Development of a facial recognition algorithm and epps detection with python for the construction industry

León R.A.L.; Rufino K.F.A. · Proceedings of the LACCEI international Multi-conference for Engineering, Education and Technology (2023) · DOI: 10.18687/leird2023.1.1.540

Resumen

This project is based on providing an alternative solution to work accidents that occur in the construction sector, through the design of a facial recognition algorithm and EPPS detection with Python for the construction industry, for the present project it was considered control and prevention of unsafe acts during working hours, with the aim of contributing to the preservation of the health and safety of workers in areas that represent high risks of occupational accidents, where the use of personal protective equipment is of vital importance; Therefore, the objective is to design a facial recognition and EPPS detection algorithm with Python, develop a prototype that can identify and in turn indicate when personnel are using EPPS (helmet and glasses) properly, for which the use of EPPS will be necessary. of the PyCharm software that will help to process images and likewise carry out different tests of the system until validating the optimal operation. tests were performed using sequences of 300, 600 and 900 frames. Reaching with this last number of frames the optimal efficiency with 95.4%, these techniques used to detect the correct use of the epps are presented through logarithms given in real time which will be captured through a camera, according to the code of programming to be developed with the data obtained and the reference points in relation to the appearance of the personnel and the use of their helmets and glasses according to the vision of the programmed system.

Instituciones: Universidad Privada del Norte