Dynamic noise map generation using machine learning

Campos, R. R. E. Programa de Engenharia Elétrica, Universidade Federal do Rio de Janeiro, Rio de Janeiro, RJ, Brasil, roberto@pads.ufrj.br

Simões, A. B. A. M. R Escola Politécnica, Universidade Federal do Rio de Janeiro, Rio de Janeiro, RJ, Brasil, anamrsimoes@poli.ufrj.br

Petraglia, M. R. Programa de Engenharia Elétrica, Universidade Federal do Rio de Janeiro, Rio de Janeiro, RJ, Brasil, mariane@pads.ufrj.br

Torres, J. C. B. Programa de Engenharia Elétrica, Universidade Federal do Rio de Janeiro, Rio de Janeiro, RJ, Brasil, julio@poli.ufrj.br

Abstract

Urban traffic is one of the most important sources of noise in cities. Noise metrics like LAeq and DNL (day-night level), are based on road traffic characteristics over long averaged periods and do not describe the traffic dynamics in these periods. This paper presents a method to produce dynamic noise maps from flow traffic data obtained by surveillance cameras. The method is implemented in a python application, where the videos are processed by machine learning algorithms. A retrained convolutional neural network is used to identify vehicles in the scene and classify them according to their size, while a tracking algorithm estimates their positions, speeds and accelerations. Such data, extracted from videos, are useful for calculating the sound pressure level at a given location by a standard model. The estimated energy noise level of all vehicles present in the scene are added to produce dynamic noise maps, representing the variation of sound pressure over time, according to the traffic flow. The levels obtained from the data extracted from the video are compared with measurements at control points to validate the model. The results of the proposed model are presented, as well as the its constraints to produce the noise maps.

Keywords:

Traffic noise

vehicle information

convolutional neural network.

References

[1] S. Fidell, T. Schultz, and D. M. Green. A theoretical interpretation of the prevalence rate of noise-induced annoyance in residential populations. The Journal of the Acoustical Society of America, 84(6):2109–2113, 1988. doi:10.1121/1.397056.

[2] SA Stansfeld, B Berglund, C Clark, I Lopez Bar rio, P Fischer, E Ohrstrom, MM Haines, J Head, S Hygge, I van Kamp, and BF Berry. Air craft and road traffic noise and children’s cogni tion and health: a cross-national study. The Lancet, 365(9475):1942– 1949, 2005. 0140-6736. ISSN doi:http://doi.org/10.1016/S0140 6736(05)66660-3.

[3] Yvonne de Kluizenaar, Sabine A. Janssen, Frank J. van Lenthe, Henk M. E. Miedema, and Johan P. Mackenbach. Long-term road traffic noise exposure is associated with an increase in morning tiredness. TheJournalofthe AcousticalSociety ofAmerica, 126 (2):626–633, 2009. doi:10.1121/1.3158834. URL http://dx.doi.org/10.1121/1.3158834.

[4] Shinichi Sakamoto. Road traffic noise predic tion model asj rtn-model 2018: Report of the re search committee on road traffic noise. Acousti cal Science and Technology, 41(3):529–589, 2020. doi:10.1250/ast.41.529.

[5] Sh. Givargis and M. Mahmoodi. Con verting the uk calculation of road traffic noise (cortn) to a model capable of calculating laeq,1h for the tehran’s roads. Applied Acoustics, 69(11):1108–1113, 2008. ISSN 0003-682X. doi:https://doi.org/10.1016/j.apacoust.2007.08.003.

[6] Diego Murillo Gomez, Juan Gil Carvajal, Valentina Zapata Rodriguez, and José Téllez García. Assess ment of the rls 90 calculation method for predicting road traffc noise in colombian conditions. Revista Facultad de Ingeniería Universidad de Antioquia, 1, 04 2015. doi:10.17533/udea.redin.n75a17.

[7] F Anfosso Ledee, M Paviotti, and S Kephalopou los. Common noise assessment methods in Eu rope (CNOSSOS-EU) : to be used by the EU Mem ber States for strategic noise mapping following adoption as specified in the Environmental Noise Directive 2002/49/EC. Publications Office, 2012. doi:doi/10.2788/32029.

[8] J. Quartieri, Nikos Mastorakis, Gerardo Iannone, Claudio Guarnaccia, Salvatore D’Ambrosio, Anto nio Troisi, and T Lenza. A review of traffic noise predictive noise models. pages 72–80, 01 2009.

[9] Marketa Dubska, Adam Herout, and Jakub Sochor. Automatic camera calibration for traffic understand ing. Proceedings of the British Machine Vision Conference, 2014.

[10] Kaiming He, Georgia Gkioxari, Piotr Dollár, and Ross Girshick. Mask r-cnn. 2017 IEEE Interna- tional Conference on Computer Vision (ICCV), pages 2980-2988, 2017. doi:10.1109/ICCV.2017.322.

[11] Tsung-Yi Lin, Michael Maire, Serge Belongie, LubomirBourdev, RossGirshick, James Hays, Pietro Perona, Deva Ramanan, C. Lawrence Zitnick, and Piotr Dollár. Microsoft coco: Common objects in context, 2014. URL http://arxiv.org/abs/1405.0312. cite arxiv:1405.0312Comment: 1) updated annota tion pipeline description and figures; 2) added new section describing datasets splits; 3) updated author list.

[12] Alex Bewley, Zongyuan Ge, Lionel Ott, Fabio Ramos, andBenUpcroft. Simple online andrealtime tracking. 2016 IEEE International Conference on Image Processing (ICIP), pages 3464–3468, 2016. doi:10.1109/ICIP.2016.7533003.

[13] J. Quartieri, Gerardo Iannone, and Claudio Guarnac cia. On the improvement of statistical traffic noise prediction tools. Proceedings of the 11th WSEAS International Conference on Acoustics and Music: Theory and Applications, AMTA ’10, 06 2010.