KI-iREPro: Artificial Intelligence for Innovative Yield Prediction in Viticulture

Funding: Federal Ministry of Food and Agriculture (BMEL), 2021 - 2024

Aim: Yield Prediction in Viticulture based on Data from Sensors, Process Management and Environment

Role: Principal investigator of the work package: Machine Learning, Feature Selection and Explainability

Methods: Deep Learning, Feature Selection, Importance Metrics

Partners: Julius Kühn Institute of Grapevine Breeding, Siebeldingen, Fraunhofer Institute of Optronics, System Technologies and Image Exploitation (IOSB), Vineyard Cloud, Deutsches Weintor eG

Automatized visual monitoring and image analyses - a sub-project of the AMMOD consortium

Funding: Federal Ministry of Education and Research (BMBF), 2019 - 2023

Aim: Object tracking in depth data from camera traps

Role: Principal investigator of the work package: Processing and interpreting depth data

Methods: Segmentation of depth data, object tracking in depth data, deep learning

Partners: Faculty of Informatics, Techn. Univ. Munich (Prof. Dr. B. Radig), Department of Mathematic and Computer Science, Univ. Jena (Prof. Dr. J. Denzler, Dr. P. Bodesheim), Faculty of Engineering, Univ. Bristol (Dr. T. Burghardt), Max Planck Institute for Evolutionary Anthropology, Leibzig (Dr. H. Kühl), The project is part of AMMOD, a research consortium for constructing Automated Multisensor Stations for Monitoring of Biodiversity (Speaker: Prof. Dr. W. Wägele)

Computational Sensing with Applications to Plant Breeding

Funding: Inst. of Comp. Science, Bonn Univ., 2017 - 2019

Aim: Exploiting approaches to light separation for plant phenotyping

Role: Principal investigator together with Prof. Dr. Matthias Hullin

Methods: Illumination Control, Subsurface Scattering, Image Decomposition and Manipulation, Object Recognition, ConvNets

Partners: Inst. of Computer Science II, Bonn Univ.

Automated Evaluation and Comparision of Grape Genotypes with respect to Cluster Architecture

Funding: German Research Council (DFG), 2016 - 2019

Aim: QTL mapping, marker development and identification of candidate genes for important traits to achieve Botrytis resilience; automated model-based high-throughput phenotyping of cluster architecture traits using high-resolution optical 3D scanning data

Role: Principal investigator of the modeling subgroup: interpretation of sensor data, pattern recognition, classification

Methods: Procedural Modeling, L-Systems, Relational Growth Grammars, CSG, Probabilistic Sampling

Partners: Julius Kühn Institute of Grapevine Breeding, Siebeldingen (Prof. Dr. R. Töpfer)

Object Classification in 3D Laser Range Data

Funding: Fraunhofer Gesellschaft (FhG), 2008 - 2015

Aim: Object Classification in 3D Laser Range Data of Dynamic Scenes

Role: Principal investigator together with Prof. Dr. A. B. Cremers

Methods: range data segmentation, feature selection, logistic regression, spectral hashing

Partners: Fraunhofer Institute for Communication, Information Processing and Ergonomics (FKIE), Prof. Dr. P. Martini, Dr. D. Schulz

Derivation of 3D Plant Architectures - a sub-project of CROP.SENSe.net

Funding: German Ministry of Education and Research (BMBF) within the scope of the competitive grants program Networks of excellence in agricultural and nutrition research (Funding code: 0315529) and from the European Union Funds for regional development (Funding code: z1011bc001a), 2010 - 2014

Aim: Multi-scale structural 3D plant reconstruction in vinegrowing

Role: Principal investigator of the modeling subgroup: interpretation of sensor data, pattern recognition, classification

Methods: Procedural modeling, L-Systems, relational growth grammars, CSG, probabilistic sampling

Partners: Julius Kühn Institute of grapevine Breeding, Siebeldingen (Prof. Dr. R. Töpfer), Institute of Geodesy and Geoinformation, Bonn Univ. (Prof. Dr. W. Förstner, Prof. Dr. H. Kuhlmann), Faculty of Informatics, Techn. Univ. Munich (Prof. Dr. D. Cremers). The project is part of Crop.Sense.net, an agronomic competence network for networking sensor technology R&D for crop breeding and management

ABIS: Automated Bee Identification System

Funding: German Research Council (DFG), 1996 - 2001, German Ministry of Education and Research (BMBF) 2000 - 2003

Aim: Automated identification of bee species by image analysis of their wings

Role: Principal investigator of the computer science group: image processing, pattern recognition, classification, geo-referencing, visualization

Methods: Geometrical image analysis, template matching, affine projection, discriminant analysis, kernel functions, GIS

Partners: Institute of Agricultural Zoology and Bee Keeping, Prof. Dr. W. Drescher, Prof. Dr. D. Wittman, Dr. S. Schröder: Taxonomical Research, Identification of Training Data, Zoological Evaluation

SMB: Automated 3D Reconstruction of Buildings from Digital Images

Funding: German Research Council (DFG), 1995 - 2000

Aim: Fully automated reconstruction of buildings from digital images

Role: Principal investigator of the modeling subgroup: modeling in 3D and 2D, model aggretation in 3D, visualization, animation

Methods: Hybrid Modeling (CSG/B-Rep), Aspect Graphs, Shape Grammars, Tecture Mapping

Partners: Department of Computer Science III, Prof. Cremers, Dr. Plümer: Validation by Constraint Logic Programming, Institute of Photogrammetry, Prof. Förstner: Image Processing, Pattern Recognition, Reconstruction of Building Corners

The project was part of the bundle project Semantic Modeling and Extraction of Spatial Objects from Images and Maps

NESA: Low-Energy Solar Architecture

Funding: 1995 - 1996 by the Ministry of Economy and Energy of the State North Rhine-Westphalia

Aim: Computer Aided Desing and Evaluation of Low-Energy-Solar-Architecture

Role: Principal investigator of the modeling group: modeling in 3D and 2D, user interface, visualization, animation

Principle: Computer-Aided 3D-Construction of Low Energy Buildings

Methods: Boundary Representation (B-Rep), User Interface, Visualization

Partners: Buero of Low Energy Building Construction, Dipl.-Ing. A. Lohr, Department of Computer Science III, Prof. Cremers: Modeling and User Interface of Roof Construction