Noncontact vision-based deformation measurement of a large-span prestressed concrete rigid-frame bridge under object occlusion

Abstract

Deformation monitoring at cantilever ends of large-span prestressed concrete rigid-frame bridges is vital for ensuring structural safety during symmetrical cantilever casting operations. Traditional contact-based measurement techniques are typically time-consuming and labor-intensive, whereas noncontact vision-based methods offer significant benefits in terms of multipoint deformation measurement and cost-effectiveness . However, their implementation in complex construction environments presents challenges including susceptibility to object occlusion, illumination variations, and reduced detection accuracy for various shaped artificial targets. To address these limitations, this study proposes an enhanced vision-based deformation measurement methodology for large-span prestressed concrete rigid-frame bridges under construction scenarios including partial target occlusion. The proposed methodology initially employs the U²-net, a learning-based background segmentation network, is combined with an incremental image repair network to automatically detect and repair occluded images. Afterward, an enhanced target detection algorithm, which integrates the Convolutional Block Attention Module (CBAM) with the You Only Look Once (YOLO)v8 neural network, is utilized to simultaneously extract deformation data from multiple targets attached to the bridge. The robustness and efficacy of the proposed method have been thoroughly verified through field tests on a prestressed concrete rigid-frame bridge during the symmetrical cantilever casting process. The results demonstrate that our proposed method greatly minimizes deformation anomalies due to object occlusion and efficiently captures deformation from targets of various shapes, such as circular and chessboard patterns. This method demonstrates significant potential for accurately measuring multipoint deformations of large-scale bridges in complex construction environments, thereby providing essential data for bridge safety assessment and construction strategy decision-making.