This proposal requests an Expedited Award for Novel Research to investigate the feasibility of using optical range imaging to perform non-contact coordinate measuring tasks. Recently developed optical 3-D sensors have the speed and resolution characteristics to make an optical coordinate measuring machine (CMM) theoretically feasible. Use of the proposed optical range imaging approach should produce a speedup of several orders of magnitude over conventional mechanical contacting CMMs, and provide a density of sampling points sufficient to make "reverse engineering" practical. "Reverse engineering" is the production of a CAD description of a part from an actual physical example of such a part. This capability can lead to a dramatic improvement in productivity. Before attempting to design an optical replacement for a CMM, it is necessary to determine what artifacts, if any, may be due to the optical nature of the sensor, how these artifacts may corrupt the accuracy of a measurement, and how such corruptions may be corrected. The research plan is to develop an accurate model of the camera, the surface being scanned, and the noise in the system; and then to determine how to correct for the resulting noise and distortion. The proposed approach is novel because it is mathematically unique, it will use differential geometry to reconstruct range imagery rather than using traditional Fourier methods. The ideas of using an optical range camera to replace the more traditional coordinate measuring machine is a fairly new idea. The idea of applying reconstruction theory to correct for anomalies in this context appears to be completely novel. If successful, this approach could lead to a large impact on industrial efficiency, because the capability to quickly produce accurate solid model representations from a collection of range images of a subject part is badly needed.
