pith. sign in

arxiv: 2604.28064 · v1 · submitted 2026-04-30 · 💻 cs.CV

3D Reconstruction Techniques in the Manufacturing Domain: Applications, Research Opportunities and Use Cases

Chialoon Cheng (1) , Kaijun liu (2) , Zhiyang Liu (1) , Marcelo H Ang Jr (1) ((1) Advanced Robotics Centre , National University of Singapore , Singapore (2) Independent Researcher) This is my paper

Pith reviewed 2026-05-07 05:57 UTC · model grok-4.3

classification 💻 cs.CV
keywords 3D reconstructionmanufacturingpoint cloud generationhybrid systemsquality inspectiondeep learningstructured lightstereo vision
0
0 comments X

The pith

Review of 106 publications classifies manufacturing 3D reconstruction techniques and highlights the rise of hybrid systems.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The authors perform a systematic review of 106 recent papers to chart the state of 3D reconstruction for manufacturing. They sort the techniques into data acquisition, point cloud generation, post-processing, and application areas, finding non-contact methods such as structured light and stereo vision in heavy use for quality inspection tasks that make up nearly half the applications. Deep learning is credited with boosting both accuracy and speed in feature handling. The review identifies persistent problems with shiny surfaces and moving scenes, leading to the observation that combining different sensors and methods is the growing solution. Readers interested in industrial automation would find this useful for spotting where current tools fall short and where research effort is heading next.

Core claim

By reviewing 106 publications, the paper establishes that 3D reconstruction techniques in manufacturing are best understood through the categories of data acquisition, point cloud generation, post-processing, and applications. Non-contact approaches dominate, particularly for the 40 percent of uses in quality inspection, and deep learning aids in feature extraction. Despite sub-millimeter accuracy in controlled conditions, issues with reflective surfaces and dynamic environments remain, prompting a shift to hybrid systems that integrate multiple sensor types and processing techniques.

What carries the argument

The systematic categorization of reconstruction techniques into data acquisition, point cloud generation, post-processing, and applications, derived from the 106-paper survey, which exposes adoption rates and the movement toward hybrid configurations.

If this is right

  • Quality inspection comprises 40% of applications, far ahead of design at 13% and machining at 8%.
  • Structured light scanning and stereo vision are the most adopted non-contact methods.
  • Deep learning improves processing speed and accuracy in feature extraction and matching.
  • Hybrid systems are the emerging response to limitations in single-method approaches.
  • Sub-millimeter accuracy is achievable but mainly in controlled rather than real-world dynamic factory settings.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Manufacturers could use this classification to audit their current 3D scanning setups and identify gaps for hybrid upgrades.
  • The call for unified frameworks implies value in creating open standards for data exchange between different reconstruction modules.
  • Testing hybrid systems specifically on reflective metal parts common in manufacturing would be a direct next step to validate the trend.
  • Tracking how many new papers adopt hybrid methods in coming years could measure the predicted shift's speed.

Load-bearing premise

The 106 publications selected for review represent a fair cross-section of the field and the chosen categories accurately reflect the main divisions in current 3D reconstruction practice without overlooking major alternative approaches.

What would settle it

A follow-up review using a larger or differently filtered set of publications that fails to detect a trend toward hybrid systems or reports different percentages for application areas would indicate the findings are not general.

Figures

Figures reproduced from arXiv: 2604.28064 by Chialoon Cheng (1), Kaijun liu (2), Marcelo H Ang Jr (1) ((1) Advanced Robotics Centre, National University of Singapore, Singapore (2) Independent Researcher), Zhiyang Liu (1).

Figure 1
Figure 1. Figure 1: Criteria executed during the search process for relevant literature. omitted to ensure relevance. Non-English publications were also excluded due to translation constraints and the potential for misinterpretation. Applying the selection criteria outlined in view at source ↗
Figure 2
Figure 2. Figure 2: An overview of 3D Reconstruction pipeline. from resolution limits and high memory usage, especially for large-scale scenes. Boundary representation (B-rep) models 3D solids by explicitly defining their boundaries—faces, edges, and vertices—along with topological relationships[41]. This representation enables precise geometric and topological modeling widely used in CAD for design, simulation, and manufactu… view at source ↗
Figure 3
Figure 3. Figure 3: Different sensor types based on population from survey papers object to acquire three-dimensional point cloud data [118]. This approach typically employs motion-enabled devices such as Coordinate Measuring Machines (CMMs) or robotic manipulators, wherein a measurement probe is affixed to a multi-degree-of-freedom actuation platform, as illustrated in view at source ↗
Figure 4
Figure 4. Figure 4: The sensor types and their data acquisition methods are categorized into active and passive systems. The left section represents active systems, while the right section represents passive systems. Active systems include tactile probes, laser scanners, X-ray, Time-of-Flight (ToF), and structured light. Passive systems include monocular and binocular vision. measurement. According to the classification intro… view at source ↗
Figure 5
Figure 5. Figure 5: Comparison of automated scanning strategies for 3D reconstruction. methodologies to enhance accuracy, efficiency, and adapt￾ability. While contact-based methods provide high precision, their limitations in scalability and efficiency have led to the widespread adoption of non-contact techniques such as structured light, ToF, and laser scanning. Additionally, multi￾modal approaches and CAD/dataset-driven met… view at source ↗
Figure 6
Figure 6. Figure 6: Coordinate transformation among the sensor, welding, and workpiece coordinate frames [57]. address challenges in capturing curved surfaces by main￾taining a stable scanning distance. Such systems typically operate with a moving local scanner [67, 158, 61, 164] or by integrating the scanner onto a stationary robotic base or a mobile platform following a predetermined track [59, 49]. The benefits of robot ar… view at source ↗
Figure 7
Figure 7. Figure 7: Schematic diagram of global calibration and station calibration [59]. Handheld scanning [93, 111, 39, 58, 104, 83] is a versatile technique that employs manually operated portable devices, offering high flexibility and an unrestricted scanning range. This method accounts for approximately 10% of research using this strategy. Despite these advantages, it is inherently limited by the operator’s reach and phy… view at source ↗
Figure 8
Figure 8. Figure 8: Sparse and dense point cloud generation for the robot brick BIM task model [179]. Feature Detection and Description. Robust feature extraction underpins registration and reconstruction. De￾tectors such as SIFT[171], SURF[172], ORB[173], and AKAZE[174] identify scale- and rotation-invariant keypoints [62, 50, 145, 92, 84]. Shape-based techniques (e.g., Laplacian filtering, edge/circle detection [52, 175, 64… view at source ↗
Figure 9
Figure 9. Figure 9: Using NeRF to grasp transparent objects Given a scene with transparent objects (left column), the pipeline is on the right to compute grasps (middle column). [34]. synthesized and real images. This enables high-fidelity novel view synthesis but requires extensive per-scene optimization and lacks explicit surface geometry. A notable example of extending NeRF beyond novel view synthesis is Dex-NeRF [34], whi… view at source ↗
Figure 10
Figure 10. Figure 10: Industrial specific Focus. Where aerospace (23), automotive (7), consumer goods (13), general manufacturing (40), marine (4) and industrial machinery (13). error detection and correction. Kim et al. [89] introduced laser trackers for ship block assembly, replacing labor-intensive traditional measurement methods view at source ↗
Figure 12
Figure 12. Figure 12: illustrates the increasing adoption of 3D reconstruction technology across six major industrial sectors. The chart highlights a significant acceleration in the technology’s integration, aligning with Industry 4.0’s focus on automation, connectivity, and real-time data analytics. By 2024–2025, nearly all sectors demonstrate a marked increase in adop￾tion, indicating the transition of 3D reconstruction from… view at source ↗
read the original abstract

This comprehensive review examines the evolution and the current state of the art in three-dimensional (3D) reconstruction techniques in manufacturing applications. The analysis covers both traditional approaches and emerging deep learning methods, showing a critical research gap in unified 3d reconstruction frameworks. Through systematic review of 106 recent publications, we classify reconstruction techniques into three primary categories: data acquisition, point cloud generation, post-processing and applications. Non-contact methods, particularly structured light scanning and stereo vision, have shown significant adoption in manufacturing, with 47% of surveyed applications focusing on quality inspection. The integration of deep learning has enhanced reconstruction accuracy and processing speed, particularly in feature extraction and matching. Key applications span design and development (13%), machining (8%), process (17%), assembly (22%), and quality inspection (40%). While current technologies achieve sub-millimeter accuracy in controlled environments, challenges persist in handling reflective surfaces and dynamic environments. Our findings indicate a trend toward hybrid systems combining multiple sensor types and processing methods to overcome individual limitations. This survey provides a structured framework for understanding current capabilities and future directions in manufacturing-focused 3D reconstruction.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript is a survey of 3D reconstruction techniques in manufacturing. It reviews 106 recent publications and classifies the techniques into three primary categories: data acquisition, point cloud generation, post-processing and applications. The paper reports adoption statistics (47% non-contact methods with emphasis on quality inspection; application-area breakdown of 13% design, 8% machining, 17% process, 22% assembly, 40% quality inspection), notes accuracy gains from deep learning in feature extraction, identifies persistent challenges with reflective surfaces and dynamic scenes, and concludes that hybrid multi-sensor systems are an emerging trend while unified frameworks remain a critical gap.

Significance. If the review methodology proves rigorous, the sample representative, and the classification scheme clearly defined and consistently applied, the manuscript would supply a practical, statistics-supported overview of manufacturing-oriented 3D reconstruction. The concrete percentages, application breakdowns, and explicit identification of the hybrid-systems trend and unified-framework gap could serve as a useful reference for both researchers and practitioners seeking to locate current capabilities and open problems.

major comments (2)
  1. [Abstract] Abstract: the text states that techniques are classified into 'three primary categories' yet immediately enumerates four items ('data acquisition, point cloud generation, post-processing and applications'). This numerical mismatch renders the classification framework ambiguous. Because the reported statistics, the hybrid-systems trend, and the unified-framework gap are all derived from the application of this scheme to the 106 papers, the framework must be unambiguously defined (including whether 'applications' is a fourth category or an orthogonal dimension) and shown to have been applied uniformly.
  2. [Abstract] Abstract: the claim of a 'systematic review of 106 recent publications' is presented without any description of the search strategy, databases, date range, or inclusion/exclusion criteria. In the absence of these details the representativeness of the sample cannot be evaluated and the reproducibility of the reported percentages and gap analysis is compromised; this information is load-bearing for the central descriptive claims.
minor comments (1)
  1. [Abstract] Abstract: 'unified 3d reconstruction frameworks' uses lowercase '3d'; standard capitalization '3D' should be used consistently.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed comments on the abstract and classification framework. These observations identify genuine issues of clarity and methodological transparency that we will address in the revised manuscript. We respond to each major comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the text states that techniques are classified into 'three primary categories' yet immediately enumerates four items ('data acquisition, point cloud generation, post-processing and applications'). This numerical mismatch renders the classification framework ambiguous. Because the reported statistics, the hybrid-systems trend, and the unified-framework gap are all derived from the application of this scheme to the 106 papers, the framework must be unambiguously defined (including whether 'applications' is a fourth category or an orthogonal dimension) and shown to have been applied uniformly.

    Authors: We agree that the abstract contains an internal inconsistency by referring to 'three primary categories' while listing four items. In the full manuscript the technical classification of reconstruction methods is organized into three categories (data acquisition, point cloud generation, and post-processing), with applications treated as an orthogonal dimension used to map the 106 papers to manufacturing use cases. The reported percentages (e.g., 40 % quality inspection) are obtained by assigning each paper to its dominant application area, independent of the technical taxonomy. We will revise the abstract to state the three technical categories explicitly and to clarify that applications constitute a separate analytical dimension. In addition, we will insert a short subsection (or expanded caption for the classification figure) that defines the scheme, lists the criteria applied to each paper, and illustrates uniform application with representative examples from the surveyed set. revision: yes

  2. Referee: [Abstract] Abstract: the claim of a 'systematic review of 106 recent publications' is presented without any description of the search strategy, databases, date range, or inclusion/exclusion criteria. In the absence of these details the representativeness of the sample cannot be evaluated and the reproducibility of the reported percentages and gap analysis is compromised; this information is load-bearing for the central descriptive claims.

    Authors: We acknowledge that neither the abstract nor the current introduction supplies the explicit search protocol required for a fully reproducible systematic review. The 106 papers were assembled through a targeted literature search emphasizing manufacturing applications of 3D reconstruction, drawing primarily from IEEE Xplore, Scopus, and Web of Science for the period 2018–2023, with inclusion limited to peer-reviewed works that report concrete manufacturing use cases. To meet the referee’s requirement we will add a dedicated “Literature Search and Selection” subsection that documents the databases, exact search strings, date range, inclusion criteria (peer-reviewed, manufacturing focus, quantitative or qualitative 3D-reconstruction results), and exclusion criteria (purely theoretical papers, non-manufacturing domains, conference abstracts without full text). This addition will allow readers to evaluate sample representativeness and will support the reproducibility of the statistics and gap analysis. revision: yes

Circularity Check

0 steps flagged

No circularity: descriptive literature review with no derivations or self-referential reductions

full rationale

This is a systematic review paper that summarizes and categorizes findings from 106 external publications. No mathematical derivations, equations, fitted parameters, or model predictions are present. Claims about trends (e.g., hybrid systems, adoption rates) are aggregated from the cited literature rather than generated internally via any self-definitional loop, fitted-input prediction, or self-citation chain. The noted phrasing inconsistency in the abstract (three categories listed with four items) is a potential clarity or enumeration issue but does not reduce any derivation to its own inputs by construction. The paper remains self-contained against external benchmarks as a descriptive synthesis.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

As a survey paper the work introduces no new free parameters, axioms, or invented entities; it aggregates and categorizes results already present in the referenced literature.

pith-pipeline@v0.9.0 · 5534 in / 1193 out tokens · 54158 ms · 2026-05-07T05:57:45.003711+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

209 extracted references · 16 canonical work pages

  1. [1]

    A Comprehensive Review of Vision-Based 3D Reconstruction Methods,

    L. Zhou, G. Wu, Y. Zuo, X. Chen, and H. Hu, “A Comprehensive Review of Vision-Based 3D Reconstruction Methods,”Sensors, vol. 24, no. 7, p. 2314, Apr. 2024. [Online]. Available: https://www.mdpi.com/1424-8220/24/7/2314

    2024

  2. [2]

    Machine perception of three-dimensional solids,

    L. G. Roberts, “Machine perception of three-dimensional solids,” Ph.D.dissertation,MassachusettsInstituteofTechnology,Cambridge, MA, USA, 1963

    1963

  3. [3]

    Measurement of the 3-d,

    S. Kiyasu, H. Hoshino, K. Yano, and S. Fujimura, “Measurement of the 3-d,”IEEE Transactions on Instrumentation and Measurement, pp. 775–778, 1995

    1995

  4. [4]

    Structured-light 3d surface imaging: A tutorial,

    J. Geng, “Structured-light 3d surface imaging: A tutorial,”Advances inOpticsand Photonics, vol. 3, no. 2, pp. 128–160, 2011

    2011

  5. [5]

    Local surface interpolation with bézier,

    L. A. Shirman and C. H. Séquin, “Local surface interpolation with bézier,”ComputerAided Geometric Design, pp. 279–295, 1987

    1987

  6. [6]

    Reverse engineering: a roadmap,

    H. A. Müller, J. H. Jahnke, D. B. Smith, M.-A. Storey, S. R. Tilley, and K. Wong, “Reverse engineering: a roadmap,” inProceedingsof the Conference. Limerick Ireland: ACM, 2000, pp. 47–60

    2000

  7. [7]

    Coordinatemeasuringmachines:Amoderninspection tool in manufacturing,

    M.R.Mantel,“Coordinatemeasuringmachines:Amoderninspection tool in manufacturing,” Master’s thesis, New Jersey Institute of Technology, 1993

    1993

  8. [8]

    Nist’s 3d structured light scanner performance eval- uation project,

    “Nist’s 3d structured light scanner performance eval- uation project,” 2020, accessed: 10 August 2025. [Online]. Available: https://www.nist.gov/programs-projects/ 3d-structured-light-scanner-performance-evaluation

    2020

  9. [9]

    Accurate, dense, and robust multiview stereopsis,

    Y. Furukawa and J. Ponce, “Accurate, dense, and robust multiview stereopsis,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 32, no. 8, pp. 1362–1376, 2010

    2010

  10. [10]

    Cyber-physical production systems: roots from manufacturing science and technology,

    L. Monostori, “Cyber-physical production systems: roots from manufacturing science and technology,” at - Automatisierungstechnik, pp. 766–776, 2015

    2015

  11. [11]

    Imagenet,

    A. Krizhevsky, I. Sutskever, and G. E. Hinton, “Imagenet,” in Advancesin Neural. Curran Associates, Inc., 2012

    2012

  12. [12]

    Nerf: Representing scenes as neural radiance fields for view synthesis,

    B. Mildenhall, P. P. Srinivasan, M. Tancik, J. T. Barron, R. Ramamoorthi, and R. Ng, “Nerf: representing scenes as neural radiance fields for view synthesis,”Commun. ACM, pp. 99–106, 2021, accessed: 11 August 2025. [Online]. Available: https://doi.org/10.1145/3503250

    work page doi:10.1145/3503250 2021

  13. [13]

    3d gaussian splatting for real-time radiance field rendering,

    B. Kerbl, G. Kopanas, T. Leimkühler, and G. Drettakis, “3d gaussian splatting for real-time radiance field rendering,”ACMTransactions onGraphics, vol. 42, no. 4, pp. 1–19, 2023

    2023

  14. [14]

    Neu- ralradiancefieldsintheindustrialandroboticsdomain:Applications,

    E.Šlapak,E.Pardo,M.Dopiriak,T.Maksymyuk,andJ.Gazda,“Neu- ralradiancefieldsintheindustrialandroboticsdomain:Applications,” RoboticsandComputer-IntegratedManufacturing, p. 102810, 2024

    2024

  15. [15]

    Digital-twin deep dynamic camera position optimisation for the v,

    L.Wang,Z.Wang,P.Kendall,K.Gumma,A.Turner,andS.Ratchev, “Digital-twin deep dynamic camera position optimisation for the v,” International Journal of ProductionResearch, pp. 3932–3951, 2024

    2024

  16. [16]

    Computer vision for automated industrial inspection: A comprehensive review,

    L. Zhou, L. Zhang, and N. Konz, “Computer vision for automated industrial inspection: A comprehensive review,”IEEETransactions onSystems,Man,andCybernetics:Systems, vol. 53, no. 1, pp. 105– 117, 2023

    2023

  17. [17]

    A review of 3d reconstruction techniques in civil engineering and their applications,

    Z. Ma and S. Liu, “A review of 3d reconstruction techniques in civil engineering and their applications,”Advanced Engineering Informatics, vol. 37, pp. 163–174, 2018

    2018

  18. [18]

    3d reconstruction and measurementtechniquesforstructuralhealthmonitoringusingvision sensors: A review,

    W. Flores-Fuentes, G. Trujillo-Hernández, I. Y. Alba-Corpus, J. C. Rodríguez-Quiñonez, J. E. Mirada-Vega, D. Hernández-Balbuena, F. N. Murrieta-Rico, and O. Sergiyenko, “3d reconstruction and measurementtechniquesforstructuralhealthmonitoringusingvision sensors: A review,”Measurement, vol. 210, p. 112321, 2023

    2023

  19. [19]

    A. M. K. Thomas and A. K. Banerjee,The History of Radiology. Oxford, UK: Oxford University Press, 2013

    2013

  20. [20]

    Random sample consensus: a paradigm for model fitting with applications to image analysis and automatedcartography,

    M. A. Fischler and R. C. Bolles, “Random sample consensus: a paradigm for model fitting with applications to image analysis and automatedcartography,”CommunicationsoftheACM,pp.381–395, 1981

    1981

  21. [21]

    Fast Point Feature Histograms (FPFH) for 3D registration,

    R. B. Rusu, N. Blodow, and M. Beetz, “Fast Point Feature Histograms (FPFH) for 3D registration,” in 2009 IEEE International Conference on Robotics and Automation. Kobe: IEEE, May 2009, pp. 3212–3217. [Online]. Available: http://ieeexplore.ieee.org/document/5152473/

    work page arXiv 2009

  22. [22]

    A method for registration of 3-d,

    P. Besl and N. D. McKay, “A method for registration of 3-d,”IEEE TransactionsonPatternAnalysisandMachineIntelligence,pp.239– 256, 1992

    1992

  23. [23]

    Deepsdf: Learning continuous signed distance functions for shape representation,

    J. J. Park, P. Florence, J. Straub, R. Newcombe, and S. Lovegrove, “Deepsdf: Learning continuous signed distance functions for shape representation,” inCVPR, 2019

    2019

  24. [24]

    Avolumetricmethodforbuildingcomplex models from range images,

    B.CurlessandM.Levoy,“Avolumetricmethodforbuildingcomplex models from range images,” inProceedings of the 23rd Annual Conference on Computer Graphics and Interactive Techniques (SIGGRAPH), 1996, pp. 303–312

    1996

  25. [25]

    Kinectfusion: Real-time dense surface mapping and tracking,

    R. A. Newcombe, S. Izadiet al., “Kinectfusion: Real-time dense surface mapping and tracking,” inISMAR, 2011, pp. 127–136

    2011

  26. [26]

    Fronts propagating with curvature- dependentspeed:Algorithmsbasedonhamilton-jacobiformulations,

    S. Osher and J. A. Sethian, “Fronts propagating with curvature- dependentspeed:Algorithmsbasedonhamilton-jacobiformulations,” Journal of Computational Physics, vol. 79, no. 1, pp. 12–49, 1988

    1988

  27. [27]

    Point cloud enhancement optimization and high-fidelity texture reconstruction methods for air material via fusion of 3d,

    Q. Hu, X. Wei, X. Zhou, Y. Yin, H. Xu, W. He, and S. Zhu, “Point cloud enhancement optimization and high-fidelity texture reconstruction methods for air material via fusion of 3d,”Expert Systemswith Applications, p. 122736, 2024

    2024

  28. [28]

    Implicit multi-sensorreconstructionbasedonneuralsigneddistancefunctions for reverse engineering,

    G. Chen, Y. Li, C. Mehdi-Souzani, W. Yang, and X. Liu, “Implicit multi-sensorreconstructionbasedonneuralsigneddistancefunctions for reverse engineering,”ProcediaCIRP, pp. 552–557, 2023

    2023

  29. [29]

    Marching cubes: A high resolution 3d surface construction algorithm,

    W. E. Lorensen and H. E. Cline, “Marching cubes: A high resolution 3d surface construction algorithm,”ACM SIGGRAPH Computer Graphics, vol. 21, no. 4, pp. 163–169, August 1987

    1987

  30. [30]

    Graspnerf,

    Q.Dai,Y.Zhu,Y.Geng,C.Ruan,J.Zhang,andH.Wang,“Graspnerf,” IEEEInternationalConferenceonRoboticsandAutomation(ICRA), 2023

    2023

  31. [31]

    Rgbgrasp: Real-time grasp detection using rgb image,

    C.Liu,K.Shi,K.Zhou,H.Wang,J.Zhang,andH.Dong,“Rgbgrasp: Real-time grasp detection using rgb image,” 2023, accessed: 10 August 2025. [Online]. Available: https://arxiv.org/abs/2303.09227

    work page arXiv 2023

  32. [32]

    Volumetric grasping network: Real-time 6 dof grasp detection in clutter,

    M. Breyer, J. J. Chung, L. Ott, R. Siegwart, and J. Nieto, “Volumetric grasping network: Real-time 6 dof grasp detection in clutter,” 2021, accessed: 10 August 2025. [Online]. Available: https://arxiv.org/abs/2105.06476

    work page arXiv 2021

  33. [33]

    Neuralrendering-enabled 3d,

    J.Zhang,S.Liu,R.X.Gao,andL.Wang,“Neuralrendering-enabled 3d,”CIRPAnnals, pp. 93–96, 2023

    2023

  34. [34]

    Dex-nerf:Usinga neuralradiancefieldtograsptransparentobjects,

    J.Ichnowski,Y.Avigal,J.Kerr,andK.Goldberg,“Dex-nerf:Usinga neuralradiancefieldtograsptransparentobjects,”2021,accessed:10 August 2025. [Online]. Available: https://arxiv.org/abs/2110.14217

    work page arXiv 2021

  35. [35]

    Free- viewpointimagegenerationusingneuralradiancefieldsforindustrial applications,

    K. Ito, S. Ueda, S. Mori, J. Sugano, H. Adachi, and H. Saito, “Free- viewpointimagegenerationusingneuralradiancefieldsforindustrial applications,” inProceedings of the 49th Annual Conferenceof the IEEE Industrial ElectronicsSociety(IECON). Chicago, IL, USA: IEEE, 2024, pp. 1–6

    2024

  36. [36]

    Deep Learning on Point Clouds and Its Application: A Survey,

    W. Liu, J. Sun, W. Li, T. Hu, and P. Wang, “Deep Learning on Point Clouds and Its Application: A Survey,” Sensors, vol. 19, no. 19, p. 4188, Sep. 2019. [Online]. Available: https://www.mdpi.com/1424-8220/19/19/4188

    2019

  37. [37]

    Grasp pose detection in point clouds,

    A. ten Pas, M. Gualtieri, K. Saenko, and R. Platt, “Grasp pose detection in point clouds,” 2017, accessed: 10 August 2025. [Online]. Available: https://arxiv.org/abs/1706.09911

    work page arXiv 2017

  38. [38]

    Toleranceanaly- sis—formdefectsmodelingandsimulationbymodaldecomposition andoptimization,

    L.Homri,E.Goka,G.Levasseur,andJ.-Y.Dantan,“Toleranceanaly- sis—formdefectsmodelingandsimulationbymodaldecomposition andoptimization,”Computer-AidedDesign,vol.91,pp.46–59,2017

    2017

  39. [39]

    Aircraft skin inspection using a high-precision 3d measurement system,

    Q. Xie, D. Lu, K. Du, J. Xu, J. Dai, H. Chen, and J. Wang, “Aircraft skin inspection using a high-precision 3d measurement system,” Computer-AidedDesign, vol. 127, p. 102805, 2020

    2020

  40. [40]

    Voxel-based representation of 3d point clouds,

    Y. Xu et al., “Voxel-based representation of 3d point clouds,” Automation in Construction, vol. 127, p. 103709, 2021. [Online]. Available: https://www.sciencedirect.com/science/article/abs/pii/ S0926580521001266

    2021

  41. [41]

    ComputerScience Press, 1988

    M.Mäntylä, AnIntroductiontoSolidModeling. ComputerScience Press, 1988

    1988

  42. [42]

    X. Liu, Z. Wang, S. N. Melkote, and D. W. Rosen, “Manufacturing process identification from 3D point cloud models Preprint Page 19 of 24 3D Reconstruction Techniques in the Manufacturing Domain: Applications, Research Opportunities and Use Cases using semantic segmentation,”Journal of Manufacturing Systems, vol. 82, pp. 858–873, Oct. 2025. [Online]. Avail...

    2025

  43. [43]

    Advancing front mesh generation for point clouds with feature preservation,

    T. Liu, H. Ye, J. Zheng, Y. Zheng, and J. Chen, “Advancing front mesh generation for point clouds with feature preservation,” Computer-AidedDesign, vol. 168, p. 103683, 2024

    2024

  44. [44]

    Constructive solid geometry for polyhedral objects,

    D. Laidlaw, W. Trumbore, and J. Hughes, “Constructive solid geometry for polyhedral objects,” inProceedingsofthe13thAnnual Conference on Computer Graphics and Interactive Techniques (SIGGRAPH’86), 1986, pp. 161–170

    1986

  45. [45]

    Teapot Three-Dimensional Geometrical Model Reconstruction Based on Reverse Engineering and Rapid Prototyping Technology,

    S. Lv, Y. Zhu, H. Ni, X. Wang, T. Huang, and J. Zhang, “Teapot Three-Dimensional Geometrical Model Reconstruction Based on Reverse Engineering and Rapid Prototyping Technology,” in2018 3rdInternationalConferenceonMechanical,ControlandComputer Engineering (ICMCCE). Huhhot: IEEE, Sep. 2018, pp. 180–184. [Online]. Available: https://ieeexplore.ieee.org/docum...

    work page arXiv 2018

  46. [46]

    A robust registration algorithm of point clouds based on adaptive distance function for surface inspection,

    J. Ding, Q. Liu, and P. Sun, “A robust registration algorithm of point clouds based on adaptive distance function for surface inspection,” MeasurementScienceand Technology, p. 075003, 2019

    2019

  47. [47]

    Deep neural constraint satisfaction for parametric cad,

    G. Harabin, A. M. Mirzendehdel, and M. Behandish, “Deep neural constraint satisfaction for parametric cad,”Computer-AidedDesign, p. 103556, 2023

    2023

  48. [48]

    3d digital twin for real-time monitoring and control in industrial automation systems,

    K. R. da Silva Santos, W. R. de Oliveira, E. Villani, and A. Dittmann, “3d digital twin for real-time monitoring and control in industrial automation systems,”Computers in Industry, vol. 148, p. 103850, 2023

    2023

  49. [49]

    Improving accuracy reconstruction of parts through a capability study: A case study in the aeronautical industry,

    A. L. Reun, K. Subrin, A. Dubois, and S. Garnier, “Improving accuracy reconstruction of parts through a capability study: A case study in the aeronautical industry,”Roboticsand Autonomous Systems, vol. 174, p. 104564, 2024

    2024

  50. [50]

    Cad-aided automated robotic grasping system for disassembly of used products,

    L. Zhang, L. Wang, X. Du, and F. Meng, “Cad-aided automated robotic grasping system for disassembly of used products,”Scientific Programming, pp. 1–11, 2022

    2022

  51. [51]

    Application of point cloud technology in automated welding deformation prediction of large- scale structures,

    L. Ma, C. Zhang, Y. Fu, and D. Ma, “Application of point cloud technology in automated welding deformation prediction of large- scale structures,” in2019ChineseControlandDecisionConference (CCDC). Nanchang, China: IEEE, 2019, pp. 1003–1007

    2019

  52. [52]

    Structural regularity detec- tion and enhancement for surface mesh reconstruction in reverse engineering,

    A. Mu, Z. Liu, G. Duan, and J. Tan, “Structural regularity detec- tion and enhancement for surface mesh reconstruction in reverse engineering,”Computer-AidedDesign, p. 103780, 2024

    2024

  53. [53]

    Mesh-driven resampling and regularization for robust point cloud-based flow analysis directly on scanned objects,

    M. Jaiswal, A. M. Corpuz, and M.-C. Hsu, “Mesh-driven resampling and regularization for robust point cloud-based flow analysis directly on scanned objects,”ComputerMethodsin Applied Mechanicsand Engineering, p. 117426, 2024

    2024

  54. [54]

    A novel method for 3d,

    T. Li, F. Duan, X. Fu, C. Liu, C. Liang, A. Chen, and X. Li, “A novel method for 3d,”Measurement, p. 112930, 2023

    2023

  55. [55]

    A ransac-based approach for 3d registration of point clouds acquired from a low-cost 3d sensor,

    N. Mosca, C. Patruno, R. Colella, S. P. Negri, and E. Stella, “A ransac-based approach for 3d registration of point clouds acquired from a low-cost 3d sensor,” in2020 IEEE International Conference on Metrology for Industry 4.0 and IoT(MetroInd4.0&IoT). Pisa, Italy: IEEE, 2020, pp. 403–408

    2020

  56. [56]

    Density-invariant and structure-preserving point cloud registration for 3d measurement,

    Y. Wang, Y. Liu, Q. Xie, Q. Wu, X. Guo, Z. Yu, and J. Wang, “Density-invariant and structure-preserving point cloud registration for 3d measurement,”IEEE Transactions on Instrumentation and Measurement, vol. 70, pp. 1–15, 2021

    2021

  57. [57]

    Rapid surface defects detection in wire and arc additive manufacturing based on laser profilometer,

    C. Huang, G. Wang, H. Song, R. Li, and H. Zhang, “Rapid surface defects detection in wire and arc additive manufacturing based on laser profilometer,”Measurement, p. 110503, 2022

    2022

  58. [58]

    3dreconstructiontechnologyfor industrialpartsbasedonbinocularvision,

    Z.Fu,B.Qiu,J.Shen,andN.Chen,“3dreconstructiontechnologyfor industrialpartsbasedonbinocularvision,”in 20232ndInternational Conference on Robotics, Automation and Artificial Intelligence (RAAI). Xishuangbanna, China: IEEE, 2023, pp. 317–321

    2023

  59. [59]

    Aviation equipment measurement and assembly analysis method based on robotic system,

    B. Zhou, L. Zhao, T. Tian, J. Zhao, R. Xia, and D. Liu, “Aviation equipment measurement and assembly analysis method based on robotic system,”Measurement, p. 114810, 2024

    2024

  60. [60]

    Ahigh-accuracy pose measurement system for robotic automated assembly in large- scale space,

    Y.Liu,J.Zhou,Y.Li,Y.Zhang,Y.He,andJ.Wang,“Ahigh-accuracy pose measurement system for robotic automated assembly in large- scale space,”Measurement, p. 110426, 2022

    2022

  61. [61]

    Registration strategy of point clouds based on region-specific projectionsandvirtualstructuresforrobot-basedinspectionsystems,

    P. Bauer, L. Heckler, M. Worack, A. Magaña, and G. Reinhart, “Registration strategy of point clouds based on region-specific projectionsandvirtualstructuresforrobot-basedinspectionsystems,” Measurement, p. 109963, 2021

    2021

  62. [62]

    Screening method for automotive suspension parts based on image 3d,

    Z. Zhang and R. Fan, “Screening method for automotive suspension parts based on image 3d,” in2023 International. Chengdu, China: IEEE, 2023, pp. 556–560

    2023

  63. [63]

    Three-dimensional reconstruction method basedonbionicactivesensinginprecisionassembly,

    Z.Ding,H.Xu,G.Chen,Z.Wang,W.Chi,H.Zhang,Z.Wang,L.Sun, G. Yang, and Y. Wen, “Three-dimensional reconstruction method basedonbionicactivesensinginprecisionassembly,” AppliedOptics, p. 846, 2020

    2020

  64. [64]

    G. Xu, H. Shen, Y. Zhu, F. Chen, and X. Li, “3d,”Measurement, p. 110582, 2022

    2022

  65. [65]

    Geometry perception and motion planning in robotic assembly based on semanticsegmentationandpointcloudsreconstruction,

    Y. Jiang, G. Liu, Z. Huang, B. Yang, and W. Yang, “Geometry perception and motion planning in robotic assembly based on semanticsegmentationandpointcloudsreconstruction,” Engineering Applications ofArtificialIntelligence, p. 107678, 2024

    2024

  66. [66]

    Assembly task execution using visual 3d,

    M. Nigro, M. Sileo, F. Pierri, D. Bloisi, and F. Caccavale, “Assembly task execution using visual 3d,”Roboticsand Computer-Integrated Manufacturing, p. 102519, 2023

    2023

  67. [67]

    Revolutionizing robotized assembly for wire harness: A,

    T. P. Nguyen, D. Kim, H.-K. Lim, and J. Yoon, “Revolutionizing robotized assembly for wire harness: A,”Journal of Manufacturing Systems, pp. 360–372, 2024

    2024

  68. [68]

    Cfvs:Coarse-to-fine visual servoing for 6-dof object-agnostic peg-in-hole assembly,

    B.-S.Lu,T.-I.Chen,H.-Y.Lee,andW.H.Hsu,“Cfvs:Coarse-to-fine visual servoing for 6-dof object-agnostic peg-in-hole assembly,” in 2023 IEEE International Conference on Robotics and Automation (ICRA). London, UK: IEEE, 2023, pp. 12402–12408

    2023

  69. [69]

    Rearrangement planning for object reconfiguration,

    Y. Li, A. Zeng, and S. Song, “Rearrangement planning for object reconfiguration,” 2023, accessed: 10 August 2025. [Online]. Available: https://arxiv.org/abs/2303.01341

    work page arXiv 2023

  70. [70]

    Vision-based robotic grasping system for cylindrical object in cluttered environments,

    W.-C. Chang, Y.-K. Lin, and V.-T. Pham, “Vision-based robotic grasping system for cylindrical object in cluttered environments,” in 2021 9th International Conferenceon Control,Mechatronicsand Automation(ICCMA). Belval, Luxembourg: IEEE, 2021, pp. 218– 223

    2021

  71. [71]

    Posemeasurementoflarge-scale components based on improved iterative closest point algorithm,

    Z.Wang,Z.Liu,J.Fan,andF.Jing,“Posemeasurementoflarge-scale components based on improved iterative closest point algorithm,” in 2020ChineseControlandDecisionConference(CCDC). Shanghai, China: IEEE, 2020, pp. 1091–1096

    2020

  72. [72]

    Multi- robot cooperative assembly based on visual servoing and task alloca- tion,

    G. Tan, J. Du, S. Shao, R. Guo, R. Su, and K. Chang, “Multi- robot cooperative assembly based on visual servoing and task alloca- tion,” in2023 IEEE 11th International Conference on Information, CommunicationandNetworks(ICICN). Hefei,China:IEEE,2023, pp. 890–894

    2023

  73. [73]

    6d pose estimation for peg- in-hole assembly using depth images,

    O. Skeik, M. S. Erden, and X. Kong, “6d pose estimation for peg- in-hole assembly using depth images,” in2022 IEEE International Conference on Mechatronics (ICM). Genova, Italy: IEEE, 2022, pp. 1–6

    2022

  74. [74]

    Vision-enhanced peg-in-hole robotic assembly through deep learning and visual servoing,

    M. Sileo, N. Capece, M. Gruosso, M. Nigro, D. D. Bloisi, F. Pierri, and U. Erra, “Vision-enhanced peg-in-hole robotic assembly through deep learning and visual servoing,”Engineering Applications of ArtificialIntelligence, vol. 126, p. 107486, 2024

    2024

  75. [75]

    Cyber-physical systems in non-rigid assemblies: A,

    N. Theodoropoulos, E. Kampourakis, D. Andronas, and S. Makris, “Cyber-physical systems in non-rigid assemblies: A,”Journal of ManufacturingSystems, pp. 525–537, 2023

    2023

  76. [76]

    Fastregistrationof 3d point clouds for robotic assembly using improved ICP algorithm,

    J.Xu,R.Chen,H.Chen,S.Zhang,andK.Chen,“Fastregistrationof 3d point clouds for robotic assembly using improved ICP algorithm,” IEEETransactionsonIndustrial Electronics, vol. 64, no. 9, pp. 717– 726, 2017

    2017

  77. [77]

    Deep learning-based inpainting of high dynamic range fringe pattern for high-speed 3d,

    D. Xi, L. Hou, F. Wu, and Y. Qin, “Deep learning-based inpainting of high dynamic range fringe pattern for high-speed 3d,”Advanced Engineering Informatics, p. 102428, 2024

    2024

  78. [78]

    Learning-basedobjectdetectionandlocalizationforamo- bile robot manipulator in sme,

    Z. Zhou, L. Li, A. Fürsterling, H. J. Durocher, J. Mouridsen, and X.Zhang,“Learning-basedobjectdetectionandlocalizationforamo- bile robot manipulator in sme,”Robotics and Computer-Integrated Manufacturing, p. 102229, 2022

    2022

  79. [79]

    Point cloud segmentation of 3d mechanical parts for robotic assembly,

    X. Gong, M. Chen, and X. Yang, “Point cloud segmentation of 3d mechanical parts for robotic assembly,” in2017 IEEE International Preprint Page 20 of 24 3D Reconstruction Techniques in the Manufacturing Domain: Applications, Research Opportunities and Use Cases ConferenceonReal-timeComputingandRobotics(RCAR). Macau SAR, China: IEEE, 2017, pp. 1–6

    2017

  80. [80]

    Eye-in-hand vision-based robotic bin- picking with active laser projection,

    W.-C. Chang and C.-H. Wu, “Eye-in-hand vision-based robotic bin- picking with active laser projection,”The International Journal of AdvancedManufacturingTechnology, pp. 2873–2885, 2016

    2016

Showing first 80 references.