Software Releases

• Bonnetal: Easy-to-use Deep-learning Training and Deployment Pipeline
• ReFusion: 3D Reconstruction in Dynamic Environments for RGB-D Cameras Exploiting Residuals
• Bonnet: Tensorflow Convolutional Semantic Segmentation Pipeline
• Fast Change Detection
• MPR: General Framework for Flexible Multi-Cue Photometric Point Cloud Registration
• Visual place recognition: Extended Version of Visual Place Recognition using Hashing
• Online place recognition: a library for graph-based visual place recognition in changing environments
• Depth Clustering: fast and robust Velodyne laser scan segmentation library
• Software releases from the ROVINA Project
• BACS: bundle adjustment for camera systems with omnidirectional cameras (Matlab)
• FastFPS: Fast marching for robust surface segmentation (Matlab)
• Förstner operator: an interest point detector (Matlab)
• SFOP: scale-invariant keypoint detector (Matlab, C++)
• Completeness of Detectors: a measurement scheme for the completeness of a set of image feature detectors (Matlab)
• IVM: Import Vector Machine classifier (Matlab, C++)
• SUGR: geometric reasoning with uncertain elements of projective geometry (Java, Perl)
• a general adjustment procedure for constrained observations with singular covariance matrices, e.g. projective geometry entities (Matlab)
• an annotation tool for images implemented in Matlab, used for the eTRIMS Image Database
• Depth Streaming using H.264 (C++, shell scripts)
• C Implementation of the Hungarian Method: libhungarian-v0.1.3.tgz (21.09.2015)
  C-implementation of the Hungarian Method: finding the optimal assignment (assigning a set of jobs to a set of machines) in O(n^3), where n=max{#jobs, #machines}. The implementation is a sligntly enhanced version of the implementation provided by the
  Stanford GraphBase. See also: Stanford GraphBase, Hungarian Method by Brian Gerkey.
• HOG-Man – Hierarchical Optimization for Pose Graphs on Manifolds: HOG-Man is our new back-end for graph-based SLAM which is also designed for online applications.
• TORO – Tree-based netwORk Optimizer: TORO is a solution to compute low error configurations
  of constraint networks used in SLAM.
• CARMEN: Download the latest release of CARMEN.
  Binary as well as source code/API documentation of CARMEN.
• GMapping: Get GMapping, our SLAM approach using a Rao-Blackwellized particle filter (hosted at OpenSLAM.org).

Misc

• Book: Photogrammetric Computer Vision by Wolfgang Förstner and Bernhard Wrobel, 2016
• homepages for our activities within ISPRS (International Society of Photogrammetry and Remote Sensing):
  • Commission III homepage, 2004-2008
  • PCV 2006 (Commission III symposium)
  • Intercommission Working Group III/VII – Pattern Recognition for Remote Sensing, 2008-2012

Institute of Geodesy
and Geoinformation

Faculty of Agriculture

University of Bonn