| 摘要 |
A method for 3D reconstruction of an object based on back-scattered and sensed signals, including: generating, from the sensed signals, 3D points to which their back-scattering intensity is respectively assigned, which form a set A of reconstructed data, starting from A, extracting a set B of data, whose points are located within a volume containing the object, as a function of volume characteristics F2, starting from B, extracting a set C of data characterizing the external surface of the object, the surface having regions with missing parts, depending on an extraction criterion, based on C, filling in the regions with missing parts by generation of a three-dimensional surface so as to obtain a set D of completed data of the object, without having to use an external database, and identifying the object based on D. |
| 主权项 |
1. A method for three-dimensional (3D) reconstruction of an object in a scene exposed to a wave, based on signals back-scattered by the scene and sensed, the method comprises:
generating, from the sensed signals, a set of 3D points to which their back-scattering intensity by the scene is respectively assigned, the 3D points associated with their intensity being denoted set A of reconstructed data, starting from the set A of reconstructed data, extracting a set B of data, whose points are located within a volume containing the object, as a function of volume characteristics F2, further comprises the following steps: starting from the set B, extracting a set C of data characterizing an external surface of the object, the external surface having regions with missing parts, depending on an extraction criterion F1, filling in the regions, based on the data of C, with missing parts by generating a three-dimensional surface so as to obtain a set D of completed data of the object, without having to use an external database, the filling step comprising the following sub-steps:
normalizing the data of C characterizing the external surface to obtain a set D12 of the normalized data of C,extracting, based on the data of D12, relevant data denoted attractors and weight them, as a function of extraction and weighting criteria F75, to obtain a set D34 of weighted attractors,selecting significant attractors from amongst data of the set D34 as a function of a selection threshold F74,reinforcing weight of the significant attractors as a function of spatially close data of C to obtain a set D56 of reinforced significant attractors,generating a three-dimensional surface D based on the data of D12 and D56, comprising:
sending the data of D56 to several processes for 3D-surface generation, by means of a control module for choosing the 3D-surface generation processes among minimum energy search type, stochastic relaxation type, Bayesian regulation type, wavelet analysis type, approximation by generalized regression type, optimization with constraints type, or basic radial function type process,choosing a best result or combine several results supplied by the 3D-surface generation processes to obtain the three-dimensional surface D by means of a combiner, and identifying the object based on the three-dimensional surface D. |
