3D is here: Point Cloud Library (PCL)

By Radu Bogdan Rusu, Steve B. Cousins. 2011

5,388 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computational Mechanics, Geology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Computational Mechanics

In Geology

Citations in the corpus

RGB-D datasets using microsoft kinect or similar sensors: a survey (2016, background): “More details and publications about PCL can be found in [74].”
Dynamic Edge-Conditioned Filters in Convolutional Neural Networks on Graphs (2017, positive): “For a single input point cloud P , a pyramid of downsampled point clouds P (h) is obtained by the VoxelGrid algorithm [30], which overlays a grid of resolution r over the point cloud and replaces all points within a voxel with their centroid (and thus maintains subvoxel accuracy).”
Deep Sets (2017, methodology): “We produce point-clouds with 100, 1000 and 5000 particles each ( x, y, z -coordinates) from the mesh representation of objects using the point-cloud-library’s sampling routine (Rusu & Cousins, 2011).”
Learning 3D Object Categories by Looking Around Them (2017, methodology): “We extract this information using off-theshelf methods: the structure-from-motion (SFM) algorithm COLMAP for RGB sequences [33, 34], and an open-source implementation [31] of KinectFusion (KF) [24] for RGBD sequences.”
A study of the effect of noise and occlusion on the accuracy of convolutional neural networks applied to 3D object recognition (2017, background): “Those sensors, together with community efforts in terms of software like10 the Point Cloud Library (PCL)[1] project, democratized 3D information, which is now easy to obtain and process.”
6-DOF GraspNet: Variational Grasp Generation for Object Manipulation (2019, positive): “We filter the measured point cloud, remove the table plane and cluster the remaining points [26].”
Learning to place new objects in a scene (2012, methodology): “• Curvature Histograms: We estimated the curvature of every point using Point Cloud Library (Rusu and Cousins, 2011), and then compute its 12-bin histogram.”
Object discovery in 3D scenes via shape analysis (2013, methodology): “With tools such as Kinect Fusion [1], Kintinuous [2], and Open Source alternatives in the Point Cloud Library [3], it is now possible to collect detailed 3D meshes of entire scenes in real-time.”
A Comprehensive Performance Evaluation of 3D Local Feature Descriptors (2015, methodology): “1 (Rusu and Cousins 2011; Aldoma et al. 2012a), while the others were implemented in Matlab 2011b.”
Robobarista: Object Part Based Transfer of Manipulation Trajectories from Crowd-Sourcing in 3D Pointclouds (2015, methodology): “intermediate step of detecting object part does not guarantee successful transfers, we built an object part classifier using multiclass SVM [56] on point-cloud features including local shape features [26], histogram of curvatures [40], and distribution of points.”
Unsupervised Learning of Depth and Ego-Motion from Monocular Video Using 3D Geometric Constraints (2018, methodology): “3While we describe a point-to-point distance, we use the more powerful point-to-plane distance [4] as in the Point Cloud Library [24].”
3D Object Detection with Multiple Kinects (2012, methodology): “We calibrate the extrinsic matrix between diﬀerent Kinect cameras using ICP from the PCL [21].”
Fully Convolutional Geometric Features (2019, methodology): “We ran other baselines on the same workstation us-ing PCL 1.8 [25] to test SHOT, USC, and Spin Image, and Open3D [38] to test FPFH.”
A Hand Motion-guided Articulation and Segmentation Estimation (2020, methodology): “We used the implementations in [24] for the Euclidean clustering for segmentation, Voxel Grid Filter, and kd-tree.”
Analysis and Observations From the First Amazon Picking Challenge (2016, methodology): “As for the software used, 75% (18) of the teams used the “Point Cloud Library” (PCL) [8], 67% (16) used the “Open Computer Vision Library” (OpenCV) [9], and 33% (8) report using their “own” tools (Figure 9).”
Real-Time Perception Meets Reactive Motion Generation (2017, methodology): “This transformed point cloud is then converted to an occupancy grid (with 2 cm resolution), which is ﬁltered for statistical outliers using PCL [28].”
Transformation Autoregressive Networks (2018, methodology): “We produce point-clouds with 1000 particles each ( x,y,z -coordinates) from the mesh representation of objects using the point-cloud-library’s sampling routine (Rusu & Cousins, 2011).”