提供百度云盘地址下载

数据集格式：Pascal VOC格式+YOLO格式(不包含分割路径的txt文件，仅仅包含jpg图片以及对应的VOC格式xml文件和yolo格式txt文件)

图片数量(jpg文件个数)：9398

标注数量(xml文件个数)：9398

标注数量(txt文件个数)：9398

标注类别数：18

所在github仓库：firc-dataset

标注类别名称(注意yolo格式类别顺序不和这个对应，而以labels文件夹classes.txt为准):["broken_drain","concrete_leaching","crack","deep_crack","disaggregation_of_concrete","efflorescence","embankment_erosion","expansion_joint","exposed_reinforcement_steel","faunal_biodegradation","fire_damaged","formwork_niche","guardrail_damaged","infiltration","pothole_asphalt","rutting_pavement_failure","staining","steel_corrosion"]

每个类别标注的框数：

broken_drain (排水管损坏) 框数 = 201

concrete_leaching (混凝土渗析) 框数 = 6762

crack (裂缝) 框数 = 796

deep_crack (深裂缝) 框数 = 183

disaggregation_of_concrete (混凝土离析) 框数 = 1681

efflorescence (泛碱) 框数 = 5466

embankment_erosion (堤岸侵蚀) 框数 = 464

expansion_joint (伸缩缝) 框数 = 462

exposed_reinforcement_steel (钢筋裸露) 框数 = 3755

faunal_biodegradation (生物降解) 框数 = 1802

fire_damaged (火灾损坏) 框数 = 83

formwork_niche (模板孔洞) 框数 = 993

guardrail_damaged (护栏损坏) 框数 = 415

infiltration (渗漏) 框数 = 1807

pothole_asphalt (沥青坑洞) 框数 = 683

rutting_pavement_failure (路面车辙破坏) 框数 = 977

staining (污渍) 框数 = 5198

steel_corrosion (钢筋锈蚀) 框数 = 1249

总框数：32977

图片分辨率：640x640

使用标注工具：labelImg

标注规则：对类别进行画矩形框

重要说明：暂无

特别声明：本数据集不对训练的模型或者权重文件精度作任何保证

图片预览：

标注例子：