Ahmedabad: A construction site can have hundreds of workers moving across different floors, sections and work areas, making manual safety checks difficult for supervisors. SafeGuard, a student project at CEPT University, has developed an AI-based safety monitoring system that can detect whether workers at construction sites are wearing helmets and protective vests in real time.“It began as a student project using computer vision and YOLO-based deep learning to detect helmets and safety vests from live CCTV and Wi-Fi camera feeds in real time. The project demonstrated high accuracy,” said Tapan Betai, a faculty member at the Faculty of Technology, CEPT University, and guide for the studio “Industry 4.0 in Future Construction Technology”.YOLO (You Only Look Once) is a high-performance object detection algorithm that spots and labels items in images or video streams in a single pass. Developed by Kirtan Gajjar, SafeGuard recently received the Project Excellence Award.
New ideas on the ground
Dwija Shah, another faculty guide for the studio, said the project was among several that explored the intersection of futuristic technology and conventional construction practices. “The students first got acquainted with on-site realities such as inventory, measurements and the current technologies being used. They then deliberated on how technologies such as IoT, AI and ML could be implemented to develop solutions,” she said.Another project, ‘Wall-E’ by Umang Kerai, helps construction site supervisors ensure that walls have the required level, slant or undulation. The robot is fitted with a levelling tube that can measure walls, particularly during and after plastering, allowing the surface and the exact slant of the wall to be assessed without manual intervention at every stage.Similarly, Vansh Gajjar’s project uses IoT to assess inventory and a robot to draw precise lines for workers at construction sites. Computer vision reduces human intervention in counting materials such as bricks, bars and cement bags, while an Arduino-based system helps the robot navigate uneven construction sites.Most of the studio’s projects were executed at construction sites in the city. Senior officials said some of the projects are at an advanced stage of implementation, and efforts are also underway to secure intellectual property rights.
