雅萌和日立4000哪个好:那位能帮我翻译下文

3. Experimental 3.1. The acquisition of the views and the multi-view alignment The 1953 Ferrari 250MM is shown in Fig. 2. To perform the measurements, the car was placed on an overhead-travelling crane in the workshop. Since OPL-3D is portable, and does not need recalibration when it is moved around the object, the tripod was mounted on wheels to facilitate the movements around the car. Fig. 3shows an example of projection of a pattern, out of the set of 11 patterns that belong to the Gray Code-Phase Shift sequence, performed as a preliminary test of the colour calibration of the system. Due to the high number of views to be taken, the IMAlign module has been used. The digitisation of the Ferrari 250MM showed up to be critical under a number of aspects. These are (i) the ARTICLE IN PRESS G. Sansoni, F. Docchio / Robotics and Computer-Integrated Manufacturing 20 (2004) 362 359–367 high regularity of the surfaces without edges to be used as ‘anchoring’ points to align the different views, (ii) the glossy red paint, and (iii) the large overall dimensions. The measurement strategy applied to overcome all the above critical aspects is described in the following. 3.1.1. Use of physical markers IMAlign is based on a semi-automatic procedure, which allows the alignment of each ‘floating’ view with respect to those already ‘locked’ together in a common reference system. The basic mechanism requires the selection of at least three fiduciary points in each of the views. In the experimental case considered here, it was observed that, when the selection of the points was performed in correspondence with the edges of the surface, the matching was very fast and accurate (i.e., the error introduced by the alignment was lower than the uncertainty of the measurement). However, when regular portions of surfaces had to be aligned, the computation time dramatically increased, the alignment was inaccurate, and it very often failed.