Hybrid Hidden Markov Model for Marine Environment Monitoring

Kévin Rousseeuw (1, 2) , Emilie Poisson Caillault (1) , Alain Lefebvre (2) , Denis Hamad (1)

1 LISIC - Laboratoire d'Informatique Signal et Image de la Côte d'Opale
2 IFREMER - Institut Français de Recherche pour l'Exploitation de la Mer

Phytoplankton is an important indicator of water quality assessment. To understand phytoplankton dynamics, many fixed buoys and ferry boxes were implemented, resulting in the generation of substantial data signals. Collected data are used as inputs of an effective monitoring system. The system, based on unsupervised Hidden Markov Model (HMM), is designed not only to detect phytoplancton blooms but also to understand their dynamics. HMM parameters are usually estimated by an iterative Expectation-Maximisation approach. We propose to estimate HMM parameters by using spectral clustering algorithm. The monitoring system is assessed on database signals from MAREL-Carnot station (Boulogne-sur-Mer, France). Experiment results show that the proposed system is efficient to detect environmental states such as phytoplankton productive and non productive periods without a priori knowledge. Furthermore, discovered states are consistent with biological interpretation.

Mots clés: Hidden Markov models; Monitoring; Clustering algorithms; Sensors; Support vector machines; Remote sensing; spectral clustering

Domaines: Machine Learning [stat.ML]; Apprentissage [cs.LG]; Traitement du signal et de l'image [eess.SP]; Biodiversité et Ecologie; Ingénierie de l'environnement

Kévin Rousseeuw, Emilie Poisson Caillault, Alain Lefebvre, Denis Hamad. Hybrid Hidden Markov Model for Marine Environment Monitoring. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2015, 8 (1), pp.204-213. ⟨10.1109/JSTARS.2014.2341219⟩. ⟨hal-01442974⟩