An Object-Based Method for Urban Land Cover Classification Using Airborne Lidar Data

Airborne Lidar (Light detection and ranging) data have been widely used for classifying different land cover types. However, few researchers have conducted urban land cover classification using discrete airborne Lidar data as the sole data source. This research explores the possibility of applying airborne Lidar data to land cover classification in urban areas. The elevation difference and intensity difference between the first and last return, which may not work efficiently in pixel-based classification, were employed as two key attributes at the object level. Since tree objects have a much larger proportion of returns which show the elevation and intensity difference, the two indicators were used to classify the most indistinguishable land cover types, buildings and trees. In addition, height and intensity information were integrated to classify other land cover types. A case study was conducted in the city of Cambridge and eight urban land cover types were classified with an overall accuracy of 93.6%. Each land cover type was classified with an accuracy of between 80% and 100% and among these types, the accuracy of more than 90% for trees and buildings was satisfactory.