Automated Detection for Alignment of Modular Integrated Construction Modules Using YOLOv9

Modular Integrated Construction (MiC) represents an innovative approach in the construction industry, where building components are manufactured off-site and assembled on-site, offering enhanced quality control and reduced construction time. While computer vision technologies have advanced significantly in construction applications, the precise alignment of MiC modules during installation remains challenging, with traditional manual methods being time-consuming and error prone. This study develops an automated detection system using a modified YOLOv9 architecture to enhance the precision and efficiency of MiC module alignment. The research employed a three-stage methodology comprising literature analysis, data acquisition of 5,000 images, and model development implemented on a GeForce RTX 3050 GPU. The system achieved exceptional performance with a mean Average Precision (mAP50) score of 0.995 and real-time processing capabilities of 3.1 ms per frame at 640×640 resolution. These results demonstrate the system's potential for transforming MiC practices through improved installation precision and reduced human error, contributing significantly to the advancement of construction automation.