The paper deals with the registration and modelling of terrestrial laser point clouds. For both problems a non parametric regression is suitably exploited, whose unknowns are the function values and the partial derivatives of a second order Taylor’s expansion estimated for a certain number of surface points. These allow to directly estimate local curvatures, namely Gaussian, mean and principal values. Relating to the registration problem, tie points are automatically detected from point clusters having extreme Gaussian curvature values. The centroids of such clusters generate a vertexes configuration: the point to point correspondences are automatically defined by the analysis of the respective adjacency matrices. For these sets of pairs, the pre-alignment roto-translation parameters are computed by a SVD algorithm, while the final alignment is executed by an ICP method. The paper further proposes a method to directly detect the discontinuities (segmentation) and to successively estimate the parameters for each recognized surface (classification). For both goals, the algorithm exploits again the curvature values: the discontinuity contours are characterized by points having mean curvature greater than a threshold, while classification is performed by a cluster analysis of points having homogeneous curvature values. Some numerical examples show the proper applicability of the proposed method for coarse and fine registration of different scans, for edge detection, and for surface primitives classification.

Automatic non parametric procedures for the processing of terrestrial laser point clouds

BEINAT, Alberto;CROSILLA, Fabio;VISINTINI, Domenico;
2007-01-01

Abstract

The paper deals with the registration and modelling of terrestrial laser point clouds. For both problems a non parametric regression is suitably exploited, whose unknowns are the function values and the partial derivatives of a second order Taylor’s expansion estimated for a certain number of surface points. These allow to directly estimate local curvatures, namely Gaussian, mean and principal values. Relating to the registration problem, tie points are automatically detected from point clusters having extreme Gaussian curvature values. The centroids of such clusters generate a vertexes configuration: the point to point correspondences are automatically defined by the analysis of the respective adjacency matrices. For these sets of pairs, the pre-alignment roto-translation parameters are computed by a SVD algorithm, while the final alignment is executed by an ICP method. The paper further proposes a method to directly detect the discontinuities (segmentation) and to successively estimate the parameters for each recognized surface (classification). For both goals, the algorithm exploits again the curvature values: the discontinuity contours are characterized by points having mean curvature greater than a threshold, while classification is performed by a cluster analysis of points having homogeneous curvature values. Some numerical examples show the proper applicability of the proposed method for coarse and fine registration of different scans, for edge detection, and for surface primitives classification.
File in questo prodotto:
File Dimensione Formato  
PIA07_Beinat_et_al.pdf

non disponibili

Tipologia: Altro materiale allegato
Licenza: Non pubblico
Dimensione 1.33 MB
Formato Adobe PDF
1.33 MB Adobe PDF   Visualizza/Apri   Richiedi una copia
Automatic Non Parametric Procedures for Terrestrial Laser Point Clouds Processing.pdf

non disponibili

Tipologia: Altro materiale allegato
Licenza: Non pubblico
Dimensione 802.39 kB
Formato Adobe PDF
802.39 kB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11390/879946
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact