Paris-CARLA-3D is a dataset of several dense colored point clouds of outdoor environments built by a mobile LiDAR and camera system. The data are composed of two sets with synthetic data from the open source CARLA simulator (700 million points) and real data acquired in the city of Paris (60 million points), hence the name Paris-CARLA-3D. One of the advantages of this dataset is to have simulated the same LiDAR and camera platform in the open source CARLA simulator as the one used to produce the real data. In addition, manual annotation of the classes using the semantic tags of CARLA was performed on the real data, allowing the testing of transfer methods from the synthetic to the real data. The objective of this dataset is to provide a challenging dataset to evaluate and improve methods on difficult vision tasks for the 3D mapping of outdoor environments: semantic segmentation, instance segmentation, and scene completion. For each task, we describe the evaluation protocol as well as the experiments carried out to establish a baseline.

Localization in indoor environments is a technique that estimates the robot's pose by fusing data from onboard motion sensors with readings of the environment, in our case obtained by scan matching point clouds captured by a low-cost Kinect depth camera. We develop both an invariant extended Kalman filter (IEKF)-based and a multiplicative extended Kalman filter-based solution to this problem. The two designs are successfully validated in experiments and demonstrate the advantage of the IEKF design.

International audienceVirtual surgery simulators show a lot of advantages in the world of surgery training, where they allow to improve the quality of surgeons' gesture. One of the current major technical difficulties for the development of surgery simulation is the possibility to perform a real-time computation of soft tissue deformation by considering the accurate modeling of their mechanical properties. However today, few models are available, they are still time consuming and limited in number of elements by algorithm complexity. We present in this paper a new method and framework that we call 'HEML' (Hyper-Elastic Mass Links), which is particularly fast. It is derived from the finite element method, can handle visco-hyperelastic and large deformation modeling. Although developed initially for medical applications, the HEML method can be used for any numerical computation of hyperelastic material deformations based on a tetrahedral mesh. A comparison with existing methods shows a much faster speed. A comparison with Mass-Spring methods, that are particularly fast but not realistic, shows that they can be considered as a degenerate case of the HEML framework

This paper investigates the use of depth images as localisation sensors for 3D map building. The localisation information is derived from the 3D data thanks to the ICP (Iterative Closest Point) algorithm. The covariance of the ICP, and thus of the localization error, is analysed, and described by a Fisher Information Matrix. It is advocated this error can be much reduced if the data is fused with measurements from other motion sensors, or even with prior knowledge on the motion. The data fusion is performed by a recently introduced specific extended Kalman filter, the so-called Invariant EKF, and is directly based on the estimated covariance of the ICP. The resulting filter is natural, and is proved to possess strong properties. Experiments with a Kinect sensor and a three-axis gyroscope prove clear improvement in the accuracy of the localization, and thus in the accuracy of the built 3D map.

With the variety of measurement techniques available on the market today, fusing multi-source complementary information into one dataset is a matter of great interest. Target-based, point-based and feature-based methods are some of the approaches used to place data in a common reference frame by estimating its corresponding transformation parameters. This paper proposes a new linear feature-based method to perform accurate registration of point clouds, either in 2D or 3D. A two-step fast algorithm called Robust Line Matching and Registration (RLMR), which combines coarse and fine registration, was developed. The initial estimate is found from a triplet of conjugate line pairs, selected by a RANSAC algorithm. Then, this transformation is refined using an iterative optimization algorithm. Conjugates of linear features are identified with respect to a similarity metric representing a line-to-line distance. The efficiency and robustness to noise of the proposed method are evaluated and discussed. The algorithm is valid and ensures valuable results when pre-aligned point clouds with the same scale are used. The studies show that the matching accuracy is at least 99.5%. The transformation parameters are also estimated correctly. The error in rotation is better than 2.8% full scale, while the translation error is less than 12.7%.