Based on the finding that companies and institutions consume a large amount of electric energy and this has a direct impact on their budgets, it was verified the need to make corporate and institutional environments more autonomous with respect to the operation of their electrical equipment, for this, in this work will be presented and tested an embedded system, in the form of minimum viable product (MVP), that makes use of computer vision and artificial neural networks (RNA) to control the energy consumption of electrical equipment. To this end, a technical survey was carried out on the methodologies of embedded systems development, computational vision and RNA. For validation, tests were conducted to verify the efficiency of the backprogation algorithm, that is, based on RNA, as an alternative to the classification process. The tests were performed during the training stages and after the installation of the RNA trained in the proposed MVP. From the results it is possible to indicate that the solution behaves satisfactorily when shipped, in addition, the results also indicate the great potential of this tool proposed as alternative solution for the control of energy consumption.
Keywords:
|
