基于卷积神经网络的路桥设施缺陷检测研究

道路桥梁在建成投入使用阶段,会受自然、人为等各类因素影响,形成各种桥梁设施病害,如裂缝、坑凼等,对路桥的结构安全、行车行人安全、环保美观都产生一定的不良影响。传统的路桥病害或缺陷检测主要以肉眼观察为主,该方法在检查体量、检查效率、检查精度以及成本控制方面都存在很大的劣势。基于卷积神经网络的人工智能图片识别系统,可在很大程度上代替人工检测,实现路桥病害自动化检测,为路桥管养工作人员提升工作效率和科学制定养护决策提供帮助。

When the road and bridge are completed and put into use,it will be affected by various factors such as nature and man-made,resulting in various bridge facilities diseases,such as cracks and pits,which have a certain adverse impact on the structural safety,traffic and pedestrian safety,environmental protection and beauty of the road and bridge.The traditional detection of road and bridge diseases or defects is mainly based on visual observation.This method has great disadvantages in inspection volume,inspection efficiency,inspection accuracy and cost control.The artificial intelligence image recognition system based on convolution neural network can replace manual detection to a great extent,realize the automatic detection of road and bridge diseases,and provide help for road and bridge management and maintenance staff to improve work efficiency and make maintenance decisions scientifically.