An impulsive noise reduction agent for rigid body motion data using B-spline wavelets

A B-spline wavelet-based agent is proposed in this paper to remove impulsive noise embedded in noisy rigid body motion data. The motion of a rigid body consists of translation and orientation; the former is described by a space curve in three-dimensional Euclidean space, whereas the latter is represented by a curve in the unit quaternion space. Rigid body motion data acquired from electronic measurement devices usually contain noise of various patterns and magnitudes, and may contain structured noise such as impulsive spikes. In order to remove such impulsive noise from noisy motion data, the noisy motion data are decomposed using multiresolution analysis, and the noise components are identified as coefficients of high magnitude. By smoothing these high-magnitude coefficients, a smoother representation of noisy motion data is achieved.