A robust method for multi-view 3D data stitching based on pasted marked points

Hua Luo, Ke Zhang, Na Yang, Minghu Tan, Jingyu Wang

Research output: Contribution to journalArticlepeer-review

3 Scopus citations

Abstract

High-precision three-dimensional (3D) measurement techniques commonly perform 3D point cloud data stitching using pasted marked points. However, pasted marked points exhibit a high degree of randomness and are susceptible to mismatching due to their structural similarity. Therefore, this paper presents a robust method for multi-view 3D data stitching based on pasted marked points (RMSM). First, potential matched marked point pairs (MMPP) are identified according to spatial feature invariance constraints, and a matching strength calculation model is established to find the fiducial point pairs. Then, a quick public MMPP searching algorithm using the congruent triangle search method is proposed. Thus, the problem of mismatching caused by structurally similar marked points can be eliminated, and robust stitching of 3D data from different views can be achieved. In simulation and measurement experiments, the RMSM demonstrates superior accuracy compared to alternative methods and holds promise for application in various multi-view 3D data stitching methodologies.

Original languageEnglish
Article number114364
JournalMeasurement: Journal of the International Measurement Confederation
Volume228
DOIs
StatePublished - 31 Mar 2024

Keywords

  • 3D point cloud data stitching
  • Congruent triangles
  • Matching strength
  • Public marked points

Fingerprint

Dive into the research topics of 'A robust method for multi-view 3D data stitching based on pasted marked points'. Together they form a unique fingerprint.

Cite this