Using the improved YOLOv11 model to enhance computer vision applications for building crack detection algorithms

Area of Science:

• Civil Engineering
• Computer Vision
• Artificial Intelligence

Background:

• Urbanization increases the need for efficient structural safety monitoring.
• Traditional crack detection methods suffer from low efficiency and high error rates.
• Deep learning, especially YOLO algorithms, shows promise for automated crack detection.

Purpose of the Study:

• To develop an enhanced YOLOv11 model for improved building crack detection accuracy and real-time performance.
• To address limitations in detecting small and complex crack patterns.

Main Methods:

• Introduced C3K2-SG module for complex background crack feature extraction.
• Implemented FPSConv module for multi-scale crack detection.
• Utilized Inner_MPDIoU loss function for enhanced small crack localization.

Main Results:

• Achieved a detection accuracy (mAP@0.5) of 88.6%, a 4.6% improvement over YOLOv11.
• Increased precision by 3.5% and recall by 7.6%.
• Demonstrated superior performance in detecting small targets and complex crack types.

Conclusions:

• The enhanced YOLOv11 model provides an efficient and accurate solution for building crack detection.
• The model shows significant advantages for intelligent inspection and structural safety monitoring.