Real-Time video Analytics Engine is a Platform that utilizes advanced image processing algorithms to turn video into actionable intelligence. Large Scale Video Processing is the greatest challenge holding a great potential for Analytics. Real-Time Video Analytics can be used for Monitoring Traffic Control, Retail Store Monitoring, Surveillance and Security.
XenonStack's Real-Time Video Analytics Platform is real-time, low cost, accurate analysis of live videos using Open Source Big Data Technologies including OpenCV, Apache Kafka, Apache Spark, HDFS, Tensor Flow, and Amazon S3.
Traditional Computer Vision System suffers from a limitation, that a server along with CV library collects and process the data at the same time. A failure in the server causes loss of Streaming Video Data.
Detecting a node failure and switching the processing to another node may result in fragmentation of data.
Major challenge relies on running an image processing algorithms in a dynamic environment.
Thus managing and efficiently analysing this data brings us with a challenge to build a system which can eliminate the addressed problems.
XenonStack Team came up with a solution to perform real-time video analytics backed by open source big data technologies.
The system is divided into three main components -
Video stream collector
Stream data buffer
Video stream processor
Talk to Experts for Assessment on Infrastructure Automation,
DevOps Intelligence, Big Data Engineering and Decision Science