An Improved Hybrid Area and Feature-based Technique for Tracking of Moving Objects
S A Kianejad
K.N.Toosi University of Technology, Iran
s_a_kianejad@alborz.kntu.ac.ir
Hamid Ebadi
K.N.Toosi University of Technology, Iran
ebadi@kntu.ac.ir
Masood Varshosaz
K.N.Toosi University of Technology, Iran
Varshosazm@kntu.ac.ir
Barat Mojaradi
K.N.Toosi University of Technology, Iran
mojaradi@alborz.kntu.ac.ir
Abstract :
Automatic detection and tracking of moving object is an important problem in many fields such as navigation. This paper proposed a new hybrid area and feature-based image matching for tracking of a moving object in image sequence. Firstly a moving object is detected in sequential frames of video images. Then, geometric corners are extracted with Harris corner detector in each frame and matched with this matching technique. Several sets of data used to test the method and results show its acceptable accuracy and reliability.