The Improvement of SURF Point Tracking Continuity under the Dynamic Environments by using KLT Prediction Scheme

• Main Authors: Tai, Chun-Feng, 戴群鋒
• Other Authors: Lien, Cheng-Chang
• Format: Others
• Language: zh-TW
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/18156867277415095465

碩士 === 中華大學 === 資訊工程學系碩士班 === 103 === Object detection and tracking approaches are crucial to the video surveillance and ecological investigation. In these applications, PTZ cameras are often applied to extract the close-up views by controlling the pan, tilt, and zoom operations in a large area. Hence, this study aims at the developing of a robust and automatic object detection and tracking with single PTZ camera. Conventional object detection and tracking with single PTZ camera often apply feature-based traditional target tracking methods. However, these method can fail under the dynamic outdoor environments with serious variations of background and illuminations. In this study, we try to integrate the prediction scheme of KLT algorithm and SURF-based target tracking scheme to develop a robust object tracking with single PTZ camera under the dynamic outdoor environments. The experimental results show that the proposed method outperform the conventional methods.