Abstract

Bridge defect record is an important criterion for maintenance, so detection is also an essential task to maintain the health of highway bridge structure. At present, the mainstream inspection method for bridge appearance in China is manual visual inspection combined with simple tools, which has the disadvantages of high labor intensity and low automation. To solve these problems, this paper has proposed a method for detecting apparent defects of concrete bridges based on target detection and depth of field image. First of all, YOLOv4 is used to identify and locate the common apparent defects of concrete bridges, such as spalling, exposed reinforcement, voids, cavities and pits. Then, the digital image processing method is used to segment the accurate defect areas of bridges. Finally, parameters such as length and width of defect areas are extracted based on the depth of field images. The experimental results show that the mAP values of the proposed apparent defect identification method amount to 70.3%, and the error rate of parameters extraction such as length and width of defect area is less than 20%.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call