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)：4957

标注数量(xml文件个数)：4957

标注数量(txt文件个数)：4957

标注类别数：27

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):["biaoxianbuqing","chepai","daolumenjia","daoluyisa","dianzibiaozhi","guangjiaojing","jiansudai","jiaotongbiaozhi","jiaotongdeng","jingshigao","jingshitong","jingshizhui","jishui","kengdong","lajitong","liefeng","lubianxiang","ludeng","posuiban","shensuofeng","shuinihulan","suliaohulan","wangzhuangliefeng","xiubukengdong","xiubuliefeng","xiubuwangzhuang","yaojing"]

对应中文类别名：["标线不清", "车牌", "道路门架", "道路遗撒", "电子标志", "广角镜", "减速带", "交通标志", "交通灯", "警示杆", "警示桶", "警示锥", "积水", "坑洞", "垃圾桶", "裂缝", "路边箱", "路灯", "破碎板", "伸缩缝", "水泥护栏", "塑料护栏", "网状裂缝", "修补坑洞", "修补裂缝", "修补网状", "窨井"]

每个类别标注的框数：

biaoxianbuqing 框数 = 1

chepai 框数 = 131

daolumenjia 框数 = 1

daoluyisa 框数 = 36

dianzibiaozhi 框数 = 3

guangjiaojing 框数 = 1

jiansudai 框数 = 19

jiaotongbiaozhi 框数 = 25

jiaotongdeng 框数 = 2

jingshigao 框数 = 766

jingshitong 框数 = 4

jingshizhui 框数 = 41

jishui 框数 = 1

kengdong 框数 = 20

lajitong 框数 = 11

liefeng 框数 = 935

lubianxiang 框数 = 12

ludeng 框数 = 1

posuiban 框数 = 1

shensuofeng 框数 = 1811

shuinihulan 框数 = 2

suliaohulan 框数 = 7

wangzhuangliefeng 框数 = 15

xiubukengdong 框数 = 1418

xiubuliefeng 框数 = 4184

xiubuwangzhuang 框数 = 7

yaojing 框数 = 142

总框数：9597

每个类别占有图片数：

biaoxianbuqing 占有图片数 = 1

chepai 占有图片数 = 99

daolumenjia 占有图片数 = 1

daoluyisa 占有图片数 = 19

dianzibiaozhi 占有图片数 = 3

guangjiaojing 占有图片数 = 1

jiansudai 占有图片数 = 18

jiaotongbiaozhi 占有图片数 = 22

jiaotongdeng 占有图片数 = 1

jingshigao 占有图片数 = 174

jingshitong 占有图片数 = 1

jingshizhui 占有图片数 = 16

jishui 占有图片数 = 1

kengdong 占有图片数 = 18

lajitong 占有图片数 = 11

liefeng 占有图片数 = 332

lubianxiang 占有图片数 = 8

ludeng 占有图片数 = 1

posuiban 占有图片数 = 1

shensuofeng 占有图片数 = 1221

shuinihulan 占有图片数 = 1

suliaohulan 占有图片数 = 2

wangzhuangliefeng 占有图片数 = 14

xiubukengdong 占有图片数 = 1047

xiubuliefeng 占有图片数 = 2392

xiubuwangzhuang 占有图片数 = 6

yaojing 占有图片数 = 124

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：数据集没有划分训练验证测试集需自行划分

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：