Next: Contours to Splines
Up:
Industrial Inspection and
Previous: Sensing Strategy
An important step in the reverse engineering process is the accurate
description of the real-world object. We generate an 
_1model
from a combination of three types of scene information.
- Two dimensional images and feature contours.
- Stereo vision depth information.
- Touch information from the CMM (still to be implemented.)
By using each sensing method, we are able to gather enough data to
construct the accurate CAD model necessary for reverse engineering. The
two dimensional images provide feature detection that is in turn used by
the stereo system to build feature models. Finally, the CMM drastically
reduces uncertainty through the exploration of the features.
The state machine and the sensors are able to produce a set of data points and
the respective enclosure relationships. Each feature is
constructed in _1independently and the final model is a combination
of these features. This combination is performed using the recursive structure
of the object by forming the corresponding string for that object and generating
the code by parsing this string recursively. The third dimension is retrieved from
the stereo information and the illumination table as described before.
An example for a reconstructed part is shown in figure 2.
This interface is one of the most important modules in this work, since it is the
real link between inspection and reverse engineering. We have chosen _1as
the CAD language since it has very powerful features, in addition to the fact that it has interfaces with
some manufacturing machines, which allows us to actually regenerate a hard copy of
the part. This
will be our next step, so that, the output of our next experiment will be another
part, hopefully identical to the original part.