the-art accuracy on various standard benchmarks and outperforms existing memory-efficient methods at comparable map sizes. Code will be available at https://github.com/cvg/Hierarchical-Localization.