ROBUST ESTIMATION OF ROTATION ANGLES FROM IMAGE SEQUENCES USING THE ANNEALING M-ESTIMATOR

A robust method is presented for computing rotation angles of image se quences from a set of corresponding points containing outliers. Assumi ng known rotation axis, a least-squares (LS) solution are derived to c ompute the rotation angle from a clean data set of point correspondenc es. Since clean data is not guaranteed, we introduce a robust solution , based on the M-estimator, to deal with outliers. Then we present an enhanced robust algorithm, called the annealing M-estimator (AM-estima tor), for reliable robust estimation. The AM-estimator has several att ractive advantages over the traditional M-estimator: By definition, th e AM-estimator involves neither scale estimator nor free parameters an d hence avoids instabilities therein. Algorithmically, it uses a deter ministic annealing technique to approximate the global solution regard less of the initialization. Experimental results are presented to comp are the performance of the LS,M-and AM-estimators for the angle estima tion. Experiments show that in the presence of outliers, the M-estimat or outperforms the LS estimator and the AM-estimator outperforms the M -estimator.