Hanno Scharr, Dr. habil., Forschungszentrum Jülich, Germany,
Fri 07 Sep 2018, 13:00 - 14:00
AGB Seminar Room AGB Building, King’s Buildings, EH9 3JL

If you have a question about this talk, please contact: Konstantinos Voulgaris (s1789135)

Image for Understanding Plant Robots: Robots, Imaging and Computer Vision

Abstract: Plant roots are frequently called the ‘hidden half’ of plants, as they cannot be imaged under field conditions. Measuring root systems properties needs special experimental setups, where considerable engineering efforts are needed especially when medium to high throughput of plants is desired. The first part of the talk gives an overview of the most relevant systems for plant root imaging we designed and built at Forschungszentrum Jülich over the recent years, as well as current developments that use moving robots. Essential for analysis of the acquired images are well adapted algorithmic and software solutions. Robust solutions are available in cases, where the whole root system is visible. The second part of the talk focusses on rhizotron imaging and image analysis, where roots grow in soil and root systems are therefore only partly visible. Here DNN-based (deep neural network) solutions yield promising results. The talk concludes with recent work in root system analysis with DNNs.

Biography: .Hanno Scharr received the Diploma degree (1996) and the PhD degree (2000) in physics (Dr. rer. nat.) from Ruprecht-­Karls-­University Heidelberg, Germany. He received a habilitation in computer science from Johann-­Wolfgang-­Goethe-­Universit y, Frankfurt, Germany (2014). After his PhD he joined the Computational Nano-­Vision Group at Intel Corporation, Santa Clara, CA, where he worked as Senior Researcher. From 2007 to 2011 he lead the Young Investigator Group Automatic Imaging and Quantitative Image Processing in Environmental Plant Sciences at Forschungszentrum Juelich. Currently he is head of Quantitative Image Analysis at the Institute of Bio-­ and Geosciences: Plant Sciences at Forschungszentrum Juelich, Germany. His research interests include computer vision and quantitative image processing as analysis tool for scientific, medical, and industrial data. He coordinated the EU FP7 project ‘Gardening with a Cognitive System’ and is author or coauthor of more than 100 scientific publications. He served as program committee member in >25 international conferences (incl. CVPR/ICCV/ECCV) and reviewer for major computer vision (incl. IJCV, PAMI, TIP) and plant biology journals (e.g. JExBot, Frontiers in Plant Sciences).