GridFlow: Structured Latent Flow for Seamless City-Scale 3D Point Cloud Generation

• URL: http://arxiv.org/abs/2608.29793v1
• Date: Sun, 30 Aug 2026 13:39:36 GMT
• Title: GridFlow: Structured Latent Flow for Seamless City-Scale 3D Point Cloud Generation
• Abstract summary: We present ours, a framework that generates colored point clouds at city scale conditioned on satellite imagery, semantic segmentation maps, and digital surface models.<n> Experiments show that ours outperforms adapted point cloud generation baselines across all geometry metrics.
• Score: 33.289877090164275
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Generating realistic 3D city environments from remote sensing data is important for simulation, urban planning, and mixed reality, yet existing point cloud generation methods are limited to single objects or bounded indoor scenes and cannot handle the scale, seamless tiling, and partial observability challenges of city-scale generation. We present \ours{}, a multi-stage framework that generates dense, colored point clouds ($10^5$ points per $150\text{m}{\times}150\text{m}$ tile) at city scale, conditioned on satellite imagery, semantic segmentation maps, and digital surface models (DSM). A \emph{Grid-Aligned VAE} encodes each tile into a topology-preserving latent grid where tokens correspond to fixed spatial regions, enabling spatially coherent multi-modal conditioning and compact latent-space edge consistency that implicitly aligns thousands of boundary points for seamless cross-tile generation. A conditional rectified flow model synthesizes geometry latents from the fused multi-modal conditions, and an orientation-aware diffusion colorizer separately handles satellite-visible horizontal surfaces and occluded vertical façades. To support standardized evaluation, we build on public 3D data sources to introduce \emph{City3D-MultiGen}, a benchmark of $163$K densely annotated tiles from Melbourne and London with aligned point clouds, satellite images, semantic maps, and elevation data. Experiments show that \ours{} outperforms adapted point cloud generation baselines across all geometry metrics and produces visually coherent colored point clouds with seamless boundaries over arbitrarily large urban extents. Our benchmark details are available at https://huggingface.co/datasets/e32/City3D-MultiGen

Related papers

• BenthicFlow: Generating Extensible Underwater Environments via Flow Matching [7.287319339936011]
  We introduce BenthicFlow, a unified framework based on a single conditional flow-matching model that jointly generates aligned textures and depth maps.<n> Experiments across geographically distinct survey sites demonstrate that BenthicFlow preserves site-specific appearance while generating coherent, large-scale 3D scenes.
  arXiv  Detail & Related papers  (2026-08-24T12:21:41Z)
• Uncovering and Shaping the Latent Representation of 3D Scene Topology in Vision-Language Models [9.722829662835233]
  We show that modern Vision-Language Models (VLMs) possess a latent topological map of 3D scenes.<n>By isolating this spatial subspace through cross-scene linear feature extraction, we extract a clean spatial subspace that causally controls the model's spatial outputs.<n>We mathematically shape this latent representation and prove its correspondence to the Laplacian eigenmaps of the scene's 3D Gaussian- kernel graph.
  arXiv  Detail & Related papers  (2026-05-08T02:32:27Z)
• CymbaDiff: Structured Spatial Diffusion for Sketch-based 3D Semantic Urban Scene Generation [55.74642848285121]
  We introduce SketchSem3D, the first large-scale benchmark for generating 3D outdoor semantic scenes from freehand sketches and satellite images.<n>We also propose Cylinder Mamba Diffusion (CymbaDiff) that significantly enhances spatial coherence in scene generation.<n>CymbaDiff achieves superior semantic consistency, spatial realism, and cross-dataset generalization.
  arXiv  Detail & Related papers  (2025-10-15T07:47:00Z)
• GaussianAnything: Interactive Point Cloud Flow Matching For 3D Object Generation [75.39457097832113]
  This paper introduces a novel 3D generation framework, offering scalable, high-quality 3D generation with an interactive Point Cloud-structured Latent space.<n>Our framework employs a Variational Autoencoder with multi-view posed RGB-D(epth)-N(ormal) renderings as input, using a unique latent space design that preserves 3D shape information.<n>The proposed method, GaussianAnything, supports multi-modal conditional 3D generation, allowing for point cloud, caption, and single image inputs.
  arXiv  Detail & Related papers  (2024-11-12T18:59:32Z)
• ImpliCity: City Modeling from Satellite Images with Deep Implicit Occupancy Fields [20.00737387884824]
  ImpliCity is a neural representation of the 3D scene as an implicit, continuous occupancy field, driven by learned embeddings of the point cloud and a stereo pair of ortho-photos. With image resolution 0.5$,$m, ImpliCity reaches a median height error of $approx,$0.7$,$m and outperforms competing methods.
  arXiv  Detail & Related papers  (2022-01-24T21:40:16Z)
• SensatUrban: Learning Semantics from Urban-Scale Photogrammetric Point Clouds [52.624157840253204]
  We introduce SensatUrban, an urban-scale UAV photogrammetry point cloud dataset consisting of nearly three billion points collected from three UK cities, covering 7.6 km2. Each point in the dataset has been labelled with fine-grained semantic annotations, resulting in a dataset that is three times the size of the previous existing largest photogrammetric point cloud dataset.
  arXiv  Detail & Related papers  (2022-01-12T14:48:11Z)
• Semantic Segmentation on Swiss3DCities: A Benchmark Study on Aerial Photogrammetric 3D Pointcloud Dataset [67.44497676652173]
  We introduce a new outdoor urban 3D pointcloud dataset, covering a total area of 2.7 $km2$, sampled from three Swiss cities. The dataset is manually annotated for semantic segmentation with per-point labels, and is built using photogrammetry from images acquired by multirotors equipped with high-resolution cameras.
  arXiv  Detail & Related papers  (2020-12-23T21:48:47Z)
• Towards Semantic Segmentation of Urban-Scale 3D Point Clouds: A Dataset, Benchmarks and Challenges [52.624157840253204]
  We present an urban-scale photogrammetric point cloud dataset with nearly three billion richly annotated points. Our dataset consists of large areas from three UK cities, covering about 7.6 km2 of the city landscape. We evaluate the performance of state-of-the-art algorithms on our dataset and provide a comprehensive analysis of the results.
  arXiv  Detail & Related papers  (2020-09-07T14:47:07Z)
• Learning Gradient Fields for Shape Generation [69.85355757242075]
  A point cloud can be viewed as samples from a distribution of 3D points whose density is concentrated near the surface of the shape. We generate point clouds by performing gradient ascent on an unnormalized probability density. Our model directly predicts the gradient of the log density field and can be trained with a simple objective adapted from score-based generative models.
  arXiv  Detail & Related papers  (2020-08-14T18:06:15Z)
• MNEW: Multi-domain Neighborhood Embedding and Weighting for Sparse Point Clouds Segmentation [1.2380933178502298]
  We propose MNEW, including multi-domain neighborhood embedding, and attention weighting based on their geometry distance, feature similarity, and neighborhood sparsity. MNEW achieves the top performance for sparse point clouds, which is important to the application of LiDAR-based automated driving perception.
  arXiv  Detail & Related papers  (2020-04-05T18:02:07Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.