无人机视角电力巡检线路设备检测数据集7933张VOC+YOLO格式

作品简介

无人机视角电力巡检线路设备检测数据集7933张VOC+YOLO格式

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):7933

标注数量(xml文件个数):7933

标注数量(txt文件个数):7933

标注类别数:18

所在github仓库:datasets_sl

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):["glass insulator","glass insulator big shackle","glass insulator small shackle","glass insulator tower shackle","lightning rod shackle","lightning rod suspension","polymer insulator","polymer insulator lower shackle","polymer insulator tower shackle","polymer insulator upper shackle","spacer","sphere","spiral damper","stockbridge damper","tower id plate","vari-grip","yoke","yoke suspension"]

标签中文对照:玻璃绝缘子,玻璃绝缘子大线夹,玻璃绝缘子小线夹,玻璃绝缘子铁塔线夹,避雷针线夹,避雷针悬垂线夹,复合绝缘子,复合绝缘子下线夹,复合绝缘子铁塔线夹,复合绝缘子上线夹,间隔棒,均压环(或球),防振锤(螺旋式),斯托克布里奇防振锤,铁塔铭牌,万向握线器(或变径线夹),联板,悬垂联板

每个类别标注的框数:

glass insulator 框数 = 2015

glass insulator big shackle 框数 = 110

glass insulator small shackle 框数 = 128

glass insulator tower shackle 框数 = 98

lightning rod shackle 框数 = 170

lightning rod suspension 框数 = 618

polymer insulator 框数 = 2377

polymer insulator lower shackle 框数 = 1448

polymer insulator tower shackle 框数 = 44

polymer insulator upper shackle 框数 = 1311

spacer 框数 = 68

sphere 框数 = 15

spiral damper 框数 = 820

stockbridge damper 框数 = 5686

tower id plate 框数 = 198

vari-grip 框数 = 827

yoke 框数 = 1336

yoke suspension 框数 = 5254

总框数:22523

每个类别占有图片数:

glass insulator 占有图片数 = 1912

glass insulator big shackle 占有图片数 = 63

glass insulator small shackle 占有图片数 = 68

glass insulator tower shackle 占有图片数 = 53

lightning rod shackle 占有图片数 = 99

lightning rod suspension 占有图片数 = 617

polymer insulator 占有图片数 = 2332

polymer insulator lower shackle 占有图片数 = 1379

polymer insulator tower shackle 占有图片数 = 44

polymer insulator upper shackle 占有图片数 = 1310

spacer 占有图片数 = 67

sphere 占有图片数 = 15

spiral damper 占有图片数 = 744

stockbridge damper 占有图片数 = 1425

tower id plate 占有图片数 = 198

vari-grip 占有图片数 = 467

yoke 占有图片数 = 1336

yoke suspension 占有图片数 = 2197

图片分辨率:640x640

使用标注工具:labelImg

数据集是否增强:是

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




创作时间: