Addressing the need for high-quality, time efficient, and easy to use annotation tools, we propose SAnE, a semiautomatic annotation tool for labeling point cloud data. The contributions of this paper are threefold: (1) we propose a denoising pointwise segmentation strategy enabling a fast implementation of one-click annotation, (2) we expand the motion model technique with our guided-tracking algorithm, and (3) we provide an interactive, yet robust, open-source point cloud annotation tool, targeting both skilled and crowdsourcing annotators. Using the KITTI dataset, we show that the SAnE speeds up the annotation process by a factor of 4 while achieving Intersection over Union (IoU) agreements of 84%. Furthermore, in experiments using crowdsourcing services, SAnE achieves more than 20% higher IoU accuracy compared to the existing annotation tool and its baseline, while reducing the annotation time by a factor of 3. This result shows the potential of SAnE, for providing fast and accurate annotation labels for large-scale datasets with a significantly reduced price. SAnE is open-sourced at https://github.com/hasanari/sane.

Inspired by the success of deep learning techniques in dense-label prediction and the increasing availability of high precision airborne light detection and ranging (LiDAR) data, we present a research process that compares a collection of well-proven semantic segmentation architectures based on the deep learning approach. Our investigation concludes with the proposition of some novel deep learning architectures for generating detailed land resource maps by employing a semantic segmentation approach. The contribution of our work is threefold. (1) First, we implement the multiclass version of the intersection-over-union (IoU) loss function that contributes to handling highly imbalanced datasets and preventing overfitting. (2) Thereafter, we propose a novel deep learning architecture integrating the deep atrous network architecture with the stochastic depth approach for speeding up the learning process, and impose a regularization effect. (3) Finally, we introduce an early fusion deep layer that combines image-based and LiDAR-derived features. In a benchmark study carried out using the Follo 2014 LiDAR data and the NIBIO AR5 land resources dataset, we compare our proposals to other deep learning architectures. A quantitative comparison shows that our best proposal provides more than 5% relative improvement in terms of mean intersection-over-union over the atrous network, providing a basis for a more frequent and improved use of LiDAR data for automatic land cover segmentation.

We present a methodology for distinguishing between three types of animal movement behavior (foraging, resting, and walking) based on high-frequency tracking data. For each animal we quantify an individual movement path. A movement path is a temporal sequence consisting of the steps through space taken by an animal. By selecting a set of appropriate movement parameters, we develop a method to assess movement behavioral states, reflected by changes in the movement parameters. The two fundamental tasks of our study are segmentation and clustering. By segmentation, we mean the partitioning of the trajectory into segments, which are homogeneous in terms of their movement parameters. By clustering, we mean grouping similar segments together according to their estimated movement parameters. The proposed method is evaluated using field observations (done by humans) of movement behavior. We found that on average, our method agreed with the observational data (ground truth) at a level of 80.75% ± 5.9% (SE).

Advances in techniques for automated classification of pointcloud data introduce great opportunities for many new and existing applications. However, with a limited number of labeled points, automated classification by a machine learning model is prone to overfitting and poor generalization. The present paper addresses this problem by inducing controlled noise (on a trained model) generated by invoking conditional random field similarity penalties using nearby features. The method is called Atrous XCRF and works by forcing a trained model to respect the similarity penalties provided by unlabeled data. In a benchmark study carried out using the ISPRS 3D labeling dataset, our technique achieves 84.97% in term of overall accuracy, and 71.05% in term of F1 score. The result is on par with the current best model for the benchmark dataset and has the highest value in term of F1 score.