A method to recognize the contaminated area using K-means in ERT contaminated site surveys

Les algorithmes de classification permettent de gérer de grandes quantités de données. L'objectif de cette étude est de comparer deux méthodes de classification, la Classification hiérarchique Ascendante (CHA) et la méthode des Centres Mobiles (CM) sur des données de Tomographie de Résistivité Electrique (TRE) caractérisant le massif entourant la grotte de Lascaux. Le rendu de ces méthodes est analysé en tenant compte de la connaissance a priori de la géologie du site. Dans ce projet, la structuration des données en 2 à 5 classes vise à subdiviser le massif en domaines les plus homogènes possible afin de leur affecter des propriétés thermiques différentes dans un modèle 3D. Ces dernières serviront de paramètres d’entrée nécessaires aux simulations thermo-aérauliques de la grotte soumise à différentes conditions météorologiques.

show abstract

Partitioning a rock mass based on electrical resistivity data: the choice of clustering method

Lharti

Sirieix

Riss

et al. 2023

Geophysical Journal International

View full text Add to dashboard Cite

Summary The goal of data classification is to organize them into relevant groups using algorithms. In this study, two clustering algorithms are applied to classify a set of geophysical measurements performed around the Lascaux Cave (Dordogne, France). Based on a non-destructive geophysical method, electrical resistivity tomography (ERT), the data (resistivity values) are supposed to characterize the rock mass around the cave. The rock mass must be divided into an optimal number of homogeneous domains with specific thermal properties to integrate them in the future thermo-aeraulic simulations of the cave. Since the data are georeferenced, a given resistivity cluster corresponds to a specific spatial domain of the rock mass. This study aims to compare two different clustering methods, the Hierarchical Agglomerative Clustering (HAC) and the K-means methods, on the resistivity data set. Thus, the objective of this study is to determine which of the two methods leads to a partition of the massif with an optimal number of classes, allowing us to find the geological structures partially known thanks to previous studies. The results of these methods are analyzed in light of two different indices, the Silhouette Index (SI) and the Coefficient of Variation (CV). Based on these indices alone, K-means might seem to be the best algorithm, but a fine analysis of each cluster shows that the HAC method gives better results. Indeed, only the HAC method highlights areas already known to be either waterlogged or clayey. Our partial geological knowledge, which is of paramount importance in such a study, also supports the choice of the HAC method.

show abstract

scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.

Contact Info

hi@scite.ai

10624 S. Eastern Ave., Ste. A-614

Henderson, NV 89052, USA

Blog Terms and Conditions API Terms Privacy Policy Contact Cookie Preferences Do Not Sell or Share My Personal Information

Made with 💙 for researchers

Part of the Research Solutions Family.

A method to recognize the contaminated area using K-means in ERT contaminated site surveys

Cited by 3 publications

References 9 publications

Evaluation of heavy metal pollution with uneven spatial sampling distribution based on Voronoi area density

Evaluation of heavy metal pollution with uneven spatial sampling distribution based on Voronoi area density

Choix d’une méthode de classification pour la partition d’un massif rocheux à partir de TRE

Partitioning a rock mass based on electrical resistivity data: the choice of clustering method

Contact Info

Product

Resources

About