Robots are usually equipped with 3D range sensors such as laser line scanners (LLSs) or lidars. These acquire a full scan in by manner while the robot is motion. All lines can be referred to common coordinate frame using data from inertial sensors. However, errors noisy measurements and inaccuracies extrinsic parameters between scanner also projected onto shared frame. This causes deformation f...

Recently, 3D deep learning models have been shown to be susceptible adversarial attacks like their 2D counterparts. Most of the state-of-the-art (SOTA) perform perturbation point clouds. To reproduce these in physical scenario, a generated cloud needs reconstructed mesh, which leads significant drop its effect. In this paper, we propose strong attack named Mesh Attack address problem by directl...

We present a new method for extracting the feature primitives of a 3D object directly from the point cloud of its surface scan. The objective is to identify a subset of points that provides the same information about the structure and the topology of the object geometry as the point cloud itself. The entire process is carried out with the least human intervention possible. The only information ...

This paper presents a novel surface reconstruction algorithm based on 3D point cloud sliced by a series of planes vertical to Z-axis. First, the density and distribution of point cloud is analyzed in vertical direction Z-axis and a stratified point cloud model is put forward to present an object feature, in which point cloud is sliced with irregular intervals. An approach of stratified point cl...

Point cloud filtering is an important step in Digital Terrain Model (DTM) production. Despite the fact that a great body of research has been conducted this area so far, there are still some problems have not yet solved, especially complex terrains. The use user-defined parameters within presented point methods, and difficulty parameter estimation parallel to increase topography slope above-gro...

Abstract While a popular representation of 3D data, point clouds may contain noise and need filtering before use. Existing cloud methods either cannot preserve sharp features or result in uneven distributions the filtered output. To address this problem, paper introduces method that considers both distribution feature preservation during filtering. The key idea is to incorporate repulsion term ...

Efficient learning of 3D shape representation from point cloud is one the biggest requirements in computer vision. In recent years, convolutional neural networks have achieved great success 2D image learning. However, unlike images that a Euclidean structure, clouds are irregular since neighbors each node inconsistent. Many studies tried to develop various graph overcome this problem and achiev...

aerosol formation due to ions in the atmosphere is not well understood, although it seems to play an important role in cloud formation. the ions in the atmosphere can enhance cloud formation and can increase nucleation as well as impeding the evaporation from the surfaces of the droplets. recently there has been some experimental work to find the relation between ions concentrations and cloud i...