Pioneering Safety Innovations for Employee Transport and Mining Vehicles
Vedanta, one of India’s leading mining conglomerates, operates extensive coal, iron, and aluminium mining and processing facilities across the nation. As a key player in the mining industry, the company places paramount importance on the safety of its employees and assets, a responsibility diligently upheld by its management.
The safety division of IOK sought to enforce Vedanta’s safety standards within the mine. They required an AI camera system capable of alerting both the driver and management in instances of driver fatigue, distraction, mobile phone usage, and seat belt non-compliance. Additionally, they expressed interest in monitoring the vehicle’s cabin.
The same AI camera system was to be installed in passenger transport vehicles to ensure the safety of employees commuting to and from the mine, further emphasizing the company’s commitment to safety.
We recommended the SC-AI04, our premier camera solution, equipped with strategic AI capabilities:
- Cabin Camera (DSM): Utilizing AI algorithms, the driver’s behaviour was monitored for signs of fatigue, distraction, mobile phone usage, and seat belt compliance. Full HD cameras with colour were introduced for cabin view.
- Front-Facing Camera (ADAS): The AI system monitored vehicles or pedestrians in front of the vehicle, providing real-time alerts to the driver if a collision threat was detected.
- Cloud-Based Video Surveillance Software: This allowed management to receive real-time alerts from vehicles in their fleet. The ability to track and live stream video also facilitated easy fleet monitoring.
- Driver Identification: Each camera solution could store up to 10 drivers’ faces and would run an identification scan when the vehicle was started and every 5 minutes thereafter to ensure the correct employee was driving the vehicle.
The SC-AI04 integrated seamlessly with the vehicle’s power system via our hardwire kit, boasting an impressive 94% accuracy rate for reliable performance.
Connectivity and Alarms: The SC-AI04 connected to the 4G mobile network for continuous data transmission. Alarms and alerts were promptly relayed to management for review. However, due to the mine’s remote location, 4G connectivity was not always available. Our solution stored all relevant alarms and GPS data until the device connected to the mobile network, at which point all stored alarms were uploaded for management review.
Seat Belt AI: We successfully trained our AI to detect if the driver was correctly wearing his seat belt. This task was challenging due to the reflective safety jackets worn by the mine drivers. We had to train the AI to detect only the seat belt and ignore the reflection from the safety jackets.
Adjusting the Sensitivity of our ADAS AI: Different regions of India required us to modify or recalibrate our AI to prevent excessive false positive alarms. For instance, the pedestrian collision alarm would only be triggered if a pedestrian was detected on the road in cities with pavements or walkways. In rural areas, pedestrians often walk on the roads, leading to numerous alarms. We adjusted the sensitivity of our AI and narrowed the detection field for potential collision risks, reducing the number of alarms and facilitating management’s review of driver performance based on exceptions.