Journal for Geometry and Graphics 30 (2026), No. 1, 097--116
Copyright by the authors licensed under CC BY SA 4.0
Geometric Clustering of Point Clouds Using the Spheres Method in Parametric Design
Larysa Ivanova
(1) Dresden University of Technology, Dresden, Germany
(2) National University of Construction and Architecture, Kyiv, Ukraine
larysa.ivanova@mailbox.tu-dresden.de
Nataliya Sadretdinova
(1) Dresden University of Technology, Dresden, Germany
(2) State University of Technologies and Design, Kyiv, Ukraine
nataliya.sadretdinova@tu-dresden.de
Zlata Tosic
(1) Dresden University of Technology, Dresden, Germany
(2) National University of Construction and Architecture, Kyiv, Ukraine
zlata.tosic@tu-dresden.de
This article explores the use of non-hierarchical clustering methods for processing point clouds data in applied design and production tasks. The paper offers a detailed classification of clustering algorithms by their formation principles (e.g., partitioning, density-based, model-based, etc.) and introduces a geometric classification of point clouds based on their spatial distribution, dimensionality, and shape complexity.
The proposed clustering approach, referred to as the Spheres Method, is designed for the geometric interpretation of both initially spatial datasets (e.g., LiDAR, photogrammetry) and statistical or technological datasets transformed into point clouds through parameterization. The Spheres Method is particularly effective for identifying implicit groupings within uniform yet structurally diverse 2D/3D datasets. It enhances automation in point clouds segmentation and provides a scalable solution for optimizing design and manufacturing workflows.
The research focuses on two distinct cases: clustering 3D point clouds in the design of rehabilitation textile products and in the construction of a prefabricated garden pavilion shell using additive manufacturing.
Keywords: Point clouds clustering, non-hierarchical clustering, spheres Method, parametric design, classification of clustering methods, point clouds geometry.
MSC: 62H30; 68U05, 65D18, 62R40.
[ Fulltext-pdf (2340 KB)]