Robust 3D-3D pose estimation

The paper focuses on the robust 30-30 single and multiple pose estimation. The robust 30-30 single pose estimation was formulated aa a general regression in terms of a contaminated Gaussian error noise model (Haralick et al., 1989). The robust 30 -30 multiple pose estimation appears much more dificult. In the paper, the problm is formulated as a series of general regressions involving a successively decreasing data size, with each regression relating to one particular pose of interest. Since first few regressions may carry a highly contaminated Gaussian error noise model, a highly robust estimator called b y the MFestimator (Zhuang e t al., 1992) is used to 8 0 h e each regression for each desired pose estimation. Computer experiments with both real imagery and simulated data are conducted and result3 are promising.