Papers for

urban mapping teams

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

BhopalMLS dataset enables urban 3D scene understanding in unplanned cities

BhopalMLS: A LiDAR Dataset for Semantic Scene Understanding in Heterogeneous Urban Environments

Abstract: We introduce BhopalMLS, a large-scale Mobile Laser Scanning (MLS) dataset from a dense, heterogeneous urban environment in Bhopal, India, with mixed traffic, informal structures, and narrow unplanned streets. It spans approximately 10 km of road corridors with over 1.2 billion points and 19,518 annotated instances across 62 fine-grained static categories, grouped into 12-class and 21-class benchmark taxonomies; 6 optional moving-class subclasses support dynamic-object analysis. Existing 3D datasets, collected in well-planned cities, are structurally more uniform than such unplanned settings, and lack their subtype diversity. On semantic and instance segmentation at both granularities, fine-tuned foundation models surpass supervised baselines: Utonia(L) reaches 71.28% mIoU (vs. PTv3's 68.04% from scratch) and 44.07% mAP50 on the 12-class benchmark. Cross-region experiments against four established datasets (Ke-MLS, Toronto-3D, Paris-Lille-3D, SemanticKITTI) reveal a consistent asymmetric transferability: models trained on BhopalMLS transfer better than the reverse on every pair, for both Utonia(L) and Concerto(L) (mean 21.5, up to 36 mIoU points), holding even when the target is denser. Factor-controlled retraining at a fixed label budget, repeated on both models, attributes this to the coverage of BhopalMLS's subtype inventory rather than to dataset size or corridor geometry: collapsing each class to a single subtype costs 13.0 and 13.3 mIoU points for Utonia(L) and Concerto(L). BhopalMLS therefore offers a testbed for 3D perception beyond well-planned cities. Data and an interactive viewer are at https://geoai4cities.github.io/bhopalmls.

Tue 3 NovComputer Vision and Pattern Recognition
The gist
Understanding complex city environments using 3D data is harder when streets are narrow and structures are informal, as in many Indian cities. The authors created BhopalMLS, a huge dataset capturing detailed 3D scans of Bhopal’s busy and diverse streets. Their work shows that models trained on this dataset work better in other cities than models trained elsewhere. This means BhopalMLS helps computers recognize and understand urban scenes even in messy, less planned places.
Open → 2611.04654v1

Planarcityloc improves drone navigation using 3d building planes

PlanarCityLoc: Breaking Silhouette Ambiguity in LoD Aerial Localization via 3D Planes

Abstract: Aerial visual localization is crucial for autonomous drone navigation in urban environments where GNSS signals are often degraded. Lightweight textureless Level-of-Detail (LoD) city models provide a geometric map alternative to textured assets. Existing silhouette-based methods rely on image-space alignment, which can leave different camera poses indistinguishable and favor incorrect poses when segmentation errors distort the building outlines. We present PlanarCityLoc, which moves beyond silhouette overlap by leveraging the planar geometry of building facades and roofs. Following geometry-aware initialization, we match image-side planar patches to LoD model planes and use the resulting 3D-to-2D correspondences to refine pose with Perspective-n-Point (PnP). The framework uses off-the-shelf geometry and segmentation models without dataset-specific training or fine-tuning. Experiments on UAVD4L-LoDv2 and Swiss-EPFLv2 achieve state-of-the-art localization recall among the evaluated off-the-shelf LoD-based methods. By grounding image-space alignment in 3D planes, PlanarCityLoc enables more accurate aerial localization with lightweight, textureless city maps.

Mon 19 OctComputer Vision and Pattern Recognition
The gist
Locating drones accurately over cities is hard because GPS signals often don’t work well there. The authors designed PlanarCityLoc, which helps drones find their exact position by matching flat surfaces on buildings seen in drone images to 3D building shapes in simple city maps. This method goes beyond just matching building outlines, which can be ambiguous or error-prone. It uses 3D geometry to better figure out where the drone is, without needing special training on specific city data.
Open → 2610.26729v1

Urbex enables fast city-scale 3D object grounding without proposals

Urbex: Agentic Spatial Grounding for City-Scale 3D Scenes

Abstract: City-scale 3D grounding is commonly formulated as candidate ranking, which requires precomputed instance proposals and becomes inefficient in large outdoor scenes. We propose Urbex, an agentic vision-language framework that instead casts grounding as active spatial search. Urbex represents a city-scale 3D scene as an interactive multi-view environment, allowing a VLM agent to query landmarks, zoom into local regions, render oblique views, and commit a grounded bounding box. We optimize this tool-use policy with reinforcement learning on CityRefer, using smooth localization rewards and lightweight shaping rewards for efficient exploration. Experiments show that Urbex achieves the best Hit@1 and IoU on CityRefer, supports candidate-free grounding without ground-truth boxes, and is substantially faster than the strongest city-scale ranking baseline. Zero-shot results on STPLS3D-Refer further indicate improved cross-scene robustness.

Tue 13 OctComputer Vision and Pattern RecognitionArtificial Intelligence
The gist
Finding exact locations of objects in huge 3D city maps usually needs lots of pre-made guesses, which is slow and hard. The authors propose Urbex, a system that actively searches and zooms around a 3D city scene like a smart agent until it finds the right spot. Urbex learns how to explore efficiently by trying different strategies during training, making it faster and more accurate than past methods. It also works well on brand-new city scenes without extra training.
Open → 2610.18356v1

Diffusion models improve 3D city maps from satellite images

Enhancing Photogrammetric Digital Surface Models with Pretrained Diffusion Models and Multimodal Conditioning

Abstract: Large-scale Digital Surface Models (DSMs) can be produced cost-effectively from satellite images via stereo-photogrammetry. However, the resulting 3D maps are often contaminated by noise, outliers, and voids. On the other hand, aerial LiDAR provides high-accuracy elevation measurements at a substantially higher cost. In this work, we study diffusion models conditioned both on photogrammetric DSMs and Pléiades imagery to refine vertically co-registered DSMs. We introduce a modified Stable Diffusion 3 architecture with a pruned text stream and a patch-wise normalization strategy, enabling stable training on LiDAR data and transfer from natural images to elevation maps. Experiments in French cities demonstrate that multimodal conditioning improves elevation accuracy, reducing Dense Urban RMSE from 6.00 to 3.45 m in the in-context cities and from 4.16 to 2.77 m in the held-out city of Bordeaux.

Fri 25 SeptComputer Vision and Pattern Recognition
The gist
Satellite images can create 3D maps of cities, but these maps often have mistakes like noise and missing areas. The authors improved these maps by using a special type of AI called diffusion models that learn from both the 3D map data and the original satellite images. They adjusted an existing AI model so it works better for these maps and trained it using high-accuracy LiDAR data. Tests in several French cities showed the improved maps are closer to the true shape of the land.
Open → 2609.31199v1