Skip to main content

A gardening robot for rose, hedge and topiary trimming

Deliverables

Second periodic report

This is the second technical and administration summary describing the project activities since the first report.

Project Exploitation Plan

This report summarises the activities planned by each of the partners for exploitation of the Foreground IPR generated during the project.

Evaluation and dissemination of manipulator and tools version 1 plus open-loop motion planning

This is a report describing and evaluating the motion planning and clipper control relative to a plan.

Proceedings of workshop 3: Gardening and agricultural robots

This is a brief summary of a conference workshop on the given topic, including the call for papers, the programme and a summary of the attendance.

Data management plan and framework

This report describes the expected data generated by the project, how it will be curated, stored and disseminated.

Project Dissemination Plan

This report summarises the plan for engaging the world with the project, including ROS packages, web presence, events, publication plans, IPR protection plans, commercialisation plans.

Wikipedia articles, popular science and presentation events

This is a brief report on the 'public' communication opportunities by the project.

Algorithmic speedup and efficient implementations, final evaluation of all 3D reconstructions

This report gives the performance analysis of the individual and combined tasks, identifies where speed and accuracy limitations are, and documents improvements made to the speed and accuracy of the algorithms.

Third periodic report

This is the final technical and administration summary describing the project activities since the second report.

Ground-truth data definitions and acquisition

This is a report of an ongoing process, describing the different types of visual and robotic data that expect to be acquired, how that data will be annotated to create a 'ground truth' dataset, how and when the datasets will be disseminated to the public.

Proceedings of workshop 2: 3D Visual Data Analysis

This is a brief summary of a conference workshop on the given topic, including the call for papers, the programme and a summary of the attendance.

Proceedings of workshop 1: 3D Visual Data Fusion

This is a brief summary of a conference workshop on the given topic, including the call for papers, the programme and a summary of the attendance.

Final evaluation and dissemination

This is a report describing and evaluating the final full arm, clipper and control package.

First periodic report

This is the first technical and administration summary describing the project activities since the start of the project.

Intrinsic image decomposition

This report describes the theory and evaluation of algorithms for the analysis and decomposition of the RGB-D video data into intrinsic components, albedo, hue and illumination.

Project website, blog, social media presence set up

This is a brief report summarising the development of the electronic media supporting the project.

Industry network meetings

Due also in M48 This is a brief summary of a set of industry events introducing the project, integrated robot and developed technology, including the call for papers, the programme and a summary of the attendance.

Project promotion material, fact sheets and press releases

These are a set of fliers describing the project, and a summary of where they were sent.

Component and system evaluation plan

This is a system integration planning document that specifies how each component will be tested individually and with what data. Then it specifies the planned tests for the whole integrated system. Evaluation metrics will be proposed.

Report on relevant audiences and stakeholders

This is a brief summary of exploitation opportunities for the project as a whole as well as individual components.

Implementation and evaluation of SLAM, 3D from binocular and motion stereo

This report describes the theory, implementation and performance of the individual 3D shape extraction and description components.

Dynamic reconstruction progress report

State of the art, design and preliminary results of acquiring 3D surface shape of ground, obstacles and hedges from video sequences.

Evaluation and dissemination of manipulator and tools version 2 plus closed-loop path and motion planning

This is a report describing and evaluating the motion planning and clipper control relative to a plan with sensing and feedback.

Data representation design and implementation, sensor calibration

This report describes the data integration structures for the fused 3D shape data, hypothesised objects, and garden structures, including the software interfaces to these structures. Also included is the description and evaluation of the methodology for linking the different sensor reference frames.

Three demonstrator plan

This report describes in detail the planned demonstrators (equipment, scenarios and tasks).

Missing data completion

This package analyses the raw and decomposed RGB-D video data and 3D shape data from other components to identify regions of incomplete data and hypothesis shape data completions (e.g. for shape and drivable region completion). A report describes the theory and performance of the package.

Scene flow software

A software package that produces the dense 3D scene motion description based on a moving sensor, as used for navigation and trimable target localisation. A report documents the theory and evaluation.

Dense deformable mapping software

A software package that produces the dense 3D scene shape description, including where and how hedge surfaces are deforming under contact with e.g. the clipper. A report documents the theory and evaluation.

Drivable garden regions

This package analyses the raw and decomposed RGB-D video data and 3D shape data from other components to recognise and localise drivable garden regions (e.g. paths, lawn). A report describes the theory and performance of the package.

Recognition and localisation of Garden Proto-Objects

This package analyses the raw and decomposed RGB-D video data and 3D shape data from other components to recognise and localise known garden objects (e.g. hedges, walls). A report describes the theory and performance of the package.

Arm and vehicle localisation

This software package (and associated report) describes the algorithms and performance behind the vehicle and cutting head locating itself relative to a map of navigable and cutting regions.

Clipping site recognition software

A software package that matches the dense 3D scene shape description to the desired shape model - localisation of the trimming location and the desired shape relative to the actual shape. A report documents the theory and evaluation.

Integrated 3D extraction package

This is the integrated software package combining the components from D3.2, providing the analysed scene data needed by the motion planning, navigation and hedge trimming control packages.

Searching for OpenAIRE data...

Publications

CNN Based Learning Using Reflection and Retinex Models for Intrinsic Image Decomposition

Author(s): Anil S. Baslamisli, Hoang-An Le, Theo Gevers
Published in: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018, Page(s) 6674-6683
DOI: 10.1109/cvpr.2018.00698

Hybrid Multi-camera Visual Servoing to Moving Target

Author(s): Hanz Cuevas-Velasquez, Nanbo Li, Radim Tylecek, Marcelo Saval-Calvo, Robert B. Fisher
Published in: 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2018, Page(s) 1132-1137
DOI: 10.1109/iros.2018.8593652

DUGMA: Dynamic Uncertainty-Based Gaussian Mixture Alignment

Author(s): Can Pu, Nanbo Li, Radim Tylecek, Bob Fisher
Published in: 2018 International Conference on 3D Vision (3DV), 2018, Page(s) 766-774
DOI: 10.1109/3dv.2018.00092

Camera Localization in Outdoor Garden Environments Using Artificial Landmarks

Author(s): Nicola Strisciuglio, Maria Leyva Vallina, Nicolai Petkov, Rafael Munoz Salinas
Published in: 2018 IEEE International Work Conference on Bioinspired Intelligence (IWOBI), 2018, Page(s) 1-6
DOI: 10.1109/iwobi.2018.8464139

Semantic Visual Localization

Author(s): Johannes L. Schonberger, Marc Pollefeys, Andreas Geiger, Torsten Sattler
Published in: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018, Page(s) 6896-6906
DOI: 10.1109/cvpr.2018.00721

Three for one and one for three:Flow, Segmentation, and Surface Normals

Author(s): Hoang-An Le, Anil Baslamisli, Thomas Mensink, Theo Gevers
Published in: BMVC 2018 : British Machine Vision Conference, 2018

3D Reconstruction Meets Semantics. Reconstruction Challenge 2017

Author(s): Torsten Sattler, Radim Tylecek, Thomas Brox, Marc Pollefeys, Robert B. Fisher
Published in: ICCV Workshop Venice, Italy, 2017., 2017

TrimBot2020: an outdoor robot for automatic gardening

Author(s): Nicola Strisciuglio, Radim Tylecek, Michael Blaich, Nicolai Petkov, Peter Bieber, Jochen Hemming, Eldert van Henten, Torsten Sattler, Marc Pollefeys, Theo Gevers, Thomas Brox, Robert B. Fisher
Published in: 50th Int. Symposium on Robotics, 2018, 2018

Trimbot Cutting Tools and Manipulator

Author(s): Jochen Hemming, Bart van Tuijl, Toon Tielen, Dejan Kaljaca, Joris IJsselmuiden, Eldert van Henten, Angelo Mencarelli, Pieter de Visser
Published in: APPIS Conf. 2018, 2018

Structured design of a novel end-effector for a bush trimming robot

Author(s): B.A.J. van Tuijl, A.P.M. Tielen, A. Mencarelli, J. Hemming
Published in: European Society of Agricultural Engineers (EurAgEng), 2018

3D Plane Labeling Stereo Matching with Content Aware Adaptive Windows

Author(s): Luis Horna, Robert B. Fisher
Published in: Proceedings of the 12th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications, 2017, Page(s) 162-171
DOI: 10.5220/0006105401620171

Embedded real-time multi-baseline stereo

Author(s): Dominik Honegger, Torsten Sattler, Marc Pollefeys
Published in: 2017 IEEE International Conference on Robotics and Automation (ICRA), 2017, Page(s) 5245-5250
DOI: 10.1109/ICRA.2017.7989615

Plane labeling trinocular stereo matching with baseline recovery

Author(s): Luis Horna, Robert B. Fisher
Published in: 2017 Fifteenth IAPR International Conference on Machine Vision Applications (MVA), 2017, Page(s) 17-20
DOI: 10.23919/MVA.2017.7986762

FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networks

Author(s): Eddy Ilg, Nikolaus Mayer, Tonmoy Saikia, Margret Keuper, Alexey Dosovitskiy, Thomas Brox
Published in: Conference on Computer Vision and Pattern Recognition (CVPR), 2017

Comparative Evaluation of Hand-Crafted and Learned Local Features

Author(s): Johannes L. Schonberger, Hans Hardmeier, Torsten Sattler, Marc Pollefeys
Published in: Conference on Computer Vision and Pattern Recognition (CVPR), 2017, Page(s) 1482-1491

DeMoN: Depth and Motion Network for Learning Monocular Stereo

Author(s): Benjamin Ummenhofer, Huizhong Zhou, Jonas Uhrig, Nikolaus Mayer, Eddy Ilg, Alexey Dosovitskiy, Thomas Brox
Published in: Conference on Computer Vision and Pattern Regognition, 2017, Page(s) 5038-5047

A Multi-View Stereo Benchmark with High-Resolution Images and Multi-Camera Videos

Author(s): Thomas Schops, Johannes L. Schonberger, Silvano Galliani, Torsten Sattler, Konrad Schindler, Marc Pollefeys, Andreas Geiger
Published in: Conference on Computer Vision and Pattern Recognition (CVPR), 2017, Page(s) 3260-3269

Joint Learning of Intrinsic Images and Semantic Segmentation

Author(s): Anil S. Baslamisli, Thomas T. Groenestege, Partha Das, Hoang-An Le, Sezer Karaoglu, Theo Gevers
Published in: Computer Vision – ECCV 2018 - 15th European Conference, Munich, Germany, September 8–14, 2018, Proceedings, Part VI, Issue 11210, 2018, Page(s) 289-305
DOI: 10.1007/978-3-030-01231-1_18

Learning Priors for Semantic 3D Reconstruction

Author(s): Ian Cherabier, Johannes L. Schönberger, Martin R. Oswald, Marc Pollefeys, Andreas Geiger
Published in: Computer Vision – ECCV 2018 - 15th European Conference, Munich, Germany, September 8–14, 2018, Proceedings, Part XII, Issue 11216, 2018, Page(s) 325-341
DOI: 10.1007/978-3-030-01258-8_20

Uncertainty Estimates and Multi-hypotheses Networks for Optical Flow

Author(s): Eddy Ilg, Özgün Çiçek, Silvio Galesso, Aaron Klein, Osama Makansi, Frank Hutter, Thomas Brox
Published in: Computer Vision – ECCV 2018 - 15th European Conference, Munich, Germany, September 8–14, 2018, Proceedings, Part VII, Issue 11211, 2018, Page(s) 677-693
DOI: 10.1007/978-3-030-01234-2_40

Occlusions, Motion and Depth Boundaries with a Generic Network for Disparity, Optical Flow or Scene Flow Estimation

Author(s): Eddy Ilg, Tonmoy Saikia, Margret Keuper, Thomas Brox
Published in: Computer Vision – ECCV 2018 - 15th European Conference, Munich, Germany, September 8–14, 2018, Proceedings, Part XII, Issue 11216, 2018, Page(s) 626-643
DOI: 10.1007/978-3-030-01258-8_38

VSO: Visual Semantic Odometry

Author(s): Konstantinos-Nektarios Lianos, Johannes L. Schönberger, Marc Pollefeys, Torsten Sattler
Published in: Computer Vision – ECCV 2018 - 15th European Conference, Munich, Germany, September 8-14, 2018, Proceedings, Part IV, Issue 11208, 2018, Page(s) 246-263
DOI: 10.1007/978-3-030-01225-0_15

Brain-Inspired Robust Delineation Operator

Author(s): Nicola Strisciuglio, George Azzopardi, Nicolai Petkov
Published in: Computer Vision – ECCV 2018 Workshops - Munich, Germany, September 8-14, 2018, Proceedings, Part III, Issue 11131, 2019, Page(s) 555-565
DOI: 10.1007/978-3-030-11015-4_41

3DContextNet: K-d Tree Guided Hierarchical Learning of Point Clouds Using Local and Global Contextual Cues

Author(s): Wei Zeng, Theo Gevers
Published in: Computer Vision – ECCV 2018 Workshops - Munich, Germany, September 8-14, 2018, Proceedings, Part III, Issue 11131, 2019, Page(s) 314-330
DOI: 10.1007/978-3-030-11015-4_24

DeepTAM: Deep Tracking and Mapping

Author(s): Huizhong Zhou, Benjamin Ummenhofer, Thomas Brox
Published in: Computer Vision – ECCV 2018 - 15th European Conference, Munich, Germany, September 8-14, 2018, Proceedings, Part XVI, Issue 11220, 2018, Page(s) 851-868
DOI: 10.1007/978-3-030-01270-0_50

What Makes Good Synthetic Training Data for Learning Disparity and Optical Flow Estimation?

Author(s): Nikolaus Mayer, Eddy Ilg, Philipp Fischer, Caner Hazirbas, Daniel Cremers, Alexey Dosovitskiy, Thomas Brox
Published in: International Journal of Computer Vision, Issue 126/9, 2018, Page(s) 942-960, ISSN 0920-5691
DOI: 10.1007/s11263-018-1082-6

Consistent Semantic Annotation of Outdoor Datasets via 2D/3D Label Transfer

Author(s): Radim Tylecek, Robert Fisher
Published in: Sensors, Issue 18/7, 2018, Page(s) 2249, ISSN 1424-8220
DOI: 10.3390/s18072249

Point Light Source Position Estimation From RGB-D Images by Learning Surface Attributes

Author(s): Sezer Karaoglu, Yang Liu, Theo Gevers, Arnold W. M. Smeulders
Published in: IEEE Transactions on Image Processing, Issue 26/11, 2017, Page(s) 5149-5159, ISSN 1057-7149
DOI: 10.1109/TIP.2017.2731619

Color Constancy by GANs: An Experimental Survey.

Author(s): P. Das, A. S. Baslamisli, Y. Liu, S. Karaoglu, and T. Gevers.
Published in: 2018