Health and safety are very important to be applied to all companies, and one of the important elements arranged by occupational health and safety is the use of personal protective equipment (PPE). They are equipment that have to be worn to enhance the worker safeness, and they have different standard for every workplace. 8 from 10 companies have rules to control the employee to wear personal protective equipment all the time whenever they work in a specific area. The number of work accidents happened in construction sites does not decrease, in this moment. One of the causes of work accidents is human negligence in wearing personal protective equipment or what is called PPE. A system that can monitor the completeness of PPE worn by workers in an industrial environment automatically is needed. Computer vision technology is used to monitor directly and detect workers who do not wear PPE. In this project, YOLOv3 (You Only Look Once) will be used for the detection of PPE. YOLO v3 is an object detection network part of yolo family (YOLOv1, YOLOv2). It can detect various things of different sizes, runs quite fast and make real-time inference possible on various devices. It takes images as input, pass it through neural network and get a vector of bounding boxes and class predictions as output.
How to run the program?
- Make sure to have Python IDLE and OpenCV installed
- Download "yolov3_training_last.weights" and "yolov3_testing.cfg" from the given google drive link.
- Run Object_Detection.py to detect PPE in pictures.
- Run Webcam.py to detect PPE in realtime using webcam.