/yolov3-Helmet-Detection

Training a YOLOv3 model to detect the presence of helmet for intrusion or traffic monitoring.

Primary LanguagePythonGNU General Public License v3.0GPL-3.0

Helmet Detection using YOLOV3

Its a simple YOLO model to detect and count the number of people wearing helmet in a image. this model can be used to detect the intrusion or to find the people ridding bike without helmet.

Here you can read my post in detail https://medium.com/@vijaysingh_60587/train-your-own-custom-model-for-helmet-detection-object-detection-using-yolo-f53a48066d7a

Getting Started

This code is very simple and with the help of little manipulation, you can count the number of detection for a particular detection.

or extract the co-ordinates of bounding boxes. download the models, create necessary files and give full path to the models and folder names and run the python script.

training:

if you want to train your own model, follow the darknet. Prerequisites
install python3.
install pip3.
install opencv.( sudo pip3 install opencv-python ).

install other liberaries if missing.

After setting up the paths.
change the path of classfile in line 19,
change the path of configuration file in line 25,
change the path of weights in line 26
change the output folder name in line 133 where you want to keep your output files.
change the name of folder in line 150 for input images.
, it's ready to run.
open terminal and python3 Helmet_detection_YOLOV3.py

link to files.

model
cfg
obj.names
If everything went well. you will get results like this
img1
img2
img3