【目标检测】混凝土缺陷数据集7755张YOLO+VOC格式

据集格式：VOC格式+YOLO格式

压缩包内含：3个文件夹，分别存储图片、xml、txt文件

JPEGImages文件夹中jpg图片总计：7755

Annotations文件夹中xml文件总计：7755

labels文件夹中txt文件总计：7755

标签种类数：20

标签名称:["ACrack","Bearing","Cavity","Crack","Drainage","EJoint","Efflorescence","ExposedRebars","Graffiti","Hollowareas","JTape","PEquipment","Restformwork","Rockpocket","Rust","Spalling","WConccor","Weathering","Wetspot","object"]

每个标签的框数（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）：

ACrack 框数 = 419

Bearing 框数 = 1188

Cavity 框数 = 9047

Crack 框数 = 3562

Drainage 框数 = 1624

EJoint 框数 = 502

Efflorescence 框数 = 3926

ExposedRebars 框数 = 1972

Graffiti 框数 = 2176

Hollowareas 框数 = 1446

JTape 框数 = 1064

PEquipment 框数 = 1883

Restformwork 框数 = 999

Rockpocket 框数 = 608

Rust 框数 = 14481

Spalling 框数 = 9684

WConccor 框数 = 405

Weathering 框数 = 4585

Wetspot 框数 = 1641

object 框数 = 13

总框数：61225

图片清晰度（分辨率：像素）：清晰

图片是否增强：否

标签形状：矩形框，用于目标检测识别

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，数据集只提供准确且合理标注