Object-Based Classification of Airborne LiDAR Point Clouds with Multiple Echoes

A method is proposed to classify the point clouds in urban areas. Particularly, surface growing algorithm is employed to segment the point clouds, which is helpful to derive more features such as area, position, orientation, multiple echo proportion, height jump between adjacent segments, and topological relationship of neighboring segments. Additionally, echo information is employed to distinguish difference types of points. Two datasets are utilized to test our proposed method. The results suggest that our method will produce the overall classification accuracy larger than 93% and the Kappa coefficient larger than 0.89, which is very satisfying.