Forensic video analysis is a popular application used by surveillance and law enforcement for crime detection and prevention. The application uses complex image processing, computer vision, pattern recognition and machine learning algorithms for video processing, which require high computational power and storage that is impossible in a machine with limited resources. Specially that this application is only used occasionally when there is a request for investigation, otherwise dedicated hardware will stay idle. In response to meeting the above challenge we need to find a system that can process large volume of video streams on demand in a scalable and reliable manner. In this paper we introduce and implement Hadoop framework for video processing in a simple cloud environment testbed. Hadoop is an implementation of Mapreduce model, which provide scalable and parallel processing for large volume of data. Our results have shown improved performance and reasonable utilization of cluster resources.

Forensic video analysis is a popular application used by surveillance and law enforcement for crime detection and prevention. The application uses complex image processing, computer vision, pattern recognition and machine learning algorithms for video processing, which require high computationa...