Abstract
Industries move toward the replacement of labours engaged in dangerous tasks with fully automated systems. The sixth sense technology aims at achieving that by integrating different technologies in such a way that enables monitoring of industrial plants and predicting any faults that could happen. One important module of the sixth sense technology is inspection robots. This paper aims at providing the inspection robots with equipment-detection capability, resembling the human inspectors performing the customised inspection for a variety of equipment. The types of equipment, used in this study, are reactor, boiler, pump, isolated pipes, meter gauge, and valves. Given the complexity of the industrial environment, we propose a real-time deep-learning-based equipment detection model. The results show that the mean average precision is above 90%, which ensure the significant performance of the proposed solution. This work validates the practicality of our equipment-detection model and shows its potential to be employed on our inspection robot.