Target vehicle tracking under occlusion for video surveillance
Loading...
Date
2018-08
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
G.B. Pant University of Agriculture and Technology, Pantnagar - 263145 (Uttarakhand)
Abstract
With the rapid development corporate infrastructure, industrial establishments, educational institutes and residential premises are facing an increasing need to enhance their security. In this thesis, a hybrid target vehicle tracking algorithm for smart video surveillance is proposed. It aims to track an unidentified target vehicle’s motion in a secured campus even in case of occlusion (if any) and to detect if it exhibits any suspicious activity. The main contents of this thesis are as follows: Aiming at the tracking of a target vehicle and handling occlusions if any through Kalman Filter assisted occlusion handling technique, performance evaluation under different noise and illumination levels and finally suspicious activity detection technique for the tracked vehicle. The algorithm works through two periods namely tracking period (no occlusion) and detection period (in case of occlusion), thus depicting its hybrid nature. Kanade-Lucas-Tomasi (KLT) feature tracker governs the operation of algorithm during the tracking period, whereas, a Cascaded Object Detector (COD) of weak classifiers, specially trained on a large database of cars, governs the operation during detection period or occlusion with the assistance of Kalman Filter. The motion analyses and suspicious activity detection capabilities have been discussed separately. The algorithm’s tracking efficiency has been tested on six different tracking scenarios of increasing complexity in real time. Performance evaluation results under high noise and low illumination show that the tracking algorithm has good robustness. All tests have been conducted on the MATLAB platform. The validity and practicality of the algorithm are verified.
Description
Keywords
null