3-DIMENSIONAL ATTITUDE ESTIMATION AND TRACKING WITH LINEAR NORMALIZEDOPTIMAL FILTERING

We propose a 3-D target attitude estimation method based on linear opt imal filtering and prediction algorithms. Through the use of correlati on filters, the estimations of the 3-D attitude and scale of the targe t enable us to predict the temporal trajectory of the target and to pe rform a robust tracking. We propose a strategy to maintain a low numbe r of correlations and thus to reduce the computational burden and the processing time. We demonstrate the efficiency of this approach in the case of optimal trade-off filters and sequences of noisy images. (C) 1997 Society of Photo-Optical Instrumentation Engineers.