摘要
为解决传统桥梁表观病害检测方法实施效率低、安全风险高和劳动强度大等问题,自主研发了新型桥梁智能检测机器人系统,介绍了桥梁智能检测机器人的总体架构和现场工作流程;进一步,提出了基于特征的影像拼接方法和基于迁移学习的改进型YOLO_v3桥梁表观病害智能识别算法。工程应用结果表明,研发的桥梁智能检测机器人系统集成了机械控制模块、数据采集模块和数据处理模块,实现了桥梁检测的自动化、智能化和信息化。基于迁移学习和数据增强两种方法,提出了改性型YOLO_v3桥梁裂缝智能识别算法,在不损失检测效率的情况下,具有良好的检测精度和抵抗噪声能力。新型桥梁智能检测机器人系统为桥梁工程的智慧化检测提供了硬件和软件支撑。
In order to solve the problems of low efficiency,high safety risks,and high labor intensity in using traditional methods to detect bridge apparent defects,a new type of intelligent bridge detection robot system has been independently developed,and the overall architecture and on-site workflow of the intelligent bridge detection robot are introduced.Furthermore,a feature-based image stitching method and an improved YOLO_v3 intelligent identification algorithm for bridge apparent defects based on transfer learning are proposed.The engineering application results show that the developed bridge intelligent detection robot system integrates mechanical control modules,data acquisition modules,and data processing modules.Through the robot system,the automation,intelligence,and informatization of bridge detection have been achieved.Based on transfer learning and data enhancement,an intelligent crack identification algorithm for modified YOLO_v3 bridge is proposed,which has good detection accuracy and noise resistance without losing detection efficiency.The new type of intelligent bridge detection robot system provides hardware and software support for intelligent detection of bridge engineering.
作者
孙晓立
杜永潇
万智
杨军
赵亚宇
SUN Xiao-li;DU Yong-xiao;WAN Zhi;YANG Jun;ZHAO Ya-yu(Guangzhou Municipal Engineering Testing Co.Ltd.,Guangzhou 510520,China;Hunan Bridge-health Intelligent Technology Co.Ltd.,Changsha 410021,China)
出处
《公路》
北大核心
2024年第8期404-411,共8页
Highway
基金
中国博士后科学基金,项目编号2022M720898
广东省住房和城乡建设厅2023年科技计划项目,项目编号2023-K62-351914
广州市建筑集团有限公司科技计划项目,项目编号:[2023]-KJ012,[2022]-KJ023,[2021]-KJ010。
