HOLO: Homography-Guided Pose Estimator Network for Fine-Grained Visual Localization on SD Maps

TL;DR

HOLO introduces a homography-guided pose estimation framework for fine-grained visual localization between multi-view images and standard-definition maps. By embedding explicit homography priors into both BEV-map semantic alignment and pose decoding, HOLO achieves faster convergence and higher accuracy than regression-only methods while supporting cross-resolution inputs. The architecture fuses BEV perception with a rasterized SD map through a learned homography, and recovers 3-DoF pose from this geometric constraint, yielding state-of-the-art results on nuScenes with robust performance under noise and cross-resolution conditions. The approach is complemented by an OpenStreetMap based SD-map extension for nuScenes and demonstrates competitive efficiency, making it practical for real-time autonomous navigation with lightweight map data.

Abstract

Visual localization on standard-definition (SD) maps has emerged as a promising low-cost and scalable solution for autonomous driving. However, existing regression-based approaches often overlook inherent geometric priors, resulting in suboptimal training efficiency and limited localization accuracy. In this paper, we propose a novel homography-guided pose estimator network for fine-grained visual localization between multi-view images and standard-definition (SD) maps. We construct input pairs that satisfy a homography constraint by projecting ground-view features into the BEV domain and enforcing semantic alignment with map features. Then we leverage homography relationships to guide feature fusion and restrict the pose outputs to a valid feasible region, which significantly improves training efficiency and localization accuracy compared to prior methods relying on attention-based fusion and direct 3-DoF pose regression. To the best of our knowledge, this is the first work to unify BEV semantic reasoning with homography learning for image-to-map localization. Furthermore, by explicitly modeling homography transformations, the proposed framework naturally supports cross-resolution inputs, enhancing model flexibility. Extensive experiments on the nuScenes dataset demonstrate that our approach significantly outperforms existing state-of-the-art visual localization methods. Code and pretrained models will be publicly released to foster future research.

Figures (11)

• Figure 1: Comparision of attention-based direct pose estimation and homography-guided pose estimation. (a) Schematic diagram of attention-based direct pose estimation and homography-guided pose estimation. (b) The evolution of localization error across training iterations between two methods on nuScenecaesar2020nuscenes.
• Figure 2: Overall architecture of the proposed Homography-Guided Pose Estimator Network. The BEV Perception Module and Map Processing Module collaboratively build neural feature pairs with homography through semantic alignment, providing explicit geometric priors for downstream pose regression. The Homography-Guided Pose Estimation Module leverages homography priors to accomplish the final pose estimation.
• Figure 3: The localization results on nuScenes dataset. The second row illustrates localization results during a left-turn maneuver, and the third row shows the localization performance during nighttime driving. The third column shows the BEV features before and after warping, and the fourth column displays the SD map features with the warped BEV positions. Shown in the fifth row is the SD map with drivable areas and buildings. The green arrows indicate the ground-truth vehicle poses, while the blue arrows represent the estimated poses.
• Figure 4: The overall architecture of HOLO-CA. The BEV feature and the map feature are fed into a cross-attention-based feature fusion module. The fused features are then concatenated and passed to the pose decoder, which can be either a homography decoder or a 3-DOF pose decoder. The iterative strategy can also be employed by direcly warpping BEV features.
• Figure 5: Training and validation curves of different methods.