Detection and identification data set and yolov5 model of helmet reflective clothing

                                                            A safety helmet - coverall ( Reflective clothing ) Test data sets and yolov4-5 Test model

Catalog

0. Abstract

1. Open source project github link

2. Dataset details

3. coverall ( Reflective clothing ) Data set expansion scheme

4. The test results

5. explain

---

0. Abstract

This article is open source 1 individual coverall ( Reflective clothing ) Test data set ( With label ) and Pre training model , This project has been uploaded github, welcome star.

coverall ( Reflective clothing )- Helmet detection ( Practical target detection ) qq Group : 980489677  qq2 Group ：710514100

CVAT Use the annotation tutorial ： Data tagging tools CVAT course _LEILEI18A The blog of -CSDN Blog _cvat Dimensioning tools

This project can be used in the construction area or Hazardous area inspectors ！！！

1. Open source project github link

https://github.com/gengyanlei/reflective-clothes-detect

The open source model is yolov5s Model

2. Dataset details

This article open source overalls ( Reflective clothing ) Detect dataset images 1083 Zhang ( contain xml mark )： Contains different colors 、( The police - Environmental protection - construction site - Seaside ( life jacket )) Such as the scene ;

CVAT Use the annotation tutorial ： Data tagging tools CVAT course _LEILEI18A The blog of -CSDN Blog _cvat Dimensioning tools

3. coverall ( Reflective clothing ) Data set expansion scheme

(1) be based on yolov4 Training reflective clothes 1083 Zhang image , Obtain the reflective clothing detection model ;

(2) be based on (1) Your model is right SHWD（ Right click the new tab to open the link ） Data set for category expansion , get 4 Class annotation ( A safety helmet 2 class , Reflective clothing 2 class , Whole person [ from coco The model provides ])

(3) be based on SHWD Data set training is enough ！

Be careful ：

yolov4 Can be empty label Image for training ;

yolov5 It can also be empty label Image for training ;

4. The test results

newest 5 Class detection effect ：

5. explain

(1) This data set is only for academic exploration ！