Digitizing technique with sequential hypothesis testing for reverse engineering using coordinate measuring machines

Reverse Engineering of unknown components needs an easy to use, efficient and robust digitization method. We propose an overall digitizing strategy. The fundamental problem to determine the minimum number of samples and location of the samples for accurate reconstruction is solved.;We use an efficient sequential sampling strategy to determine number of samples needed to digitize features to a desired confidence level. The location of samples is clearly defined using the center of largest empty circle algorithm which works well with sequential sampling. This method is used to determine the parameters of specific fixed features on the component.;Sculptured surfaces represent a problem to conventional techniques. These surfaces are also effectively digitized using a variation on the sequential method with the computational geometric algorithm. We use a nonic polynomial representation on a triangulation to represent the class of commonly encountered CAD surfaces. The surface fit is refined iteratively until the desired confidence level needed for the reconstruction is reached.;This method is very useful for digitizing using existing conventional coordinate measuring machines and probes. The method represents a new paradigm for digitizing when compared to the offline, dense sampling and post process fitting that are done today.