| The measurement of displacement fields on three-dimensional object surfaces plays an important role in mechanical engineering. The displacement data gives insight into a component's response to loading conditions, and can be used to determine the surface strains on the object. As the measured displacement data has noise, a smoothing step is typically performed in the post-processing chain. The smoothing operation should ideally preserve the peak strain, which is directly related to the peak displacement gradient. It might not be the case for isotropic smoothing operations such as box or Gaussian filters that are commonly applied. The work is focused on three promising anisotropic diffusion filters: Perona-Malik PDE, trace-based PDE and curvature-preserving PDE. Both simulation and experimental data are used to do the testing on all the filters. Performance of these filters are evaluated and compared. |
