Open AccessDezert–Smarandache theory for multiple targets tracking in natural environment
Author(s) -
Wang Yi,
Fang Yingwu,
Zhang Jian,
Li Pengyang
Publication year - 2013
Publication title -
iet computer vision
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.38
H-Index - 37
eISSN - 1751-9640
pISSN - 1751-9632
DOI - 10.1049/iet-cvi.2012.0193
Subject(s) - computer science , tracking (education) , artificial intelligence , basis (linear algebra) , particle filter , computer vision , natural (archaeology) , pattern recognition (psychology) , data mining , machine learning , kalman filter , mathematics , psychology , pedagogy , geometry , archaeology , history
The aim of this article was to investigate multiple targets tracking in natural environment based on Dezert–Smarandache theory (DSmT). On the basis of establishing conflict strategy and combination model, the basic framework and algorithm of fusing multi‐source information were described. The multiple targets tracking platform which embedded location and colour cues into the particle filters (PFs) was developed in the framework of DSmT. Three sets of experiments with comparisons were carried out to validate the suggested tracking approach. Results showed that the conflict strategy and DSmT combination model were available, and the introduced approach exhibited a significantly better performance for dealing with high conflict between evidences than a PF. As a result, the approach was suitable for real‐time video‐based targets tracking, and it had the ability to track interesting targets. Furthermore, the approach can easily be generalised to deal with larger number of targets and additional cues in a complicated environment.