Small object detection (SOD) system for comprehensive construction site safety monitoring

Although object detection is essential for recognizing hazardous situations in construction sites where various objects coexist, existing systems fail to ensure real-time accuracy and flexibility in detecting small objects in various scene scales. Therefore, a small object detection (SOD) system was developed based on the YOLOv5 algorithm for comprehensive site monitoring. The proposed SOD simultaneously crops images into multiple segments for small object detection set by the user's desired flexibility while gaining real-time inference in edge computing environments. The SOD outperforms existing systems, especially regarding small object detection accuracy and flexibility for detecting objects of different sizes. The SOD can detect multi-scale objects not initially detected by existing methods (i.e., workers) to large construction equipment without much inference time lost in the edge device. The proposed system facilitates real-time site monitoring by correcting existing system limitations, thereby improving site monitoring and safety management. •Development of a small object detection system for comprehensive site monitoring.•Inference after cropping images into multiple segments by the user's desired flexibility.•Detection of multi-scale objects from small workers to large construction equipment.•Use of edge computing to reduce computational resources for real-time monitoring.•Contribution to the future development of safety management through site monitoring.