Automatic railway classification using surface and subsurface measurements - Robotics Institute Carnegie Mellon University

Automatic railway classification using surface and subsurface measurements

George A. Kantor, Herman Herman, Sanjiv Singh, John G. Tabacchi, and William Kaufman
Conference Paper, Proceedings of 3rd International Conference on Field and Service Robotics (FSR '01), pp. 43 - 48, June, 2001

Abstract

The proper assessment of railroad condition requires the consideration of a number of factors. Some factors, such as the condition of the ties, can be measured by inspecting features visible from the surface of the railway. Other factors, such as the condition of the ballast, require subsurface measurements. Extensive human resources are currently applied to the problem of evaluating railroad health. Here we present the results of a study in automatic railroad classification that combines surface and subsurface measurements to characterize the railroad condition. To obtain surface measurements, we generate a 3-D profile of the railroad surface with a vision system that employs a laser light stripe. Subsurface measurements were made using ground penetrating radar (GPR). Principal component analysis was used to reduce the dimension of the raw data. Classifiers were trained on the resulting data using both memory based and Bayesian methods. The results are presented.

BibTeX

@conference{Kantor-2001-16802,
author = {George A. Kantor and Herman Herman and Sanjiv Singh and John G. Tabacchi and William Kaufman},
title = {Automatic railway classification using surface and subsurface measurements},
booktitle = {Proceedings of 3rd International Conference on Field and Service Robotics (FSR '01)},
year = {2001},
month = {June},
editor = {Aarne Halme, Raja Chatila, and Erwin Prassler},
pages = {43 - 48},
publisher = {Yleisjalljennos},
address = {Helsinki, Finland},
keywords = {railroad classification, pattern recognition, principal component analysis, ground penetrating radar},
}