We are living a fast-paced life where we often don’t pause to see how far we have come as a species. We are capable of dealing with all sorts of difficulties, adapt, and overcome them. We all are witnessing the COVID-19 impact and wouldn't deny the immense efforts we all are putting at our ends to deal with the situation at hand in terms of living, working, planning, and executing things. 

We have used our IQ (intelligence quotient) as well as EQ (emotional quotient) to come up with solutions that let the show go on and sail through these times. In just one-and-a-half years, we have transitioned drastically that otherwise could have taken years together to come true. Organizations are using technologies to keep businesses afloat and keep the cycle moving. 

However, in these sensitive times, safety is the key. Without a doubt, we know technologies like AI have been the key to this transition and real-time safety. Manual monitoring of safety practices can lead to people becoming complacent, and no one can be held responsible for detrition of safety standards in an area in a manufacturing unit. 

As a solution, a wonderful example of a technology-aided tool to take care of people’s safety and make the operation go on is VEDA. VEDA (Video Enabled Decision and Alerts) is an AI-enabled video analytics system that helps create a safe environment in Aditya Birla Group factories and offices. This product uses computer vision techniques to analyze videos and provide alerts and insights to make real-time decisions.

“VEDA, an AI platform and a first-of-its-kind proprietary product, uses computer vision techniques to provide alerts and insights to take real-time decisions,” explains Deep Thomas, Group Chief Data and Analytics Officer, Aditya Birla Group.

GDNA (The group data and analytics cell) is a central function in Aditya Birla Group (the US $48.3 Billion revenue) created to provide cutting edge AI solutions to businesses and take these AI products to a larger audience. The team comprises more than 60 data scientists, product specialists, and business engagement managers from varied backgrounds, including manufacturing, finance, and retail Areas of focus include Safety, Quality, Logistics, Procurement, and Energy.

“VEDA was an idea that emerged from a hackathon in 2019. Six months later it was successfully deployed at a plant in Taloja. During the lockdown, we added new features to the platform that enhanced worker safety in the post-COVID world,” adds Naveen Xavier, Vice President, Group Data & Analytics, Aditya Birla Group.

VEDA has been successfully deployed at 3 plants of UltraTech, 7 plants of Hindalco and 4 plants of Grasim. According to company sources, “manufacturing plants lose a big amount of man-hours due to accidents and other safety-related issues. In the 2019-20 period for Hindalco, approximately 26 man-days were lost, and over 600 minor injuries and accidents were reported.” 

Safety accidents, in most cases, are avoidable if adequate safety training is provided, safety procedures are followed, and proper safety equipment is worn. With VEDA, the Aditya Birla group is trying to change fundamental behaviour by combining human behaviour with computer vision technology. 

These are not just accidents that are the sole concern for the organization; it is the amount of work that goes into identifying the root cause and updating systems so that accidents do not occur again is a time-consuming affair. Even after the process completion, the real impact numbers cannot be accurately estimated.

The underlying idea is to ensure the safety of workers and employees in a manufacturing plant by processing multiple video feeds in real-time. The system takes input in the form of multiple raw video feeds captured through CCTV, which is then analyzed 24x7 via an on-premise server using AI. The output provides safety trends on a dashboard that can send alerts through real-time voice, text, or email.

The platform is scalable and is built for quick deployment. The advanced capabilities of the system can cater to the needs of the manufacturing units and offices post COVID. 

VEDA is embedded with a social distancing monitoring module that detects workers in a frame and then calculates the distance between each worker. When any video feed records recurring violations lower than the threshold distance value for a minimum specified duration, an alert is triggered. The platform works on a transfer learning approach and fine-tuned MobileNetV2 model for person detection using the TensorFlow framework. The accuracy of the distance is based on the transformation matrix, which in turn depends on the accurate calibration of ground points. Besides temperature monitoring and face mask detection, a robust social distancing measurement algorithm has also been built into VEDA.

The platform has a competitive advantage due to factors like cost-effectiveness, scalability, and flexibility. In addition, it can work both on-premises or on the cloud, and owing to its modular nature, new models can be trained independently.

“VEDA is a powerful platform. We used it successfully for changing people’s behaviour, attitude and habits in the manufacturing plant,” says Vaishali Surawar, Unit Head, Hindalco Industries Ltd at Taloja. 

The implementation of video analytics will aid transformation in a business environment by increasing process efficiencies, understanding customer behaviour, and enhancing safety in diverse environments while considerably reducing manual intervention. In addition, VEDA delivers an intangible amount of cost benefits by monitoring the fire, oil spillage, crowd formation, people movement in unsafe zones, and use of safety gear by workers on shop floors.

The solution faced several challenges such as lack of data for modelling, annotating data, streamlining a feedback mechanism for model governance, and change management. 

Post-deployment, the system needs human intervention or surveillance as only a human can detect and differentiate between an erroneous and a genuine one. 

VEDA has two primary segments Suraksha and COVID- 19. Suraksha aims to improve productivity improvement, audit process compliance, reduction in equipment downtime, and reduced accidents. Covid 19 segment is ensuring adherence to guidelines and protocols by detecting fever and monitoring social distancing.