A Kalman filtering based data fusion for object tracking

Abstract

To solve that single camera has its limitation of field of view, this paper proposed an object tracking method using multiple camera data fusion in image sequences. In this approach, a tracking filter and a multiple-view data fusion algorithm are applied. An estimation structure, called hierarchical estimation, is used to generate local and global estimate and to combine the estimates obtained from each camera views to form a global estimate. The advantage of this approach is the data of one camera view complements that of another camera view in order to obtain better target measurement information and to make more accurate estimates. A set of image sequences from multiple views are applied to evaluate performance. Computer simulation and experimental results indicate that this approach successfully tracks objects and has good estimation.