Robust object recognition is a crucial skill for robots operating autonomously in real world environments. Range sensors such as LiDAR and RGBD cameras are increasingly found in modern robotic systems, providing a rich source of 3D information that can aid in this task. However, many current systems do not fully utilize this information and have trouble efficiently dealing with large amounts of point cloud data. In this paper, we propose VoxNet, an architecture to tackle this problem by integrating a volumetric Occupancy Grid representation with a supervised 3D Convolutional Neural Network (3D CNN). We evaluate our approach on publicly available benchmarks using LiDAR, RGBD, and CAD data. VoxNet achieves accuracy beyond the state of the art while labeling hundreds of instances per second.

/pdf/voxnet-a-3d-convolutional-neural-network-for-real-time-2dutg0m0bu.pdf

VoxNet: A 3D Convolutional Neural Network for real-time object recognition

Three-dimensional models provide a volumetric representation of space which is important for a variety of robotic applications including flying robots and robots that are equipped with manipulators. In this paper, we present an open-source framework to generate volumetric 3D environment models. Our mapping approach is based on octrees and uses probabilistic occupancy estimation. It explicitly represents not only occupied space, but also free and unknown areas. Furthermore, we propose an octree map compression method that keeps the 3D models compact. Our framework is available as an open-source C++ library and has already been successfully applied in several robotics projects. We present a series of experimental results carried out with real robots and on publicly available real-world datasets. The results demonstrate that our approach is able to update the representation efficiently and models the data consistently while keeping the memory requirement at a minimum.

/pdf/octomap-an-efficient-probabilistic-3d-mapping-framework-3dp42nnofi.pdf

OctoMap: an efficient probabilistic 3D mapping framework based on octrees

Three-dimensional models provide a volumetric representation of space which is important for a variety of robotic applications including flying robots and robots that are equipped with manipulators. In this paper, we present an open-source framework to generate volumetric 3D environ- ment models. Our mapping approach is based on octrees and uses probabilistic occupancy estimation. It explicitly repre- sents not only occupied space, but also free and unknown areas. Furthermore, we propose an octree map compression method that keeps the 3D models compact. Our framework is available as an open-source C++ library and has already been successfully applied in several robotics projects. We present a series of experimental results carried out with real robots and on publicly available real-world datasets. The re- sults demonstrate that our approach is able to update the representation efficiently and models the data consistently while keeping the memory requirement at a minimum.

An Efficient Probabilistic 3D Mapping Framework Based on Octrees

Image semantic segmentation is more and more being of interest for computer vision and machine learning researchers. Many applications on the rise need accurate and efficient segmentation mechanisms: autonomous driving, indoor navigation, and even virtual or augmented reality systems to name a few. This demand coincides with the rise of deep learning approaches in almost every field or application target related to computer vision, including semantic segmentation or scene understanding. This paper provides a review on deep learning methods for semantic segmentation applied to various application areas. Firstly, we describe the terminology of this field as well as mandatory background concepts. Next, the main datasets and challenges are exposed to help researchers decide which are the ones that best suit their needs and their targets. Then, existing methods are reviewed, highlighting their contributions and their significance in the field. Finally, quantitative results are given for the described methods and the datasets in which they were evaluated, following up with a discussion of the results. At last, we point out a set of promising future works and draw our own conclusions about the state of the art of semantic segmentation using deep learning techniques.

A Review on Deep Learning Techniques Applied to Semantic Segmentation.

https://rua.ua.es/dspace/bitstream/10045/75753/3/2018_Garcia-Garcia_etal_ApplSoftComp_accepted.pdf

A survey on deep learning techniques for image and video semantic segmentation

This paper presents a set of segmentation methods for various types of 3D point clouds. Segmentation of dense 3D data (e.g. Riegl scans) is optimised via a simple yet efficient voxelisation of the space. Prior ground extraction is empirically shown to significantly improve segmentation performance. Segmentation of sparse 3D data (e.g. Velodyne scans) is addressed using ground models of non-constant resolution either providing a continuous probabilistic surface or a terrain mesh built from the structure of a range image, both representations providing close to real-time performance. All the algorithms are tested on several hand labeled data sets using two novel metrics for segmentation evaluation.

On the segmentation of 3D LIDAR point clouds

This paper presents a method for pairwise 3D alignment which solves data association by matching scan segments across scans. Generating accurate segment associations allows to run a modified version of the Iterative Closest Point (ICP) algorithm where the search for point-to-point correspondences is constrained to associated segments. The novelty of the proposed approach is in the segment matching process which takes into account the proximity of segments, their shape, and the consistency of their relative locations in each scan. Scan segmentation is here assumed to be given (recent studies provide various alternatives [10], [19]). The method is tested on seven sequences of Velodyne scans acquired in urban environments. Unlike various other standard versions of ICP, which fail to recover correct alignment when the displacement between scans increases, the proposed method is shown to be robust to displacements of several meters. In addition, it is shown to lead to savings in computational times which are potentially critical in real-time applications.

/pdf/scan-segments-matching-for-pairwise-3d-alignment-1ori4ujsvm.pdf

Scan segments matching for pairwise 3D alignment

This paper presents an algorithm for segmenting 3D point clouds. It extends terrain elevation models by incorporating two types of representations: (1) ground representations based on averaging the height in the point cloud, (2) object models based on a voxelisation of the point cloud. The approach is deployed on Riegl data (dense 3D laser data) acquired in a campus type of environment and compared against six other terrain models. Amongst elevation models, it is shown to provide the best fit to the data as well as being unique in the sense that it jointly performs ground extraction, overhang representation and 3D segmentation. We experimentally demonstrate that the resulting model is also applicable to path planning.

https://www.shriharivasudevan.com/papers/iros2010-douillard.pdf

Hybrid elevation maps: 3D surface models for segmentation

This paper presents a novel local feature for 3D range image data called ‘the line image’. It is designed to be highly viewpoint invariant by exploiting the range image to efficiently detect 3D occupancy, producing a representation of the surface, occlusions and empty spaces. We also propose a strategy for defining keypoints with stable orientations which define regions of interest in the scan for feature computation. The feature is applied to the task of object classification on sparse urban data taken with a Velodyne laser scanner, producing good results.

An occlusion-aware feature for range images

This paper presents a 3D classifier that is shown to maintain performance whether trained with real sensor data from the field or purely trained with 3D geometric (Computer Aided Design, CAD, like) models (downloaded from the Internet for instance). The proposed classifier is a global 3D template matching technique which exploits the location of the ground surface for more accurate alignment. The segmentation and position of the ground is given by the segmentation technique in [7] (which does not assumed the ground to be flat). The proposed classifier outperforms Spin Image and Fast Point Feature Histogram (FPFH) based classifiers by up to 30% (the latter being tested at different scales), in the case of sparse 3D data acquired with a Velodyne sensor. In addition, the experimental results suggest that field samples may not be required in the training set of alignment-based 3D classifiers. This finding implies that the laborious task of gathering hand labelled field data for training may be avoidable for this type of classifier.

A. Quadros

Papers

On the segmentation of 3D LIDAR point clouds

Scan segments matching for pairwise 3D alignment

Hybrid elevation maps: 3D surface models for segmentation

An occlusion-aware feature for range images

A 3D classifier trained without field samples