Classifying airborne LiDAR point clouds via deep features learned by a multi-scale convolutional neural network

Point cloud classification plays a critical role in many applications of airborne light detection and ranging (LiDAR) data. In this paper, we present a deep feature-based method for accurately classifying multiple ground objects from airborne LiDAR point clouds. With several selected attributes of LiDAR point clouds, our method first creates a group of multi-scale contextual images for each point in the data using interpolation. Taking the contextual images as inputs, a multi-scale convolutional neural network (MCNN) is then designed and trained to learn the deep features of LiDAR points across various scales. A softmax regression classifier (SRC) is finally employed to generate classification results of the data with a combination of the deep features learned from various scales. Compared with most of traditional classification methods, which often require users to manually define a group of complex discriminant rules or extract a set of classification features, the proposed method has the ability to automatically learn the deep features and generate more accurate classification results. The performance of our method is evaluated qualitatively and quantitatively using the International Society for Photogrammetry and Remote Sensing benchmark dataset, and the experimental results indicate that our method can effectively distinguish eight types of ground objects, including low vegetation, impervious surface, car, fence/hedge, roof, facade, shrub and tree, and achieves a higher accuracy than other existing methods.