Skip to Main Content
The paper presents the assumptions and the architecture of a multi-layer system for a perimetric protection of an industrial infrastructure object, i.e. the natural gas city gas station. Perimetry, which originally refers to the systematic measurement of human's eye visual system, here is used to describe system capabilities of monitoring a wide-angle visual field located in an industrial object. Such surveillance is aimed at detecting malfunctioning elements, even if their dimensions and contrast are extremely low. In particular, the reported research was focused on detecting gas leakages. The system analyses multimodal image input, acquired from conventional video and IR thermography cameras as well as accounts for information on natural gas flow pattern inside the pipelines. The information is processed and submitted to the natural gas dispatching centre, and the acquired parameters are use to evaluate a level of failure probability. The objective of the system is to support the dispatching centre crew in taking decisions concerning appropriate maintenance or rescue actions.