桥梁和道路病害智能识别大规模数据集

针对桥梁和道路中出现的高强螺栓松动及脱落，钢结构表面锈蚀及涂装脱落，混凝土结构表面破损及露筋，沥青混凝土路面破损及坑槽，伸缩缝破损5类典型表观病害，利用无人机、智能巡检车等智能采集设备，采集不同时间、光照、拍摄角度等环境下的病害图片得到内部数据，通过互联网采集得到外部数据。运用数据清洗、图像增强等数据预处理手段对图片进行筛选、优化，并开展样本精细化标注，最终得到11万条桥梁、道路病害高质量数据集。

This dataset is developed targeting 5 typical apparent structural defects in bridges and roads: high-strength bolt loosening and shedding, steel structure surface corrosion and coating peeling, concrete structure surface damage and exposed rebar, asphalt concrete pavement damage and potholes, and expansion joint damage. Internal data was collected by collecting images of these defects under diverse environmental conditions including different times, lighting scenarios and shooting angles using intelligent acquisition equipment such as drones and intelligent inspection vehicles, while external data was gathered via Internet-based collection. Subsequently, data preprocessing methods including data cleaning and image enhancement were applied to screen and optimize the collected images, and fine-grained annotation of the samples was conducted. Finally, a high-quality dataset consisting of 110,000 entries of bridge and road defect data was established.