Naslov
Classification of 3D shape deviation using feature recognition operating on parameterization control points

Autori
Vučina, Damir ; Ćurković, Milan ; Novković, Tonči

Izvornik
Computers in industry (0166-3615) 65 (2014), 6; 1018-1031

Vrsta, podvrsta i kategorija rada
Radovi u časopisima, članak, znanstveni

Ključne riječi
B-spline surfaces; shape deviation; 3D parameterization; feature detefction; classification; wind turbine blades
(B-spline surfaces; shape deviation; 3D parameterization; feature detection; classification; wind turbine blades)

Sažetak
A numerically efficient expert system for evaluation of 3D shape based on features extracted from the parameterization control points data-set is developed. Reference shapes are stored and periodically compared with current shapes at the level of windowed control-point data grids for the purpose of detection of 3D shape deviation. Classification heuristics for the respective types of deviations which operate on the control point sets rather than the original raw point clouds are developed based on operations of windowing, coordinate transformation, filtering and singular value decomposition (SVD)/ principal component analysis (PCA). The methodology is demonstrated with the cases of detecting and computationally recognizing local impact damage and cavities, narrow gaps or fatigue cracks and surface wear based surface deterioration on a wind turbine blade.

Izvorni jezik
Engleski

Znanstvena područja
Strojarstvo, Temeljne tehničke znanosti

POVEZANOST RADA