Cross-View Localization via Redundant Sliced Observations and A-Contrario Validation

2025-08-07

AI Summary

Cross-view localization (CVL) suffers from insufficient redundant observations in GNSS-denied environments, making pose estimation reliability assessment challenging. To address this, we propose Slice-Loc, a two-stage method: first, query images are partitioned into patches, each independently yielding a 3-DoF pose estimate to construct multiple geometric observations; second, an a-contrario model is introduced to quantify localization significance via the Number of False Alarms (NFA), enabling automatic reliability verification, while multi-patch inlier results are fused to enhance robustness. This work pioneers both image slicing–driven redundant pose generation and NFA-guided significance validation. Evaluated on the cross-city DReSS dataset, Slice-Loc achieves a mean localization error of 1.86 m (↓58%) and orientation error of 1.24° (↓64%), with large errors (>10 m) occurring in less than 3% of cases—substantially outperforming state-of-the-art methods.

Technology Category

• Computer Vision: Biometrics, Face, Gesture & Pose
• Intelligent Robots: Localization, Mapping, and Navigation
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Security and Privacy: Large-scale security measurements
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Abstract

Cross-view localization (CVL) matches ground-level images with aerial references to determine the geo-position of a camera, enabling smart vehicles to self-localize offline in GNSS-denied environments. However, most CVL methods output only a single observation, the camera pose, and lack the redundant observations required by surveying principles, making it challenging to assess localization reliability through the mutual validation of observational data. To tackle this, we introduce Slice-Loc, a two-stage method featuring an a-contrario reliability validation for CVL. Instead of using the query image as a single input, Slice-Loc divides it into sub-images and estimates the 3-DoF pose for each slice, creating redundant and independent observations. Then, a geometric rigidity formula is proposed to filter out the erroneous 3-DoF poses, and the inliers are merged to generate the final camera pose. Furthermore, we propose a model that quantifies the meaningfulness of localization by estimating the number of false alarms (NFA), according to the distribution of the locations of the sliced images. By eliminating gross errors, Slice-Loc boosts localization accuracy and effectively detects failures. After filtering out mislocalizations, Slice-Loc reduces the proportion of errors exceeding 10 m to under 3%. In cross-city tests on the DReSS dataset, Slice-Loc cuts the mean localization error from 4.47 m to 1.86 m and the mean orientation error from $mathbf{3.42^{circ}}$ to $mathbf{1.24^{circ}}$, outperforming state-of-the-art methods. Code and dataset will be available at: https://github.com/bnothing/Slice-Loc.

Problem

• Enhances cross-view localization accuracy in GNSS-denied environments
• Provides redundant observations for reliable camera pose estimation
• Validates localization meaningfulness using false alarm quantification

Innovation

• Divides query image into sub-images for redundant observations
• Uses geometric rigidity to filter erroneous 3-DoF poses
• Quantifies localization meaningfulness via false alarm estimation