Storage and Processing Of Streaming Videos For Road Traffic And M Learning Applications

Abstract

Rapid growth in various technologies has led to drastic change in video technology. These video systems to be intelligent use Video Analytics (V.A). Video analytics aims to extract useful information or occurrence of any event from a video automatically. Video analytics have various applications in areas like video surveillance, traffic monitoring, education, health care etc. However, video analytics faces lot of challenges while developing efficient video analytics system for a specific domain. Some of the challenges faced by VA system in video surveillance system is handling continuously growing video data captured from numerous cameras deployed for monitoring purpose, searching a video clip from huge, stored video data when an event is triggered, streaming of lecture content from mobile phone, storage, and retrieval of lecture videos in mobile phones. newlineFirst phase of our research aims at addressing the first issue of handling continuously increasing gigantic video data captured from different cameras installed for surveillance. Storing such a growing gigantic data is a challenge. When an event occurs then manually examining the entire video data becomes tiresome as well as challenging therefore, there is a need of such a system where the system can automatically fetch the required data. Hence in this phase a video system is proposed which, performs function like handling, storage, and analytics of extensively accumulated multi streams video data captured from various cameras deployed for monitoring using Hadoop. Hadoop is a freely available software framework designed for distributed processing and distributed storage and is built on low-cost computers. Proposed algorithm not only processes but also extracts data stored in Hadoop Distributed File System (HDFS) when any event triggers. Computation time is reduced by varying number of cluster nodes and reducers (workers) performing the task. newlineSecond phase of our research deals with the challenges faced in traffic monitoring system for identifying traffic congest