Acoustic waves are commonly used to locate buried polyethylene pipes. In this preliminary study we are particularly interested in pipes depth. To obtain depth information we are moving towards a multi-sensor solution. Several estimators are implemented and tested on real data. A depth estimator according to the relative delays between sensors is proposed. We compare two relative delays estimators : the method using the cross-correlation and the one using the coherence function. We will verify on real measurements that the second method is much more efficient than the first one. Before discussing the results we will present another approach which consists in adapting the MUSIC (MUltiple SIgnals Classification) algorithm to our problem.

Localization of buried polyethylene pipes is an important issue for network managers. This study focuses on an acoustic method, which consists of vibrating the pipe and observing the signal with a receiver placed on the ground surface. This method provides an estimate of the path of the pipe but gives no information on the depth. We developed a multi-sensor method based on the principle of vibrating the pipe, which allows estimating the depth while being non-invasive and non-destructive and without a priori information on the propagation medium. These sensors are positioned perpendicular to the pipe. We developed a new estimator to estimate the depth and the propagation velocity in the medium, which is an important variable in our problem. This estimator is based on the MUSIC algorithm and is adapted to our choice of modeling. In this paper, two models of travel times in typical situations are presented. The first one represents the case where all sensors can be placed inside the trench (on the ground surface) in which the pipe is buried. The second one represents the case where sensors are placed inside and outside the trench. These travel time models aim to provide a fast result to allow the method to be used by field agents. They are compared with a full wavefield modeling by finite differences.