Forest Monitoring Using Forest Sound Identification
Svetlana Segărceanu, Elena Olteanu, George Dan Suciu · 2020
Massive deforestation represents an important global issue affecting the forest environment and wildlife. An intelligent forest environment monitoring solution with signal analysis algorithms is proposed in this paper. Environmental sounds are processed using some modelling algorithms based on which the acoustic forest events can be classified into one of the categories: chainsaw, vehicle, genuine forest background noise. The article will explore and compare several methodologies for environmental sound classification, among which the dominant Deep Neural Networks and the classical Gaussian Mixtures Molelling and Dynamic Time Warping.