RHO: Robust Holistic OSM-Based Metric Cross-View Geo-Localization

Junwei Zheng; Ruize Dai; Ruiping Liu; Zichao Zeng; Yufan Chen; Fangjinhua Wang; Kunyu Peng; Kailun Yang

cs.CVMar 29, 2026

RHO: Robust Holistic OSM-Based Metric Cross-View Geo-Localization

Authors: Junwei Zheng, Ruize Dai, Ruiping Liu, Zichao Zeng, Yufan Chen, Fangjinhua Wang, Kunyu Peng, Kailun Yang,

Organizations: Karlsruhe Institute of Technology · ETH Zurich · UCL · INSAIT, Sofia University “St. Kliment Ohridski” ·

Abstract

Metric Cross-View Geo-Localization (MCVGL) aims to estimate the 3-DoF camera pose (position and heading) by matching ground and satellite images. In this work, instead of pinhole and satellite images, we study robust MCVGL using holistic panoramas and OpenStreetMap (OSM). To this end, we establish a large-scale MCVGL benchmark dataset, CV-RHO, with over 2.7M images under different weather and lighting conditions, as well as sensor noise. Furthermore, we propose a model termed RHO with a two-branch Pin-Pan architecture for accurate visual localization. A Split-Undistort-Merge (SUM) module is introduced to address the panoramic distortion, and a Position-Orientation Fusion (POF) mechanism is designed to enhance the localization accuracy. Extensive experiments prove the value of our CV-RHO dataset and the effectiveness of the RHO model, with a significant performance gain up to 20% compared with the state-of-the-art baselines. Project page: https://github.com/InSAI-Lab/RHO.

Figures & tables

Figure 1: OrienterNet [ 34 ] fails under adverse conditions. PR and OR stand for Position Recall and Orientation Recall.

Country	City	360°-FoV Pan.	120°-FoV Pin.	OSM
USA	Chicago	12.4k	37.2k	12.4k
	Detroit	39.4k	118.1k	39.4k
	San Francisco	15.5k	46.5k	15.5k
	Washington	12.5k	37.5k	12.5k
France	Montrouge	9.5k	28.5k	9.5k
France	Toulouse	12.4k	37.1k	12.4k

Table 1: Data distribution of the CV-RHO dataset.

Dataset	Pin.	Pan.	Satellite	OSM	Weather noise	Sensor noise
CVUSA [ 49 ]	✗	✓	✓	✗	✗	✗
CVACT [ 26 ]	✗	✓	✓	✗	✗	✗
VIGOR [ 74 ]	✗	✓	✓	✗	✗	✗
CVGlobal [ 60 ]	✗	✓	✓	✓	✗	✗
KITTI [ 12 ]	✓	✗	✓	✓	✗	✗
MGL [ 34 ]	✓	✗	✗	✓	✗	✗

Table 2: Comparison between our CV-RHO dataset and other widely-used CVGL datasets.

Figure 2: Data samples of our CV-RHO dataset.

Figure 3: Overview of the RHO model in a two-branch Pin-Pan architecture. Given the current panorama as input, the panoramic branch with Split-Undistort-Merge (SUM) module produces the 360°-FoV feature, while the pinhole branch focuses on local dense feature in 120° FoV. Both generated features are matched with the OSM feature map and then fused by the proposed Position-Orientation Fusion.

Figure 4: Position-Orientation Fusion module of the RHO model. Given both 360°-FoV and 120°-FoV volumes, POF combines the pinhole volume $\mathbf{S}_{1}$ with $Prior_{uv}$ , and then the panoramic volume $\mathbf{S}_{pano}$ with $Prior_{\theta}$ .

Approach	FoV	PR @ Xm $\uparrow$			OR @ X° $\uparrow$
Approach	FoV	1m	3m	5m	1°	3°	5°
OrienterNet [ 34 ]	90°	18.02	58.37	71.04	27.72	63.86	77.50
OrienterNet [ 34 ]	120°	21.83	66.16	78.03	35.02	74.89	85.62
OrienterNet [ 34 ]	360°	3.79	19.35	28.78	10.29	28.43	36.87
RHO (ours)	360°	24.59	73.55	84.36	43.46	83.61	90.44

Table 3: Results on clean CV-RHO dataset. Position Recall (PR) and Orientation Recall (OR) are in percentages.

Method	Variation		PR @ Xm $\uparrow$			OR @ X° $\uparrow$
Method	Train	Eval	1m	3m	5m	1°	3°	5°
OrienterNet [ 34 ]	Clean	Clean	21.83	66.16	78.03	35.02	74.89	85.62
	Clean	Rain	13.98	52.98	68.12	25.51	60.49	74.68
	Clean	NT	11.44	44.55	57.77	20.71	51.13	64.84
	Clean	Fog	16.26	54.52	67.12	26.81	62.27	75.02
	Clean	Snow	14.29	50.02	63.23	24.30	57.52	70.68

Table 4: Results on CV-RHO dataset under varying conditions and noises. Position Recall (PR) and Orientation Recall (OR) are in percentages. NT: nighttime; OE: over-exposure; UE: under-exposure; MB: motion blur; ANV: all noisy variations; AV: all variations; Average Degradation is overall performance drop.

Method	Dataset	Variation	City	PR @ Xm $\uparrow$			OR @ X° $\uparrow$
Method	Dataset	Variation	City	1m	3m	5m	1°	3°	5°
OrienterNet [ 34 ]	Sim2Real	Rain	Ogori	4.85	40.20	56.36	18.59	47.47	61.62
		Snow	Detroit	11.16	47.07	57.54	22.59	52.13	63.95
		OE	Hasselt	3.97	29.71	47.84	17.62	47.67	63.04
		UE	Hasselt	2.07	17.10	28.67	10.19	27.29	39.03
		MB	Hasselt	3.87	28.62	45.29	15.99	46.63	62.12

Table 6: Sim2Real Results. Position Recall (PR) and Orientation Recall (OR) are in percentages. MV: Mount Vernon; OE: over-exposure; UE: under- exposure; MB: motion blur.

FoV	Position R @ Xm $\uparrow$			Orientation R @ X° $\uparrow$
FoV	1m	3m	5m	1°	3°	5°
40°	0.29	2.82	6.39	10.40	23.63	30.68
60°	0.34	3.04	7.34	0.90	2.55	4.07
90°	3.21	21.82	34.22	16.17	35.79	41.26
120°	14.44	68.14	80.60	18.61	51.47	69.87

Table 7: Results of different merge strategies evaluated with the clean data of Washington in our CV-RHO dataset. Position R and Orientation R stand for Position Recall and Orientation Recall, respectively. All results are in percentages.

Strategy	Position R @ Xm $\uparrow$			Orientation R @ X° $\uparrow$
	1m	3m	5m	1°	3°	5°
No Fusion	14.44	68.14	80.60	18.61	51.47	69.87
$\text{Prior}_{\text{uv}}$ Fusion	11.93	59.47	77.38	30.29	73.13	85.14
$\text{Prior}_{\theta}$ Fusion	13.66	68.63	81.76	32.15	78.21	88.69
POF (ours)	14.53	70.33	83.16	37.39	82.96	91.21

Table 8: Analysis of different fusion strategies. Position R and Orientation R stand for Position Recall and Orientation Recall, respectively. All results are in percentages.

Appendix figures & tables4 assets

Supplementary material from the paper’s appendix.

Appendix

Variation	Prompt	Hyperparameter
Rain	Change the background into a heavy rainy day, change the sky into gray overcast sky, add visible rain streaks and puddles.	Guidance scale: 7.5
Snow	Change the background into a snowy day, add visible falling snowflakes	Guidance scale: 7.7
Fog	Turn the background into a foggy day, add visible smog on the street	Guidance scale: 7.7
Nighttime	Change the background into evening, illuminate the scene brighter with streetlights.	Guidance scale: 5.0
Over Exposure	-	Brightness factor: 2.5
Under Exposure	-	Brightness factor: 0.25

Appendix

Table 9: Prompts and Parameters configuration for generating images with Flux.1 Kontext.

Figure 5: More samples of generated images. Images with red borders are original images.

Figure 6: Selected data points of our CV-RHO dataset and Sim2Real across $11$ cities. Color from red to green indicates capture date.

Panel of Figure 7 — Figure 7: One-way fusion strategy: (a) $Prior_{uv}$ fusion: fuse $S_{1}$ with position prior $Prior_{uv}$ . (b) $Prior_{\theta}$ fusion: fuse $S_{pano}$ with orientation prior $Prior_{\theta}$ .

Cross-View Geo-Localization

Cross-view geo-localization matches ground, drone, or aerial images to maps or satellite views to estimate where they were captured.

See all 94 papers

SeqLoc: Beyond the Single Frame for Cross-View Geo-Localization in Feature-Sparse Scenes
Aug 8, 2026Junwei Zheng, Yun Huang, Ruize Dai +8Cross-View Geo-Localization Visual Place Recognition
Karlsruhe Institute of Technology, Karlsruhe, Germany · Hunan University, Changsha, China · University of Cambridge, Cambridge, UK
From Horizontal to Rotated: Cross-View Object Geo-Localization with Orientation Awareness
Mar 16, 2026Chenlin Fu, Ao Gong, Yingying ZhuCross-View Geo-Localization Object Detection
College of Computer Science and Software Engineering, Shenzhen University, China
Just Zoom In: Cross-View Geo-Localization via Autoregressive Zooming
Mar 26, 2026Yunus Talha Erzurumlu, Jiyong Kwag, Alper YilmazCross-View Geo-Localization Map-Based Localization
Photogrammetric Computer Vision Lab, The Ohio State University

Aug 8, 2026J/K move · Enter open · S save

SeqLoc: Beyond the Single Frame for Cross-View Geo-Localization in Feature-Sparse Scenes

Junwei Zheng, Yun Huang, Ruize Dai, Ruiping Liu+7

Karlsruhe Institute of Technology, Karlsruhe, Germany · Hunan University, Changsha, China · University of Cambridge, Cambridge, UK

Cross-View Geo-Localization Visual Place Recognition Map-Based Localization

Open paper AskPDF arXiv

Cross-View Geo-Localization (CVGL) with OpenStreetMap (OSM) performs well in structure-rich urban environments but collapses in feature-sparse scenes such as rural roads. To study this failure mode, in this work, we introduce CV-FSS, a benchmark that pairs sequential panoramas from five rural regions with aligned OSM maps, on which single-frame methods degrade drastically. We then propose SeqLoc, an online test-time sequence aggregation mechanism that recursively maintains a log-belief volume with three key components: (1) Entropy-Tempered Uncertainty (ETU) tempers each incoming pose likelihood volume by its normalized entropy; (2) Map-Guided Relocalization (MGR) mixes a map-shaped recovery distribution into the belief so that a suppressed true pose can recover; (3) Peak-Anchored Smoothing (PAS) derives the final pose at sub-grid precision. Extensive experiments on CV-FSS and CV-RHO demonstrate that SeqLoc outperforms single-frame localization by a large margin, improving both position and orientation recall by over 50%. The benchmark and source code are publicly available at https://zhengjunwei.com/publications/SeqLoc/SeqLoc.html.

Abstract

Cross-View Geo-Localization

Explore similar work

SeqLoc: Beyond the Single Frame for Cross-View Geo-Localization in Feature-Sparse Scenes

From Horizontal to Rotated: Cross-View Object Geo-Localization with Orientation Awareness

Just Zoom In: Cross-View Geo-Localization via Autoregressive Zooming