Niezawodność identyfikacji danych w systemach monitoringu

Systemy monitoringu wideo od szeregu lat wykorzystywane są w celu zwiększenia poziomu bezpieczeństwa nadzorowanych obiektów i dóbr materialnych. W każdym z wymienionych zastosowań niezwykle istotnym zagadnieniem jest niezawodność procesów nadzoru i monitorowania. W artykule przedstawiono analizę wpływu oświetlenia, odległości obiektu od sensora czy rozdzielczości sensora na poprawną identyfikację. Przeprowadzone badania umożliwiają określenie przydatności systemu monitoringu w określonych warunkach środowiskowych i technicznych.

Nowadays, many areas have to be protected. Monitoring systems (MS) effectively support this kind of processes. A monitoring system provides an image of the area containing the monitored objects. These systems are also used to increase the safety of people. That is why the efficiency is important in the reliable monitoring system (RMS). These issues are interpreted as the possibility of correct identification of the collected data. The reliability of elements, the environmental exposure as well as the process of object detection and face identification analysis have important impact on the MS. Therefore, the main problem of the research is to answer the question: what is the reliability of the identification data (i.e. people) in the real-time monitoring system? The paper presents the main results of the study. The study was conducted in the MS demonstrator consisting of a telecommunication network with implemented video monitoring sensors. Various experiments were performed to adjust the lighting, the distance from the camera and the number of face images in the database. The developed method has to determine the reliability of the identification data. It takes into account the influence of degrading factors which are present in the monitoring system environment. On the basis of the obtained results we specify a set of recommendations for the reference MS.