Abnormality Detection and Traffic FlowMeasurement using a Hybrid Scheme of SIFT inDistributed Multi camera intersection Monitoring

This paper presents an unsupervised abnormalitydetection method using a multi camera system with clustering inreal time. Among the most important research in intelligenttransportation systems (ITS), automatically intersection flowmonitoring is one of the critical and challenging tasks. Theproposed work addresses anomaly detection by means oftrajectory analysis based on single support vector machine(single-SVM) clustering. The main problem associated withvehicle tracking is the occlusion effect. Using multiple views ofcameras for producing a uniform tracking configuration is moresuitable for vehicle’s behaviour extraction. We use a hybridscheme of scale invariant feature transform (SIFT) to detect andrecognize vehicles in multi view system, so behaviour extractionis done more accurately and conveniently. The main focus of thispaper is to extract traffic flows which assists in regulating trafficlights based on smart cameras.

کلیدواژه ها:

Unsupervised abnormality detection ، Multi camera ، Trajectory clustering ، Invariant feature transform ، Intersectionmonitoring ، behaviour extraction