Incremental 3D Cuboid Modeling with Drift Compensation

Abstract

This paper presents a framework of incremental 3D cuboid modeling by using the mapping results of an RGB-D camera based simultaneous localization and mapping (SLAM) system. This framework is useful in accurately creating cuboid CAD models from a point cloud in an online manner. While performing the RGB-D SLAM, planes are incrementally reconstructed from a point cloud in each frame to create a plane map. Then, cuboids are detected in the plane map by analyzing the positional relationships between the planes, such as orthogonality, convexity, and proximity. Finally, the position, pose, and size of a cuboid are determined by computing the intersection of three perpendicular planes. To suppress the false detection of the cuboids, the cuboid shapes are incrementally updated with sequential measurements to check the uncertainty of the cuboids. In addition, the drift error of the SLAM is compensated by the registration of the cuboids. As an application of our framework, an augmented reality-based interactive cuboid modeling system was developed. In the evaluation at cluttered environments, the precision and recall of the cuboid detection were investigated, compared with a batch-based cuboid detection method, so that the advantages of our proposed method were clarified.

Evaluation of the use of box size priors for 6D plane segment tracking from point clouds with applications in cargo packing

“…Such information, readily available in manual packing applications, has not been previously incorporated into 6D plane segment tracking algorithms. The choice of image type for processing, spatial sensor array configuration, and processing approach were selected for their suitability in industrial packing scenarios, specifically: (1) depth maps, point clouds generated from these maps, and their derived features exhibit low sensitivity to object texture, (2) the inside-out sensor array configuration [27] facilitates extensive coverage of the workspace and enables the exploration of less occluded viewpoints during workspace exploration, and (3) the incremental approach enables refinement of detections as frames of the same object accumulate; moreover, it allows for the integration of multiple detections into a single global map of detected elements [5].…”

“…1. The increase in odometric error in techniques that utilize environment reconstruction [5][6][7] leading to a subsequent decrease in tracking performance. This is a consequence of the extensive area that must be scanned during manual packing operations.…”

“…Moreover, the method is sensitive to the texture and reflectivity level of the object to be tracked, properties that cannot be controlled in the cargo packing operation. Finally, SLAM-based tracking algorithms have been validated in multiple-box scenarios [5,24,25]. Although their precision may not be the highest, evidence indicates that performance can be improved by incorporating prior knowledge of the tracked objects [26].…”

Direct and accurate feature extraction from 3D point clouds of plants using RANSAC

“…Cuboid shape: As shown in Fig. 3(a), a rectangular cuboid is represented by eight corner points C = {C 1 , ..., C 8 } [48]. To generate these parameters, we need three perpendicular planes.…”

Voxar Labs

“…As examples of the lab effort in long term research, two are very representative: the 3D Tracking area and the Sicure project ( Figure 5). The 3D Tracking research area has powered international cooperation projects (i.e., INRIA/UChile/ UFPE and TUGraz/Kyushu/UFPE), industry projects and patents, various collaborative PostDoc and Ph.D.'s (e.g., with Microsoft Research, EPFL, TUGraz), research papers and prizes in high quality venues (ISMAR [5] [6], ICRA [7], C&G [8][9], IJCNN [10], Sensors [11]).…”