SemCityLoc: Aerial 6DoF Localization Using Semantic 3D City Models

2026-06-25

AI Summary

This work addresses the challenge of six-degree-of-freedom aerial localization in urban canyon environments, where repetitive structures and severe occlusions hinder traditional methods that rely on high-precision GNSS or radiometrically rich 3D reconstructions. To overcome these limitations, the authors reformulate pose estimation as a semantic geometric alignment problem and introduce, for the first time, a semantic structured surface registration mechanism. This approach fuses semantics extracted by foundation vision models with monocular depth priors to enable lightweight matching against standardized LoD1–LoD3 semantic urban scene models—without requiring photorealistic reconstruction. Evaluated on SemCityLockeD, the first real-world benchmark introduced in this study, the method significantly enhances pose discriminability in complex urban settings, achieving up to a 36% improvement in recall and reducing the average position error from 9.89 meters to 2.62 meters compared to existing map-based approaches.

Abstract

Aerial 6DoF localization typically relies on precise GNSS signals or radiometrically rich 3D reconstructions, limiting scalability and on-board deployment. We propose SemCityLoc, a semantic-geometric alignment system that reframes aerial pose estimation as structured surface registration between foundation-model-derived visual priors and standardized LoD-compliant 3D city models. Instead of matching sparse contours or dense texture, our method aligns semantic surfaces and monocular depth with lightweight semantic 3D building models, increasing pose discriminability in repetitive and occluded urban environments. To enable accurate evaluation, we introduce SemCityLockeD, the first real-world benchmark combining centimeter-accurate UAV poses with standardized LoD1--LoD3 semantic city models and challenging low-altitude imagery. Experiments demonstrate substantial improvements over existing map-based approaches, improving recall by up to 36% and reducing mean positional error from 9.89m to 2.62m in challenging urban canyons. Our results indicate that semantically structured geometry provides sufficient and scalable constraints for high-precision aerial localization without radiometric scene reconstructions. The code and data are available at https://albertchen98.github.io/SemCityLoc.

Problem

• aerial localization
• 6DoF pose estimation
• semantic 3D city models
• urban canyons
• scalable localization

Innovation

• semantic 3D city models
• 6DoF aerial localization
• structured surface registration
• foundation-model priors
• LoD-compliant modeling