Identification and analysis of airborne laser swath mapping data in a novel feature space

Several novel features are examined to determine their effectiveness in separating buildings from trees in airborne laser swath mapping (ALSM) data. New one- and two-dimensional distance measures are created to quantify the separability of the classes using the different features. Several features involving the intensity of the laser returns were found to be highly effective at separating the classes. The new distance measure provides insight into what makes a good/bad feature when discriminating between classes. It also lays the groundwork for future classification of ALSM data by providing a systematic method of ranking features to be used for classification.