Abstract
Extracting horizontal planes in heavily cluttered three-dimensional (3D) scenes is an essential procedure for many robotic applications. Aiming at the limitations of general plane segmentation methods on this subject, we present HoPE, a Horizontal Plane Extractor that is able to extract multiple horizontal planes in cluttered scenes with both organized and unorganized 3D point clouds. It transforms the source point cloud in the first stage to the reference coordinate frame using the sensor orientation acquired either by pre-calibration or an inertial measurement unit, thereby leveraging the inner structure of the transformed point cloud to ease the subsequent processes that use two concise thresholds for producing the results. A revised region growing algorithm named Z clustering and a principal component analysis (PCA)-based approach are presented for point clustering and refinement, respectively. Furthermore, we provide a nearest neighbor plane matching (NNPM) strategy to preserve the identities of extracted planes across successive sequences. Qualitative and quantitative evaluations of both real and synthetic scenes demonstrate that our approach outperforms several state-of-the-art methods under challenging circumstances, in terms of robustness to clutter, accuracy, and efficiency. We make our algorithm an off-the-shelf toolbox which is publicly available.
Highlights
Modeling and understanding three-dimensional (3D) scenes have been hot research topics in the computer vision and robotics communities
Aiming at the limitations of conventional methods, we present horizontal plane extractor (HoPE) (Horizontal Plane Extractor), a solution that robustly yet rapidly extracts multiple horizontal planes from both organized and unorganized 3D point clouds of cluttered scenes, and at the same time preserves the identities of each plane segments in successive frames
Since the RGB-D camera is one of the most used range sensors, we provide an algorithm in HoPE that generates point clouds using camera calibration parameters and synchronized image pairs obtained from the camera following the stereo registration paradigm
Summary
Modeling and understanding three-dimensional (3D) scenes have been hot research topics in the computer vision and robotics communities. It is hard to distinguish points of a plane from outliers belonging to the objects atop the plane or proximal planes of similar height Fitting such points to a global model as done by RANSAC [9,10,11], Hough Transform (HT) [12,13] and Expectation-Maximization (EM) methods [14] commonly leads to producing sloped planes as depicted, which is counter-factual regarding extracting horizontal planes and complicates the computation of robotic tasks involving the surface’s pose, such as retrieving the objects upon the surface, determining where to step on during stair climbing or orienting the end-effector for picking or placing objects Fitting such points to a global model as done by RANSAC [9,10,11], Hough Transform (HT) [12,13] and Expectation-Maximization (EM) methods [14] commonly leads to producing sloped planes as depicted in Figure 1b, which is counter-factual regarding extracting horizontal planes and complicates the computation of robotic tasks involving the surface’s pose, such as retrieving the objects upon the surface, determining where to step on during stair climbing or orienting the end-effector for picking or placing objects
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
