Civil construction sites have a high-risk environment that exposes workers to mishaps such as falls or accidents arising from equipment failures or not wearing proper safety gear during work. Worker safety is an essential aspect of civil construction sites. Therefore, construction managers must regularly monitor the safety protocols. Using conventional methods for monitoring safety at work can be ineffective and time-consuming. A leading construction company had approached ThinkPalm to support them with an AI-based objection detection system to verify objects in real time to avoid accidents on their construction site.
ThinkPalm used its advanced AI for object detection with the help of computer vision and machine learning to avoid security incidents in civil construction sites. The designed detection system is based on the YOLOv8 model, an advanced object detection algorithm that identifies and localizes objects in real-time. Also, ThinkPalm installed multiple surveillance cameras to provide a feasible solution. The model trained on custom datasets detected personal protective equipment (PPE) items such as hard hats, gloves, boots, safety vests, goggles, and many more.
The model’s architecture included the following elements.
The model uses YOLOv8 to identify workers through the video feeds from the cameras in real-time.
When a worker is detected, the model uses a re-identification (ReID) algorithm to monitor his/her movement.
The AI-based object detection model analyses the worker’s outfit to identify the PPE used. Further, an immediate alert is sent to the safety personnel in case a worker is identified without the essential PPE.
The AI model offers real-time monitoring and notifications during safety incidents. Also, it enabled quick response to avert accidents.
Tools and technologies used for developing the Object Detection Model include:
PyTorch
OpenCV
Python
YOLOv8
The object detection model helped in achieving the following results:
ThinkPalm’s AI-based object detection model exhibited high accuracy in terms of identifying workers and PPE items.
A GUI application was developed to monitor the system and receive notifications from the safety superintendents.
ThinkPalm’s AI-based object detection system for civil construction sites reinforces the endless power of artificial intelligence and machine learning for avoiding accidents and ensuring high standards of worker safety. Using YOLOv8 technology, the model assists in real-time monitoring and sends alerts for faster response before safety incidents happen. Apart from this, the AI-based object detection model can be integrated with existing infrastructure so that it can be used as a fail-safe solution for worker safety compliance monitoring in construction sites.
Contact us and we'll have one of our experts reach out to you and discuss how we can lead your project to success.
