Akademik Veri Yönetim Sistemi
2023 Innovations in Intelligent Systems and Applications Conference (ASYU), Sivas, Türkiye, 11 - 13 Ekim 2023, ss.1-6
The robotics sector is one of the areas most affected by the developments in technology. Today, robots become autonomous in industrial workplaces such as factories and take over some jobs that humans can do. One of the most important issues in these fields of work is undoubtedly occupational health and safety. Protective helmets are indispensable for the occupational health and safety of employees. Thanks to the helmet, the worker is protected from major mishaps. It is tiring for a person to check that all employees are wearing helmets during long working hours. It is aimed to offer a more practical solution to this problem with an autonomous robot that can detect workers who do not wear helmets during the tour and return to the charging station automatically when the battery is low, which maps the indoor environment of the work area and goes for a walk in the field. Indoor environment mapping was performed with the SLAM (Simultaneous Localization and Mapping) algorithm used with LIDAR. Vision-Based Positioning Method is used for the robot to reach the charging station perfectly. In this context, the use of ArUco marker was preferred, and the openCV platform was used for the production detection and distance measurement of the marker. In this way, the possibility of shifting the position of the robot on the map during the turning movements and the problem in reaching the charging station as a result have been eliminated by the use of a vision-based positioning method to be followed independently from the map. For helmet detection, it was preferred to prepare a model with the YOLOv7 (You Only Look Once version 7) algorithm. Based on 5000 data, our study has been determined to have an accuracy rate of 94.4%.