What lies behind: Recovering hidden shape in dense mapping

In mobile robotics applications, generation of accurate static maps is encumbered by the presence of ephemeral objects such as vehicles, pedestrians, or bicycles. We propose a method to process a sequence of laser point clouds and back-fill dense surfaces into gaps caused by removing objects from the scene - a valuable tool in scenarios where resource constraints permit only one mapping pass in a particular region. Our method processes laser scans in a three-dimensional voxel grid using the Truncated Signed Distance Function (TSDF) and then uses a Total Variation (TV) regulariser with a Kernel Conditional Density Estimation (KCDE) “soft” data term to interpolate missing surfaces. Using four scenarios captured with a push-broom 2D laser, our technique infills approximately 20 m2 of missing surface area for each removed object. Our reconstruction's median error ranges between 5.64 cm - 9.24 cm with standard deviations between 4.57 cm - 6.08 cm.

[1]  Alexander G. Gray,et al.  Fast Nonparametric Conditional Density Estimation , 2007, UAI.

[2]  Andrew Zisserman,et al.  Multiple View Geometry , 1999 .

[3]  Paul Newman,et al.  Dense and Swift Mapping with Monocular Vision , 2015, FSR.

[4]  Paul Newman,et al.  Practice makes perfect? Managing and leveraging visual experiences for lifelong navigation , 2012, 2012 IEEE International Conference on Robotics and Automation.

[5]  Andrew W. Fitzgibbon,et al.  KinectFusion: Real-time dense surface mapping and tracking , 2011, 2011 10th IEEE International Symposium on Mixed and Augmented Reality.

[6]  Ingmar Posner,et al.  Wrong Today, Right Tomorrow: Experience-Based Classification for Robot Perception , 2015, FSR.

[7]  Steve Marschner,et al.  Filling holes in complex surfaces using volumetric diffusion , 2002, Proceedings. First International Symposium on 3D Data Processing Visualization and Transmission.

[8]  Paul Newman,et al.  Distraction suppression for vision-based pose estimation at city scales , 2013, 2013 IEEE International Conference on Robotics and Automation.

[9]  Manuel Menezes de Oliveira Neto,et al.  A hole-filling strategy for reconstruction of smooth surfaces in range images , 2003, 16th Brazilian Symposium on Computer Graphics and Image Processing (SIBGRAPI 2003).

[10]  Shahram Izadi,et al.  MonoFusion: Real-time 3D reconstruction of small scenes with a single web camera , 2013, 2013 IEEE International Symposium on Mixed and Augmented Reality (ISMAR).

[11]  F. Kahl.  Multiple View Geometry and the -norm , 2005 .

[12]  Tim Weyrich,et al.  Real-Time 3D Reconstruction in Dynamic Scenes Using Point-Based Fusion , 2013, 2013 International Conference on 3D Vision.

[13]  Paul Newman,et al.  Dense mono reconstruction: Living with the pain of the plain plane , 2015, 2015 IEEE International Conference on Robotics and Automation (ICRA).

[14]  Paul Newman,et al.  BOR ^2 2 G: Building Optimal Regularised Reconstructions with GPUs (in Cubes) , 2015, FSR.

[15]  Horst Bischof,et al.  CP-Census: A Novel Model for Dense Variational Scene Flow from RGB-D Data , 2014, BMVC.

[16]  Marc Pollefeys,et al.  A Patch Prior for Dense 3D Reconstruction in Man-Made Environments , 2012, 2012 Second International Conference on 3D Imaging, Modeling, Processing, Visualization & Transmission.

[17]  Paul Newman,et al.  Too much TV is bad: Dense reconstruction from sparse laser with non-convex regularisation , 2015, 2015 IEEE International Conference on Robotics and Automation (ICRA).

[18]  William E. Lorensen,et al.  Marching cubes: A high resolution 3D surface construction algorithm , 1987, SIGGRAPH.

[19]  Kevin Curran,et al.  OpenStreetMap , 2012, Int. J. Interact. Commun. Syst. Technol..

[20]  Paul Newman,et al.  What could move? Finding cars, pedestrians and bicyclists in 3D laser data , 2012, 2012 IEEE International Conference on Robotics and Automation.

[21]  Paul Newman,et al.  LAPS - localisation using appearance of prior structure: 6-DoF monocular camera localisation using prior pointclouds , 2012, 2012 IEEE International Conference on Robotics and Automation.

[22]  Hugh F. Durrant-Whyte,et al.  Gaussian Process modeling of large scale terrain , 2009, ICRA.

[23]  Marc Levoy,et al.  A volumetric method for building complex models from range images , 1996, SIGGRAPH.

[24]  John J. Leonard,et al.  Real-time large-scale dense RGB-D SLAM with volumetric fusion , 2014, Int. J. Robotics Res..

[25]  Antonin Chambolle,et al.  A First-Order Primal-Dual Algorithm for Convex Problems with Applications to Imaging , 2011, Journal of Mathematical Imaging and Vision.

[26]  丸山 徹.  Convex Analysisの二,三の進展について , 1977 .