Automatic bridge detection in airborne laser scanned data

Abstract

Collecting information about objects or areas without being in physical contact with them is referred to as remote sensing. Several techniques for acquisition of remote sensing data are available, one of the most common techniques used today is airborne laser scanning where a laser mounted on an airplane or helicopter is used to acquire remote sensing data. The remote sensing data can for example be used to create digital terrain models used in hydrological modelling. Objects like bridges can potentially disturb hydrological modelling, therefore bridges should be removed from such terrain models in order to provide accurate hydrological models. This thesis describes a method for removing bridges from remote sensing data acquired by airborne laser scanning. The method is shown to be able to detect bridges of varying sizes in several types of landscapes while maintaining a high detection rate.