Abstract
MADRAS is an acoustic monitoring system that features the automatic recognition of noise sources. It is based on a tree classifier able to use different classification techniques like signal processing, morphology, fractional analysis, and neural networking. A very high accuracy of recognition is achieved due to the optimization of the choice of real-time technique applied and to the intrinsic performance of each algorithm. MADRAS can learn the characteristics of new sources and also use a database of previously processed cases and typical events. The paper presents the results of multiple source recognition in a complex acoustic environment. Implemented on a real time PC-based measuring analyzer called Symphonie, MADRAS has been able to distinguish all noise sources. Acquired data may then be processed to obtain the energy contribution of each source and a list of the frequency of each appearance. Simulations may be viewed and modified.
Paper version not known (
Free)
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call