Identification of Spatially Extended Pollution Sources by Means of Blind Sources Separation Algorithms

Blind Source Separation (BSS) is able to recover original source signals from their mixtures. Well-known sources are brain electric activity, acoustic phenomena, earthquakes. Here we apply BSS to air pollution analysis identificating gas emissions using only sensor measurements mutually and statistically independent. This application consents an independent, low cost and real time environmental monitoring system. Besides, a real-world pollution case from an industrial region of central Italy is presented and processed by BSS.