Selective sampling in the reverse engineering of free form surfaces

Focusing the attention on the first step of the reverse engineering cycle, the digitisation, it is possible to denote that the major part of the systems operate with punctual information measured with different technical approaches, light or laser beam or mechanical pointer. Anyway, even if the measuring technology employed could be different, these points, reached on a working plane, always come from a discretisation process. In fact working in a cartesian space and considering the orthogonal plane (XY) to the measuring direction (Z), the procedure divides the entire surface of the object in a grid deciding two step sizes and consequently the number of reachable points during the acquisition. So considering that the major part of the systems could employ one only step size value for the entire acquisition phase its choice became a big deal. From one side if the value is too small too many redundant information could occur, causing the unjustified growing up of the computational costs of the entire reverse engineering cycle. From the other side if the value is too large, there is the possibility to have few information about the real shape, causing in this way an inefficient reconstruction of the final mathematical model of the object. In order to solve this ambiguity the idea of the paper, working with a piezoelectric reverse engineering system, is to develop an operative methodology able to selectively subdivide the entire working plane in different zones in relation with the different step size that the specific object regions need. Working with an industrial application the procedure has been experimented in order to obtain an operative validation.