Artificial Intelligence and Learning Systems Methods in Supporting Long-Term Acoustic Climate Monitoring
Maciej Kłaczyński, Tadeusz Wszołek · Acta Physica Polonica A · 2013
Developing eective methods for automatic identication of noise sources is currently one of the most important tasks in long-term acoustical climate monitoring of the environment.Manual verication of recorded data, when it comes to proper determination of noise levels, is time-consuming and costly.A possible solution is to use pattern recognition techniques for acoustic signal recorded by a monitoring station.This paper presents usefulness of special directed measurement techniques, acoustic signal processing, and classication methods using articial intelligence (the Sammon mapping) and learning systems methods (Support Vector Machines) in the recognition of corona audible noise from ultra-high voltage AC transmission lines.