三维线框生成的研究发展与综述

点击次数：

三维线框生成的研究发展与综述

三维线框因其轻量化、可编辑性强以及能准确刻画模型尖锐与关键特征的优势，已广泛应用于工业设计、遥感测量、城市建模等领域。三维线框在保留几何与形状特征的同时，显著降低了存储空间与计算成本，在三维重建、几何编辑、曲面分割等任务中发挥着重要作用。近年来，三维线框的实用性与有效性已得到广泛验证，其生成技术正逐渐成为计算机视觉与计算机图形学领域的研究热点。

常用的点云特征提取backbone

预览	论文名称	相关链接
	Pointnet: Deep learning on point sets for 3d classification and segmentation	Code
	Pointnet++: Deep hierarchical feature learning on point sets in a metric space	Code
	Dynamic graph cnn for learning on point clouds	Code
	Pointcept	Code

常用的数据集

预览	论文名称	相关链接
	Abc: A big cad model dataset for geometric deep learning	Project
	Building3d: A urban-scale dataset and benchmarks for learning roof structures from point clouds	Project
	Semantic3d. net: A new large-scale point cloud classification benchmark	Project
	Large-scale data for multiple-view stereopsis	Project
	Blendedmvs: A large-scale dataset for generalized multi-view stereo networks	Project
	Nef: Neural edge fields for 3d parametric curve reconstruction from multi-view images	Project

部分技术和算法源代码

预览	论文名称	相关链接
	Fast and robust edge extraction in unorganized point clouds	Code
	Voronoi-based feature curves extraction for sampled singular surfaces	Code
	Edge-aware point set resampling	Code
	Multiscale feature line extraction from raw point clouds based on local surface variation and anisotropic contraction	Code
	Rfeps: Reconstructing feature-line equipped polygonal surface	Code
	Ec-net: an edge-aware point set consolidation network	Code
	Def: Deep estimation of sharp geometric features in 3d shapes	Code
	MSL-Net: Sharp feature detection network for 3D point clouds	Code
	Complexgen: Cad reconstruction by b-rep chain complex generation	Code
	Surface and edge detection for primitive fitting of point clouds	Code
	Brepgen: A b-rep generative diffusion model with structured latent geometry	Code
	Split-and-fit: Learning b-reps via structure-aware voronoi partitioning	Code
	Sharp feature consolidation from raw 3D point clouds via displacement learning	Code
	STAR-Edge: Structure-aware Local Spherical Curve Representation for Thin-walled Edge Extraction from Unstructured Point Clouds	Code
	Pie-net: Parametric inference of point cloud edges	Code
	Pc2wf: 3d wireframe reconstruction from raw point clouds	Code
	Nerve: Neural volumetric edges for parametric curve extraction from point cloud	Code
	Deep Point Cloud Edge Reconstruction Via Surface Patch Segmentation	Code
	Generating 3D House Wireframes with Semantics	Code
	Clr-wire: Towards continuous latent representations for 3d curve wireframe generation	Code
	Robust and accurate feature detection on point clouds	Code
	Point2Roof: End-to-end 3D building roof modeling from airborne LiDAR point clouds	Code
	Detecting vertices of building roofs from ALS point cloud data	Code
	BWFormer: Building Wireframe Reconstruction from Airborne LiDAR Point Cloud with Transformer	Code
	3D Edge Sketch from Multiview Images	Code
	Nef: Neural edge fields for 3d parametric curve reconstruction from multi-view images	Code
	Neat: Distilling 3d wireframes from neural attraction fields	Code
	3d neural edge reconstruction	Code
	EdgeGaussians-3D Edge Mapping via Gaussian Splatting	Code
	Curve-Aware Gaussian Splatting for 3D Parametric Curve Reconstruction	Code
	SGCR: Spherical Gaussians for Efficient 3D Curve Reconstruction	Code

一、基于网格的三维线框生成方法

1.1 基于曲率估计的谷脊线检测

1. Solid shape

2. Ridge curves and shape analysis

3. Extraction of feature lines on triangulated surfaces using morphological operators

4. Detection of salient curvature features on polygonal surfaces

5. Normal vector voting: crease detection and curvature estimation on large, noisy meshes

6. The implicit structure of ridges of a smooth parametric surface

7. Ridge-valley lines on meshes via implicit surface fitting

8. Fast and Robust Detection of Crest Lines on Meshes

9. Smooth Feature Lines on Surface Meshes

10. Estimating Curvatures and Their Derivatives on Triangle Meshes

11. Functional Optimization for Fair Surface Design

12. Large Deformable Splines, Crest Lines and Matching

13. Finding Ridges and Valleys in a Discrete Surface Using a Modified MLS Approximation

14. Fast, Robust, and Faithful Methods for Detecting Crest Lines on Meshes

15. Separatrix Persistence: Extraction of Salient Edges on Surfaces Using Topological Methods

16. Feature Detection of Triangular Meshes Based on Tensor Voting Theory

17. Demarcating Curves for Shape Illustration

18. Thin nets and crest lines: Application to satellite data and medical images

19. The 3D marching lines algorithm

20. Extracting line representations of sulcal and gyral patterns in MR images of the human brain

21. Using a statistical shape model to extract sulcal curves on the outer cortex of the human brain

22. Crest lines for surface segmentation and flattening

23. Learning line features in 3D geometry

24. Feature curve co-completion in noisy data

1.2 基于网格分割的特征曲线检测

1. A general and efficient method for finding cycles in 3D curve networks

2. Flow aligned surfacing of curve networks

3. FlowRep: Descriptive curve networks for free-form design shapes

4. Cassie: Curve and surface sketching in immersive environments

5. Strokes2Surface: Recovering curve networks from 4D architectural design sketches

6. Fibermesh: Designing freeform surfaces with 3D curves

7. Segmentation through variable-order surface fitting

8. High-level CAD model acquisition from range images

9. Variational shape approximation

10. Structure Recovery via Hybrid Variational Surface Approximation

11. Quadric surface extraction by variational shape approximation

12. Variational mesh segmentation via quadric surface fitting

13. Interactive segmentation of scanned mechanical models based on quadratic surfaces fitting

14. Extraction and remeshing of ellipsoidal representations from mesh data

15. D-Charts: Quasi-Developable Mesh Segmentation

16. Blending surface segmentation and editing for 3D models

17. Hierarchical mesh segmentation based on fitting primitives

18. Hierarchical mesh decomposition using fuzzy clustering and cuts

19. A new CAD mesh segmentation method, based on curvature tensor analysis

20. Abstraction of man-made shapes

21. Adapting feature curve networks to a prescribed scale

22. iWIRES: An analyze-and-edit approach to shape manipulation

23. Exoskeleton: Curve network abstraction for 3D shapes

24. Recognition of feature curves on 3D shapes using an algebraic approach to Hough transforms

25. Patch layout generation by detecting feature networks

26. Extracting cycle-aware feature curve networks from 3D models

27. Feature-aligned segmentation using correlation clustering

28. Seg-mat: 3D shape segmentation using medial axis transform

29. 3D Shape Segmentation with Potential Consistency Mining and Enhancement

30. Patch layout from feature graphs

31. Extract feature curves on noisy triangular meshes

二、基于点云的三维线框生成方法

2.1 基于几何分析的尖锐特征检测

1. Feature Extraction From Point Clouds

2. Multi‐scale feature extraction on point‐sampled surfaces

3. Spline-based feature curves from point-sampled geometry

4. Feature preserving mesh generation from 3D point clouds

5. Sharp feature preserving MLS surface reconstruction based on local feature line approximations

6. Sharp feature detection in point clouds

7. Fast and robust edge extraction in unorganized point clouds

8. Feature‐preserving reconstruction of singular surfaces

9. Robust moving least-squares fitting with sharp features

10. Data-dependent MLS for faithful surface approximation

11. Voronoi-based feature curves extraction for sampled singular surfaces

12. Patch-graph reconstruction for piecewise smooth surfaces

13. Feature preserving point set surfaces based on non‐linear kernel regression

14. Multi-scale tensor voting for feature extraction from unstructured point clouds

15. A feature-preserving framework for point cloud denoising

16. Voronoi-based curvature and feature estimation from point clouds

17. SGLBP: Subgraph‐based local binary patterns for feature extraction on point clouds

18. Extracting feature lines from point clouds based on smooth shrink and iterative thinning

19. Neighbor reweighted local centroid for geometric feature identification

20. An adaptive normal estimation method for scanned point clouds with sharp features

21. A statistical approach for extraction of feature lines from point clouds

22. Edge and corner detection for unorganized 3d point clouds with application to robotic welding

23. Multiscale feature line extraction from raw point clouds based on local surface variation and anisotropic contraction

24. Rfeps: Reconstructing feature-line equipped polygonal surface

2.2 基于深度学习的CAD模型三维线框生成

1. Contour detection in unstructured 3D point clouds

2. Edge-aware point set resampling

3. Ec-net: an edge-aware point set consolidation network

4. Learning part boundaries from 3D point clouds

5. Def: Deep estimation of sharp geometric features in 3d shapes

6. MSL-Net: Sharp feature detection network for 3D point clouds

7. Complexgen: Cad reconstruction by b-rep chain complex generation

8. Surface and edge detection for primitive fitting of point clouds

9. Brepgen: A b-rep generative diffusion model with structured latent geometry

10. Split-and-fit: Learning b-reps via structure-aware voronoi partitioning

11. Hola: B-rep generation using a holistic latent representation

12. B-repLer: Semantic B-rep Latent Editor using Large Language Models

13. CADDreamer: CAD Object Generation from Single-view Images

14. Cmt: A cascade mar with topology predictor for multimodal conditional cad generation

15. Sharp feature consolidation from raw 3D point clouds via displacement learning

16. Sepicnet: Sharp edges recovery by parametric inference of curves in 3d shapes

17. Thin-walled aircraft panel edge extraction from 3-D measurement surfaces via feature-aware displacement learning

18. STAR-Edge: Structure-aware Local Spherical Curve Representation for Thin-walled Edge Extraction from Unstructured Point Clouds

19. Sketch2cad: Sequential cad modeling by sketching in context

20. Free2cad: Parsing freehand drawings into cad commands

21. Pie-net: Parametric inference of point cloud edges

22. Pc2wf: 3d wireframe reconstruction from raw point clouds

23. Coarse-to-fine pipeline for 3D wireframe reconstruction from point cloud

24. WireframeNet: A novel method for wireframe generation from point cloud

25. EDWG: Efficient Edge Detection and Wireframe Generation from Point Clouds

26. Nerve: Neural volumetric edges for parametric curve extraction from point cloud

27. Deep Point Cloud Edge Reconstruction Via Surface Patch Segmentation

28. Generating 3D House Wireframes with Semantics

29. Clr-wire: Towards continuous latent representations for 3d curve wireframe generation

2.3 面向激光雷达点云的三维线框生成

2.3.1 面向大型建筑物三维点云的线段检测

1. Edge detection and feature line tracing in 3D-point clouds by analyzing geometric properties of neighborhoods

2. Line segment extraction for large scale unorganized point clouds

3. Geometric feature enhanced line segment extraction from large-scale point clouds with hierarchical topological optimization

4. Facet segmentation-based line segment extraction for large-scale point clouds

5. Vision-based localization using an edge map extracted from 3D laser range data

6. Line-based extrinsic calibration of range and image sensors

7. Topologically aware building rooftop reconstruction from airborne laser scanning point clouds

8. Robust and accurate feature detection on point clouds

9. Large-scale point cloud contour extraction via 3D guided multi-conditional generative adversarial network

10. Superline3d: Self-supervised line segmentation and description for lidar point cloud

11. Feature line generation and regularization from point clouds

12. Detection of closed sharp edges in point clouds using normal estimation and graph theory

13. PECo: A Point-Edge Collaborative Framework for Global-Aware Urban Building Contouring From Unstructured Point Clouds

14. Extracting 3-D structural lines of building from ALS point clouds using graph neural network embedded with corner information

15. LCE-NET: Contour extraction for large-scale 3-D point clouds

16. Accurate and complete line segment extraction for large-scale point clouds

2.3.2 面向屋顶点云的线框生成

1. 3D building detection and modeling from aerial LIDAR data

2. Model driven reconstruction of roofs from sparse LIDAR point clouds

3. Flexible building primitives for 3D building modeling

4. Self-supervised pre-training for 3-D roof reconstruction on LiDAR data

5. Point2Roof: End-to-end 3D building roof modeling from airborne LiDAR point clouds

6. Detecting vertices of building roofs from ALS point cloud data

7. Self-Supervised Pretraining Framework for Extracting Global Structures From Building Point Clouds via Completion

8. PBWR: Parametric-building-wireframe reconstruction from aerial LiDAR point clouds

9. RR-Net: 3D Roof Reconstruction from Airborne LiDAR Point Clouds via Edge Segmentation and Wireframe Generation

10. BWFormer: Building Wireframe Reconstruction from Airborne LiDAR Point Cloud with Transformer

11. EdgeDiff: Edge-aware Diffusion Network for Building Reconstruction from Point Clouds

三、基于图像的三维线框生成方法

3.1 基于运动恢复结构的三维线段检测

1. Structure-from-motion using lines: Representation, triangulation, and bundle adjustment

2. Exploiting global connectivity constraints for reconstruction of 3D line segments from images

3. Line-based 3D reconstruction of wiry objects

4. Incremental Line-based 3D Reconstruction using Geometric Constraints

5. Semi-Global 3D Line Modeling for Incremental Structure-from-Motion

6. Improving sparse 3D models for man-made environments using line-based 3D reconstruction

7. Efficient 3D scene abstraction using line segments

8. ELSR: Efficient line segment reconstruction with planes and points guidance

3.2 基于多视图立体视觉的三维线段生成

1. Structure and motion from line segments in multiple images

2. Matching, reconstructing and grouping 3d lines from multiple views using uncertain projective geometry

3. Multi-view stereo 3D edge reconstruction

4. SOLD2: Self-supervised occlusion-aware line description and detection

5. Deeplsd: Line segment detection and refinement with deep image gradients

6. 3d line mapping revisited

7. Mv2cyl: Reconstructing 3d extrusion cylinders from multi-view images

8. 3D curve sketch: Flexible curve-based stereo reconstruction and calibration

9. 3D Edge Sketch from Multiview Images

3.3 基于可微渲染的三维线框生成

1. Nef: Neural edge fields for 3d parametric curve reconstruction from multi-view images

2. 3D wireframe model reconstruction of buildings from multi-view images using neural implicit fields

3. Neat: Distilling 3d wireframes from neural attraction fields

4. 3d neural edge reconstruction

5. EdgeGaussians-3D Edge Mapping via Gaussian Splatting

6. Curve-Aware Gaussian Splatting for 3D Parametric Curve Reconstruction

7. SGCR: Spherical Gaussians for Efficient 3D Curve Reconstruction