Deep learning for conifer/deciduous classification of airborne LiDAR 3D point clouds representing individual trees

The purpose of this study was to investigate the use of deep learning for coniferous/deciduous classification of individual trees segmented from airborne LiDAR data. To enable processing by a deep convolutional neural network (CNN), we designed two discrete representations using leaf-off and leaf-on LiDAR data: a digital surface model with four channels (DSM × 4) and a set of four 2D views (4 × 2D). A training dataset of tree crowns was generated via segmentation of tree crowns, followed by co-registration with field data. Potential mislabels due to GPS error or tree leaning were corrected using a statistical ensemble filtering procedure. Because the training data was heavily unbalanced (~8% conifers), we trained an ensemble of CNNs on random balanced sub-samples. Benchmarked against multiple traditional shallow learning methods using manually designed features, the CNNs improved accuracies up to 14%. The 4 × 2D representation yielded similar classification accuracies to the DSM × 4 representation (~82% coniferous and ~90% deciduous) while converging faster. Further experimentation showed that early/late fusion of the channels in the representations did not affect the accuracies in a significant way. The data augmentation that was used for the CNN training improved the classification accuracies, but more real training instances (especially coniferous) likely results in much stronger improvements. Leaf-off LiDAR data were the primary source of useful information, which is likely due to the perennial nature of coniferous foliage. LiDAR intensity values also proved to be useful, but normalization yielded no significant improvement. As we observed, large training data may compensate for the lack of a subset of important domain data. Lastly, the classification accuracies of overstory trees (~90%) were more balanced than those of understory trees (~90% deciduous and ~65% coniferous), which is likely due to the incomplete capture of understory tree crowns via airborne LiDAR. In domains like remote sensing and biomedical imaging, where the data contain a large amount of information and are not friendly to human visual system, human-designed features may become suboptimal. As exemplified by this study, automatic, objective derivation of optimal features via deep learning can improve prediction tasks in such domains.