PROCESS FOR ASSESSING MOVEMENTS RECURRING IN TIME AND FOR TRACKING MOVING OBJECTS IN THE ANALYSIS OF TIME SEQUENCES OF DIGITAL IMAGES

摘要

It discloses a method for motion estimation recursive in time and to the tracking of objects moving through the analysis of digital image sequences, with which there corresponds (DS) for moving segments in the images ( I) of a succession of digital images of the binary segment functions (S) and are measured movement parameters (Z) of these segments in motion by a maximization (SM) of an adaptation function (C). These movement parameters (Z) measured are used in a recursive scheme in time for correction (MU) of predicted state values ​​(X), which are calculated (TU, SEL) in turn from the values ​​of corrected state (Y) of the instant which precedes or which have been initialized (INIT) to the top of the execution of the method. The difficulties inherent in the movement of the evaluation known methods are overcome and thus the accuracy of the motion estimation is increased with a concomitant reduction in cost.