supasswd / reflective-clothes-detect

reflective-clothes-detect-dataset、helemet detection、工作服(反光衣)检测数据集、安全帽检测、施工人员穿戴检测

reflective-clothes-detect and dataset

工作服(反光衣)检测数据集

• author is leilei
• reflective-clothes-detect qq群: 980489677
• 如果此项目对您有所帮助，请给个star，您的star是对我的鼓励！

Some details

• reflective-clothes-detect-dataset (with xml annotations 1083) download: BaiDuYunPan 提取码->(dooh)
• yolov5s's weight is in reflective-clothes-detect-dataset !

Applicable instructions

0. download BaidDuYunPan's data and weight file
1. put yolov5s's weight file into yolov5 folder
2. cd yolov5, and excuting an order:

   python detect.py --source ***/aaa.jpg --weights ./best.pt
   

3. convert VOC2021 to YOLO format：

   call yolov5's voc_label_Re.py
   

How to use dataset?

• We annotate the reflective-clothes-detect-dataset as Pascal VOC format:

  --VOC2021(反光衣)
      --Annotations (xml_num: 1083)
      --ImageSets(Main)
      --JPEGImages (image_num: 1083)
  
      --images (yolov5 need, the same with JPEGImages)
      --labels (yolov4-5 need)
      --txt_yolov5 (yolov5 need 2021_train.txt)
      --2021_train.txt (yolov4 need)
  
      --yolov5_weight (yolov5s's weight)
  
      --label_name: reflective_clothes、other_clothes
  

• If you want to crawl some images

  Please refer to this crawler code on github:
  https://github.com/gengyanlei/fire-detect-yolov4 -> crawl_baidu.py
  

实际应用(Practical application)

• 基于SHWD数据集进行安全帽-反光衣-整体人 5类标注训练yolov4-yolov5，实现施工区域or危险区域检测
• Based on the SHWD data set, perform the five-category labeling training yolov4-yolov5 for helmet-reflective clothing-holistic people to achieve detection of construction areas or dangerous areas.

demo

• ./result:

----
----

How to expand reflective clothing data?

0. Based on the trained yolov4 model of 1083 reflective clothing images, expand the reflective clothing category of the SHWD data set;
1. Based on the data set expanded by 0 steps, train yolov4.

如何基于此项目进行反光衣数据扩充，并进行自动标注？

0. 基于1083张反光衣图像训练好的yolov4模型，对SHWD数据集进行反光衣类别扩充；
1. 基于0步骤扩充得到的数据集，训练yolov4.

Cite

• yolov5 (supports empty-labeled images)
• yolov4 (supports empty-labeled images)
• 博客blog
• 本数据仅学术探索！！！

Other

• building-segmentation-dataset
• fire-smoke-detect-datatset