Customer is one of India’s leading providers of Intelligent Transport System, Electronic Toll Collection (ETC), enforcement and traffic telematic solutions is managing a TMC (Traffic Management Center) along with law enforcement for watching traffic flows and incidents/accidents on Expressway.
The Customer’s TMC collects traffic footage from over 200 cameras which are monitored real time by equipped staff and their footage is saved continuously to ensure post-analysis if needed. The installed cameras are both Fixed and PTZ (Pan, Tilt and Zoom) and manufactured by Hikvision and Polixel. The facility has 10 TB Storage area networks (SAN) with 8TB provisions for video.
A single camera generates somewhere between 500MB to 1GB worth of video every hour, for 200+ cameras, which roughly equals to 5TB of video footage generated every day. Due to this the company was recycling the storage every 36 hours and needed to add more storage to extend the recycle time from 36 hours to more.
• Achieve at least 50% space saving over existing video without any visual loss in quality of video.
• System should be fast enough to achieve all re-encoding on a single server with support of GPU.
OUR APPROACH :
• Understanding video extensions and formats.
• Understanding performance characteristics of video re encoding on CPU and GPU.
• Developing re-encoding solution – We customized our video compression solution based on standard video format and libraries using NDIVIA cuvid (hardware acceleration for video) support in C#.
• Testing Solution on real data.
Our Video Compression solution was able to achieve significant compression without compromising the visual quality of video increasing the Customer’s TMC’s capacity by 2x without any major IT costs. Compression speed achieved upto 100x using GPUs which greatly reduced cost of hardware and implementation complexity.