Vehicle Clustering Under Surveillance Video

• Main Authors: Guan-Jie Huang, 黃冠傑
• Other Authors: Chun-Rong Huang
• Format: Others
• Language: en_US
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/9474ye

碩士 === 國立中興大學 === 資訊科學與工程學系所 === 101 === In this paper, we propose an automatic vehicle event clustering method for visual surveillance. Firstly, we retrieve trajectories of vehicles as features to represent motion information. To solve the length problem of computing similarity be-tween trajectories, we apply information theory to model trajectories as entropy under variant motions models. Finally, a heterogeneous feature clustering method is proposed to cluster trajectories with similar behaviors. As shown in the experiments, trajectories with similar behaviors are grouped together in the same cluster. In the future, we will apply the method to separate abnormal events from various normal events.