Geometric identification of dynamic objects on image sequences

Abstract

As a result of the work, an approach to fast identification of dynamic objects described as points in metric spaces has been developed, the identification process is based on the assumption of the proximity of points corresponding to different states of the same object in a given metric space. The presented approach is based on the comparison of distances between points describing the state of the object. The results of the work are implemented in the form of software for the identification of dynamic objects on sequences of digital images and visualization of such experiments. The approach was tested on synthetic data and on real data for the identification of objects found in the video. Figs.: 4. Refs.: 12 titles.