Abstract-With the advent of new, low-cost 3D sensing hardware such as the Kinect, and continued efforts in advanced point cloud processing, 3D perception gains more and more importance in robotics, as well as other fields.In this paper we present one of our most recent initiatives in the areas of point cloud perception: PCL (Point Cloud Library -http://pointclouds.org). PCL presents an advanced and extensive approach to the subject of 3D perception, and it's meant to provide support for all the common 3D building blocks that applications need. The library contains state-ofthe art algorithms for: filtering, feature estimation, surface reconstruction, registration, model fitting and segmentation. PCL is supported by an international community of robotics and perception researchers. We provide a brief walkthrough of PCL including its algorithmic capabilities and implementation strategies.

No abstract

No abstract

ENVIRONMENT models serve as important resources for an autonomous robot by providing it with the necessary task-relevant information about its habitat. Their use enables robots to perform their tasks more reliably, flexibly, and efficiently. As autonomous robotic platforms get more sophisticated manipulation capabilities, they also need more expressive and comprehensive environment models: for manipulation purposes their models have to include the objects present in the world, together with their position, form, and other aspects, as well as an interpretation of these objects with respect to the robot tasks.This thesis proposes Semantic 3D Object Models as a novel representation of the robot's operating environment that satisfies these requirements and shows how these models can be automatically acquired from dense 3D range data. The thesis contributes in two important ways to the research area acquisition of environment models.The first contribution is a novel framework for Semantic 3D Object Model acquisition from Point Cloud Data. The functionality of this framework includes robust alignment and integration mechanisms for partial data views, fast segmentation into regions based on local surface characteristics, and reliable object detection, categorization, and reconstruction. The computed models are semantic in that they infer structures in the data that are meaningful with respect to the robot task. Examples of such objects are doors and handles, supporting planes, cupboards, walls, or movable smaller objects. The second key contribution is point cloud representations based on 3D point feature histograms (3D-PFHs), which model the local surface geometry for each point. 3D-PFHs distinguish themselves from alternative 3D feature representations in that they are very fast to compute, robust against variations in pose and sampling density, and cope well with noisy sensor data. Their use substantially improves the quality of the Semantic 3D Object Models acquired, as well as the speed with which they are computed. 3D-PFHs come with specific software tools that allow for the learning of surface characteristics based on their underlying geometry, the assembly of most distinctive 3D points from a given cloud, as well as limited view-invariant correspondence search for 3D registration. IIIThe contributions presented in this thesis have been fully implemented and empirically evaluated on different robots performing different tasks in different environments. The first demonstration relates to the problem of cleaning tables by disposing the objects on them into a garbage bin with a personal robotic assistant in the presence of humans in its working space. The framework for Semantic 3D Object Model acquisition is demonstrated and used to construct dynamic 3D collision maps, annotate the surrounding world with semantic labels, and extract object clusters supported by tables in real-time performance. The second demonstration presents an on-the-fly model acquisition system for door and handle identification from n...

No abstract