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

摘要

Electrical resistivity tomography (ERT) is widely used in environmental investigation such as contaminated sites surveys since it is fast and economic. Analysis and processing of the ERT data using computer have become one of the key issues to promote the application of ERT. A method is proposed to recognize contaminated area in ERT profiles using K-means algorithm. The effectiveness of the proposed method is evaluated using a synthetic model of typical contaminated sites. The study results show that when the resistivity difference between the contaminated area and the background area is large (more than 30%), the contaminated area identified by the K-means algorithm is similar to the contaminated area set in the synthetic model; when the difference between the contaminated area and the background area is small (less than 30%), the recognition accuracy of the method is reduced. The main reason is that when the resistivity of the contaminated area is close to the background, the difference between the two categories obtained by K-means algorithm is small, so some background areas may be misjudged as contaminated areas because of the data errors in the acquisition and inversion process.