Severity classification of abnormal traffic events at intersections

The purpose of this work is to investigate the severity characteristics of abnormal events at intersections by using video processing techniques and statistical deviation analysis methods. In order to detect the abnormal events, trajectory of normal vehicle motions are clustered and common route models are learned by Continuous Hidden Markov Model. In the second part, the abnormal spatio-temporal deviations are detected by extracting partial vehicle motion observations using Maximum Likelihood. Next, the severity definition and classification is done for abnormal events using k-Nearest Neighborhood and Support Vector Machines methods. The two-class event classifier is built to classify abnormal observations into one of the low or high severe event classes. The results indicate that abnormal events can be detected and represented by likelihood probabilities, and depending on these probabilities, severity analysis can be done successfully.