Timely alerts provided to the communities at risk of landslides can prevent casualties and costly damages to people, buildings and infrastructures. The rainfalls are one of the primary triggering causes for landslides so that empirical approaches based on the correlation between landslides occurrence and rainfall characteristics, are considered effective for warning systems. This research work has intended to develop a landslide alerting system by using a Sensor Fusion method based on the SVC (Support Vector Classification) techniques. This method fuses rainfall data gathered in continuous by networked rain gauges and returns confidence degrees associated to the not occurrence of the landslide event as well as to the occurrence of one. By using a k-fold validation technique, an SVC-model, with AUC (Area Under the Curve) mean of 0,964733 and variance of 0,001243, has been defined. The proposed method has been tested on the regional rain gauges network, deployed in Calabria (Italy). © Springer International Publishing AG 2018.

A Sensor Fusion Method Applied to Networked Rain Gauges for Defining Statistically Based Rainfall Thresholds for Landslide Triggering

Di Francia, G.;Buonanno, Antonio;De Vito, S.;Fattoruso, G.
2018

Abstract

Timely alerts provided to the communities at risk of landslides can prevent casualties and costly damages to people, buildings and infrastructures. The rainfalls are one of the primary triggering causes for landslides so that empirical approaches based on the correlation between landslides occurrence and rainfall characteristics, are considered effective for warning systems. This research work has intended to develop a landslide alerting system by using a Sensor Fusion method based on the SVC (Support Vector Classification) techniques. This method fuses rainfall data gathered in continuous by networked rain gauges and returns confidence degrees associated to the not occurrence of the landslide event as well as to the occurrence of one. By using a k-fold validation technique, an SVC-model, with AUC (Area Under the Curve) mean of 0,964733 and variance of 0,001243, has been defined. The proposed method has been tested on the regional rain gauges network, deployed in Calabria (Italy). © Springer International Publishing AG 2018.
9783319668017
Support vector machine;Landslides warning system;Sensing systems;Networked rainfall gauges;Sensor fusion methods
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.12079/3778
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
social impact