140 likes | 280 Views
Student: Dane Brown 2713985 Supervisor : James Connan Co-Supervisor : Mehrdad Ghaziasgar . SUSPICIOUS ACTIVITY DETECTION. OVERVIEW. INTRODUCTION USER INTERFACE CHANGES DESIGN DECISIONS IMPLEMENTATION TOOLS USED PROJECT PLAN DEMO. INTRODUCTION .
E N D
Student: Dane Brown 2713985 Supervisor : James Connan Co-Supervisor : Mehrdad Ghaziasgar SUSPICIOUS ACTIVITY DETECTION
OVERVIEW • INTRODUCTION • USER INTERFACE CHANGES • DESIGN DECISIONS • IMPLEMENTATION • TOOLS USED • PROJECT PLAN • DEMO
INTRODUCTION • What does the system regard as normal activity • Park car, get out, walk away, get back in, drive away • What does the system regard as suspicious? • Loitering next to a vehicle is suspicious
DESIGN DECISIONS • Haar feature-extraction • Typically the training 1000+ sample frames containing normal activity and suspicious activity • What not haarfeature-extraction? • Performance is good only on a very fast machine • There are simpler and more robust ways to differentiate suspicious and normal behaviour.
IMPLEMENTATION • Gray Scale and Frame differencing
IMPLEMENTATION cont. • Thresholding and Motion History Image (MHI)
IMPLEMENTATION cont. • Blob and movement detection
IMPLEMENTATION cont. • Suspicious activity detected!
TOOLS USED cont. • Kubuntu 10.04 • Opencv with ffmpeg – video manipulation • VirtualDub – open source video editor
REFERENCES • Davis, J. W. (2005). Motion History Image. Retrieved 2010, from The OhiaState University. • Bouakaz, S. (2003). Image Processing and Analysis Reference. Retrieved 2010, from Université Claude Bernard Lyon 1. • Green, B. (2002). Histogram, Thresholding and Image Centroid Tutorial. Retrieved 2010, from Drexel University site.
DEMO • 1.Introduction – normal car driving past • 2. Normal activity – typical drive away • 3. Suspicious – Two men loitering