Illegal Deforestation Detection in the Amazon Rainforest Based on Audio Processing

• Waldomiro J. G. Seabra UFAM
• Juan G. Colonna UFAM
• Eduardo F. Nakamura UFAM

Abstract

In this work, we present a sound detection method for chainsaw engines to combat illegal logging. Our approach uses the Wavelet Decomposition in high frequencies, which divides optimally the natural sounds class from the other classes and divides very well, although there is still overlap, the chainsaw class from the artificial sounds class. We used a one-class classifier method, the Support Vector Data Description (SVDD), that creates a hypersphere around the points representing the target class (chainsaws) into the feature space, thus being able to differentiate chainsaws from other natural and artificial sounds. It is shown that the method is efficient discriminating sounds of chainsaws from natural sounds (AUC = 96%), discriminating sounds of chainsaws and artificial sounds from natural sounds (AUC = 91%), but loses efficiency when tested against artificial sounds (AUC = 87%).

References

Bertrand, A. (2011). Applications and trends in wireless acoustic sensor networks: a signal processing perspective. In Communications and Vehicular Technology in the Benelux (SCVT), 2011 18th IEEE Symposium on, pages 1–6. IEEE.

Chang, W.-C., Lee, C.-P., & Lin, C.-J. (2013). A revisit to support vector data description. Dept. Comput. Sci., Nat. Taiwan Univ., Taipei, Taiwan, Tech. Rep.

Colonna, J. G., Gama, J., & Nakamura, E. F. (2016a). How to Correctly Evaluate an Automatic Bioacoustics Classification Method, volume 9868 of Advances in Artificial Intelligence. Lecture Notes in Computer Science (LNCS), pages 37–47. Springer International Publishing.

Colonna, J. G., Gatto, B., Dos Santos, E. M., & Nakamura, E. F. (2016b). A framework for chain-saw detection using one-class kernel and wireless acoustic sensor networks into the amazon rainforest. In Mobile Data Management (MDM), 2016 17th IEEE International Conference on, volume 2, pages 34–36. IEEE.

Fawcett, T. (2006). An introduction to roc analysis. Pattern Recognition Letters, 27(8):861–874.

Graps, A. (1995). An introduction to wavelets. IEEE computational science and engineering, 2(2):50–61.

Greenpeace (2016). Desmatamento dispara na amazônia. disponível em: [link].

Khan, S. S. (2010). Kernels for One-Class Nearest Neighbour Classification and Comparison of Chemical Spectral Data. PhD thesis, College of Engineering and Informatics, National University of Ireland.

Lin, J. (1991). Divergence measures based on the shannon entropy. IEEE Transactions on Information theory, 37(1):145–151.

Papán, J., Jurečka, M., & Púchyová, J. (2012). Wsn for forest monitoring to prevent illegal logging. In Computer Science and Information Systems (FedCSIS), 2012 Federated Conference on, pages 809–812. IEEE.

Seabra, W. J. G., Colonna, J. G., & Nakamura, E. F. (2017). Sensor acústico para detecção de desmatamento ilegal na floresta amazônica. IX Simpósio Brasileiro de Computação Ubíqua e Pervasiva (SBCUP).

Soisiinthorn, T. & Rujipattanapong, S. (2007). Deforestation detection algorithm for wireless sensor networks. In Communications and Information Technologies, 2007. ISCIT’07. International Symposium on, pages 1413–1416. IEEE.

Tax, D. & Duin, R. (2004). Support vector data description. Machine learning, 54(1):45–66.