CMA

35 facility through CCTV, thermal and drone camera feeds; its analysis up-till now had mostly been manual. There was need to transform business environments by using real time video analytics so as to enhance process efficiency, understand employee behavior and enhance safety in diverse environment. To address this, the team thought of implementing VEDA, a Video Analytics and Decisions Alerts platform which had use- cases in varied fields as process efficiency, understanding consumer behavior, adherence to safety norms, visual quality inspection etc. It had already been pilot tested at Hindalco in July 2019, a subsidiary of Aditya Birla Group specializing in manufacturing of aluminum and copper. The platform deployed custom built AI/ML models for analyzing video feeds to provide real-time, actionable insights and alerts and had the capacity to process multiple streams in parallel. 24*7 automated analysis of feeds using AI in an on-premise server was carried out, which raised alarms through real-time voice, text and e-mail alerts in case of any deviation, while enabling monitoring of safety trends through the dashboard. Commonly termed as ‘Suraksha’ (Safety) feature list, the most commonly tracked indicators were safety gear detection, object detection, person detection, crowd formation and movement of people in unsafe zones, vehicle detection, detection of untoward incidents as fire, multi- channel alerts etc. VEDA was also instrumental in identifying safety violations resulting due to non-adherence of arc flash suit in substation or while working with hot materials, and provided real time intervention through audio-visual alarm. VEDA through its advanced features and capabilities not only proved to be a game changer in ensuring adherence to safety guidelines during COVID but also delivered intangible amount of cost benefits. In the first phase, 10 plants were made VEDA enabled. The deployment of VEDA delivered various tangible and intangible benefits and provided real-time actionable insights to work on. VEDA was also instrumental in remote monitoring of the production process and sites, and enabled supervisors and managers to remotely manage their teams to ensure seamless operations. It also helped in identifying procedural violations and predominant violators.