Obstructive sleep apnea (OSA) is a disease suffered by more than one hundred million people in the world. To be diagnosed, an assessment is required where people are monitored overnight while they sleep, these exams are uncomfortable, and their scheduling is delayed. This degree work develops a system that allows the monitoring and identification of the respiratory frequency of a person in a dream state from the technique of video magnification through video-analytics in the cloud, for support in the assessment and diagnosis of OSA Sleep. In methodological terms, an analysis is performed on a remote system for the recognition of respiratory rate by means of video magnification of a recording of a person in a dream state. For its part, the state of the art allows us to show that there are previous works of which video analysis has been implemented for the detection of vital signs. The tests carried out show that the system has a degree of precision of 97.87% compared to a sensor made up of commercial electronic devices. Finally, it is concluded in general terms that the proposed system has the potential to consolidate as a system that contributes to the detection of OSA in the health area.