“PPE Detection” seeks to identify flaws and bring more safety to companies and employees
A key sector in companies, Occupational Safety aims to protect the physical integrity of workers, with standards, activities and preventive actions. And thinking about automating processes in the segment, students of the Computer Science course at Ufal, who are part of Edge Academy, are developing an application aimed at work activities, the ‘PPE Detection’, which identifies, through a trained model, the use of safety equipment in companies.
“The idea came about while talking to someone who is a safety technician, who explained that it is an area with very manual processes, and therefore prone to errors. So, in light of this, we had the idea of trying to automate the process,” explains student Edvar Neto, who is responsible for the project together with students João Costa and José Dantas, all in the 5th period of Computer Science at Ufal.
They conducted research that proved that the safety process is, in fact, still very manual – in the interviews they conducted, they received reports of several failures, such as the lack of use of equipment when there is a larger volume of employees, as well as the use of protective equipment incorrectly.
“In addition to the research and interviews, we conducted API tests and found that it would be possible to identify the safety equipment and, in this way, automate the process,” points out Edvar. The students used the Yolo network architecture, which is the basis of the project’s neural network. As for the programming language, the students adopted Python and, for the Requests, Flask.
The initial idea was to process static images sent to the software to identify where basic safety equipment was missing, which would be marked by the model. However, the students went further and divided the project into two distinct functions – ‘Falcão’, for static images, and ‘Big Brother’.
Big Brother consists of a ‘live stream’ model – that is, through security cameras installed in a company, the software can identify the safety equipment of all people who pass by the cameras in real time, checking whether it is being used properly.
Edge Academy – Made possible with resources from Softex, Edge Academy is the talent training project of the Edge Innovation Center, aimed at students of the Computer Science Institute of the Federal University of Alagoas (UFAL), which aims to promote academic excellence, develop skills and competencies, and prepare students for professional life from the beginning of their undergraduate studies.
