Detection and tracking process of moving objects in a sequence of digital images with moving background

摘要

A method for detecting and tracking moving objects in digital image sequence having unmoved background by comparing the image sequence with a calculated background image sequence in which a sequence is calculated from the binary object masks whose segments reflect the shapes and positions of the moving objects, and by matching these segments the motion vectors of the moving objects are determined. In addition, a sequence of background images is calculated by spatially selective and temporal recursive averaging the input picture sequence, are not included in the moving objects, however, other changes of the background that are not caused by moving objects are included. Furthermore, a sequence of binary object masks by binarizing the differential image sequence from the input image sequence and the background image sequence is calculated using a threshold whose value is used to control the spatial selectivity in the calculation of the background image sequence and the segments are determined along with their sizes, and center of gravity. Finally, motion vectors of the detected objects are determined by minimizing the above the intersection of the shifted mask segments averaged squared gray value differences, the differences of the centroid vectors corresponding segments are used as starting vectors for the minimization.