Operator-Splitting Surface Evolution with Event-Driven Hard Constraints for Watertight Point-Cloud Reconstruction
Mathematical Modeling and Computing · 2026 · 13 (2) · 598-613
DOI: 10.23939/mmc2026.02.598 ↗Kurz gefasst
Fachartikel in der Zeitschrift Mathematical Modeling and Computing, 2026. Verfasst von Moriya Netzer. Die Arbeit nennt 35 Literaturangaben.
Abstract (Originalsprache)
We introduce a hybrid surface-reconstruction scheme that couples a smooth pressure-tension surface evolution to an event-driven hard-constraint mechanism for point-cloud attachment. The key additional element is an active-set formulation in which constraints are activated by a thresholded capture rule and persist as hard pins during subsequent relaxation, enabling robust enclosure and feature locking under extreme non-uniform point density while maintaining a watertight triangulated surface. Adaptive remeshing is coupled to the pinned set to preserve mesh quality without violating hard constraints. We validate the framework on geometric primitives and apply it to a ray-tracing-generated Focal Body point cloud extracted from a volumetric ray-density field above a spherical mirror. By incorporating adaptive attractive weights and dynamic constraint projections, the method reconstructs the resulting mushroom-like high-density structure for which convex-hull and alpha-shape pipelines are unreliable under the observed density gradients.
Zitieren
APA
Netzer, M. (2026). Operator-Splitting Surface Evolution with Event-Driven Hard Constraints for Watertight Point-Cloud Reconstruction. Mathematical Modeling and Computing, 13(2), 598-613. https://doi.org/10.23939/mmc2026.02.598
Vancouver
Netzer M. Operator-Splitting Surface Evolution with Event-Driven Hard Constraints for Watertight Point-Cloud Reconstruction. Mathematical Modeling and Computing. 2026;13(2):598-613. doi:10.23939/mmc2026.02.598
BibTeX
@article{Netzer2026,
author = {Netzer, Moriya},
title = {Operator-Splitting Surface Evolution with Event-Driven Hard Constraints for Watertight Point-Cloud Reconstruction},
journal = {Mathematical Modeling and Computing},
year = {2026},
volume = {13},
number = {2},
pages = {598--613},
doi = {10.23939/mmc2026.02.598}
}
EUVIMED