Creating and sharing knowledge for telecommunications

Novel approach for simultaneous sediment classification and concentration determination of water turbidity

Duarte, D. ; Prats, S. ; Keizer, J. J. ; Georgieva, P. ; Nogueira, R.N. ; Bilro, L.

Novel approach for simultaneous sediment classification and concentration determination of water turbidity, Proc International Conference on Optical Fibre Sensors 24th International Conference on Optical Fibre Sensors OFS24, Curitiba, Brazil, Vol. 9634, pp. 96342U-1 - 96342U-4, September, 2015.

Digital Object Identifier: 10.1117/12.2194578

Abstract
A new approach for data analysis and classification for datasets obtained by a multiparameter optical turbidity sensor is proposed. This approach is based on the combination of statistical or machine learning methods such as linear regressions and clustering analysis. A case study is presented using a 6 dimensional fiber optic sensor to simultaneously classify types of sediments and concentration. Results show a 79% of success for the used training data sets. The methodology proposed is flexible because can be easily adapted to other physical scenarios.